Effect of Si on Precipitation Behavior of Nb-Laves Phase

and Amount of Nb in Solid Solution at Elevated Temperature

in High Purity 17%Cr-0.5%Nb Steels

Yasushi Kato1, Masatoshi Ito2, Yoshimine Kato2 and Osamu Furukimi2

1Steel Research Lab., JFE Steel Corp., Chiba 260-0836, Japan2Department of Materials Science & Engineering, Kyushu University, Fukuoka 819-0395, Japan

The effect of Si was investigated on the precipitation behavior and the amount of Nb in solid solution at temperature ranging from 1073 to1173K in high purity 17Cr-0.5Nb steels. Adding Si promoted the precipitation of the Nb Laves phase and then decreased the solubility of Nb insteels. The Nb Laves phase which was composed of Fe, Cr and Nb could be expressed as (Fe, Cr)2Nb in a 0.002mass% Si steel. On the otherhand, in a 0.5mass% Si steel the Nb Laves phase which was composed of Fe, Si, Cr and Nb could be expressed as (Fe, Si, Cr)2Nb. Based oncalculations from the experimental results assuming the Laves phase is Fe2Nb, the standard free energy change of the Fe2Nb precipitationreaction was about �61 k J/mol for a 0.002mass% Si steel. [doi:10.2320/matertrans.M2010112]

(Received March 29, 2010; Accepted June 28, 2010; Published August 25, 2010)

Keywords: stainless steels, ferritic steels, Laves phases, precipitation

1. Introduction

The solubility of carbon and nitrogen in ferritic stainlesssteels, which have a bcc structure, is so low that corrosionresistance tends to decline due to the formation of a Cr-depleted zone along the grain boundaries resulting fromprecipitation of Cr-carbonitrides at the grain boundariesduring heat treatment.1) Sensitization can be suppressed byadding Nb, which has greater chemical affinity to C and Nthan Cr.2) Niobium is also an effective alloying element forenhancing strength at elevated temperatures in solid solu-tion.3) Therefore, Nb-bearing ferritic stainless steels havebeen developed and used mainly as materials for automotiveexhaust parts.

Niobium forms a hexagonal Laves phase, a C14 type, ofFe2Nb at high temperature, and various properties of Nb-bearing ferritic stainless steels are significantly influencedby the precipitation of the Laves phase.4–6) For instance,high temperature strength of steels decreases as a result ofthe reduced amount of Nb in solid solution due to theprecipitation of the Laves phase.5) The Laves phase also actsas a notch in the matrix, deteriorating the toughness ofsteels.6) Therefore, clarification of the precipitation behaviorof the Laves phase is an important research issue fordeveloping Nb-bearing ferritic stainless steels with higherheat resistance than the conventional steels. Estimation ofthe amount of Nb in solid solution in steels during hightemperature service is required in order to analyze theirstrength quantitatively.

It has been reported that the Laves phase, MX typecarbonitrides and M6C type carbide (Fe3Nb3C) can precip-itate in Nb-bearing ferritic stainless steels.7,8) At hightemperature, precipitation of the Laves phase competes withthat of M6C. Concerning this competition, Fujita et al.7)

reported that M6C was a stable precipitate at 973–1123K ina 19Cr-0.4Nb-0.014C-0.017N (mass%) steel and at 1023–1173K in a 19Cr-0.8Nb-0.014C-0.016N (mass%) steel. Onthe other hand, in a 14Cr-0.3Nb-0.15Ti-0.5Mo-0.011C-0.010N (mass%) steel, the Laves phase was a stable

precipitate at 973–1173K when M6C precipitation wassuppressed by Ti addition, and the solubility product of theLaves phase was obtained experimentally under certainassumptions.7) Because Si is a useful alloying element foroxidation property,9) a moderate amount of Si, 0.3–1mass%,is added to commercial Nb-bearing ferritic stainless steels.10)

The effect of Si on the precipitation behavior of the Lavesphase in a martenstic 9Cr-2Mo steel was reported by Isedaet al.11) The precipitation of the Mo Laves phase (Fe2Mo)was promoted and the solubility of Mo was lowered withincreasing Si contents when the material was subjected totemperatures in the range of 773–973K for a long period.However, the effect of Si on the precipitation behavior ofthe Nb Laves phase in ferritic stainless steels has not beenrevealed.

In this study, in order to clarify the inherent precipitationbehavior of the Nb Laves phase, high purity 17Cr-0.5Nbsteels with small amounts of C and N were used, as both ofthese elements tend to form precipitates of Nb in forms ofMX and M6C. The objective of this study is to clarify theeffect of Si on the precipitation behavior of the Nb Lavesphase and the amount of Nb in solid solution at elevatedtemperatures in 17%Cr-0.5%Nb ferritic stainless steels.

2. Experimental Procedures

Two kinds of steel were used in this study, one is a 17Cr-0.5Nb steel (base steel) and the other is a 17Cr-0.5Nb-0.5Sisteel (0.5Si steel). The steels were melted by high-frequencyinduction heating in a vacuum and cast as 30 kg ingots. Thechemical compositions of the ingots obtained in this mannerare shown in Table 1. Amounts of impurities were rather lowlevel than those in commercial ferritic stainless steels. Inparticular, contents of carbon and nitrogen in both steels wereless than 20massppm. The ingots were heated at 1443K for3.6 ks in an Ar atmosphere and hot-rolled to a thickness of4mm, followed by annealing at 1373K for 60 s in an Aratmosphere and mechanical descaling. These sheets werecold-rolled to a thickness of 1mm and then degreased.

Materials Transactions, Vol. 51, No. 9 (2010) pp. 1531 to 1535#2010 The Japan Institute of Metals

Finally, the sheets were annealed at 1323K for 60 s in an Aratmosphere for recrystallization and then water-cooled. Thecold-rolled and annealed sheets were subjected to aging at1073K and 1173K for various times in an Ar atmosphere inorder to investigate the change in precipitates. Quantitativeanalysis of Nb precipitates was performed as describedbelow.12) The Nb precipitates were totally extracted fromthe specimens by means of electrochemical dissolution of10 vol% acetylaceton and 1 vol% tetramethylammomiumchloride in methanol. The residue was filtered with 200 nmdiameter pores to trap fine particles. The residue wasdissolved in a solvent and analyzed using Inductive CoupledPlasma (ICP). The results corresponded to the total amountof precipitates. Furthermore, the residue was immersed in a25 vol% sulfuric hydroric acid solution and then dissolved ina 1mass% potassium permanganate and 25 vol% sulfurichydroric acid solution in order to dissolve the Laves phaseonly. The residue was analyzed by the same methodmentioned above. The results were equivalent to the amountof precipitates except the Laves phase. Therefore, the amountof the Laves phase in the precipitates was obtained as thedifference between the analytical results of this residueand the total amount of precipitates. Identification of theprecipitates was made with extracted residue using X-raydiffraction spectroscopy (XRD). The precipitates wereobserved with extracted residue using scanning electronmicroscopy (SEM) and with thin foil using transmissionelectron microscopy (TEM).

3. Results and Discussion

Figure 1 shows the results of XRD of the total extractedresidues samples before and after the aging at 1073K for360 ks. The Nb(C,N) and M6C were identified as precip-itates of the specimen before aging. Furthermore, Nb Laves

phase was detected as precipitates in the aged specimensother than carbonitrides. Figure 2 shows the amounts of Nb,Fe, Cr and Si as the precipitates with aging time at 1073K.As the Laves phase precipitated, the total amount ofprecipitates increased after aging at 1073K in both steels.In the base steel, the amount of the Laves phase increasedafter aging for more than 0.3 ks, and became constant after10 ks. On the other hand, in the 0.5Si steel, the amount ofthe Laves phase increased after aging for more than 0.1 ks,and became constant at 0.9 ks. This result revealed that theprecipitation reaction of the Nb Laves phase was promotedby adding Si.

Figure 3 shows the comparison for the amount of Nb, Fe,Cr and Si as the precipitates between 1073K–360 ks aged

Table 1 Chemical compositions of specimens.

(mass%)

Steel Cr Si Nb Mn C N S O P

base 16.9 0.002 0.51 0.01 0.0013 0.0018 0.0007 0.0023 0.006

0.5Si 17.0 0.50 0.50 0.01 0.0013 0.0012 0.0007 0.0011 0.005

(a)

(c) (d)

2000

cps

(b)

2θθ / deg

2θ / deg

42 44 46 4838 403634

42 44 46 4838 403634

Fig. 1 X-ray diffraction pattern of residues extracted. : Fe2Nb, :

Fe3Nb3C, : Nb(C,N) (a) base steel (before aging) (b) base steel (after

aging at 1073K for 360 ks) (c) 0.5Si steel (before aging) (d) 0.5Si steel

(after aging at 1073K for 360 ks)

0

0.1

0.2

0.3

0.4

0.50

0.1

0.2

0.3

0.4

0.5

NbFe

CrSi

NbFe

CrSi

NbFe

CrSi

NbFe

CrSi

NbFe

CrSi

NbFe

CrSi

NbFe

CrSi

before aging

aged for 0.1ks

aged for 0.3ks

aged for 0.9ks

aged for 10ks

aged for 360ks

aged for 720ks

as Laves

as carbonitrides(a)

(b)

Am

ount

of

prec

ipita

tes

(mas

s%)

Am

ount

of

prec

ipita

tes

(mas

s%)

Element

Treatment

Fig. 2 Changes in amount of Nb, Fe, Cr and Si as precipitates with aging in

17%Cr-0.5%Nb steels at 1073K. (a) base steel, (b) 0.5Si steel

0

0.1

0.2

0.3

0.4

0.5

as Lavesas carbonitrides

NbFe

CrSi

NbFe

Cr NbFe

CrSi

NbFe

Cr

1073K 1173K 1073K 1173K

base Steel 0.5Si steel

Am

ount

of

prec

ipita

tes

(mas

s%)

Element

Aging temp.

Fig. 3 Effect of aging temperature on the amount of Nb, Fe, Cr and Si as

precipitates in 17%Cr-0.5%Nb steels after 360 ks aging.

1532 Y. Kato, M. Ito, Y. Kato and O. Furukimi

specimens and 1173K–360 ks aged specimens. In both steels,the amounts of Nb in total precipitates and in the Laves phaseof the specimens aged at 1173K were smaller than those ofthe specimens aged at 1073K. The amounts of the carboni-trides precipitates aged at 1173K is larger compared withthose in the specimens aged at 1073K, in spite of the Sicontents. Taking into account the Fe3Nb3C solubility productreported by Fujita et al.,7) the increase in carbonitrides likelyto correspond to the increase in the M6C. From the resultsmentioned above, it was clear that the equilibrium amount ofNb in solid solution for 17Cr-0.5Nb steels decreased byadding Si with the temperatures from 1073K to 1173K dueto increasing of the Laves phase precipitated.

The Laves phase was observed not only at grain bounda-ries but also within the grains in both steels after the aging.There seemed to be no difference in the precipitation sitesbetween the base steel and 0.5Si steel. Figures 4(a)–(d) showthe results of SEM observation of residues extracted from thespecimens before and after the aging at 1073K for 360 ks forboth steels. There was no distinct difference in size ofprecipitates (MX, M6C) between the base steel and the 0.5Sisteel before aging the specimens. On the other hand, asshown in Fig. 4(d) size of the Laves phase in the 0.5Si steelwas slightly smaller than that of the base steel (see Fig. 4(c))after aging at 1073K for 360 ks, in spite of the fact that theamount of the Laves phase precipitated in the 0.5Si steel wasmore than that in the base steel as shown in Fig. 3. This

seemed to indicate that the nucleation sites of the Lavesphase precipitates increased by adding Si. Figures 5(a)–(d)show the results of TEM observation of the specimens agedat 1073K for 360 ks. The results of EDS analysis clearlyrevealed that the peaks corresponded to Fe, Nb and Cr in thebase steels and to Fe, Nb, Cr and Si in the 0.5Si steel. Table 2shows the amounts of each element as total precipitates, theLaves phase and carbonitrides in the specimens aged for360 ks. The Laves phase included small amount of Cr in bothsteels and also much larger amount of Si than that of Cr in the0.5Si steel.

The Laves phase, which occurs at or around a fixedstoichiometry A2B, are well known as the size compounds.The composition range of Nb in the Fe2Nb Laves phase is31–37 at% (AxBy: x=y ¼ 1:7{2:2) in Fe-Nb binary system.13)

It was shown by Kaloev et al.14) that in Fe-Cr-Nb ternarysystem as much as 60 at% of Cr can be dissolved in the Fe2NbLaves phase, and formation of (Fe, Cr)2Nb as Laves phasewas confirmed by Mansour et al.15) Effects of Si on thestability and the composition of the C14 Laves phase such asFe2Nb were also reported for various alloy systems. It wasreported that the Laves phase does not form in the binaryNickel and Cobalt systems, such as Ni-Ti, Ni-Ta, Co-Mo andCo-W. However, in ternary systems the type of (A3Si)B2 canbe formed. In (A3Si)B2 Si is substituted for 25% of theA-component in forms of ternary systems, such as Ni-Ti-Si,Ni-Ta-Si, Co-Mo-Si and Co-W-Si.16) Goldschmidt et al.17)

(a) (b)

(c) (d)

Fig. 4 SEM micrographs of residues extracted from specimens. (a) base steel (before aging), (b) 0.5Si steel (before aging), (c) base steel

(aged at 1073K for 360 ks), (d) 0.5Si steel (aged at 1073K for 360 ks)

Effect of Si on Precipitation Behavior of Nb-Laves Phase and Amount of Nb in Solid Solution at Elevated Temperature 1533

showed that the Fe2Nb Laves phase possesses a veryextensive solubility of Si in Fe-Nb-Si ternary system. Itwas also reported that 25 at% of Si can be dissolved into theFe2Nb Laves phase with the temperature range of 1273 to1573K, and Si substitutes only Fe in the Laves phase.17–19)

The reason of Si stabilization in the C14 Laves phaseappeared to be due to decrease of the effective electronconcentration.16) Zhao et al.20) reported that Si couldstabilized the Laves phase at lower temperatures and itsLaves phase was expressed as (Cr, Si)2Nb in Cr-Nb-Siternary system. Maziasz21) and Yamamoto et al.22) reportedthat adding Si to the heat-resisting austenitic steel promotesthe formation of the Fe2(Mo, Nb) Laves phase, but does notchange the phase equilibrium between �-Fe and Laves phase.In a study of a 9Cr-2Mo-Si steel, Iseda et al.11) reported thatSi occupy the A site of the Mo-Laves phase, and the Lavesphase could be expressed as (Fe, Cr, Si)2Mo according to theexperimental results of analyzing the Laves phase usingenergy dispersive spectrometry (EDS).

Taking those previous research results into consideration,discussion on the Laves phase formation was made in thisstudy. Since the ratios of (Fe + Cr) to Nb in the Laves phaseprecipitates were 2.08 at 1073K and 1.82 at 1173K aging,the Laves phase could be expressed as (Fe, Cr)2Nb in thebase steel. If A-sites are composed of Fe, Si and Cr, it ispossible to consider that the ratios of (Fe + Si + Cr) to Nbwere 2.3 at 1073K and 2.1 at 1173K aging, as shown inTable 2. It seems reasonable that the Laves phase of the0.5Si steel could be expressed as (Fe, Cr, Si)2Nb.

Laves phase is a prominent precipitate in both steels. Inparticular, major components of Laves phase are Fe and Nbfor the base steel. If small amount of Cr in Laves phase isdealt rigorously, it must be complicated for calculating thestandard free energy changes for the precipitation of the NbLaves phase. To avoid difficulty, some assumptions are setas the Fujita’s research.7) Namely, it is assumed that Lavesphase is dealt as Fe2Nb, because the amount of Cr inprecipitates is small enough compare with those of Fe and

cps

(c) (d)

Fe

Cr

Nb

Si

Fe

Fe

Fe

Cr

Nb

Fe

Fe

200nm

(a)

Laves

(b)

Laves

Energy / keV Energy / keV

Fig. 5 TEMmicrographs and EDS analysis results of Laves phase of specimen aged at 1073K for 360 ks. (a), (c): base steel, (b), (d): 0.5Si

steel.

Table 2 Amount of Nb, Fe, Cr and Si as precipitates in specimens aged at 1073K and 1173K for 360 ks.

(mol%)

Steel Base 0.5Si

Aging condition 1073K–360 ks 1173K–360 ks 1073K–360 ks 1173K–360 ks

Total-Nb 0.18 0.10 0.24 0.16

Total-Fe 0.32 0.15 0.41 0.28

Total-Cr 0.043 0.025 0.054 0.041

Total-Si <0:004 <0:004 0.092 0.050

Nb as Laves 0.16 0.064 0.24 0.12

Fe as Laves 0.30 0.10 0.41 0.21

Cr as Laves 0.040 0.017 0.053 0.030

Si as Laves <0:004 <0:004 0.090 0.036

Nb as carbonitrides 0.017 0.032 0.004 0.046

Fe as carbonitrides 0.020 0.050 0.002 0.070

Cr as carbonitrides 0.003 0.008 0.001 0.008

Si as carbonitrides <0:004 <0:004 <0:004 0.008

1534 Y. Kato, M. Ito, Y. Kato and O. Furukimi

Nb. The standard free energy change for the precipitation ofFe2Nb was calculated based on the amount of Nb in theprecipitates as shown in Table 2. The precipitation reactionof Fe2Nb can be expressed as follows:

2½Fe� þ ½Nb� ¼ Fe2Nb ð1Þ

Assuming that the activities of the other elements areequivalent to the concentrations, the solubility products ofFe2Nb in the matrix can be expressed as follows:

ðXFeÞ2ðXNbÞ ¼ K0 expð��G0=RTÞ ð2Þ

where XFe and XNb are the concentrations in solution in thematrix, K0 is a constant, �G0 is the standard free energychange for the precipitation reaction, R is a gas constant andT is reaction temperature. Because the materials studied hereare Fe base alloys, XFe is regarded as one. The solubilityproduct is expressed as the solubility of Nb by the followingequation.

lnðXNbÞ ¼ lnK0 ��G0=RT ð3Þ

In this study, MX and M6C existed in base steel before theaging. Not the Laves phase, but M6C was reported to be theprominent precipitate at temperatures from 973 to 1223K ina 19Cr-0.4Nb-0.014C-0.017N steel.7) Therefore, the equi-librium concentration of Nb in solution was calculated bysubtracting the concentration of Nb in the carbonitrides andthe Laves phase from that of Nb added in the specimensaged at 360 ks. Figure 6 shows the relationship betweenlnðXNbÞ and 1=T of the base steel. In Fig. 6, natural log ofthe Nb concentration in solid solution of the 0.5Si steel isplotted. The data of a 14Cr-0.3Nb-0.15Ti-0.5Mo-0.011C-0.010N steel in which the precipitation of M6C did notoccur, is also plotted as ‘‘Fujita et al.’’.7) The calculatedstandard free energy change of Fe2Nb precipitation wasabout �61 kJ/mol at temperatures from 1073 to 1173K.The solubility product of Fe2Nb for the base steel could beexpressed by mol% as follows:

lnðXNbÞ ¼ �7300=T þ 0:05 ð4Þ

The calculated value of solubility product of the base steelin this study differs from that of the Fujita’s research. Itseems that this is because of the difference not only in

compositions of the steel studied, but also in assumptions ofthe calculation.

4. Conclusions

The effect of Si on the precipitation behavior of the NbLaves phase and the solubility of Nb in high purity 17Cr-0.5Nb steels was investigated. The following conclusionswere obtained.

(1) Adding Si promoted the precipitation of the Nb Lavesphase and decreased the amount of Nb in solid solution attemperature range from 1073 to 1173K in high purity 17Cr-0.5Nb steels.

(2) The Nb Laves phase was composed of Fe, Cr and Nband could be expressed as (Fe, Cr)2Nb in the base steel. Onthe other hand, in the 0.5Si steel the Nb Laves phase wascomposed of Fe, Si, Cr and Nb and could be expressed as(Fe, Si, Cr)2Nb.

(3) Based on calculations from the experimental results byassuming the Laves phase as Fe2Nb, the standard free energychange of the Fe2Nb precipitation reaction was about�61 k J/mol for the base steel and the solubility productcould be expressed as below:

lnðXNbÞ ¼ �7300=T þ 0:05:

REFERENCES

1) M. A. Streicher: Corrosion 29 (1973) 337.

2) H. Abo, T. Nakazawa, S. Takamura, M. Onoyama, H. Ogawa and H.

Okada: Stainless Steel ’77, Climax Molybdenum Co (1978) p. 97.

3) N. Fujita, K. Ohmura, M. Kikuchi, T. Suzuki, S. Funaki and I.

Hiroshige: Scr. Mater. 35 (1996) 705.

4) A. Miyazaki, K. Takao and O. Furukimi: ISIJ Int. 42 (2002) 916.

5) N. Fujita, K. Ohmura and A. Yamamoto: Mater. Sci. Eng. A 351 (2003)

272.

6) T. Tomita and M. Oku: Nissin Technical Report 87 (2006) p. 11.

7) N. Fujita, M. Kikuchi and K. Ohmura: ISIJ Int. 43 (2003) 1999.

8) A. Miyazaki, M. Gunji and K. Yoshioka: Kawasaki Steel Technical

Report 31 (1994) p. 21.

9) G. C. Wood, J. A. Richardson, M. G. Hobby and J. Boustead: Corros.

Sci. 9 (1969) 659.

10) K. Ishii, A. Miyazaki and S. Satoh: Kawasaki Steel Technical Report

40 (1999) p. 39.

11) A. Iseda, H. Teranishi and K. Yoshikawa: Tetsu-to-Hagane 76 (1990)

2190.

12) M. Inose, K. Fujimoto and Y. Kato: CAMP-ISIJ 20 (2007) 1355.

13) O. Kubaschewski: Iron-Binary Phase Diagrams, (Springer-Verlag

Heidelberg New York Verlag Stahleisen m.b.H. Dusseldorf, 1982)

p. 40.

14) N. I. Kaloev, A. Kh. Abram’yan, S. V. Kabanov and L. K. Kulova:

Izvest. Akad. Nauk SSSR Metally 1 (1988) 205.

15) S. Mansour, N. Boutarek, H. Aid and S. E. Amara: JEEP 00003 (2009)

p. 1.

16) D. I. Bardos, K. P. Gupta and P. A. Beck: Trans. AIME 221 (1961)

1087.

17) H. J. Goldschmidt: J. Iron Steel Inst. 194 (1960) 169.

18) A. W. Denham: J. Iron Steel Inst. 205 (1967) 435.

19) B. N. Singh and K. P. Gupta: Metall. Trans. 3 (1972) 1427.

20) J.-C. Zhao, M. R. Jackson and L. A. Peluso: Acta Mater. 51 (2003)

6395.

21) P. J. Maziasz: J. Met. 41 (1989) 14.

22) Y. Yamamoto, M. Takeyama, Z. P. Lu, C. T. Liu, N. D. Evans, P. J.

Maziasa and M. P. Brady: Intermetallics 16 (2008) 453.

-8.0

-7.5

-7.0

-6.5

-6.0

8.0E-04 9.0E-04 1.0E-03

base steel

0.5Si steel

Fujita, et al.

1/T /K

ln(X

Nb)

Fig. 6 Relationship between lnðXNbÞ and 1=T .

Effect of Si on Precipitation Behavior of Nb-Laves Phase and Amount of Nb in Solid Solution at Elevated Temperature 1535