Effect of Inclusion Size and Type on the Nucleation of Acicular Ferrite in High Strength Ship Plate Steel

Abstract

The mechanisms of acicular ferrite formation on non-metallic inclusion have been studied in high strength ship plate steel. The effect of inclusion size and type on the formation of acicular ferrite was studied by metallographic microscope and SEM-EDS. The inclusion composition was analyzed by Raman spectra. Experimental results showed that all the inclusions were nearly spherical in shape. Inclusion size was mainly from 1 µm to 4 µm in high strength ship plate steel. With increase of the inclusion size, the acicular ferrite nucleating rate increased. The probability of inducing nucleation increased gradually with the inclusion size increasing. The inclusions had an appropriate size to become nucleation center for acicular ferrite. This optimal size was about 3 µm. The inclusion was composed by Ti, Mn, S, Si, O, Al and Mg. The composition of Ti–O–Mn inclusion was Ti₂O₃ and MnS, but no MnTiO₃. The surface of inclusion as an inert surface played an important role in nucleating of acicular ferrite. The lower low mismatch was also the important factors to promote the nucleation of acicular ferrite. Based on the two mechanisms, it would make it easier for acicular ferrite to nucleate.

1. Introduction

The microstructure of acicular ferrite in steel has been extensively studied, because it provided optimum mechanical properties, both in terms of strength and toughness. The formation conditions of acicular ferrite were created in order to obtain a lot of acicular ferrites, so as to meet the requirements of high strength and toughness in the high strength ship plate steel.^1,2) Therefore, the effect factors of acicular ferrite formation on non-metallic inclusions need further research to reveal the transformation mechanisms.

For now, many experiments have been done on the formation conditions of acicular ferrite in steels and lots of early workers have reported that inclusion often played an essential role in the nucleation of acicular ferrite.^3,4,5) Li X C et al.⁶⁾ researched the role of inclusions for acicular ferrite formation in micro-alloy steels and found that the complex inclusion V (C, N) with MnS in the core could induce nucleation of acicular ferrite. Li P et al.⁷⁾ found that MgO–Al₂O₃–TiO_x–SiO₂–MnS complex inclusions which size was about 1.25 μm could induce nucleation of acicular ferrite through systemic observation by SEM.

However, there are many unresolved issues about the major role of non-metallic inclusions for acicular ferrite formation. Especially, the size and type of inclusions for the formation of acicular ferrite is still far from being understood. And how can inclusions promote acicular ferrite nucleation? This problem is not resolved yet. A lot of studies about the mechanisms of inclusions for acicular ferrite formation were undertaken by early workers. Cheng J H et al.⁸⁾ recently noted that Mn-depletion around the Ti–Al–O–S–Mn complex inclusion was attributable to the formation of acicular ferrite. They considered that Mn-depleted zone could improve the driving force of acicular ferrite formation, thus promoting the nucleation of acicular ferrite. However, Yu S F et al.⁹⁾ thought that acicular ferrite could be nucleated from CuS because of the low mismatch between CuS and acicular ferrite.

From this brief review of the literature, it is clear that no consensus has been reached on the mechanisms of inclusions for the formation of acicular ferrite. Therefore, this study was intended to explore the influence of non-metallic inclusions on the formation of acicular ferrite by metallographic microscope, scanning electron microscopy and Raman spectra. Meanwhile, the possible nucleation mechanisms of acicular ferrite were further discussed.

2. Experiment

The high-strength ship plate steel was selected to be the reference material in this investigation. The samples were prepared from a commercial, continuous cast slab. Oxygen potential was left in the steel, when the molten steel was deoxidized with aluminium. And then the element of titanium was added to the molten steel for the sake of the formation of titanium oxide. The main chemical compositions of sample was shown in Table 1. The steps of experiment were as following:

Table 1. Main chemical compositions of sample (mass fraction/%).

C	Si	Mn	S	P	Mo	Ti	Nb	Als	Mg	Fe
0.06–0.09	0.20–0.30	0.90–1.50	0.005	0.015	0.10	0.02–0.04	0.04	0.010–0.040	Trace	Bal

(1) The specimens were polished as metallographic samples with an automatic grinder-polisher. The distribution of the inclusion size was observed by metallographic microscope. Inclusion sizes were measured from image analysis software;

(2) The specimens were mechanically polished and were etched in 4% nital. These specimens were observed with higher magnification by metallographic microscope. The size of inclusion for acicular ferrite nucleation was measured by Image J;

(3) To study the morphology and type of inclusions, the details of analysis were conducted in SEM-EDS;

(4) The details of inclusion composition were analyzed by Raman spectra.

3. Effect of Inclusion Size on the Nucleation of Acicular Ferrite

It is well known that the inclusion size has an important effect on acicular ferrite nucleation. In recent years, there are many papers studying on the inclusion size on the effectiveness of inducing acicular ferrite nucleation. For instance, Lee T. K. et al.¹⁰⁾ found that the inclusion which size was 0.25–0.8 μm could induce nucleation of acicular ferrite easily. They believed that when the inclusion size was less than 1.1 μm, the probability of nucleation increased markedly with increasing the inclusion size. But with increasing the size over 1.1 μm, the probability was basically constant. Barbaro et al.¹¹⁾ believed that when the inclusion size was between 0.4 μm and 0.6 μm, the probability of nucleation was very strong. Piao Y et al.¹²⁾ thought that the inclusions which sizes were between 0.1 μm and 0.6 μm were most effective in nucleating the ferrite. And preferable size for the formation of acicular ferrite is also affected by the cooling rate.¹³⁾ The sample was cooled down to ambient temperature at a rate of 7°C/s after heating and heat preservation.

The metallographic microscope was employed to study the distribution of inclusion size. And then the inclusion size of twenty images was analyzed by Image J software. The statistics showed that the average size of the total inclusions was 3.4 μm. The distribution of inclusion size was shown in Fig. 1. As can be seen from Fig. 1, the number of inclusions which the size was between 1 μm and 2 μm was the largest and the quantity percentage of the inclusions was 24.81%. The quantity percentage of the inclusions of a 2–3 μm diameter changed into 20.3%. The quantity percentage of the inclusions with the size of <1 μm, 3–4 μm, 4–5 μm, 5–6 μm and >6 μm was 9.02%, 15.79%, 7.52%, 14.29%, 8.27%, respectively. It is clear that with increase of the inclusion size, the number of inclusions increased firstly and then decreased. Therefore, the size of the inclusions in high strength ship plate steel was very small, mainly concentrated in 1–4 μm (61% of the total inclusions).

Fig. 1.

The distribution of inclusion size.

A total of 127 inclusions which was connected with acicular ferrite were observed using metallographic microscope. According to the morphological difference of acicular ferrite, these inclusions were classified into four types as shown in Fig. 2. Type 1 (Fig. 2(a)) were the inclusions located within the acicular ferrite and type 2 (Fig. 2(b)) were the inclusions on the acicular ferrite boundary. Inclusions classified as type 3 (Fig. 2(c)) were connected with two acicular ferrites. And these two acicular ferrites grow at different angles with inclusion. The last type of inclusions classified as type 4 (Fig. 2(d)) were surrounded by acicular ferrites in all directions, therefore, almost the entire surface of inclusions was covered with ferrite.

Fig. 2.

4 types of inclusions for the formation of ferrite. (Online version in color.)

The nucleation site of acicular ferrite was classified into three types: the boundary of austenite, the inclusions and the surface of primary ferrite. The ferrite on the austenite grain boundaries might meet inclusion in the growth progress, and the inclusion might locate everywhere of ferrite. However, the ferrite could not grow along different directions on inclusion. The sympathetic ferrite nucleated at the side of the primary ferrite. The sympathetic ferrite might meet inclusion in the growth progress, but there were not many different angles between ferrite and inclusion. Only the acicular ferrite formation on inclusion might grow along different directions on inclusion (as shown in Figs. 2(c) and 2(d)). The acicular ferrites as shown in Fig. 2(a) were considered as two ferrites formation on inclusion, and the orientation of two ferrites was the same.¹⁴⁾ It can be believed that the inclusions classified as type 2 (as shown in Fig. 2(b)) could not cause nucleation of acicular ferrite because of the absence of acicular ferrite growing direction.

The distribution of the inclusion size of each type was shown in Fig. 3. Among these inclusions, 33.07% was classified to type 1. The quantity percentage of type 2, 3 and 4 were 18.9%, 22.05% and 25.98%, respectively.

Fig. 3.

Proportions of inclusions of each type.

These inclusions for acicular ferrite formation were observed using metallographic microscope and the sizes of it were measured by software. The results reflected that the mean size of the total was 4.91 μm. The average size of each type was 4.80 μm, 4.98 μm, 4.92 μm and 5.00 μm, respectively. Thus the average size of each type was basically the same, having a slight upward tendency. Figure 4 showed the size distribution of each type. As shown in this figure, each type of inclusions changed greatly in size. For type 1, 2 and 3, the quantity percentage of inclusions increased firstly and then decreased with the inclusion size increasing, while the quantity percentage of inclusions increased significantly when the size of inclusions was larger than 6 μm. The inclusion size with the largest number ranged from 3 μm to 4 μm. However, for type 4 with the inclusion size increasing, the quantity percentage of inclusions enhanced gradually. When the inclusion size was larger than 6 μm, the quantity percentage of inclusion markedly decreased. The inclusion size was mainly concentrated in 3–6 μm.

Fig. 4.

The distribution of inclusion size of various types. (Online version in color.)

The probability of inclusion for nucleating acicular ferrite at a given size, d_i, could be expressed by the following equation:¹⁵⁾   

$$\begin{array}{l}{N}_{d_i}=\\ \frac{(\mathrm{t}\mathrm{h}\mathrm{e}\mathrm{\hspace{0.17em} }\mathrm{n}\mathrm{u}\mathrm{m}\mathrm{b}\mathrm{e}\mathrm{r}\mathrm{\hspace{0.17em} }\mathrm{o}\mathrm{f}\mathrm{\hspace{0.17em} }\mathrm{t}\mathrm{y}\mathrm{p}\mathrm{e}\mathrm{\hspace{0.17em} }1\mathrm{\hspace{0.17em} }\mathrm{i}\mathrm{n}\mathrm{c}\mathrm{l}\mathrm{u}\mathrm{s}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{s}{)}_{d_i}+(\mathrm{t}\mathrm{h}\mathrm{e}\mathrm{\hspace{0.17em} }\mathrm{n}\mathrm{u}\mathrm{m}\mathrm{b}\mathrm{e}\mathrm{r}\mathrm{\hspace{0.17em} }\mathrm{o}\mathrm{f}\mathrm{\hspace{0.17em} }\mathrm{t}\mathrm{y}\mathrm{p}\mathrm{e}\mathrm{\hspace{0.17em} }3\mathrm{\hspace{0.17em} }\mathrm{i}\mathrm{n}\mathrm{c}\mathrm{l}\mathrm{u}\mathrm{s}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{s}{)}_{d_i}+(\mathrm{t}\mathrm{h}\mathrm{e}\mathrm{\hspace{0.17em} }\mathrm{n}\mathrm{u}\mathrm{m}\mathrm{b}\mathrm{e}\mathrm{r}\mathrm{\hspace{0.17em} }\mathrm{o}\mathrm{f}\mathrm{\hspace{0.17em} }\mathrm{t}\mathrm{y}\mathrm{p}\mathrm{e}\mathrm{\hspace{0.17em} }4\mathrm{\hspace{0.17em} }\mathrm{i}\mathrm{n}\mathrm{c}\mathrm{l}\mathrm{u}\mathrm{s}\mathrm{i}\mathrm{o}\mathrm{n}\mathrm{s}{)}_{d_i}}{\mathrm{T}\mathrm{h}\mathrm{e}\mathrm{\hspace{0.17em} }\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}\mathrm{\hspace{0.17em} }\mathrm{n}\mathrm{u}\mathrm{m}\mathrm{b}\mathrm{e}\mathrm{r}\mathrm{\hspace{0.17em} }\mathrm{o}\mathrm{f}\mathrm{\hspace{0.17em} }\mathrm{i}\mathrm{n}\mathrm{c}\mathrm{l}\mathrm{u}\mathrm{s}\mathrm{i}\mathrm{o}\mathrm{n}{\mathrm{s}}_{d_i}}\end{array}$$

Where N_di is the probability of nucleation at d_i, and d_i is the size of inclusions.

The probability of 127 inclusions was calculated through the above formula. The typical result was shown in Fig. 5. It is clear that the inclusion size affected acicular ferrite nucleation in a certain way, showing a transition from non-nucleant inclusion to nucleant inclusion with the inclusion size increasing. Overall, the probability of inclusions for nucleating acicular ferrite increased gradually with the inclusion size increasing. When the inclusion size was less than 1 μm, the probability of inclusions for nucleating acicular ferrite was almost zero. When the inclusion size was between 1 μm and 2 μm, the probability of inclusions for nucleating ferrite enhanced and reached 3.94%. The probability of nucleation inclusions with the size of 2–3 μm, 3–4 μm, 4–5 μm, 5–6 μm and >6 μm was 7.87%, 19.69%, 14.96%, 13.39%, 20.47%, respectively. It is evident that when the inclusions size was larger than 1 μm, the probability of inclusions for nucleating acicular ferrite increased significantly. The probability of nucleation inclusions with the size over 2 μm was very strong. However, the mechanical properties and process performance of steels will be endangered, when the inclusion size is too large. So, the inclusions with the size of about 3 μm were the most advantageous to the formation of acicular ferrite. This is consistent with the work of Du S L et al.¹⁶⁾

Fig. 5.

Effect of inclusion size on the probability of inclusions for formation of acicular ferrite. (a): Micrograph of inclusion; (b): Si; (c): Mg; (d): S; (e): Mn; (f): Ti; (g): Al; (h): O.

4. Impact of Inclusion Types on the Nucleation of Acicular Ferrite

4.1. Composition and Morphology of Inclusions

The morphology and composition of inclusion for nucleation of acicular ferrite were analyzed by SEM and EDS. The result of typical inclusions was shown in Fig. 6. As can be seen from Fig. 6, the inclusion was composed by Ti, Mn, S, Si, O, Al and Mg. Obviously, the morphology of the inclusions was nearly spherical and the edge was smooth. From the element mapping figures, it can be seen that MnS attached to the surface of Titanium oxide.

Fig. 6.

The morphology and distribution of elements for the typical inclusion. (a): Micrograph of inclusion; (b): Si; (c): Mg; (d): S; (e): Mn; (f): Ti; (g): Al; (h): O. (Online version in color.)

The inclusion could be penetrated by electron beams of SEM, because the penetrating depth of the electron beams of SEM was about 10 μm and the size of inclusion for acicular ferrite formation was 2–5 μm. Not only did the EDX surface scanning reflect the surface details of inclusion, but the EDX surface scanning could reflect the internal information. The lighter region of spectra should be a higher element contents part. The darker spectra might be this element located deep in the sample, which would result in an additional absorption of X-ray in the exit progress, so the X-ray signal didn’t work properly (as shown in Fig. 7).

Fig. 7.

Influence of X-ray intensity.

As can be seen from Fig. 6(g), the signal of aluminum was not very good, but this was not to suggest that aluminum content was very low. Al might be situated at a greater depth of inclusion, so the signal was low. Therefore, it can be inferred that Al was in the middle of inclusion. Conversely, the signal of Mn was strong (as shown in Fig. 6(e)), so it can be inferred that Mn attached to the surface of inclusion.

4.2. The Phase Composition of Inclusions

Titanium oxide is mainly TiO, TiO₂, Ti₂O₃ and Ti₃O₅, but which kind will be formed is not known. Therefore, in order to further study the characteristics of inclusions, the Raman Spectra were applied to reveal the phase composition of Ti–O–Mn complex inclusion for nucleation of acicular ferrite.

Raman spectra of MnS and MnTiO₃ were illustrated in no. 17 literature. Raman spectrum of titanium oxide was illustrated in no. 18 literature.

Sample preparation of Raman spectra was very simple. The sample was polished carefully and etched in 4% nital solution. And then the sample could be performed by Raman spectra under normal temperature. The result was shown in Fig. 8. Compared to the standard pattern, it can be found that the phase composition of Ti–O–Mn complex inclusion for nucleation of acicular ferrite included Ti₂O₃, but no MnTiO₃. It is showed that titanium oxide was Ti₂O₃.

Fig. 8.

Constituent phases of Ti–O–Mn inclusions.

5. Research of the Mechanism of Acicular Ferrite Nucleation

There is no consensus on the mechanisms of acicular ferrite formation on non-metallic inclusions currently. But the following four mechanisms were to achieve most of the scholars’ recognition: 1) the manganese-depleted zone mechanism;⁸⁾ 2) the inert interface mechanism;⁹⁾ 3) thermal strain energy mechanism;¹²⁾ 4) low mismatch mechanism.¹⁹⁾

So far, it is widely accepted that inclusion size often played an essential role in the nucleation of acicular ferrite. Furthermore, they believed that the probability of inclusions for nucleating acicular ferrite increased gradually with the inclusion size increasing in the suitable size range. The reason was that non-metallic inclusions acted as an inert interface, which could reduce the nucleation barrier. At the same time, this surface could provide suitable nucleation sites and nucleation driving force for acicular ferrite. This mechanism believed that with the increase of the inclusion size, the nucleation energy barrier of inclusions decreased rapidly. It is showed that the larger inclusion was more conducive to the acicular ferrite nucleation within the suitable size range. This was consistent with the test result. The analysis about the probability of inclusion for nucleating acicular ferrite showed that the probability of inducing acicular ferrite indeed increased gradually with the inclusion size increasing. In addition, the inclusions with the size of about 3 μm were the most advantageous to the formation of acicular ferrite (as shown in Fig. 5). Therefore, the inclusion as an inert surface could provide nucleation sites and nucleation driving forcefor acicular ferrite. This would be attributable to the nucleation of acicular ferrite.

What’s more, the melting point of MnS was lower that MnS would precipitate on the high- melting-point oxides. At the same time, the element distribution of Ti–O–Mn inclusion was analyzed by mapping functions of EDS (as shown in Fig. 6), and the detail of Ti–O–Mn complex inclusion composition was analyzed by Raman spectra (as shown in Fig. 8). It could be inferred that the inclusion included Ti₂O₃ and MnS. So, it was easy for MnS to precipitate on the surface of Ti₂O₃ because of the high melting-point of Ti₂O₃. This would form Ti–O–Mn complex inclusion. In addition, the lattice mismatch between MnS and ferrite was only 8.8%. This low mismatch was attributable to the formation of acicular ferrite. Therefore, the low mismatch mechanism was one of the mechanisms of acicular ferrite formation.

In summary, there were two mechanisms of acicular ferrite formation in high strength ship plate steel. The inclusion as an inert surface could provide nucleation sites and nucleation driving force for acicular ferrite; the low mismatch between MnS and ferrite was effective in nucleating the acicular ferrite. Based on these mechanisms, it would make it easier for acicular ferrite to nucleate.

6. Conclusion

(1) The average size of each type was basically the same. The probability of inducing nucleation was different with the difference of inclusion size. The probability increased gradually with the inclusion size increasing. The inclusions with the size of about 3 μm were the most advantageous to the formation of acicular ferrite.

(2) The main inclusions was composed by Ti, Mn, S, Si, O, Al and Mg. The Ti–O–Mn complex inclusion composition was Ti₂O₃ and MnS, but no MnTiO₃. It is showed that titanium oxide was Ti₂O₃.

(3) The surface of inclusion as inert surface played an important role in nucleating of acicular ferrite. The lower low mismatch was also the important factors to promote the nucleation of acicular ferrite. Based on the twomechanisms, it would make it easier for acicular ferrite to nucleate.

Acknowledgements

The authors would like to acknowledge the National Natural Science Foundation of China (Grant No: 51474089) for the financial support.

References

• 1)   T.  Yamada,  H.  Terasaki and  Y. I.  Komizo: Trans. Mater. Res. Soc. Jpn., 34 (2009), 241.
• 2)   H.  Jin,  J.  Shim,  Y.  Cho and  H.  Lee: ISIJ Int., 43 (2003), 1111.
• 3)   X.  Li,  Y.  Min and  C.  Liu: J. Iron Steel Res. Int., 23 (2016), 415.
• 4)   C.  Zhang,  Z.  Xia,  Z.  Yang and  Z.  Liu: J. Iron Steel Res. Int., 17 (2010), 36.
• 5)   X.  Wan,  K.  Wu,  L.  Cheng and  R.  Wei: ISIJ Int., 55 (2015), 679.
• 6)   X.  Li,  X.  Wang and  X.  Feng: J. Wuhan Univ. Technol., 25 (2010), 228.
• 7)   P.  Li,  G.  Li and  W.  Zheng: J. Iron Steel Res. Int., 25 (2013), 49.
• 8)   J.  Cheng,  H.  Yu,  H.  Yu and  W.  Wang: Hot Work. Technol., 39 (2010), 76 (in Chinese).
• 9)   S.  Yu,  Y.  Zhang and  W.  Lu: Trans. China Weld. Inst., 23 (2002), 72.
• 10)   T.  Lee,  H.  Kim,  B.  Kang and  S.  Hwang: ISIJ Int., 40 (2000), 1260.
• 11)   F.  Barbaro,  P.  Krauklis and  K.  Easterling: Mater. Sci. Technol., 11 (1989), 1057.
• 12)   Y.  Bo,  B.  Hu,  F.  Yin and  X.  Chen: J. Univ. Sci. Technol. Beijing, 28 (2006), 357.
• 13)   A.  Takada,  H.  Terasaki and  Y.  Komizo: Sci. Technol. Weld. Join., 18 (2013), 91.
• 14)   K.  Wu,  Y.  Inagawa and  M.  Enomoto: Mater. Charact., 52 (2004), 121.
• 15)   L.  Xiong and  H.  Li: J. Iron Steel Res. Int., 22 (2010), 53.
• 16)   S.  Du,  Y.  Jin,  Z.  Gao,  Y.  Yang and  Y.  Bao: J. Univ. Sci. Technol. Beijing, 32 (2010), 574.
• 17)   J.  Qi: Ph.D. thesis, Wuhan University of Science and Technology, (2013), http://kns.cnki.net/KCMS/detail/detail.aspx, (accessed 2013-10-01).
• 18)   P.  Xiao,  S.  Zheng and  J.  You: Spectrosc. Spectr. Anal., 27 (2007), 936.
• 19)   A.  Gourgues,  H.  Flower and  T.  Lindley: Mater. Sci. Technol., 16 (2000), 26.