1. Introduction

In addition to the known classic sites for heterogeneousnucleation of ferrite, such as grain boundaries and disloca-tions, in recent years a third way is being studied whichconsists of the intragranular nucleation of ferrite on precipi-tates and inclusions. Some authors have studied the intra-granular nucleation of acicular ferrite on MnS–CuS inclu-sions1) and also on Ti2O3 particles.2–5) Others have studiedthe intragranular nucleation of idiomorphic ferrite on com-plex precipitates formed by V(C, N) type inclusions andprecipitates.6–8)

The mechanisms that promote intragranular nucleationvary according to the type of particle or inclusion. On theone hand, the presence of MnS type inclusions causes alocal Mn depletion that increases the Ar3 value, favouringthe nucleation of ferrite. On the other hand, VC or VN havea low interfacial energy in relation to ferrite but a relativelyhigh interfacial energy in relation to austenite for the(001)V(C, N) boundary compared to MnS. These advantagesof VC and VN over MnS in the interphase boundary energybalance presumably promote the intragranular ferrite trans-formation for complex precipitates.9)

However, the intragranular nucleation of ferrite has notyet been clearly demonstrated and its contribution to ferritegrain refinement is still to be quantified. The present workrepresents the continuation of previous work by the authorson induced precipitation and on the intragranular nucleationof ferrite in vanadium microalloyed steels, using an ownmethodology developed over recent years, which seeks to

demonstrate and to quantify the contribution of intragranu-lar nucleation of ferrite on V(CN) precipitates.10–13)

2. Materials and Experimental Procedure

The steel, whose chemical composition is shown inTable 1, was manufactured industrially and is a low carbonV-microalloyed structural steel. It has been denominated Y5in order to distinguish it from other microalloyed steelsmanufactured by the same factory.

The specimens for torsion tests had a gauge length of50 mm and a diameter of 6 mm. The austenitisation temper-ature was 1 200°C for 10 min. The recrystallised fractionwas determined for several temperatures between 1 100°Cand 800°C applying the double deformation technique, andin particular the back extrapolation method.14,15) The ap-plied strains were 0.20 and 0.35, respectively, which wereinsufficient to promote dynamic recrystallisation (Fig. 1).The ferritic microstructures were resolved by optical mi-croscopy and the analysis of V(C, N) precipitates was per-formed by Electron Microscopy (TEM, SEM-FEG).

3. Results and Discussion

3.1. Static Recrystallisation

First of all the solubility temperatures of vanadium car-

1603 © 2008 ISIJ

ISIJ International, Vol. 48 (2008), No. 11, pp. 1603–1608

Intragranular Nucleation of Ferrite on Precipitates and GrainRefinement in a Hot Deformed V-microalloyed Steel

Sebastian F. MEDINA,1) Lucia RANCEL,1) Manuel GÓMEZ,1) Randa ISHAK2) and Massimo De SANCTIS2)

1) National Centre for Metallurgical Research (CENIM-CSIC), Av. Gregorio del Amo 8, 28040-Madrid, Spain.E-mail: smedina@cenim.csic.es 2) Department of Chemical Engineering, Industrial Chemistry and Materials Science,Pisa University, Via Diotisalvi 2, Pisa 56126, Italy.

(Received on April 23, 2008; accepted on August 19, 2008 )

The intragranular nucleation of ferrite has been studied in a V microalloyed steel (C�0.103; Mn�1.463;V�0.139; N�0.0100% mass). Using torsion tests and applying the double deformation method known as“back extrapolation”, the recrystrallised fraction of austenite has been determined for several deformationtemperatures and two strain values (0.20 and 0.35) and has been plotted as a function of time. Recrystalli-sation-precipitation-time-temperature (RPTT) diagrams have then been drawn. The RPTT diagrams depictprecipitation kinetics as a function of the temperature and time and this information has been used to studythe intragranular nucleation of ferrite, cooling specimens from programmed temperatures and moments forwhich the precipitated volume and the average precipitate size (determined by TEM) are known. The resultshave allowed us to determine the contribution of intragranular nucleation to ferritic grain refinement, whichwas approximately 20%.

KEY WORDS: intragranular nucleation; ferrite; precipitates.

Table 1. Chemical composition of steel Y5 (% mass).

bides and nitrides were calculated in accordance with Turk-dogan’s solubility product16) and the precipitated volumefraction at 950°C according to procedure of Manohar et al.,17)

yielding the values shown in Table 2.Austenitisation at 1 200°C�10 min was sufficient to

completely dissolve the vanadium precipitates. With theaim of measuring the austenite grain size at the austenitisa-tion temperature, one specimen was quenched in water. Thegrain size was measured applying ASTM standard E-112,obtaining a value of 127 mm.

The recrystallised fraction (Xa) determined by torsiontests at several temperatures was plotted against time, ob-taining curves such as those included in Figs. 2 and 3,which correspond to strains of 0.20 and 0.35, respectively.The curves corresponding to temperatures equal to orbelow 1 000°C present a plateau, or the inhibition of recrys-tallisation, due to strain-induced precipitation of V(C, N)particles.

The curve that does not present a plateau has the typicalsigmoidal shape of Avrami’s law. The length of the plateauis not unlimited, and after a short inhibition time recrystalli-sation once again progresses until reaching completion. Therecommencement of recrystallisation is due fundamentallyto the reduction of pinning forces by the growth of particlesduring heterogeneous precipitation, versus the drivingforces for recrystallisation. The start and end of the plateauare approximately identified with the start and end of pre-cipitation. It is conventionally accepted to estimate a 5%recrystallised fraction for the start and a 95% fraction forthe end.

It is known that once the plateau has ended, and recrys-tallisation is again progressing, the pinning forces exertedby the precipitates are now lower than the driving forces forrecrystallisation.

However, while the start of the plateau seems to coincidewith good exactness with the start of strain-induced precipi-tation,18,19) recent studies have demonstrated that at the endof the plateau a small fraction of precipitates conserve thesame size as the precipitates formed at the start of precipi-tation. Therefore, this suggests that the end of the plateau,or the Pf curve, coincides approximately with the end of the

precipitation.20)

On the other hand, once the plateau has started, the ex-perimental points that define it are not exactly on it but arenormally below it. This is due to the fact that the double de-formation technique, and in particular back extrapolation, isa mechanical method and the start of precipitation increasesthe flow stress of the steel, which is translated into a pseudo-reduction in the recrystallised fraction. Obviously, as theholding time after deformation increases, so the recrys-tallised fraction cannot decrease.

3.2. RPTT Diagrams

Recrystallised fraction versus time curves were used toplot recrystallisation-precipitation-time-temperature (RPTT)diagrams. The points defining the start and end of theplateau were taken to plot curves for the start (Ps) and end(Pf) of precipitation, respectively. On the other hand, thelines corresponding to different recrystallised fractions weredetermined by regression of the points resulting from theintersection of horizontal lines corresponding to differentXa values with the recrystallised fraction curves. In thisway, RPTT diagrams were obtained for the two strains of0.20 and 0.35, and are shown in Figs. 4 and 5, respectively.The recrystallised fraction does not change between Ps andPf, and for times in excess of Pf the slope of the straight linefor any Xa value is lower, which means that recrystallisationis more difficult after precipitation. Physically this meansthat the activation energy for the recrystallisation of austen-ite is higher when precipitation has taken place.21)

The RPTT diagrams provide ample information on the

ISIJ International, Vol. 48 (2008), No. 11

1604© 2008 ISIJ

Fig. 1. Schedule of double deformation.Fig. 2. Recrystallised fraction against time. Strain�0.20.

Fig. 3. Recrystallised fraction against time. Strain�0.35.

Table 2. Solubility temperatures (Ts) and maximum precipi-tated volume fraction (Xp) at 950°C.

recrystallisation–precipitation interaction, apart from show-ing the precipitation kinetics. It is seen that the nose of thePs curve, or the minimum incubation time of strain-inducedprecipitation, interacts approximately with the 50% recrys-tallised fraction line. In other words, the optimum mi-crostructure for precipitation to take place is that corre-sponding to the 50% recrystallised fraction, where satura-tion and diffusion reach a maximum. When Xa is equal to orless than 30%, precipitation is very difficult. On the otherhand, the temperature of the nose of the Ps curve corre-sponded to a temperature close to 950°C for a strain of 0.20and somewhat lower, approximately 940°C, for a strain of0.35.

At the moment when precipitation starts, whatever thetemperature (Ps curve), it is assumed that the precipitatedfraction corresponds to a value of 5%. In the same way,when the Pf curve is reached, the precipitated volume isclose to 95%. Once the Pf curve has been reached, recrys-tallisation starts to progress again due to fact that the pin-ning forces exerted by the precipitates are lower than thedriving forces for recrystallisation.

The RPTT diagrams, and especially the Ps and Pf curves,define a time interval, whatever the temperature, duringwhich the precipitation state (size and precipitated volume)is changing. For times after Pf, the precipitated fractiondoes not vary but a coarsening of the precipitates occursdue to the effect of Ostwald ripening.20)

Finally, Figs. 4 and 5 show a series of arrows that indi-cate the holding times at 950°C, after deformation, in new

tests performed with the aim of studying the influence ofprecipitation on intragranular nucleation, as will be seen inthe following section.

According to the RPTT diagrams, the deformation con-ditions were selected to coincide approximately with timesbefore the nose of the Ps curve, close to Ps and Pf, and afterPf, respectively. In this way it was possible to evaluate theeffect of precipitation kinetics on the intragranular nucle-ation of ferrite.

The aim of these tests is to verify whether two austeniticmicrostructures subjected to the same deformation at a cer-tain temperature, with holding times after deformation situ-ated between the Ps and Pf curves, equivalent to a plateau,and cooled at the same cooling rate, would yield the sameor a different ferrite grain size. It should be noted that twoor more holding times corresponding to the Pf–Ps intervalwould have the same microstructure, i.e. the same recrys-tallised fraction, which means the same dislocation densityand the same average austenite grain size, and the only dif-ference would be the precipitated volume. In other words,both microstructures will be differentiated by the precipi-tated volume, conventionally from 5% corresponding to Ps,which corresponds to the start of the plateau, and 95% cor-responding to Pf.

The deformation conditions are shown in Table 3. Afterthe holding time at the deformation temperature, the speci-mens were cooled at a rate (dT/dt)800–500°C of 3.5 K/s. Table3 also reports the recrystallised fraction and the precipitatedvolume, calculated in accordance with the RPTT diagrams.

Metallographic preparation of the specimens allowed themeasurement of the ferrite grain size, as is shown in the fer-ritic microstructures obtained in the conditions noted belowthe figure (Figs. 6(a)–6(g)). Ferrite grain sizes were meas-ured applying the linear intersection method and the valuesobtained were graphically represented as a function of theaustenite holding time after the deformation temperature(Fig. 7).

An increase in the holding time (D t) after deformationreduces the ferrite grain size, but is insufficient to suggestthat grain refinement has been due exclusively to VN pre-cipitates acting as nucleation sites for ferrite during cool-ing. In this sense, Fig. 6 does not adequately clarify thecontribution of intragranular nucleation of ferrite on precip-itates, and thus it would be necessary to set out the resultsin another way.

As has been noted at the start of this work, when austen-ite is deformed at a given temperature and held for a certaintime at that temperature, the resulting microstructure canrecrystallise either partially or completely. A hardening ofaustenite, originated by incomplete recrystallisation, in-

ISIJ International, Vol. 48 (2008), No. 11

1605 © 2008 ISIJ

Fig. 5. RPTT diagram. Strain�0.35.

Fig. 4. RPTT diagram. Strain�0.20.

Table 3. Testing conditions: Deformation temperature (Td),strain (e), recrystallisation/precipitation time (D t), re-crystallised fraction (Xa), precipitated fraction (Xp)and ferrite grain size (Da).

creases the density of dislocations and consequently thenumber of nucleation sites. Therefore, the contribution ofintragranular nucleation on precipitates to ferrite grain re-finement must be evaluated in conditions where both theaustenite grain size and the dislocation density remain con-stant. In this sense, it is supposed that two austenitic mi-crostructures would have the same dislocation density if therecrystallised fraction is the same.

The recrystallised fractions and precipitated fractionsmay be deduced from the RPTT diagrams (Figs. 4 and 5),making the opportune interpolations. Thus, for a strain of0.20 the values of Xa (Table 3) corresponding to holding

times (D t) of 50 and 700 s were 53%. The precipitated frac-tions for these times were 5%, coinciding with Ps, and90%, slightly before Pf (Fig. 4), and thus close to the startand the end of precipitation. Between these two conditions,both the average austenite grain size and the dislocationdensity are approximately the same and only the precipi-tated fraction varies. The latter has been calculated bearingin mind that the precipitation kinetics between Ps and Pf

obey Avrami’s law.10,22) Nevertheless, a linear approxima-tion between Ps and Pf would yield very similar results.

These results are shown in Fig. 8, where the ferrite grainsize has now been represented as a function of the recrys-

ISIJ International, Vol. 48 (2008), No. 11

1606© 2008 ISIJ

Fig. 6. Ferritic microstructures. Td�940°C; cool. rate�3.5 K/s; e�0.20: (a) D t�10 s, (b) D t�50 s, (c) D t�700 s;e�0.35: (d) D t�10 s, (e) D t�20 s, (f) D t�100 s; (g) D t�500 s.

Fig. 7. Ferrite grain size against holding time after deformation. Fig. 8. Ferrite grain size against austenite recrystallised fraction.

tallised fraction, related with the dislocation density, notingon each point the corresponding precipitated fraction. Thecontribution of ferrite intragranular nucleation would bemeasured on the approximately vertical segment of eachrepresentation corresponding to the strains of 0.20 and0.35, respectively, if necessary making extrapolations to the0% and 100% precipitated fraction, respectively.

The reduction in the ferrite grain size (Da) obtained be-tween the two aforementioned conditions (holding times of50 s and 700 s) was 3.5 mm, which extrapolated to a 95%precipitated fraction would mean approximately 4 mm, whichcompared to the grain size corresponding to a 5% precipi-tated fraction (Da�17 mm) implies a 24% reduction. In thesame way, for a strain of 0.35 and holding times (D t) of 50and 700 s, the Xa values were 55%. In these conditions ofthe same average austenite grain size and the same disloca-tion density, the precipitated volume varies between 5%and 80%. Between these conditions the reduction obtainedfor Da was 2 mm, which extrapolated to a 95% precipitatedfractions would mean 2.5 mm, representing a reduction ofapproximately 16% from the size corresponding to a 5%precipitated fraction (Da�15.5 mm).

Comparing the above results it is seen that the recrys-tallised fraction between Ps and Pf is approximately 53%and 55% for strains of 0.20 and 0.35, respectively. Thismeans that both the recrystallised and the non-recrys-tallised fractions are similar, and in particular that theaustenite hardening reached in the latter case on the non-re-crystallised fraction is obviously greater. On the other hand,the recrystallised grain size decreases as the strain in-creases. Thus it may be concluded that in the case of thestrain of 0.35, the dislocation density was greater and theaverage austenite grain size was smaller than with the 0.20strain. Both aspects favour the nucleation of ferrite, both ataustenite grain boundaries and on subgrains formed by thestacking or cluster of dislocations.23) Therefore, it was ex-pectable that the contribution of intragranular nucleation toferrite grain refinement would be less in the case of the 0.35strain than in the case of the 0.20 strain, as has indeed beenthe case.

Finally, it is not possible to measure the percentage con-tribution of each of the nucleation sites separately, since asthe recrystallised fraction increases so the average grainsize decreases, favouring nucleation at the grain bound-aries, and at the same time the dislocation density de-creases, reducing the capacity of dislocations to serve asnucleation sites for ferrite. This is observed very well whenthe holding time after the 0.20 strain increases from 10 to50 s (Fig. 7), causing an increase in the recrystallised frac-tion from 20 to 53% (Fig. 8) and an undetermined reduc-tion in the dislocation density.

Finally, in other work reported by the authors13,24) it wasshown that in V-microalloyed steels with a greater precipi-tated volume fraction, which is obviously achieved withhigher V and N contents, the percentage reduction in theferrite grain size due to intragranular nucleation on VN pre-cipitates rises notably, and in a steel with C�0.100; V�0.151 and N�0.0227 (%) has achieved a 45% reduction. Inthis case24) the total precipitated fraction at the deformationtemperature was 1.1�10�3, more than one order of magni-tude higher than in the present case (Table 2).

3.3. Precipitate Analysis

The precipitation states were characterised by studyingthe nature, size and distribution of vanadium nitrides–car-bonitrides formed in hot-deformed specimens. Transmis-sion Electron Microscopy (TEM) techniques on carbon extraction replicas have been used. The particle size distri-bution of second phase particles and determination of theaverage particle diameter have been obtained by measuringan average number of 200 particles on each specimen. Fig-ure 9(a) shows a TEM micrograph in which the VN precip-itates can be seen. The electron energy dispersive X-rayspectrum (Fig. 9(b)) shows the presence of V and N and thelattice parameter determined revealed a f.c.c. cubic latticewith a value of 0.414 nm (Fig. 9(c)), which corresponds toV(C, N) precipitates.

Figures 10(a) and 10(b) show the particle size distribu-tions obtained for the steel used. The specimens were de-formed (e�0.35) at 950°C and held for times of 20 s and100 s, which coincide with the nose of the Ps curve and be-fore Pf, respectively, in the RPTT diagram. It can be seenthat the average particle size was 6.4 nm at the start of pre-cipitation and that after 100 s (end of plateau) the averageparticle size had reached 9.5 nm, which was still small. Theaverage precipitate size is relatively small and a larger size

ISIJ International, Vol. 48 (2008), No. 11

1607 © 2008 ISIJ

Fig. 9. TEM images of steel used. (a) Image showing precipi-tates for specimen tested at reh. temp. of 1 200°C; deformationtemp. of 950°C, e�0.35; D t�20 s; (b) EDAX spectrum of precipi-tate. (c) Electron diffraction pattern.

could be expected to improve intragranular nucleation.With the aim of revealing possible intragranular ferrite

nucleation on V-precipitates, a further microscopy studywas carried out using a scanning electron microscopeequipped with a field emission gun (SEM-FEG). Figure 11shows a small ferrite grain, well below the average size, in-side which a relatively large VN precipitate and othersmaller precipitates can be seen; it being highly possiblethat the biggest precipitate, or indeed any of the smallerones, may have served as a source for the nucleation of thisgrain. The observation of this and other microscopic fieldsconfirms the presence of many small ferrite grains on thisspecimen.

4. Conclusions

(1) The applied methodology has allowed a quantitativedetermination of the contribution of ferrite intragranularnucleation on V(C, N) precipitates.

(2) The contribution of ferrite intragranular nucleationon V(C, N) particles is quantitatively smaller than the con-tribution of nucleation at grain boundaries and on disloca-tion stacking (subgrains). Nevertheless, in the present studyit has been seen that the ferrite grain size was notably re-duced by the intragranular nucleation of ferrite on precipi-tates.

(3) Intragranular nucleation led to a reduction of ap-proximately 25% in the ferrite grain size with a strain of0.20 and 16% with a strain of 0.35.

(4) In accordance with the above, the contribution ofintragranular nucleation decreases as the applied strain in-creases. An increase in the strain favours nucleation of fer-rite at austenite recrystallised grain boundaries, as theaustenite grain is finer, and on subgrains, in preference tointragranular nucleation on VN-precipitates.

Acknowledgements

The authors are grateful to the European Union for fund-ing through the RFCS Programme.

REFERENCES1) I. Madariaga and I. Gutiérrez: Acta Mater., 47 (1999), 951.2) J. H. Shim, Y. J. Oh, J. Y. Such, Y. W. Cho, J. D. Shim, J. S. Byun and

D. N. Lee: Acta Mater., 49 (2001), 2115.3) J. H. Shim, Y. W. Cho, S. H. Cheng, J. D. Shim and D. N. Lee: Acta

Mater., 47 (1999), 2751.4) J. S. Byun, J. H. Shim, Y. W. Cho and D. N. Lee: Acta Mater., 51

(2003), 1593.5) K. F. Al Hajari, C. I. Garcia, M. Hua and A. DeArdo: ISIJ Int., 46

(2006), 1233.6) T. Furuhara, T. Shinyoshi, G. Miyamoto, J. Yamaguchi, N. Sugita, N.

Kimura, N. Takemura and T. Maki: ISIJ Int., 43 (2003), 2028.7) G. Miyamoto, T. Shinyoshi, J. Yamaguchi, T. Furuhara, T. Maki and

U. Uemori: Scr. Mater., 48 (2003), 371.8) T. Yokomizo, M. Enomoto, O. Umezawa, G. Spanos and R. O.

Rosenberg: Mater. Sci. Eng. A, A344 (2003), 261.9) T. Furuhara, J. Yamaguchi, N. Sugita, G. Miyamoto and T. Maki:

ISIJ Int., 43 (2003), 1630.10) S. F. Medina and A. Quispe: Mater. Sci. Forum, 426–432 (2003),

1139.11) A. Quispe, S. F. Medina and P. Valles: ISIJ Int., 37 (1997), 783.12) A. Quispe, S. F. Medina, J. M. Cabrera and J. M. Prado: Mater. Sci.

Tech.-London, 15 (1999), 635.13) S. F. Medina, M. Gómez, J. I. Chaves, P. P. Gómez and P. Adeva:

Mater. Sci. Forum, 500–501 (2005), 371.14) H. L. Andrade, M. G. Akben and J. J. Jonas: Metall. Trans. A, 14A

(1993), 1967.15) J. S. Pertula and L. P. Karjalainem: Mater. Sci. Technol., 14 (1998),

626.16) E. T. Turkdogan: Iron Steelmaker, 16 (1989), 61.17) P. A. Manohar, D. P. Dunne, T. Chandra and C. R. Killmore: ISIJ

Int., 36 (1996), 194.18) A. Laarasoui and J. J. Jonas: Metall. Trans. A 22A (1991), 151.19) P. D. Hodgson and R. K. Gibbs: ISIJ Int., 32 (1992), 1329.20) S. F. Medina, A. Quispe, P. Valles and J. L. Baños: ISIJ Int., 39

(1999), 913.21) S. F. Medina, A. Quispe and M. Gómez: Mater. Sci. Technol., 19

(2003), 99.22) S. F. Medina, A. Quispe and M. Gómez: Steel Res. Int., 76 (2005),

527.23) J. Liu and J. Yanagimoto: ISIJ Int., 47 (2007), 1188.24) S. F. Medina, M. Gómez and L. Rancel: Scr. Mater., 58 (2008), 1110.

ISIJ International, Vol. 48 (2008), No. 11

1608© 2008 ISIJ

Fig. 10. Relative frequency of precipitate size for steel used de-formed at 950°C; e�0.35. (a) D t�20 s; (b) D t�100 s.

Fig. 11. SEM-FEG image at different magnifications showingferrite small grains and VN-precipitates. A large precipitate is indi-cated by the arrow. reh. temp.�1 200°C�10 min; def. temp.�950°C; e�0.35; holding time�500 s.