Effect of Silicon Addition on Carbide Transition in Tempered Martensite of Middle Carbon Steels

Abstract

Mechanical properties of tempered martensitic steel are controlled by precipitation conditions of carbides. In high strength medium carbon martensitic steel, carbide precipitation at low tempering temperature is strongly affected by silicon addition. Silicon addition prevent carbide transition from ε carbide to cementite, also prevent tempering softening and low temperature temper embrittlement. In this paper, we investigated the carbide transition in medium carbon steel with various Si content, using chemical analysis and XRD analysis of residue carbide after electrolytic extraction, DSC thermal analysis and FE-SEM observation. We found three step carbide transition from ε carbide to para-cementite and ortho-cementite. Especially, para-ortho transition of cementite is not only change in chemical composition of carbide, but also change in the cementite morphology and precipitation site from platelet shape inside of martensite block to granular shape at martensite block boundary. Silicon addition inhibited the formation of para-cementite, and shifted higher the transition temperature of ortho-cementite.

1. Introduction

Carbon steels for machine structural use are generally used as quenched and tempered martensite microstructure in order to improve mechanical properties.¹⁾ Steels transform to martensite microstructure with supersaturated carbon after quenching process, and some types of carbides precipitate after tempering process. Tempering process has been researched in detail, and it is known to have mainly three stages based on dilatometry in tempering process.^2,3,4,5,6,7) Figure 1 shows tempering process of carbon steel martensite. First stage is ε carbide^4,5,6,8,9) or η carbide^10,11,12,13) precipitation about 423 K. These carbides are reported to have good lattice fitting relationship with BCC-Fe.^8,12,13) Second stage is decomposition of retained austenite. Third stage is carbide transition to cementite (θ carbide) about 573 K. Cementite is orthorhombic structure, and some types of lattice fitting relationships with BCC-Fe, such as Bagaryatsky relationship, are reported.^7,14,15) In addition, χ carbide is also reported to precipitate before cementite,^{4,5,9,12,13,16)} and it is often mentioned to precipitate on twin interface.^4,12,13) After third stage, chemical composition of cementite shifts to equilibrium composition.¹⁷⁾ And some alloyed steel shows secondary hardening behavior around 773 K, related to precipitation of ally carbides such as V carbides and Mo carbides.^18,19)

Fig. 1. Tempering process of middle carbon martensite.

This carbide transition from ε carbide to cementite is obvious especially in medium carbon steel and high carbon steel, not only by microstructure observation, but by electrical resistivity⁵⁾ and thermal analysis.^6,20) Carbide transition is also known to be kinetic phenomenon based on carbide precipitation dependance on holding time^21,22) and heating rate.^23,24)

Silicon is known to retard the carbide transition by delaying cementite formation in tempering.^{21,22,24,25,26)} Miyamoto et al.²¹⁾ observed Si partitioning between ferrite and cementite using three-dimensional atom probe (3DAP), and clearly showed that Si is less contained in cementite. Nagao et al.²⁰⁾ investigated carbide transition using Differential Scanning Calorimetry (DSC) and showed peak of reaction heat in carbide transition shifts to higher temperature by Si addition to steel.

Retardation of cementite precipitation by Silicon is also investigated in various transformation process, such as bainite transformation,²⁷⁾ austemper process^28,29) and quench & partitioning (Q&P) process.³⁰⁾

Carbide transition affects on mechanical properties. Teramoto et al. analyzed carbide transition in tempered martensite of about 2% Si added medium carbon steel using Transmission Electron Microscopy (TEM).³¹⁾ They showed three step carbide transition, ε carbide → χ carbide → cementite, about 723 K. This χ carbide precipitation temperature was higher than former works using plane carbon steel,^16,17) then Si was thought to retard χ carbide precipitate. They also reported the effect of carbide transition on mechanical properties. Yield strength of steel was dropped at the tempering temperature of ε to χ carbide transition.

Suzuki et al. researched relationship between carbide transition and mechanical properties in medium carbon steels with various Si contents.³²⁾ The temperature range of low temperature temper embrittlement shifts to higher temperature by the increase of Si content in steel. This embrittlement corresponds to growth of intragranular carbide, such as ε carbide and cementite precipitate in martensite block. And the end of embrittlement corresponds to carbide transition to grain boundary carbide, such as cementite precipitate in martensite block boundary and former austenite grain boundary. Carbide transition from ε carbide to cementite is obvious in DSC analysis, but it was difficult to distinguish intragranular cementite from ε carbide in Field Emission Scanning Electron Microscopy (FE-SEM) observation.

In this work, we analyzed the carbide precipitated in each step of tempering of medium carbon steels with various Si contents. Carbides were separated from steel by electrolytic extraction and identified by both chemical analysis and X-Ray Diffractometry (XRD) analysis. These results were compared with SEM observation and DSC analysis in former works,³²⁾ and effect of Si addition on the mechanism of carbide transition was discussed.

2. Experimental Procedure

2.1. Material

Test steels are middle carbon steel (about 0.55 mass% C) with various content of silicon (0.31–1.99 mass% Si), the same as previous work.³²⁾ Chemical compositions of test steels are shown in Table 1. The steels were vacuum induction melted and casted into 150 kg ingots. The ingots were heated above 1473 K and hot forged into round bars of 40 mm in diameter. The round bars were homogenized by heat treatment of 1473 K for 3600 s. Cylinder test pieces of 16 mm in diameter and 220 mm in length were machined from the round bars. To avoid central segregation of ingots, center of the cylinder is the point of 1/4 in diameter of round bars. The cylinders were normalized by heat treatment of 1273 K for 900 s and air cooling. And the cylinders were quenched and tempered by heat treatment of 1273 K for 900 s, quenching into 337 K oil, tempering from 473 to 823 K for 1800 s and air cooling. Tempering temperatures were differed from steels as shown in Table 2.

Table 1. Chemical compositions of test steels.

(mass%)
	C	Si	Mn	Cr	P	S
0.3Si steel	0.55	0.31	0.70	0.70	0.001	0.001
0.8Si steel	0.54	0.81	0.70	0.70	0.001	0.001
1.5Si steel	0.56	1.50	0.70	0.70	0.002	0.001
2.0Si steel	0.56	1.99	0.71	0.70	0.001	0.001

Table 2. Tempering temperatures of test steels.

	473 K	523 K	573 K	623 K	673 K	723 K	773 K	823 K
0.3Si steel	○	○	○	○	○	○	–	–
0.8Si steel	○	○	○	○	○	○	–	–
1.5Si steel	–	○	○	○	○	○	○	–
2.0Si steel	–	–	○	○	○	○	○	○

○: investigated –: not investigated

2.2. Carbide Identification

Carbides in quenched and tempered steels were tested in various method.

Crystal structure and chemical composition of carbides were investigated using electrolytical extraction method. Block test pieces of 15 × 10 × 5 mm were cut from quenched and tempered cylinder steels. Electrolytic extraction was conducted by constant current electrolysis method, using electrolyte solution of methanol - 10% acetylacetone - 1% tetramethylammonium chloride. Solutions containing electrolyzed steel about 0.8 g were suction filtrated using 0.2 μm membrane filter. These methods were conducted two times for each quenched and tempered steels, one is for XRD analysis and the other is for chemical analysis.

XRD analysis of 2θ method was conducted for whole carbides on membrane filter, then diffraction plane was parallel to membrane filter. X-ray source was Cu-Kα 45kV-200mA, 2θ angle scan was from 30 to 80 degrees.

Chemical analysis was conducted using Inductively Coupled Plasma (ICP) emission spectroscopy, and mass fraction of Fe, Si, Mn, Cr were calculated based on the electrolyzed mass of steel specimens.

Carbide precipitation temperature was measured using DSC analysis. Disk shaped test pieces of 3 mm in diameter and 1 mm in thickness were cut from the quenched and tempered cylinder steels. DSC analysis was conducted in N₂ atmosphere and alumina powder was used as reference material. Thermal cycle of DSC was heading from R.T. to 823 K for 0.25 K/s and keeping 823 K for 900 s. This thermal cycle was conducted two times for each test piece sequentially, carbide precipitation reaction heat was measured in first cycle, and background without chemical reaction was measured in second cycle. Reaction heat of carbide precipitations were shown as difference between first and second.

Carbide precipitation morphology was observed using FE-SEM. Test piece were cut from the quenched and tempered cylinder steels, polished and etched electrolytically. Microstructure and carbides around former austenite grain boundary triple point was observed.

3. Results

3.1. XRD Analysis of Electrolytically Extracted Carbides

XRD analysis peaks of electrolytically extracted carbides were compared with JCPDS cards of three types iron carbides, PDF#035-0772 Fe₃C (Orthorhombic), PDF#020-0508 Fe₅C₂ (Monoclinic) and PDF#036-1249 Fe₂C (Hexagonal). Hereafter, each carbide is described as cementite, χ carbide and ε carbide. But χ carbide was not identified in this work. Table 3 shows lattice parameter, 2θ angle, intensity, and (hkl) index of these carbides within 2θ is from 35 to 55 degrees. Figure 2 shows schematic peaks of these carbides in 2θ method.

Table 3. Material data of identified carbides (only 35< 2θ <55).

PDF#035-0772
Fe3 C		Orthorhombic
Lattice: 5.091 × 6.7434 × 4.526 <90.0 × 90.0 × 90.0>
2θ		Intensity	h	k	l
37.63		43	1	2	1
37.74		41	2	1	0
39.80		22	0	0	2
40.63		22	2	0	1
42.88		57	2	1	1
43.74		67	1	0	2
44.57		56	2	2	0
44.99		100	0	3	1
45.86		53	1	1	2
48.59		32	1	3	1
49.12		43	2	2	1
51.81		19	1	2	2
54.40		15	2	3	0

PDF#020-0508
Fe5 C2		Monoclinic
Lattice: 11.56 × 4.56 × 5.03 <90.0 × 98.05 × 90.0>
2θ		Intensity	h	k	l
36.191		20	0	0	2
37.604		20	−2	0	2
39.856		50	0	2	0
41.384		50	2	0	2
43.916		100	0	2	1
44.6		100	−4	0	2
45.789		20	−5	1	1
47.569		50	6	0	0
50.674		70	3	1	2
51.91		10	4	0	2
53.211		10	−4	2	1

PDF#036-1249
Fe2 C		Hexagonal
Lattice: 2.754 × 2.754 × 4.349 <90.0 × 90.0 × 120.0>
2θ		Intensity	h	k	l
37.686		25	1	0	0
41.484		25	0	0	2
43.232		100	1	0	1

Fig. 2. Schematic peaks of identified carbides. (Online version in color.)

Carbide transition temperature in XRD was compared with DSC results.⁸⁾ In DSC, carbide precipitate at some tempering temperature was judged by peak disappearance of carbide formation reaction in DSC heating.

XRD results of each steel of various tempering temperatures were compared with DSC results of previous works.³²⁾ DSC peak correspond to heat of formation of carbide precipitation in continuous heating condition at 0.25 K/s of DSC measurement. If DSC peak disappeared at higher temperature tempering, carbides were judged as already precipitated at this tempering temperature.

Figure 3 shows XRD peaks and DSC peaks of 0.3Si steel after various tempering temperature.

Fig. 3. XRD peaks and DSC peaks of 0.3Si steels. (Online version in color.)

In 0.3Si steel, XRD of 473 K tempering showed peaks too small to identify. XRD of 523 K–723 K tempering showed peaks of cementite. Cementite peaks of 723 K tempering was almost identical with that of database as shown in Fig. 2, but peaks of 623 K and lower tempering temperatures were shifted to higher 2θ angles, indicating lower lattice parameter. These XRD peaks of 623 K and lower tempering temperatures also showed irregular order of peak intensity. In database, (031) peak about 2θ = 45 deg. should be the strongest, but maybe (102) peak about 2θ = 44 deg. was the strongest in these specimens. (002) peak about 2θ = 40 deg. was also stronger than database. These results indicate some structural change in cementite between precipitated at 623 K and at 723 K.

DSC analysis also showed that cementite started to precipitate at 523 K tempering. But structural change in cementite was not noticed in DSC.

Figure 4 shows XRD peaks and DSC peaks of 0.8Si steel after various tempering temperature.

Fig. 4. XRD peaks and DSC peaks of 0.8Si steels. (Online version in color.)

In 0.8Si steel, XRD of 523 K tempering showed peak at 2θ≒43° corresponding to ε carbide. XRD of 573 K tempering, this ε carbide peak disappeared but irregular cementite peaks were observed instead. XRD peak shape of cementite were continuously change with increase of tempering temperature, and only XRD of 723 K tempering was regular cementite peaks.

DSC peaks were characteristic in 0.8Si steel. DSC of 523 K tempering showed two peaks of carbide precipitation. First peak (about 620 K) disappeared at 573 K tempering, and second peak (about 700 K) disappeared at 673 K tempering. First peak corresponds to irregular cementite precipitation in XRD. Second peak is speculated to correspond regular cementite precipitation. XRD peak of 673 K tempering is still irregular, but it is regarded to the mixture of irregular and regular cementite. For example, main (031) peak about 2θ = 45 deg. intensified and other minor peaks were also observed at 673 K.

Figure 5 shows XRD peaks and DSC peaks of 1.5Si steel after various tempering temperature.

Fig. 5. XRD peaks and DSC peaks of 1.5Si steels. (Online version in color.)

In 1.5Si steel, XRD of 623 K and lower tempering showed peak about 2θ = 43 deg. corresponding to ε carbide. XRD of 673 K tempering, but ε carbide peak disappeared and very broadened peaks were observed. XRD of 723 K tempering showed broadened cementite peaks and XRD of 773 K tempering showed identical cementite peaks.

DSC of 623 K and lower tempering showed cementite precipitation peak about 730 K. DSC of 673 K showed smaller 730 K peak, indicating cementite precipitation was started but not finished at 673 K tempering. Based on DSC, broadened XRD peak of 673 K tempering is estimated to be cementite.

Figure 6 shows XRD peaks and DSC peaks of 2.0Si steel after various tempering temperature.

Fig. 6. XRD peaks and DSC peaks of 2.0Si steels. (Online version in color.)

In 2.0Si steel, XRD of 673 K and lower tempering showed peak about 2θ =43 deg. corresponding to ε carbide. XRD of 723 K tempering, ε carbide peak disappeared and broadened cementite peak was observed. XRD of 773 K and 823 K tempering showed identical cementite peaks of regular peak intensity order.

DSC of 673 K and lower tempering showed cementite precipitation peak about 750 K. DSC of 723 K showed no peak, indicating cementite precipitation was finished at 723 K tempering.

As shown above, carbide transition tempering temperatures based on DSC peak are almost corresponds to carbide transition tempering temperatures based on XRD peak. XRD analysis of electrolytically extracted carbides showed detailed information about structural change, and also explained the mechanism of dual cementite peaks in DSC of 0.8Si steel.

3.2. Chemical Analysis of Electrolytically Extracted Carbides

Chemical analysis of electrolytically extracted carbides from each tempered steels are shown in Fig. 7.

Fig. 7. Chemical analysis of electrolytically extracted carbides of each steel. (Online version in color.)

In 0.3Si steel, carbide mass of 473 K tempering was very small. Carbide mass of 523 K and higher tempering temperature were increased with tempering temperature. At 623 K and lower tempering temperature, Mn/Fe ratio and Cr/Fe ratio was less than 1/100, almost the same as chemical composition of steel. But Si was little contained in carbide. These carbides were identified as cementite by XRD. In this paper, these cementite without Mn–Cr enrichment is defined as ‘para-cementite’. Para-cementite corresponded to irregular cementite in XRD analysis.

At 723 K tempering, Mn and Cr was concentrated into cementite than chemical composition of steel. This Mn and Cr concentration temperature is almost the same as transition temperature from irregular structure cementite to regular structure cementite in XRD. In this paper, these cementite with Mn–Cr enrichment is defined as ‘ortho-cementite’.

In 0.8Si steel, chemical composition of carbides at 473 K and 523 K tempering contained Si, and these were identified as ε carbide by XRD. Chemical composition of carbides at 573 K and 623 K tempering did not contain Si but almost para composition. Based on XRD, these carbides were identified as para-cementite. Chemical composition of carbides at 723 K tempering showed higher Cr and Mn. Based on XRD, these carbides were identified as ortho-cementite.

In 1.5Si steel, chemical composition of carbides at 523 K–623 K tempering contained Si, and these were identified as ε carbide by XRD. Mass of these ε carbides was not increased with tempering temperature. Mass of carbide started to increase from 673 K tempering, and chemical composition were changed to ortho-cementite, rich with Mn and Cr, and less Si. In this steel, para-cementite was not identified. Chemical composition of 723 K and 773 K tempering were richer with Mn and Cr, and mass of carbide was also increasing.

In 2.0Si steel, chemical composition of carbides at 573 K–673 K tempering contained Si, and these were identified as ε carbide by XRD. Mass of these ε carbides was not increased with tempering temperature. Mass of carbide started to increase from 673 K tempering, and chemical composition were changed to ortho-cementite, rich with Mn and Cr, and less Si. Chemical composition of 773 K and 823 K tempering were richer with Mn and Cr, and mass of carbide was also increasing.

In each steel, mass of carbides was not increased while ε carbide precipitated, but increased after cementite precipitation. If all carbon of 0.55C steel forms cementite and all the cementite is extracted, mass of cementite is 8.25 mass% and Fe in cementite is 7.7 mass%. But, in these experiment, maximum Fe mass in carbide was about 2 mass% in 0.3Si steel, and about 1 mass% in other steels. The possible causes are 1) tempering temperature was low and insufficient for full cementite precipitation, 2) electrolytic extraction method is not adequate for ε carbide and cementite then some of carbides are path through membrane. Cause 1) is possible because mass of residue of each steel were increased within measured tempering temperature. Cause 2) is not investigated precisely in this work. But carbide transition from ε carbide to cementite was confirmed by XRD, then both carbides were extracted in this method. To confirm the loss of carbide in suction filtration, we should analyze the chemical composition of solution after filtration.

4. Discussion

4.1. Carbide Transition and Morphology

Based on the results of XRD and chemical analysis of electrolytically extracted carbides, carbide transition in tempering of middle carbon martensite is three steps as follows, from ε carbide to para-cementite and ortho-cementite.

Morphology and precipitation sites of carbides in tempering were reported in previous work,³²⁾ and FE-SEM images of microstructure were cited in Fig. 8. At lower tempering temperature, carbides precipitated at both inside of martensite block and grain boundary of martensite. Intragranular carbides were line and plate shape with identical orientation in martensite. These intragranular carbides increased and grew up with temperature. But, at higher tempering temperature, 673 K in 0.8Si steel and 1.5Si steel, 723 K in 2.0Si steel, these intragranular carbides disappeared and shifted grain boundary carbides.

Fig. 8. SEM images of microstructure and carbides in each steel.⁸⁾

In 1.5Si steel and 2.0Si steel, these temperatures of carbide precipitation site change almost correspond to carbide transition temperature from ε carbide to ortho-cementite investigated.

In 0.8Si steel, this temperature of carbide precipitation site change corresponds to the temperature of carbide transition from para-cementite to ortho-cementite. This indicates that grain boundary carbides are ortho-cementite, and intragrain carbides are ε carbide and para-cementite. It is difficult to distinguish ε carbide at 523 K tempering from para-cementite at 573 K in FE-SEM observation. But based on this work, we can distinguish para-cementite form ε carbide using DSC analysis.

In this work, χ carbide was not observed. χ carbide is mainly reported mainly in high carbon steel (C > 1.0 mass%) and is observed to precipitate on twin interface.^4,12,13) Teramoto et al. also observed χ carbide in middle C-high Mn steel.³¹⁾ Based on these reports, middle C-middle Mn steels in this work may be not so preferable for χ carbide precipitation.

4.2. Structural Change in Cementite

Based on XRD, para-ortho transition of cementite both in structure and chemical compositions were observed in 0.3Si and 0.8Si steel. Para-cementite showed irregular order of peak intensity in XRD analysis, relatively strong (102) and (002) peaks. This mechanism is discussed in the view of lattice fitting between cementite and bcc-Fe.

Lattice fitting between bcc-Fe and cementite has been researched using TEM analysis.^7,14,15) Andrews¹⁴⁾ showed good lattice fitting between three axis of bcc-Fe [110] [111] [112] and three axes of cementite [001] [100] [010] as shown in Table 4. Lattice parameter of Table 4 is calculated based on JCPDS database, #006-0696 for bcc-Fe and #035-0772 for cementite.

Table 4. Lattice fitting relation between bcc-Fe and cementite ls.

Ferrite PDF#006-0696					Cementite PDF#035-0072
Indices	Interplanar spacing [Å]	Number of planes	Length [Å]	Expansion or Contraction [%]	Indices	Interplanar spacing [Å]
[110]	2.0269	2	4.0537	11.65	[001]	4.526
[111]	1.6549	3	4.9648	2.54	[100]	5.091
[112]	1.1702	6	7.0212	−3.96	[010]	6.7434

Figure 9 shows schematic of this lattice fitting. Lattice misfit between [111]bcc-Fe//[100]Fe₃C and [112]bcc-Fe//[010]Fe₃C are small, and [110]bcc-Fe//[001]Fe₃C is large. If cementite precipitate semi-coherently at low temperature, shape of cementite is expected to be platelet with lattice fitting of Fig. 9, thin for [001]Fe₃C and large area of (001)Fe₃C. In this relationship, cementite platelet precipitates on (110)bcc-Fe. This plane is slip plane of martensite transformation then dislocation induced in transformation may work as nucleation sites.

Fig. 9. Schematic of lattice fitting between bcc-Fe and cementite. (Online version in color.)

XRD analysis of these platelet carbides are expected that diffraction from lattice plane of large area (001)Fe₃C is relatively strong. This mechanism can explain the reason of relatively weak diffraction from main peak (031)Fe₃C, and relatively strong diffraction from (002)Fe₃C and (102)Fe₃C. At the same time, thickness of platelet carbide can be decreased. And this may be the reason of higher 2θ angle and smaller lattice parameter of these relatively strong XRD peaks.

On the other hand, precipitation site of ortho-cementite at higher tempering temperature were block boundary, packet boundary and austenite grain boundary. These sites are structurally not tight. These boundaries are also works as diffusion path for flow out of Si from cementite, and flow in of Mn and Cr into cementite. These conditions are thought to be the reason of structure and chemical composition ortho-cementite.

4.3. Effect of Si Addition on Carbide Transition

Based on the results of XRD and chemical analysis of electrolytically extracted carbides, carbide transition from ε carbide to cementite was retarded by Si addition. In 1.5Si steel and 2.0Si steel, para-cementite was suppressed to form, and ε carbide transited to ortho-cementite directly. These relationships between carbon transition and Si addition are schematically described as Fig. 10.

Fig. 10. Schematic of carbide transition in tempering process of low and high Si steel. (Online version in color.)

Principal of carbide transition is as follows. Chemical composition of ε carbide can contain Si, as shown in Fig. 7. On the other hands, cementite does not contain Si, then Si flow out from carbide precipitation site is trigger of cementite precipitation.

In the case of low content of Si (0.3Si and 0.8Si steel), ε carbide precipitate at first. Because of low content of Si, short range diffusion at low tempering temperature is enough for cementite precipitation. Then fine cementite can precipitate within martensite block, but it is forced to keep lattice coherency with matrix. This fine cementite precipitates only after Si diffusion out, then chemical composition is almost the same as matrix. Such ‘para-cementite’ is secondary carbide. At higher tempering temperature, Mn–Cr can diffuse and enrich in carbide. Then ‘ortho-cementite’ precipitates at the martensite block boundary and former austenite grain boundary, where is the diffusion path of alloying element.

In the case of high content of Si (1.5Si and 2.0Si steel), ε carbide precipitate at first. But para-cementite does not precipitate because short range diffusion is not sufficient to Si rejection. Then ε carbide containing Si remains until higher tempering temperature. At higher tempering temperature enough for rapid Si diffusion through grain boundary, cementite can precipitate at grain boundary. Size of cementite is small at first because of Si diffusion restriction, but chemical composition is nearly ideal. This is ortho-cementite.

In this model, discontinuous precipitation of cementite from ε carbide is also explained as follows. Precipitation site of ε carbide is intragranular, because ε carbide can contain any alloying element in matrix and only C diffusion is needed. Para-cementite is also precipitate in intragranular. Para-cementite may be possible to precipitate continuously from ε carbide if Si content is low enough to flow out by short range diffusion. But typical cementite with ortho composition and high thermodynamical stability precipitate at the grain boundaries, diffusion path of alloying element.

5. Conclusion

Carbide transition in medium carbon tempered martensite and the effect of Si content was investigated using XRD analysis and chemical analysis of electrolytically extracted carbides, combined with SEM observation and DSC thermal analysis in previous work.

In low Si steels, carbide was transited in three steps, from ε carbide to para-cementite and ortho-cementite. Para-cementite showed almost the same composition with steel matrix except for Si and irregular order of XRD peaks intensity indicating anisotropy of crystal structure. Ortho-cementite showed enrichment of Mn and Cr in chemical composition, and regular order of XRD peaks intensity.

In high Si steels, carbide was transited in two steps, from ε carbide to ortho-cementite directly. This mechanism is estimated as follows. Cementite precipitation was suppressed by Si and it was difficult to flow out Si atoms by short range intragranular diffusion at low tempering temperature. At high tempering temperature, Si can diffuse path through grain boundaries, then intergranular ortho-cementite can precipitate.

References

• 1)   G.  Krauss: ISIJ Int., 35 (1995), 349. https://doi.org/10.2355/isijinternational.35.349
• 2)   D. P.  Antia,  S. G.  Fletcher and  M.  Cohen: Trans. Am. Soc. Met., 32 (1944), 290.
• 3)   Y.  Imai: Trans. JIM, 16 (1975), 721. https://doi.org/10.2320/matertrans1960.16.721
• 4)   Y.  Imai,  T.  Ogura and  A.  Inoue: Trans. ISIJ., 15 (1975), 79. https://doi.org/10.2355/isijinternational1966.15.79
• 5)   Y.  Ohmori and  S.  Sugisawa: Trans. JIM, 12 (1971), 170. https://doi.org/10.2320/matertrans1960.12.170
• 6)   I.  Eguchi,  K.  Kato,  S.  Tokunaga and  K.  Suzuki: Trans. JIM, 2 (1961), 218. https://doi.org/10.2320/matertrans1960.2.218
• 7)   Y.  Ohmori: ISIJ Int., 41 (2001), 554. https://doi.org/10.2355/isijinternational.41.554
• 8)   Y.  Ohmori and  I.  Tamura: Metall. Trans. A, 23 (1992), 2737. https://doi.org/10.1007/BF02651753
• 9)   Y.  Ohmori: Trans. JIM., 13 (1972), 119. https://doi.org/10.2320/matertrans1960.13.119
• 10)   Y.  Hirotsu and  S.  Nagakura: Trans. JIM, 15 (1974), 129. https://doi.org/10.2320/matertrans1960.15.129
• 11)   W.  Lu,  M.  Herbig,  C. H.  Liebscher,  L.  Morsdorf,  R. K. W.  Marceau,  G.  Dehm and  D.  Raabe: Acta Mater., 158 (2018), 297. https://doi.org/10.1016/j.actamat.2018.07.071
• 12)   Y.  Tanaka and  K.  Shimizu: Trans. JIM, 22 (1981), 779. https://doi.org/10.2320/matertrans1960.22.779
• 13)   Y.  Nakamura and  S.  Nagakura: Trans. JIM, 27 (1986), 842. https://doi.org/10.2320/matertrans1960.27.842
• 14)   K. W.  Andrews: Acta Met., 11 (1963), 939. https://doi.org/10.1016/0001-6160(63)90063-4
• 15)   Y.  Ohmori,  A. T.  Davenport and  R. W. K.  Honeycombe: Transactions ISIJ., 12 (1972), 112. https://doi.org/10.2355/isijinternational1966.12.112
• 16)   Y.  Imai,  T.  Ogura and  A.  Inoue: Transactions ISIJ, 13 (1973), 183. https://doi.org/10.2355/isijinternational1966.13.183
• 17)   T.  Kunitake: Trans. JIM, 7 (1966), 253. https://doi.org/10.2320/matertrans1960.7.253
• 18)   Y.  Zhang,  C.  Zhao,  M.  Sato,  G.  Miyamoto and  T.  Furuhara: ISIJ Int., 61 (2021), 1641. https://doi.org/10.2355/isijinternational.ISIJINT-2020-412
• 19)   G.  Kafadar,  A.  Kalkanli,  A. T.  Özdemir and  B.  Ögel: ISIJ Int., 61 (2021), 1679. https://doi.org/10.2355/isijinternational.ISIJINT-2020-419
• 20)   M.  Nagao,  T.  Kochi,  T.  Tsuchida,  H.  Yaguchi and  T.  Watanabe: CAMP-ISIJ, 17 (2004), 359 (in Japanese).
• 21)   G.  Miyamoto,  J. C.  Oh,  K.  Hono,  T.  Furuhara and  T.  Maki: Acta Mater., 55 (2007), 5027. https://doi.org/10.1016/j.actamat.2007.05.023
• 22)   J.  Kobayashi,  S. M.  Song and  K.  Sugimoto: ISIJ Int., 52 (2012), 1124. https://doi.org/10.2355/isijinternational.52.1124
• 23)   T.  Furuhara,  K.  Kobayashi and  T.  Maki: ISIJ Int., 44 (2004), 1937. https://doi.org/10.2355/isijinternational.44.1937
• 24)   V. K.  Euser,  D. L.  Williamson,  A. J.  Clarke and  J. G.  Speer: ISIJ Int., 60 (2020), 2990. https://doi.org/10.2355/isijinternational.ISIJINT-2020-263
• 25)   S.  Xia,  F.  Zhang,  C.  Zhang and  Z.  Yang: ISIJ Int., 57 (2017), 558. https://doi.org/10.2355/isijinternational.ISIJINT-2016-505
• 26)   G. A.  Thomas,  F.  Danoix,  J. G.  Speer,  S. W.  Thompson and  F.  Cuvilly: ISIJ Int., 54 (2014), 2900. https://doi.org/10.2355/isijinternational.54.2900
• 27)   T.  Suzuki,  Y.  Ono,  G.  Miyamoto and  T.  Furuhara: ISIJ Int., 50 (2010), 1476. https://doi.org/10.2355/isijinternational.50.1476
• 28)   O.  Matsumura,  Y.  Sakuma and  H.  Takechi: Trans. ISIJ, 27 (1987), 570. https://doi.org/10.2355/isijinternational1966.27.570
• 29)   K.  Sugimoto,  N.  Usui,  M.  Kobayashi and  S.  Hashimoto: ISIJ Int., 32 (1992), 1311. https://doi.org/10.2355/isijinternational.32.1311
• 30)   J.  Tobata,  K. L.  Ngo-Huynh,  N.  Nakada,  T.  Tsuchiyama and  S.  Takaki: ISIJ Int., 52 (2012), 1377. https://doi.org/10.2355/isijinternational.52.1377
• 31)   S.  Teramoto,  M.  Imura,  Y.  Masuda,  T.  Ishida,  M.  Ohnuma,  Y.  Neishi and  T.  Suzuki: ISIJ Int., 60 (2020), 182. https://doi.org/10.2355/isijinternational.ISIJINT-2019-331
• 32)   T.  Suzuki,  S.  Teramoto and  Y.  Neishi: Transactions of JSSE, 64 (2019), 1 (in Japanese). https://doi.org/10.5346/trbane.2019.1