Research Doc

4/21/2014 HO ME

Max-Planck-Institut für Eisenforschung: Microstructural Banding Effects Clarified Through Micrographic Digital Image Correlation A bout

Res earc h

P ublic ations

D epartments

M embers

C ontac t and Servic es

D oc toral P rogramme

D is c laimer

Dire ct Acce ss

Microstructural Banding Effects Clarified Through Micrographic Digital Image Correlation Hom e

MA

Inte r-de partm e ntal R e se arch

Microstructural Banding Effects Clarified Through Micrographic Digital Image Correlation C ollaborators: J.P.M.Hoe fnage ls & M.G.D. Ge e rs (@TU/e )

Introduction Ne arly a ce ntury ago, re se arche rs re fe rre d to m icrostructural bands with a rathe r intriguing nam e : ghosts [1, 2]. Although this nom e nclature se e m s surprising at first, it can be unde rstood upon close r inspe ction: the se bands which typically appe ar upon hot work ing can only be e rase d with care fully de signe d he at tre atm e nts, only to m yste riously re -appe ar upon he ating [3] or de form ation [4].

Microstructure Physics and Alloy De sign Inte rdisciplinary R e se arch Fie lds Inter-departmental Research C ollaborative R e se arch C e ntre SFB R e se arch

Awards, Honours

Publications

Le cture s, Movie s

C onfe re nce s, C olloquia, Se m inars Groups

Equipm e nt

Facilitie s

Me m be rs

Alum ni

Visitors

Se rvice s

In the following de cade s, inte nse re se arch has e lucidate d the form ation m e chanism s of the se bands, as illustrate d for the case the fe rritic-pe arlitic bands in footnote 1 . De spite this qualitative unde rstanding, the proce ssing route s to com ple te ly avoid the form ation of bands [6-8] or to pe rm ane ntly re m ove form e d bands [9, 10] m ay not be e conom ically fe asible for all case s [4], or in som e case s be com e the rm odynam ically im possible [11]. Ye t, the band m orphology (thick ne ss, continuity, ge om e try e tc.) de pe nds on he at tre atm e nt [6] and de form ation proce sse s [5], and thus can be m odifie d within e conom ically fe asible lim its to re m ove de trim e ntal e ffe cts of the band. Such an approach calls for cle ar unde rstanding of the influe nce of band prope rtie s on the global m e chanical be havior, which is curre ntly unavailable . W hile a conse nsus is re ache d on the de trim e ntal e ffe cts on the fracture toughne ss, (with sm all e ffe cts on yie ld or te nsile stre ngths), contradictory re sults e x ist for the ductility, varying from no influe nce [12], a slight influe nce [4, 13], to a significant influe nce [9, 14]. Eve n advantage ous e ffe cts of m icrostructural banding we re re porte d, e .g., incre ase in fatigue life [15]. This appare nt inconsiste ncy in the lite rature e sse ntially re sults from the e x pe rim e ntal m e thodology com m only e m ploye d. The influe nce of banding on m e chanical prope rtie s is typically inve stigate d by com paring the global m ate rial be haviour of a bande d m icrostructure with an unbande d m icrostructure [e .g., 4, 9, 12-14, 16]. The unbande d m ate rial is ge ne rally produce d by a hom oge nizing he at tre atm e nt from the bande d m ate rial, although diffe re nt starting m ate rials have also be e n use d [e .g., 9, 14, and 16]. Howe ve r, as e m phasize d by som e authors [14], the e x clusive influe nce of banding can only be e lucidate d by k e e ping othe r m icrostructural param e te rs (e .g. com position, phase fractions, grain size , inclusion m orphology, e tc.) unalte re d. In this re spe ct, it was pointe d out that hom oge nizing he at tre atm e nts also change the m orphology of inclusions, the re by changing the m e chanical be haviour [4, 12]. This illustrate s the com ple x re lation be twe e n the m icrostructure and the m e chanical be havior in bande d m ate rials, e ve n for care fully de signe d he at tre atm e nts [14]. Anothe r e x planation for the above -m e ntione d inconsiste ncy in the lite rature e m e rge s from the diffe re nce in the bande d-phase and/or band m orphologie s in the studie d m icrostructure s. Bande d structure s occur in m any type s of ste e ls (e .g., fe rritic-bainitic, fe rritic-m arte nsitic, pe arlitic-bainitic, pe arlitic-m arte nsitic, bainitic-m arte nsitic [5]) and the obvious diffe re nce in the m e chanical prope rtie s of the se phase s le ad to a diffe re nt ne t e ffe ct of the band. The re fore a ge ne ral unde rstanding of the m ain influe nce of banding cannot be acquire d by e x am ining a single case of bande d m icrostructure . To ove rcom e the afore m e ntione d difficultie s, we carry out an in-situ, local analysis of the de form ation of two distinct bande d m icrostructure s, couple d with m icro-scale strain fie ld m e asure m e nts. The be haviour of the bande d phase is com pare d with unbande d re gions in the sam e m icrostructure , avoiding unde sire d variations introduce d by band-re m oving he at tre atm e nts. In-situ m icrostructural analyse s also provide im prove d insight in the unde rlying de form ation m e chanism s, which gove rn the global m e chanical re sponse . Furthe rm ore , to probe the influe nce of the bande d phase and its m orphology, the local m e chanical be havior of two care fully-chose n lim it case s of bande d m icrostructure s are com pare d in de tail: a m icrostructure containing a continuous, hard band (the m arte nsitic-fe rritic syste m ) and a m icrostructure containing non-continuous, softe r bands (the pe arlitic-fe rritic syste m ), whe re by both m icrostructure s are proce sse d from the sam e starting m ate rial for optim al com parison. The obtaine d inform ation yie lds inde pth m e chanistic unde rstanding of the e ffe ct of band m orphology, providing ge ne ral re com m e ndations for industry to fine -tune ste e l proce ssing param e te rs, accordingly.

Methodology For the m arte nsitic-fe rritic bande d m icrostructure , a dual-phase ste e l (DP600) is chose n, which is com pose d of fe rritic grains surrounde d by m arte nsitic islands at the fe rritic grain boundarie s (Fig.1a). A ∼5 µm thick m arte nsitic band is obse rve d in the ce nte r of the 1 m m thick she e t . Te nsile te st sam ple s are cut with e le ctro-discharge m achining, and the cross se ctions are m e tallographically pre pare d using succe ssive grinding and m e chanical polishing ste ps, followe d by e le ctropolishing and e tching with 2 vol. % nital solution. This protocol give s optim al contrast, i.e . the be st subse t track ing, for the digital im age corre lation (DIC ). The fe rritic-pe arlitic bande d m ate rial is produce d by auste nizing the DP600 (at 1000°C ) followe d by furnace cooling to room te m pe rature . A m orphologically com parable m icrostructure is obtaine d, but with pe arlite re placing the m arte nsite (Fig. 3a), i.e . with the non-continuous pe arlite band re placing the continuous m arte nsite band. Inte rm itte nt te nsile te sts of the se m ate rials are carrie d out using a Kam m rath-W e iss m icro-te nsile stage ,

http://www.mpie.de/index.php?id=4191

1/5

4/21/2014

Max-Planck-Institut für Eisenforschung: Microstructural Banding Effects Clarified Through Micrographic Digital Image Correlation

O pe n Positions

place d in a FEI Q uanta 600 scanning e le ctron m icroscope (SEM). High re solution, low m agnification m icrographs are obtaine d in se condary e le ctron m ode at succe ssive stage s of de form ation (∼Δεglobal ≈ 0.03%). The gray value histogram s of the obtaine d im age s we re adjuste d for m ax im um ove rlap, and re gions of inte re st are chose n for local strain fie ld analysis (using Aram is software , GO M Gm bH.). The subse t size and spacing we re care fully chose n in orde r to re alize the be st trade -off be twe e n spatial and strain re solution. Good im age corre lations we re ge ne rally achie ve d, although a fe w subse ts are ine vitably lost as a re sult of se ve re surface roughe ning, spe cifically at large de form ations in the fe rrite grains.

Results and Discussion The analysis starts by e x am ining the de form ation-induce d m orphological change s in the DP ste e l m icrostructure (Fig. 1). Note that the global stre ss-strain re sponse is also provide d in the inse ts.

Fig. 1 . T ypic al example of the deformation of D P s teel: SE M images (left) and von M is es effec tive s train field overlays (right) at four different s tages of deformation. T he global mec hanic al res pons e is als o s hown in the ins ets (true s tres s (M P a) vs . D I C averaged s train). T he rolling direc tion, i.e. the tes ting direc tion, is s hown in s ubplot (a). White arrows indic ate void nuc leation s ites .

Alre ady in the e arly stage s of de form ation, the form ation of local slip line s within individual fe rritic grains is obse rve d (Fig. 1b). As the de form ation proce e ds, the num be r of slip line s incre ase s, and the strain distribution be com e s incre asingly he te roge ne ous (Fig. 1c). This he te roge ne ity is cause d by intra-granular form ation of she ar bands at approx im ate ly 45° to the te sting dire ction. C are ful inspe ction of highe r m agnification im age s re ve als that the m arte nsitic grains are acting as obstacle s to the propagation of the slip line s, and he nce to the plastic de form ation of the fe rritic grains. In fact, the local strain fie lds re ve al that the fe rrite grains e x pe rie nce significant plastic de form ation, whe re as the m arte nsite islands are straine d only e lastically at this stage (εglobal ≤ 0.05). In the m arte nsite band, on the othe r hand, local re gions of high strain are obse rve d at those positions whe re the m arte nsite band is the thinne st, as m ark e d with white arrows in Fig. 1c and 1d. As can be ve rifie d from the m icrographs, the m arte nsitic band is se ve re ly de form e d at the se particular re gions up to its fracture strain, le ading to void nucle ation. From care ful e x am ination of the com ple te se rie s of strain m aps, it be com e s cle ar that the intra-granular she ar band inte rce pts the m arte nsitic band e x actly at the se critical locations, m ak ing the m the first site s whe re voids nucle ate , while the m arte nsite islands (and the re st of the m arte nsite band) show hardly any de form ation or void form ation. The sam e tre nds in the strain distribution within the bands are obse rve d for othe r spe cim e ns of the sam e DP ste e l (Fig.2), i.e ., the highe st local strains are obse rve d in the narrowe st se ction of the band, le ading ine vitably to void nucle ation at ‘hot spots’ in the se continuous band m orphologie s.

http://www.mpie.de/index.php?id=4191

2/5

4/21/2014

Max-Planck-Institut für Eisenforschung: Microstructural Banding Effects Clarified Through Micrographic Digital Image Correlation

Fig. 2 . V oid nuc leation within the band, s hown here for different martens itic band morphologies , for undeformed s tate (above) and deformed s tate (below). N ote that the s c ale is identic al to Fig. 1 . White arrows indic ate the void nuc leation s ites .

From the obse rve d se que nce of m icro-e ve nts, it is cle ar that the she ar bands are de ve lope d in pre fe re ntial paths within the m icrostructure . In orde r to accom m odate the plasticity of the surrounding fe rritic m atrix , the m arte nsite in the narrow re gions of the band carrie s m ost of the de form ation, the re by com pe nsating the lack of de form ability in the thick e r parts of the band. Since m arte nsite has a lim ite d strain-to-fracture , the re sulting high strains ine vitably le ad to e arly void nucle ation. The se obse rvations also re ve al the significant role of the m arte nsite m orphology on the global m ate rial be havior, which te nds to be e ve n m ore im portant than the m arte nsite volum e fraction [18]. Sim ilar to DP m icrostructure s with highly-conne cte d m arte nsite ne twork s, the m arte nsite bands analyze d he re are also force d to plastically de form at an e arly stage of de form ation, de viating from “ide al” DP ste e l be haviour , and le ading to a de cre ase in strain-to-fracture . Le t us ne x t conce ntrate on the de form ation-induce d m icrostructure e volution in the pe arlitic-fe rritic (PF) ste e l (with the sam e ove rall che m ical com position as the DP ste e l). A typical e x am ple is shown in Figure 3, whe re the de form ation stage s of the re sulting PF ste e l are shown.

Fig. 3 . T ypic al example of the deformation of P F s teel: SE M images (left) and von M is es effec tive s train field overlays (right) at four different s tages of deformation. T he global mec hanic al res pons e is als o s hown in the ins ets (true s tres s (M P a) vs . D I C averaged s train). N ote that all s c ales are identic al to thos e in Fig. 1 . T he rolling direc tion, i.e. the tes ting direc tion, is s hown in s ubplot (a). T he white arrow in (b) points at the gap in the pearlitic band, and the white arrow in (d) s hows the narrowes t s ec tion of the pearlitic band, where the s hear band penetrates through the mic ros truc tural band.

As se e n in the m icrographs, the distribution of the pe arlite phase in the fe rrite -base d m icrostructure is qualitative ly sim ilar to that of m arte nsite in the DP ste e l. O n the othe r hand, the global re sponse of the m ate rial is significantly diffe re nt, as se e n from the stre ss strain inse t figure s (Fig. 3). As e x pe cte d, e ffe ctive ly re placing m arte nsite with pe arlite re sults in a de cre ase in stre ngth, whe re as the ove rall ductility im prove s. The m ain point of inte re st he re , howe ve r, is the local de form ation patte rns at the le ve l of the bande d m icrostructure .

http://www.mpie.de/index.php?id=4191

3/5

4/21/2014

Max-Planck-Institut für Eisenforschung: Microstructural Banding Effects Clarified Through Micrographic Digital Image Correlation At the first glance , e x am ination of the m icrographs and the strain m aps at diffe re nt stage s of de form ation re ve als sim ilar tre nds in the de form ation as found for DP ste e l (Fig. 1 and 2): initially form ation of the slip line s in the fe rritic grains, localize d de form ation at approx im ate ly 45° to the te sting dire ction, and the e volution of the strain distribution from hom oge ne ous to incre asingly he te roge ne ous with incre asing de form ation. Howe ve r, m ore de taile d analyse s re ve als a significant diffe re nce at a rathe r e arly stage , se e Fig. 3b. The first she ar band initiate s at a discontinuity in the pe arlitic band (shown with a white arrow in Fig. 3b). As confirm e d by additional m e asure m e nts, whe n the band (m orphology) is not continuous, the pre fe re ntial path for the she ar bands is through the gaps in the band. W ith furthe r de form ation, this she ar band de form s ste adily while the re st of the pe arlitic band e x pe rie nce s only a lim ite d am ount of de form ation (Fig. 3c). O nly at m uch highe r global strain, anothe r she ar band is de ve lope d that pe rcolate s through the narrowe st se ction of the pe arlitic band (shown with an arrow in Fig. 3d), in the abse nce of any void nucle ation. This m ark e d diffe re nce in m icrostructural be haviour be twe e n DP and PF ste e ls is quantifie d in Fig. 4, which com pare s the local strain e volution of m arte nsite and pe arlite to the local strain e volution in the fe rrite grains outside the band, as a function of DIC ave rage d strain.

Fig. 4 . L oc al s train evolution in martens itic and ferritic grains in D P s teel (green), and pearlitic and ferritic grains in P F s teels (red). H igh and low deformation zones within the bands in eac h s teel are als o s hown, with s quare and c irc le markers res pec tively. M artens ite c rac king is indic ated with a X. N ote that the error bars repres ent the s tandard deviation of the average s train in the s pec ific c ons tituent.

First conce ntrate on the phase s outside the band. The strain partitioning be twe e n m arte nsite and fe rrite is obviously m ore pronounce d than be twe e n pe arlite and fe rrite . For the DP ste e l, the narrow band se ctions show high de form ations and e ve ntually crack ing to com pe nsate for the re lative ly low de form ation in the wide se ctions of the band. Note that the strains in the se critical band re gions are significantly highe r than e ve n the m ost se ve re ly de form ing fe rritic grains. For the PF ste e l, the strain partitioning is le ss significant since the pe arlite phase we ll accom m odate s the fe rritic de form ation. The re fore , the am ount of e x ce ss de form ation of the narrow se ctions of the pe arlitic band re m ains re lative ly lim ite d. He nce , the probe d pe arlitic band (which has a large variance in its thick ne ss sim ilar to the m arte nsitic band) e x pe rie nce s no void nucle ation at the sam e le ve l of (global) strain. Eve n though this obse rvation is consiste nt with the large r ductility of pe arlite com pare d to m arte nsite , it cle arly shows that the influe nce of band m orphology on global m e chanical be haviour is strongly de pe nde nt on the m e chanical prope rtie s of the bande d phase .

Conclusions In sum m ary, the obtaine d re producible re sults of the local analysis on two lim it case s of bande d structure s support the following conclusion for the m e chanical be haviour of bande d m icrostructure s: The influe nce of a bande d structure on the global prope rtie s (YS, UTS, ductility, e tc.) of m e tals is critically de pe nde nt on the m orphology of the band, as we ll as the m e chanical be haviour of the phase that com pose s the band. In m icrostructure s whe re the re is a continuous m icrostructural band, she ar bands are force d to de ve lop through the band, the re fore the y pe rcolate through the narrowe st se ction of the band. This force s the bande d phase to de form be yond its plastic lim it, e spe cially if the re is a significant diffe re nce in the ultim ate strains of the phase s com posing the bande d m icrostructure (e .g. the case of m arte nsitic-fe rritic dual phase ste e l). For discontinuous m icrostructural bands, she ar bands naturally cross at the gaps within the band, the re by de laying e arly dam age initiation. O bviously, the stre ngth of the bande d phase also plays an im portant role , e .g. a continuous band of a high fracture stre ngth phase m ay accom m odate high local stre sse s without dam age initiation, whe re as a low fracture stre ngth phase m ay not.

http://www.mpie.de/index.php?id=4191

4/5

4/21/2014

Max-Planck-Institut für Eisenforschung: Microstructural Banding Effects Clarified Through Micrographic Digital Image Correlation From an industrial point of vie w, it is obviously com ple x to e lim inate the bande d phase . Howe ve r, e ve n for case s whe re com ple te re m oval of banding is not e conom ically fe asible , pre se nte d re sults re ve al that the de trim e ntal influe nce of a bande d m icrostructure can be significantly re duce d by alte ring the m orphology of the band in orde r to (i) avoid m icrostructure s with continuous bands, and (ii) de cre ase the thick ne ss variation of the band. For the critical case of re ducing the se ve rity of banding in DP ste e ls, for e x am ple , this m ay be achie ve d by optim izing the re porte d critical production param e te rs (e .g. rate of cooling during hot rolling, inte rcritical anne aling te m pe rature , soak ing duration, e tc. [6-8]) that have a significant influe nce on the de gre e of banding.

Acknowledgements This re se arch was carrie d out unde r the proje ct num be r MC 2.05205 in the fram e work of the R e se arch Program of the Mate rials innovation institute M2i (www.m 2i.nl), the form e r Ne the rlands Institute for Me tals R e se arch. The authors would lik e to thank Bart Vosse n for his contribution.

References [1] H.M. Howe , The Me tallography of Ste e l and C ast Iron, McGraw-Hill, Ne w York , 1916. [2] J.H. W hite ly, J. Iron. Ste e l. I., 113 (1926) 213. [3] L.E. Sam uals, O ptical Microscopy of C arbon Ste e ls, ASM, O hio, 1980. [4] R .A. Grange , Me tall. Mate r. Trans. B., 2 (1971) 417. [5] J.D. Ve rhoe ve n, J. Mate r. Eng. Pe rform ., 9 (2000) 286. [6] F.G. C aballe ro e t al., Mate r. Trans., 47 (2006) 2269. [7] P.E.J. R ive ra-Diaz-de l-C astillo, S. van de r Zwaag, Ste e l. R e s. Int., 75 (2004) 711. [8] W . Xu, P.E.J. R ive ra-Diaz-de l-C astillo, S. van de r Zwaag, ISIJ. Int., 45 (2005) 380. [9] A.R . Marde r, Me tall. Trans. A., 13 (1982) 85. [10] X.L. C ai e t al., Me tall. Trans. A., 16 (1985) 543. [11] C .F. Jatczak e t al., Trans. ASM., 48 (1956) 279. [12] W .A. Spitzig, Me tall. Trans. A., 14 (1983) 271. [13] A.S. Bor, ISIJ. Int., 31 (1991) 1445. [14] X.L. C ai e t al., Me tall. Trans. A., 16 (1985) 1405. [15] A.A. Korda e t al., Scripta. Mate r., 54 (2006) 1835. [16] M. Mazinani, W .J. Poole , Me tall. Mate r. Trans. A., 38 (2007) 328. [17] C .C . Tasan e t al., Int. J. Me ch. Mate r. 41 (2009) 1264. [18] S. Kim , S. Le e , Me tall. Mate r. Trans. A., 31 (2000) 1753.

Other group activities ● othe r proje cts:

http://www.mpie.de/index.php?id=4191

Adaptive Structural Mate rials

5/5