At

HK

a

ARRA

KSMUD

1

oapmmpcdhvdteec

aotmtf

(

0d

Materials Science and Engineering A 527 (2010) 5430–5434

Contents lists available at ScienceDirect

Materials Science and Engineering A

journa l homepage: www.e lsev ier .com/ locate /msea

n ultrahigh strength steel with ultrafine-grained microstructure producedhrough superplastic forming

an Zhang ∗, Xiaole Cheng, Lei Zhang, Bingzhe Baiey Laboratory for Advanced Materials, Department of Materials Science and Engineering, Tsinghua University, 100084 Beijing, China

r t i c l e i n f o

rticle history:eceived 12 January 2010eceived in revised form 24 April 2010

a b s t r a c t

A novel procedure combined with superplastic forming was applied to produce an ultrahigh strength steelwith ultrafine-grained microstructure. Although the hardness reached HRC65 at room temperature, thevalue of m = 0.38 and flow stress of 57 MPa were found during superplastic forming at 1023 K and strain

ccepted 10 June 2010

eywords:uperplasticityartensite

rate of 10−4 s−1. Due to reasonable composition design, ultrahigh strength was ensured within a widerange of cooling rate after superplastic forming without the need of supplementary heat treatment.

© 2010 Elsevier B.V. All rights reserved.

ltrafine-grained microstructureynamic equilibrium

. Introduction

The application of grain refinement has long been the focusf material research. In the metallic field, ultrafine-grained (UFG)nd nanocrystalline (NC) microstructures usually exhibit superiorroperties and performance [1]. Generally speaking, there are twoajor routes to obtain UFG and NC microstructures for metallicaterial. One is to refine coarse-grained bulk metal through severe

lastic deformation (SPD), including equal channel angular pro-essing (ECAP) [2–4], high pressure torsion (HPT) [5], and wirerawing. The other is to compact NC powders via hot-pressing,ot isostatic pressing (HIP), etc. However, there exist some disad-antages on these methods: grain refinement is restricted by theifficulty of heat dissipation during ECAP; continuous productionhrough ECAP and HPT is hard to realize because of the limitedquipment capacity; preparation of NC powders is complicated,tc. Therefore, the above two routes do not seem suited for theonventional mass-production of metals.

In recent years, in the field of steel, several studies havettempted to achieve UFG microstructures by rolling and annealingf steels with various carbon contents and starting microstruc-

ures. In the cases of ferrite–pearlite [6], martensite [7], tempered

artensite [8] and ferrite-martensite [9] as starting microstruc-ures before rolling and subsequent annealing, submicrometererrite grains have been obtained. Unfortunately, these studies are

∗ Corresponding author. Fax: +86 10 62771160.E-mail addresses: zhanghan06@mails.tsinghua.edu.cn, zhangaoxuan@163.com

H. Zhang).

921-5093/$ – see front matter © 2010 Elsevier B.V. All rights reserved.oi:10.1016/j.msea.2010.06.027

mostly referred to low carbon steels and ferrite grains are prone togrow coarser during annealing at high temperature due to less pin-ning effect by a small amount of carbides. In addition, a few decadesago, Sherby et al. studied a ferrite (˛) UFG structure with a uni-form distribution of spheroidized cementite (�) particles (i.e. (˛ + �)microduplex structure) [10,11], with the purpose of exploring thesuperplasticity of ultrahigh carbon steels. However, the methodsof obtaining the (˛ + �) microduplex structure are also more or lesscomplicated because of the necessity of divorced eutectoid trans-formation with associated deformation. On the other hand, an ideaof forming the (˛ + �) microduplex structure without thermome-chanical processing in ultrahigh carbon steels was addressed [12],but the heat treatment by multiple pass in this idea is a little time-consuming.

The present paper is to reveal a Mn-alloyed ultrahigh strengthsteel with UFG microstructure produced by a novel procedure. Thekey of the idea is selecting an optimal warm deformation con-dition to realize superplasticity of the steel and then ensuringultrahigh strength after deformation without the need of supple-mentary heat treatment. Besides, special attention is paid to theeffect of composition design and the microstructure refinementthrough predeformation. In this way, the procedure combined withsuperplastic forming for producing the ultrahigh strength steel issimplified.

2. Experimental

The Fe–0.56C–2.12Mn–1.76Si–0.95Cr–0.25Mo (mass%) steelwas designed and used in this study. Fig. 1(a) shows schematic dia-

H. Zhang et al. / Materials Science and Engineering A 527 (2010) 5430–5434 5431

Fig. 1. (a) Schematic diagram of the novel procedure and (b) strain–time curve of the warm deformation step.

F withinc color

gwtcislcpcff(

ig. 2. (a) The SEM morphology of specimen 1. (b) The magnification of the zoneomposition analysis on the blue spot in (b). (For interpretation of the references to

ram of the procedure for producing the ultrahigh strength steelith UFG microstructure. A1 and Acm represent pearlite–austenite

ransformation (in equilibrium) temperature and pro-eutectoidementite–austenite transformation (in equilibrium) temperaturen Fe–C phase diagram, respectively. Here we consider the designedteel as a hypereutectoid steel since the alloying in the present workeads to a lower eutectoid carbon content (∼0.5%C) [13]. The pro-edure in Fig. 1(a) mainly includes three steps, i.e. austenitizing,

redeformation and warm deformation. For simulation of this pro-edure, long cylinders 8 mm in diameter and 90 mm in length cutrom forging billets were firstly used for austenitizing and prede-ormation. Then specimens 6 mm in diameter and 10 mm in lengthhereafter, specimen 1) machined from the central deformation

the red dashed line frame in (a). (c) The SEM morphology of specimen 3. (d) Thein this figure legend, the reader is referred to the web version of this article.)

part of the long cylinders underwent warm deformation, leadingto the formation of final product (hereafter, specimen 2). In addi-tion, a procedure without the predeformation step was discussedfor the purpose of comparison with the above procedure (the othersteps kept unchanged, as shown briefly in the upper right side ofFig. 1(a)). The simulation of this procedure is as follows: long cylin-ders 8 mm in diameter and 90 mm in length cut from forging billetswere used for austenitizing and then specimens 6 mm in diame-

ter and 10 mm in length (hereafter, specimen 3) machined fromthe central part of the long cylinders underwent warm deforma-tion, leading to the formation of specimen 4. Specimens 3 and 4obtained during this procedure were used to compare with spec-imens 1 and 2, respectively. Water quenching (W.Q.) was applied

5432 H. Zhang et al. / Materials Science and Engineering A 527 (2010) 5430–5434

en 3

fm

Mrc0s(TbNttBu(ma[fwm

pMem

3

3

wiiAsfbpswhpT

Fig. 3. The stress–strain condition of (a) specim

or the purpose of convenient experimental analysis on all speci-ens.The strain–time curve of the warm deformation step (i.e. period

N in Fig. 1(a)) is shown in Fig. 1(b). True strain of 0.04 at strainates of 10−4, 2 × 10−4, 4 × 10−4, 10−3 and 3 × 10−3 s−1 was appliedorresponding to stages I, II, III, IV and V, respectively. True strain of.3 at 10−3 s−1 (i.e. period AB) was introduced with the purpose ofhowing the development of flow stress during warm deformationfirst period and second period are defined as shown in Fig. 1(b)).he predeformation and warm deformation steps were performedy uniaxial compression test on Gleeble 1500D thermal simulator.otably, as an indispensable step for obtaining UFG microstructure,

he warm deformation was also used for evaluation of superplas-ic behavior because strain at different strain rates was applied.esides, the warm deformation is here referred to low rate man-facturing procedure. Therefore, classical strain rate experimentsi.e. Backofen method) can be employed for the measurement of

value. The m value is strain rate sensitivity exponent, which isn important parameter connected with the macroscopic necking14]. For metallic materials, a high m value usually indicates a dif-use neck development and thus a delay of the onset of fracture,hich leads to high tensile elongation [15]. Generally, a value of≥ 0.3 is considered as a superplastic characteristic.In order to analyze microstructural evolution during the two

rocedures, the microstructure of specimen 1–4 was focused on.icrostructural observation was conducted using transmission

lectron microscopy (TEM, JEOL JEM-2011) and scanning electronicroscopy (SEM, JEOL JSM-6301F).

. Results

.1. Microstructure after predeformation

Fig. 2(a) shows the SEM morphology of specimen 1. The net-ork carbide is located along original austenite boundary, which

ndicates that carbide precipitates during predeformation. The orig-nal austenite grains are inhomogeneous and large (80–350 �m).fter water quenching, austenite transforms to martensite. Fig. 2(b)ignifies the magnification of the zone within the red dashed linerame, from which it can be seen that relatively fine martensitelocks are obtained. This may be attributed to the refining effect ofredeformation addressed in our previous work [16]. The compo-

ition analysis on the blue spot in Fig. 2(b) is shown in Fig. 2(d),hich further identifies the formation of carbide. On the otherand, no carbide exists in the microstructure cooled to room tem-erature at 0.5 K/s after the same austenitizing as shown in Fig. 1(a).his might be caused by the suppressing effect of silicon on car-

and (b) specimen 1 during warm deformation.

bide precipitation during cooling [17]. Therefore, it is supposedthat the predeformation used to refine microstructure in this studyaccelerates the formation of network carbide which is harmful forthe steel performance. Fortunately, this kind of network carbidecan be eliminated during subsequent warm deformation, whichwill be seen later from the microstructure of specimen 2. Fig. 2(c)shows the SEM morphology of specimen 3. Compared with speci-men 1, specimen 3 without undergoing predeformation has coarsermartensite blocks, in accordance with our previous work [16]. Herewe focus on the martensite morphology of specimen 1 and 3 dueto its great effect on the superplastic characteristic during subse-quent warm deformation (since many austenite nuclei exist fromthe prior martensite).

3.2. Superplastic characteristic during warm deformation

Fig. 3(a) and (b) shows the warm deformation steps of specimen3 and 1, respectively. The stress of specimen 1 is lower than thatof specimen 3 under an identical strain. From the stress downwardtrend during period AB, it could be supposed that the microstruc-tural evolution of the two specimens during warm deformation isbeneficial for improving deformability. Noticeably, the stress at Bpoint is almost the same as that at stage IV of second period for thetwo specimens, which indicates that the microstructure reaches adynamic equilibrium at B point. The dynamic equilibrium is herereferred to the equilibrium of average grain size [18], which will bediscussed later. The m value and flow stress during second periodfor specimen C and A are listed in Table 1. It can be seen thatspecimen 1 after undergoing predeformation exhibits superplas-tic characteristic (m > 0.3) at strain rate of 10−4–10−3 s−1. However,superplastic characteristic of specimen 3 is not found. Interestingly,although the hardness of specimen 1 reaches HRC65 at room tem-perature, the stress value during second period is quite low (evenclose to 57 MPa at 10−4 s−1). In addition, the stress of specimen 1is obviously lower than that of specimen 3 for the correspondingstage of second period. This indicates that the predeformation stepis crucial for realizing the superplasticity during warm deformation.

3.3. Microstructure after warm deformation

Fig. 4(a) and (b) shows the SEM micrographs of specimen 2and 4, respectively. It can be easily seen that a kind of phase in

nearly equiaxial shape (grain size about 0.5–1.5 �m) distributeson the martensite matrix. The light zone of the TEM micrographin Fig. 4(c) denotes the phase in specimen 2. Through SAED pat-tern analysis, the phase is identified as ferrite. The blue and redcurves in Fig. 4(a) and (b) represent the grain boundary of austen-

H. Zhang et al. / Materials Science and Engineering A 527 (2010) 5430–5434 5433

mic

idfowqfifamdcGol4eets

austenite grain size in order to realize superplasticity of ultra-

TT

Fig. 4. The SEM micrographs of (a) specimen 2 and (b) specimen 4. The TEM

te formed from the martensite of specimen 1 and 3 during warmeformation, respectively. It is seen that austenite grains originatedrom specimen 1 with fine martensite blocks are smaller than thoseriginated from specimen 3 with coarse martensite blocks duringarm deformation. Therefore, after warm deformation and subse-

uent water quenching, the martensite blocks of specimen 2 arener than those of specimen 4. Another interesting finding is that

errite in specimen 2 distributes more uniformly and also in a largermount than that in specimen 4. Besides, the small ferrite grains areainly located at the grain boundary of austenite (at temperature)

uring warm deformation, which could be considered as “lubri-ant” for austenite grain boundary sliding (GBS). As “lubricant” forBS, ferrite might also has the function of suppressing the growthf austenite grains and coordinating the warm deformation step,eading to a dynamic equilibrium of austenite grain size (about–8 �m in specimen 2) during warm deformation. The dynamic

quilibrium could be expected as the main reason why specimen 2xhibits superplastic characteristic in the present work, similar tohe dynamic equilibrium of ferrite grain size in (˛ + �) microduplextructure [16]. From the TEM micrograph of local ferrite in speci-

able 1he m value and stress during second period for specimen 3 and 1.

Strain rate (s−1)

10−4–2 × 10−4 2 ×Specimen3

m value 0.23 0.2Stress (MPa) 86–101 10

Specimen1

m value 0.38 0.3Stress (MPa) 57–74 7

rographs of specimen 2 focusing on (c) a ferrite grain and (d) a ferrite pair.

men 2 (as shown in Fig. 4(d)), it can be seen that a clear boundaryis located between two ferrite grains (i.e. a ferrite pair). Here it isnot an accidental phenomenon because such ferrite grain pairs arefrequently found at the austenite (at temperature) grain juncture,as shown by the green arrows in Fig. 4(a). It could be supposed thatferrite grains might relieve stress concentration through formingsub-boundary and finally disintegrate into smaller ones with theincrease of deformation degree. The light ferrite zone in Fig. 4(c)shows that dislocation density within ferrite is low, which couldalso suggest the ferrite discussed in the present work has effect oncoordinating warm deformation.

4. Discussion

In the past, it was found that cementite could maintain fine

high carbon steel during austenite plus cementite range [11] andsometimes silicon was added to decrease the carbon content ofeutectoid composition, thereby ensuring a larger amount of cemen-tite available for the pinning of austenite grain boundary [19].

10−4–4 × 10−4 4 × 10−4–10−3 10−3–3 × 10−3

2 0.23 0.191–118 118–146 146–180

5 0.32 0.234–94 94–126 126–163

5 nd En

Up(ottn

dtaohc(aftmdb(aatI(b

5

fwspifibtbmf

[[[[

[

[[[

434 H. Zhang et al. / Materials Science a

nfortunately, ultrahigh strength is difficult to ensure after super-lastic forming without additional heat treatment since austeniteat temperature) mainly transforms to pearlite within a wide rangef cooling rate. Maybe water quenching soon after superplas-ic forming can be applied to ensure ultrahigh strength directlyhrough obtaining martensite. But technically, water quenching isot a promising way to produce ultrahigh carbon steel.

In the present work, the steel composition is reasonablyesigned. Alloy elements like Mn significantly retards pearliteransformation [20] during cooling after warm deformation. As

result, martensite (or bainite) plus ferrite microstructure isbtained even at very low cooling rate (like 0.5 K/s). Generally,igh hardness of steels with martensite (or bainite) microstructureorresponds with high strength [21]. Therefore, ultrahigh strengthHRC ≥ 61) of the designed steel may be ensured during air coolingfter superplastic forming. Besides, two requirements are neededor the purpose of realizing superplasticity during warm deforma-ion in the present work: (1) adequate predeformation at A1–Acm

ust be applied in order to obtain fine microstructure before warmeformation; (2) the warm deformation should be conducted at justelow A1 temperature in order to form a large amount of austeniteat temperature) through strain induced austenite transformationnd simultaneously retain some ferrite to suppress the growth ofustenite. In this way, a novel procedure combined with superplas-ic forming is developed to produce an ultrahigh strength steel.n addition, the abovementioned ferrite located at the austeniteat temperature) boundary remains in the final product, which iseneficial for the steel toughness [22].

. Conclusion

In summary, a novel procedure combined with superplasticorming for producing ultrahigh strength steel has been put for-ard, whose schematic diagram is shown in Fig. 1(a). In fact, the

econd W.Q. process can be changed to relatively slow coolingrocess (like air cooling) in real production. The predeformation

s introduced before the first W.Q. process in order to obtainne microstructure. Notably, the predeformation at A1–Acm could

e simplified by adjusting the deformation condition as long ashe refining effect is realized. During warm deformation at justelow A1 temperature, strain induced austenite transformationay occur and a dynamic equilibrium of austenite grain size is

ormed gradually which is beneficial for superplasticity. Although

[[[[

[

gineering A 527 (2010) 5430–5434

the microstructure after predeformation and water quenching con-tains network carbide located along the boundary of very largeaustenite grains (80–350 �m at temperature), the warm defor-mation at just below A1 temperature may eliminate the networkcarbide and lead to the formation of relatively small austenitegrains (4–8 �m) whose growth is suppressed by the ferrite dis-tributed at the austenite boundary. Furthermore, due to reasonablecomposition design (Mn, Cr, etc), pearlite transformation duringcooling is retarded. Therefore, UFG microstructure (martensite orbainite) could be directly obtained during relatively slow cool-ing process after warm deformation. At the same time, ultrahighstrength (HRC ≥ 61) is ensured without the need of supplementaryheat treatment after superplastic forming.

Acknowledgement

Electron microscopy analysis was supported by Beijing NationalCenter for Electron Microscopy, Department of Materials Scienceand Engineering, Tsinghua University.

References

[1] K.S. Kumar, H. Van Swygenhoven, S. Suresh, Acta Mater. 51 (2003)5743–5774.

[2] R.Z. Valiev, R.K. Islamgaliev, I.V. Alexandrov, Prog. Mater. Sci. 45 (2000)103–189.

[3] V.M. Segal, Mater. Sci. Eng. A 197 (1995) 157–164.[4] Y. Iwahashi, J. Wang, Z. Horita, et al., Scripta Mater. 35 (1996) 143–146.[5] S.X. McFadden, R.S. Mishra, R.Z. Valiev, et al., Nature 398 (1999) 684–686.[6] R. Song, D. Ponge, D. Raabe, et al., Acta Mater. 53 (2005) 845–858.[7] N. Tsuji, R. Ueji, Y. Minamino, et al., Scripta Mater. 46 (2002) 305–310.[8] A. Ohmori, S. Torizuka, K. Nagai, ISIJ Int. 44 (2004) 1063–1071.[9] Y. Okitsu, N. Takatab, N. Tsuji, Scripta Mater. 60 (2009) 76–79.10] O.D. Sherby, B. Walser, C.M. Young, et al., Scripta Metall. 9 (1975) 569–573.11] B. Walser, O.D. Sherby, Metall. Trans., A 10 (1979) 1461–1471.12] K. Tsuzaki, E. Sato, S. Furimoto, et al., Scripta Mater. 40 (1999) 675–681.13] J.E. Zhou, G.S. Wang, Y.M. Xu, et al., Handbook of Heat Treatment, China Machine

Press, Beijing, 2008.14] T.G. Nieh, J. Wadsworth, O.D. Sherby, Superplasticity in Metals and Ceramics,

Cambridge University Press, New York, 2005.15] E.W. Hart, Acta Metall. 15 (1967) 351–355.16] H. Zhang, L. Zhang, X. Cheng, et al., Scripta Mater. 62 (2010) 798.17] Y. Tomita, T. Okawa, Mater. Sci. Eng. A 172 (1993) 145–151.

18] G. Krauss, Deformation, Processing and Structure, Metals Park, Ohio, 1984.19] O.D. Sherby, T. Oyama, D.W. Kum, et al., J. Met. 37 (1985) 50–56.20] L. Yang, H.S. Fang, Z.H. Meng, Acta Metall. Sin. 28 (1992) 16–20.21] M.Y. Tu, C.A. Hsu, W.H. Wang, et al., Mater. Chem. Phys. 107 (2008)

418–425.22] P.G. Xu, H.S. Fang, B.Z. Bai, J. Iron Steel Res. Int. 9 (2002) 33–38.