- Email: [email protected]
Cr phase decomposition process and age-hardening in Fe–15Cr ferritic alloys
Journal of Nuclear Materials 455 (2014) 436–439
Contents lists available at ScienceDirect
Journal of Nuclear Materials journal homepage: www.elsevier.com/locate/jnucmat
Correlation of Fe/Cr phase decomposition process and age-hardening in Fe–15Cr ferritic alloys Dongsheng Chen a,⇑, Akihiko Kimura b, Wentuo Han b a b
Graduate School of Energy Science, Kyoto University, Gokasho, Uji, Kyoto 611-0011, Japan Institute of Advanced Energy, Kyoto University, Gokasho, Uji, Kyoto 611-0011, Japan
a r t i c l e
i n f o
Article history: Available online 7 August 2014
a b s t r a c t The effects of thermal aging on the microstructure and mechanical properties of Fe–15Cr ferritic model alloys were investigated by TEM examinations, micro-hardness measurements and tensile tests. The materials used in this work were Fe–15Cr, Fe–15Cr–C and Fe–15Cr–X alloys, where X refers to Si, Mn and Ni to simulate a pressure vessel steel. Specimens were isothermally aged at 475 °C up to 5000 h. Thermal aging causes a significant increase in the hardness and strength. An almost twice larger hardening is required for embrittlement of Fe–15Cr–X relative to Fe–15Cr. The age-hardening is mainly due to the formation of Cr-rich a0 precipitates, while the addition of minor elements has a small effect on the saturation level of age-hardening. The correlation of phase decomposition process and age-hardening in Fe–15Cr alloy was interpreted by dispersion strengthened models. Ó 2014 Elsevier B.V. All rights reserved.
1. Introduction High-Cr ferritic steels have been considered to be candidates for structure material of nuclear power plants as well as advanced fusion blankets. One of the critical issues in the use of the ferritic steels is the 475 °C aging embrittlement [1–5] that is well known to occur in ferritic, dual phase and martensitic stainless steels with Cr contents higher than 12 wt.% [6–10]. The age-hardening is caused by the decomposition of the ferrite phase to chromium-rich phase, a0 , and iron-rich phase, a, in the temperature range of 280– 500 °C. The a0 precipitation leads to a progressive hardening and deterioration of fracture toughness. Williams [11] carried out Mössbauer measurements on the effect of 475 °C aging in a series of iron-chromium binary alloys with a varying the chromium content, and concluded that the alloys with chromium content 12–30 at.% decomposed via nucleation and growth mechanisms. An atomistic study on the hardening behavior of iron-chromium alloys under thermal aging has been done by Bonny et al. [12] showing the different stages of a–a0 unmixing. No study have been done to discuss the a0 size dependence of strengthening factor for evaluation of the hardening. Courtnall and Pickering [13] found that the embrittlement was enhanced by addition of an interstitial element. However, only a limited data is available on the effects of carbon on the phase ⇑ Corresponding author. Tel.: +81 774 38 3478; fax: +81 774 38 3479. E-mail address: [email protected] (D. Chen). http://dx.doi.org/10.1016/j.jnucmat.2014.07.069 0022-3115/Ó 2014 Elsevier B.V. All rights reserved.
decomposition process in ferritic steels during thermal aging. Our preliminary research [14] showed that the early hardening was also attributed to the precipitations of carbides. In this work, the age-hardening process is correlated with the evolution of phase decomposition with focusing on the dispersion model, and effects of alloying on the age-hardening of Fe–15Cr ferritic alloys were also investigated. 2. Experimental The materials used in this work are Fe–15Cr, Fe–15Cr–C and Fe–15Cr–Xs alloys, where Xs refers to Si, Mn and Ni to simulate a pressure vessel steel. The chemical compositions of the alloys are shown in Table 1. The model alloys were prepared using high purity base metals (>99.995%) by arc melting method in argon atmosphere. The alloy button ingots were homogenized at 1100 °C for 72 h followed by furnace cooling and then cold rolling to 80% at room temperature. All specimens were punched out from each alloy sheet with 0.3 mm thick, and annealed at 900 °C for 2 h followed by ice water quenching. Some of the specimens were sealed in quartz tube in high vacuum conditions (104 Torr) and isothermally aged at 475 °C for up to 5000 h. Micro-hardness measurements and tensile tests were carried out before and after thermal aging. Specimens for hardness measurements were mechanically grinded and polished. Vickers hardness was measured with a 1 kg load. At least ten measurements were made and averaged. The dimensions of tensile test specimens are 16 mm 4 mm 0.25 mm. Room temperature tensile
D. Chen et al. / Journal of Nuclear Materials 455 (2014) 436–439
437
Table 1 Chemical compositions of the alloys (wt.%). Materials
Cr
C
Mn
Si
Ni
Fe
Fe–Cr Fe–Cr–C Fe–Cr–Xs
14.55 14.83 14.96
0.001 0.022 0.002
– – 0.83
– – 0.51
– – 0.49
Bal. Bal. Bal.
tests were carried out by using an INSTRON testing machine with a crosshead velocity of 0.2 mm/min. Disk-type specimens of 3 mm diameter were punched out from a 0.3 mm thick plate and mechanically grounded to about 50 lm in thickness. Final TEM specimens were prepared by electrolytic polishing in a 10 vol.% perchloric acid and 90 vol.% acetic acid using a twin-jet polisher at a voltage of 20 V at room temperature. The size and number density of a0 particles were measured to investigate the effect of a0 phase on the age-hardening behavior. The chemical compositions of precipitates were measured using a TEM JEM2200FS equipped with an energy-dispersive X-ray spectroscopy (EDS) system. 3. Results 3.1. Vickers hardness measurement The Vickers hardness of the model alloys after thermal aging treatment at 475 °C is shown in Fig. 1. As aging time increases up to 5000 h, the Vickers hardness values increase abruptly after an incubation period in all the materials. The increments of the hardness, DHV, in all the alloys are also shown in Fig. 1. For aging times up to 2000 h, larger hardness values were observed in Fe–Cr and Fe–Cr–X than Fe–Cr, with the largest difference (about 40 HV) at an aging time of around 1000 h. However, after aging for 5000 h, the DHV of each alloy became similar with an average increment about 120 HV.
Fig. 2. Yield stress (YS), ultimate tensile stress (UTS) and elongation as a function of aging time of each alloy.
3.2. Tensile properties The changes in the tensile properties is summarized in Fig. 2, showing a significant increase in both yield stress (YS) and ultimate tensile stress (UTS), with a concomitant decrease of elongation. In Fe–Cr, the yield stress increased from 197 MPa to 398 MPa after 5000 h aging. The early stage of aging (up to 2000 h) leads to about two thirds increment in yield stress, while the increasing rate of yield stress becomes slower for longer aging time. Linear relations
Fig. 3. Linear relations between the change in yield stress, Dr, and the change in elongation, De, of each alloy after aging for 500, 2000 and 5000 h.
between the change in yield stress, Dr, and the change in elongation, De, of each alloy are shown in Fig. 3. The slopes of the plots, Dr/De of each alloy are classified into two groups, one is Fe–Cr and Fe–Cr–C and the other is Fe–Cr–Xs. It is considered that the ductility loss of Fe–Cr–Xs needs a larger age-hardening than Fe– Cr and Fe–Cr–C. 3.3. Microstructural features
Fig. 1. Vickers hardness (HV) and increment in Vickers hardness (DHV) as a function of aging time in each model alloy, thermally aged at 475 °C.
Phase decomposition of the ferrite to a0 /a phases is identified by TEM EDS line analysis, as shown in Fig. 4, which shows the composition fluctuation of Cr in the matrix. The fluctuation amplitude of
438
D. Chen et al. / Journal of Nuclear Materials 455 (2014) 436–439
Fig. 4. FE-TEM EDS line analysis of Fe and Cr in Fe–Cr alloy after aging for 5000 h, bright field images, finely-dispersed, nanometer size a0 particles in matrix.
Fig. 6. Average diameter and number density of a0 particles as a function of aging time in Fe–Cr alloy.
calculated the activation energy of the phase decomposition in Fe– 14Cr alloy, which is about 122 kJ mol1 [17]. The nucleation of a0 particle thus requires a certain thermal aging time to gain enough nucleation energy. This agrees with the present results of mechanical property measurement because there is almost no change of Vickers hardness and yield stress until 500 h aging. Thermal aging for 2000 h causes the formation of quite a large number of a0 particles in matrix, with an average diameter of 5.2 nm and a number density of 4.9 1022 m3. After 5000 h aging, however, the a0 precipitates coarsen to an average diameter of 8.3 nm, and the number density of a0 particles decreases, to a number density of 3.8 1022 m3. With increasing aging time, Vickers hardness and yield stress increase due to the a/a0 phase decomposition. The increase in yield stress can be given by the following equations as a function of the diameter and number density of a0 particles:
Dry ¼ MaGbðNdÞ Fig. 5. Bright field transmission electron micrographs, showing a microstructure evolution of Fe–Cr alloy (a) before aging, (b) aged for 500 h, (c) aged for 2000 h and (d) aged for 5000 h, black dots indicate Cr-rich a0 particles.
a0 particles is about 8 wt.%, with a minimum value of 11 wt.% and maximum 19 wt.%. The Cr-rich a0 precipitates, which have a lattice parameter aCr = 0.288 nm, are found to be coherent body-centered cubic (bcc) particles [15]. As is shown in Fig. 4, finely-dispersed, nanometer size a0 particles can be observed after 5000 h aging. The microstructure evolution of Fe–Cr alloy during thermal aging is presented in Fig. 5. Only a few a0 particles are found after 500 h aging, while larger size and higher number density of a0 particles are observed as aging time increases. 4. Discussion 4.1. Correlation of microstructure and mechanical properties Phase decomposition process can be described in terms of the average diameter, d, and number density, N, of a0 particles. Fig. 6 shows the average diameter and number density of a0 particles as a function of aging time, which were estimated from TEM observations. The underlying phase decomposition process of Fe–15Cr alloy is nucleation and growth of a0 precipitates. This mechanism agrees with a Mëssbauer spectroscopy study by Dubiel and Zukrowski [16], who reported that nucleation and growth mechanisms worked in Fe–15 at.% Cr alloy during annealing at 415 °C. They also
1=2
2=3
Dry ¼ aMGbðNÞ
ð1Þ
d
ð2Þ
Eq. (1) is based on the dispersed barrier hardening model [18] and Eq. (2) is on the Friedel–Kroupa–Hirsch (FKH) model [19]. In the two equations, ry is the increment in yield stress, M is Taylor factor (3.06), a is strength factor, G is the shear modulus (75.3 GPa), b is the Burgers vector (0.248 nm). In using Eq. (1), a is assumed to be varying from 0.11 to 1 depending on barrier type [18]. While, in using Eq. (2), a is assumed to be 45 for strong obstacles and 18 for small dislocation loops [20]. Both models can be applied for weak precipitates. Another model for impenetrable obstacles is an expression derived from continuum model simulations by Scattergood and Bacon (SB) [21]:
"
Dry ¼
lb ln 2pL
1 R1 L
þ Rd1
!
# þ 0:7
ð3Þ
where R1 is cut-off radius of the dislocation core (taken to be equal to b) and L is the free passage distance L = Ly 2RPRP, RPRP is radius of precipitates). However, it should be only applied for reaction satisfying the Orowan bypass model (the presence of a screw dislocation dipole drawn out from the obstacle). As for strength factor, a, of obstacles, a molecular dynamics (MD) simulation work [22] provided those values of the coherent Cr precipitate of 2 nm diameter, dislocation loops and voids as a function of temperature in bcc Fe, and the obtained values were in the range 0.17–0.37 at 300 K. In this research, the strength factor a of Cr precipitates at room temperature is assumed to be 0.2 that
D. Chen et al. / Journal of Nuclear Materials 455 (2014) 436–439
439
5. Conclusions The effect of thermal aging (at 475 °C, up to 5000 h) on Fe–15Cr ferritic alloys has been investigated. The following results were obtained:
Fig. 7. The relation between the calculated increment in yield stress Drcal: and the y measured increment in yield stress Drmea: in Fe–Cr alloy. y
is similar to the estimated value of Cr precipitates estimated in the resent MD work [22]. The relationships between the calculated increment in yield stress, Drcal: and the measured increment in y yield stress, Drmea: , are shown in Fig. 7. Precipitation hardening y was often interpreted in terms of the Orowan type dispersed barrier hardening model, although no dislocation loop was left surrounding precipitates. In this work, however, Orowan type model does not reproduce the monotonic increase in the hardness with increasing aging time. The same occurs with the FKH model, which is considered to be applied to rather weak obstacles for dislocation motion. Although no linear relationship is observed for any of the models, both Orowan type model and SB model can be applied to explain the age-hardening caused by a–a0 phase separation after longer aging time (5000 h). More detailed experimental works are necessary to clear the hardening mechanism. Another explanation can be given that the strengthening factor depends on the size of a0 particles. When the strengthening factor increases with the size of the a0 particles, a linear dependence can be observed for the relation between Drcal: and Drmea: in Fig.7. y y 4.2. Aging hardening evolution As reported in the previous research [14], during short time aging (aging time up to 1000 h), carbon and other alloying elements play a role in the hardening by the formation of carbides and potential segregation of alloying elements on grain boundaries. The addition of carbon leads to some amount of carbides formed nearby grain boundaries, which can influence the stability of ferrite by increasing the probabilities of nucleation and growth of a0 precipitates. The faster age-hardening rate in Fe–Cr–C alloys indicates the addition of alloying elements may affect the kinetics of a–a0 phase decomposition. However, the addition of alloying elements has a small effect on the total age-hardening; the Cr-rich a0 particles dominate hardening of Fe–15Cr alloys after aging beyond 1000 h at 475 °C.
(1) Thermal aging causes a significant increase in Vickers hardness, yield stress and ultimate tensile stress, with a concomitant decrease of total elongation. No remarkable effect of addition of the alloying elements, Mn, Si and Ni and impurity C was observed. (2) Twice larger hardening is required to cause the same amount of reduction in the total elongation for Fe–15Cr–Xs than the other alloys, suggesting that the alloy has a lower susceptibility to age-hardening embrittlement. (3) Based on the TEM observation, the age-hardening is mainly due to the formation of Cr-rich a0 precipitates. Although the addition of minor elements may affect the kinetics of a–a0 phase decomposition, it has a small effect on total agehardening. (4) The age-hardening in Fe–Cr alloy was interpreted in terms of the dispersed barrier models. Both Orowan type and SB models can be applied to explain the age-hardening caused by a–a0 phase decomposition after longer aging time. (5) It is considered that the strengthening factor may increase with increasing the size of a0 precipitates.
Acknowledgment The authors would like to extend gratitude to Dr. Chonghong Zhang, visiting professor of Institute of Advanced Energy, Kyoto University, for his assistance of the TEM observations. References [1] A. Kimura, R. Kasada, A. Kohyama, H. Tanigawa, J. Nucl. Mater. 367–370 (2007) 60–67. [2] S. Ukai, M. Fujiwara, J. Nucl. Mater. 307–311 (1) (2002) 749–757. [3] S. Ukai, T. Nishida, T. Okuda, T. Yoshitake, J. Nucl. Sci. Technol. 35 (4) (1998) 294–300. [4] T. Yoshitake, T. Ohmori, S. Miyakawa, J. Nucl. Mater. 307–311 (2002) 788. [5] G.R. Odette, M.J. Alinger, B.D. Wirth, Ann. Rev. Mater. Res. 38 (2008) 471–503. [6] H.S. Cho, A. Kimura, S. Ukai, J. Nucl. Mater. 329–333 (2004) 387. [7] P.J. Grobner, R.F. Steigerwald, J. Metals 28 (1977) 17–23. [8] F. Danoix, P. Auger, Mater. Charact. 44 (2000) 177–201. [9] H. Peter, B. Saeed, P. liu, O. Joakim, Mater. Sci. Eng. A 534 (2012) 552–556. [10] J.W. Cahn, Acta Metall. 9 (1961) 795–801. [11] R.O. Williams, Trans. TMS-AIME 212 (1958) 497–502. [12] G. Bonny, D. Terentyev, L. Malerba, J. Nucl. Mater. 385 (2009) 278–283. [13] M. Courtnall, F.B. Pickering, Metal Sci. 10 (1976) 273–276. [14] D. Chen, A. Kimura, C. Zhang, W. Han, in: Proc. 8th PRICM. Symp. C, 2013, pp. 529–536. [15] J. Ribis, S. Lozano, Mater Lett. 74 (2012) 143–146. [16] S.M. Dubiel, J. Zukrowski, Mater. Chem. Phys. 141 (2013) 18–21. [17] S.M. Dubiel, J. Zukrowski, Acta Mater. 61 (2013) 6207–6212. [18] G.E. Lucas, J. Nucl. Mater. 206 (1993) 287–305. [19] J. Friedel, Philos. Mag. 46 (1995) 1169. [20] K. Suganuma, H. Kayano, J. Nucl. Mater. 118 (1983) 234–241. [21] R.O. Scattergood, D.J. Bacon, Acta Metall. 30 (1982) 1665. [22] S.M. Hafez Haghighat, R. Schaubin, D. Raabe, Acta Mater. 64 (2014) 24–32.
Contents lists available at ScienceDirect
Journal of Nuclear Materials journal homepage: www.elsevier.com/locate/jnucmat
Correlation of Fe/Cr phase decomposition process and age-hardening in Fe–15Cr ferritic alloys Dongsheng Chen a,⇑, Akihiko Kimura b, Wentuo Han b a b
Graduate School of Energy Science, Kyoto University, Gokasho, Uji, Kyoto 611-0011, Japan Institute of Advanced Energy, Kyoto University, Gokasho, Uji, Kyoto 611-0011, Japan
a r t i c l e
i n f o
Article history: Available online 7 August 2014
a b s t r a c t The effects of thermal aging on the microstructure and mechanical properties of Fe–15Cr ferritic model alloys were investigated by TEM examinations, micro-hardness measurements and tensile tests. The materials used in this work were Fe–15Cr, Fe–15Cr–C and Fe–15Cr–X alloys, where X refers to Si, Mn and Ni to simulate a pressure vessel steel. Specimens were isothermally aged at 475 °C up to 5000 h. Thermal aging causes a significant increase in the hardness and strength. An almost twice larger hardening is required for embrittlement of Fe–15Cr–X relative to Fe–15Cr. The age-hardening is mainly due to the formation of Cr-rich a0 precipitates, while the addition of minor elements has a small effect on the saturation level of age-hardening. The correlation of phase decomposition process and age-hardening in Fe–15Cr alloy was interpreted by dispersion strengthened models. Ó 2014 Elsevier B.V. All rights reserved.
1. Introduction High-Cr ferritic steels have been considered to be candidates for structure material of nuclear power plants as well as advanced fusion blankets. One of the critical issues in the use of the ferritic steels is the 475 °C aging embrittlement [1–5] that is well known to occur in ferritic, dual phase and martensitic stainless steels with Cr contents higher than 12 wt.% [6–10]. The age-hardening is caused by the decomposition of the ferrite phase to chromium-rich phase, a0 , and iron-rich phase, a, in the temperature range of 280– 500 °C. The a0 precipitation leads to a progressive hardening and deterioration of fracture toughness. Williams [11] carried out Mössbauer measurements on the effect of 475 °C aging in a series of iron-chromium binary alloys with a varying the chromium content, and concluded that the alloys with chromium content 12–30 at.% decomposed via nucleation and growth mechanisms. An atomistic study on the hardening behavior of iron-chromium alloys under thermal aging has been done by Bonny et al. [12] showing the different stages of a–a0 unmixing. No study have been done to discuss the a0 size dependence of strengthening factor for evaluation of the hardening. Courtnall and Pickering [13] found that the embrittlement was enhanced by addition of an interstitial element. However, only a limited data is available on the effects of carbon on the phase ⇑ Corresponding author. Tel.: +81 774 38 3478; fax: +81 774 38 3479. E-mail address: [email protected] (D. Chen). http://dx.doi.org/10.1016/j.jnucmat.2014.07.069 0022-3115/Ó 2014 Elsevier B.V. All rights reserved.
decomposition process in ferritic steels during thermal aging. Our preliminary research [14] showed that the early hardening was also attributed to the precipitations of carbides. In this work, the age-hardening process is correlated with the evolution of phase decomposition with focusing on the dispersion model, and effects of alloying on the age-hardening of Fe–15Cr ferritic alloys were also investigated. 2. Experimental The materials used in this work are Fe–15Cr, Fe–15Cr–C and Fe–15Cr–Xs alloys, where Xs refers to Si, Mn and Ni to simulate a pressure vessel steel. The chemical compositions of the alloys are shown in Table 1. The model alloys were prepared using high purity base metals (>99.995%) by arc melting method in argon atmosphere. The alloy button ingots were homogenized at 1100 °C for 72 h followed by furnace cooling and then cold rolling to 80% at room temperature. All specimens were punched out from each alloy sheet with 0.3 mm thick, and annealed at 900 °C for 2 h followed by ice water quenching. Some of the specimens were sealed in quartz tube in high vacuum conditions (104 Torr) and isothermally aged at 475 °C for up to 5000 h. Micro-hardness measurements and tensile tests were carried out before and after thermal aging. Specimens for hardness measurements were mechanically grinded and polished. Vickers hardness was measured with a 1 kg load. At least ten measurements were made and averaged. The dimensions of tensile test specimens are 16 mm 4 mm 0.25 mm. Room temperature tensile
D. Chen et al. / Journal of Nuclear Materials 455 (2014) 436–439
437
Table 1 Chemical compositions of the alloys (wt.%). Materials
Cr
C
Mn
Si
Ni
Fe
Fe–Cr Fe–Cr–C Fe–Cr–Xs
14.55 14.83 14.96
0.001 0.022 0.002
– – 0.83
– – 0.51
– – 0.49
Bal. Bal. Bal.
tests were carried out by using an INSTRON testing machine with a crosshead velocity of 0.2 mm/min. Disk-type specimens of 3 mm diameter were punched out from a 0.3 mm thick plate and mechanically grounded to about 50 lm in thickness. Final TEM specimens were prepared by electrolytic polishing in a 10 vol.% perchloric acid and 90 vol.% acetic acid using a twin-jet polisher at a voltage of 20 V at room temperature. The size and number density of a0 particles were measured to investigate the effect of a0 phase on the age-hardening behavior. The chemical compositions of precipitates were measured using a TEM JEM2200FS equipped with an energy-dispersive X-ray spectroscopy (EDS) system. 3. Results 3.1. Vickers hardness measurement The Vickers hardness of the model alloys after thermal aging treatment at 475 °C is shown in Fig. 1. As aging time increases up to 5000 h, the Vickers hardness values increase abruptly after an incubation period in all the materials. The increments of the hardness, DHV, in all the alloys are also shown in Fig. 1. For aging times up to 2000 h, larger hardness values were observed in Fe–Cr and Fe–Cr–X than Fe–Cr, with the largest difference (about 40 HV) at an aging time of around 1000 h. However, after aging for 5000 h, the DHV of each alloy became similar with an average increment about 120 HV.
Fig. 2. Yield stress (YS), ultimate tensile stress (UTS) and elongation as a function of aging time of each alloy.
3.2. Tensile properties The changes in the tensile properties is summarized in Fig. 2, showing a significant increase in both yield stress (YS) and ultimate tensile stress (UTS), with a concomitant decrease of elongation. In Fe–Cr, the yield stress increased from 197 MPa to 398 MPa after 5000 h aging. The early stage of aging (up to 2000 h) leads to about two thirds increment in yield stress, while the increasing rate of yield stress becomes slower for longer aging time. Linear relations
Fig. 3. Linear relations between the change in yield stress, Dr, and the change in elongation, De, of each alloy after aging for 500, 2000 and 5000 h.
between the change in yield stress, Dr, and the change in elongation, De, of each alloy are shown in Fig. 3. The slopes of the plots, Dr/De of each alloy are classified into two groups, one is Fe–Cr and Fe–Cr–C and the other is Fe–Cr–Xs. It is considered that the ductility loss of Fe–Cr–Xs needs a larger age-hardening than Fe– Cr and Fe–Cr–C. 3.3. Microstructural features
Fig. 1. Vickers hardness (HV) and increment in Vickers hardness (DHV) as a function of aging time in each model alloy, thermally aged at 475 °C.
Phase decomposition of the ferrite to a0 /a phases is identified by TEM EDS line analysis, as shown in Fig. 4, which shows the composition fluctuation of Cr in the matrix. The fluctuation amplitude of
438
D. Chen et al. / Journal of Nuclear Materials 455 (2014) 436–439
Fig. 4. FE-TEM EDS line analysis of Fe and Cr in Fe–Cr alloy after aging for 5000 h, bright field images, finely-dispersed, nanometer size a0 particles in matrix.
Fig. 6. Average diameter and number density of a0 particles as a function of aging time in Fe–Cr alloy.
calculated the activation energy of the phase decomposition in Fe– 14Cr alloy, which is about 122 kJ mol1 [17]. The nucleation of a0 particle thus requires a certain thermal aging time to gain enough nucleation energy. This agrees with the present results of mechanical property measurement because there is almost no change of Vickers hardness and yield stress until 500 h aging. Thermal aging for 2000 h causes the formation of quite a large number of a0 particles in matrix, with an average diameter of 5.2 nm and a number density of 4.9 1022 m3. After 5000 h aging, however, the a0 precipitates coarsen to an average diameter of 8.3 nm, and the number density of a0 particles decreases, to a number density of 3.8 1022 m3. With increasing aging time, Vickers hardness and yield stress increase due to the a/a0 phase decomposition. The increase in yield stress can be given by the following equations as a function of the diameter and number density of a0 particles:
Dry ¼ MaGbðNdÞ Fig. 5. Bright field transmission electron micrographs, showing a microstructure evolution of Fe–Cr alloy (a) before aging, (b) aged for 500 h, (c) aged for 2000 h and (d) aged for 5000 h, black dots indicate Cr-rich a0 particles.
a0 particles is about 8 wt.%, with a minimum value of 11 wt.% and maximum 19 wt.%. The Cr-rich a0 precipitates, which have a lattice parameter aCr = 0.288 nm, are found to be coherent body-centered cubic (bcc) particles [15]. As is shown in Fig. 4, finely-dispersed, nanometer size a0 particles can be observed after 5000 h aging. The microstructure evolution of Fe–Cr alloy during thermal aging is presented in Fig. 5. Only a few a0 particles are found after 500 h aging, while larger size and higher number density of a0 particles are observed as aging time increases. 4. Discussion 4.1. Correlation of microstructure and mechanical properties Phase decomposition process can be described in terms of the average diameter, d, and number density, N, of a0 particles. Fig. 6 shows the average diameter and number density of a0 particles as a function of aging time, which were estimated from TEM observations. The underlying phase decomposition process of Fe–15Cr alloy is nucleation and growth of a0 precipitates. This mechanism agrees with a Mëssbauer spectroscopy study by Dubiel and Zukrowski [16], who reported that nucleation and growth mechanisms worked in Fe–15 at.% Cr alloy during annealing at 415 °C. They also
1=2
2=3
Dry ¼ aMGbðNÞ
ð1Þ
d
ð2Þ
Eq. (1) is based on the dispersed barrier hardening model [18] and Eq. (2) is on the Friedel–Kroupa–Hirsch (FKH) model [19]. In the two equations, ry is the increment in yield stress, M is Taylor factor (3.06), a is strength factor, G is the shear modulus (75.3 GPa), b is the Burgers vector (0.248 nm). In using Eq. (1), a is assumed to be varying from 0.11 to 1 depending on barrier type [18]. While, in using Eq. (2), a is assumed to be 45 for strong obstacles and 18 for small dislocation loops [20]. Both models can be applied for weak precipitates. Another model for impenetrable obstacles is an expression derived from continuum model simulations by Scattergood and Bacon (SB) [21]:
"
Dry ¼
lb ln 2pL
1 R1 L
þ Rd1
!
# þ 0:7
ð3Þ
where R1 is cut-off radius of the dislocation core (taken to be equal to b) and L is the free passage distance L = Ly 2RPRP, RPRP is radius of precipitates). However, it should be only applied for reaction satisfying the Orowan bypass model (the presence of a screw dislocation dipole drawn out from the obstacle). As for strength factor, a, of obstacles, a molecular dynamics (MD) simulation work [22] provided those values of the coherent Cr precipitate of 2 nm diameter, dislocation loops and voids as a function of temperature in bcc Fe, and the obtained values were in the range 0.17–0.37 at 300 K. In this research, the strength factor a of Cr precipitates at room temperature is assumed to be 0.2 that
D. Chen et al. / Journal of Nuclear Materials 455 (2014) 436–439
439
5. Conclusions The effect of thermal aging (at 475 °C, up to 5000 h) on Fe–15Cr ferritic alloys has been investigated. The following results were obtained:
Fig. 7. The relation between the calculated increment in yield stress Drcal: and the y measured increment in yield stress Drmea: in Fe–Cr alloy. y
is similar to the estimated value of Cr precipitates estimated in the resent MD work [22]. The relationships between the calculated increment in yield stress, Drcal: and the measured increment in y yield stress, Drmea: , are shown in Fig. 7. Precipitation hardening y was often interpreted in terms of the Orowan type dispersed barrier hardening model, although no dislocation loop was left surrounding precipitates. In this work, however, Orowan type model does not reproduce the monotonic increase in the hardness with increasing aging time. The same occurs with the FKH model, which is considered to be applied to rather weak obstacles for dislocation motion. Although no linear relationship is observed for any of the models, both Orowan type model and SB model can be applied to explain the age-hardening caused by a–a0 phase separation after longer aging time (5000 h). More detailed experimental works are necessary to clear the hardening mechanism. Another explanation can be given that the strengthening factor depends on the size of a0 particles. When the strengthening factor increases with the size of the a0 particles, a linear dependence can be observed for the relation between Drcal: and Drmea: in Fig.7. y y 4.2. Aging hardening evolution As reported in the previous research [14], during short time aging (aging time up to 1000 h), carbon and other alloying elements play a role in the hardening by the formation of carbides and potential segregation of alloying elements on grain boundaries. The addition of carbon leads to some amount of carbides formed nearby grain boundaries, which can influence the stability of ferrite by increasing the probabilities of nucleation and growth of a0 precipitates. The faster age-hardening rate in Fe–Cr–C alloys indicates the addition of alloying elements may affect the kinetics of a–a0 phase decomposition. However, the addition of alloying elements has a small effect on the total age-hardening; the Cr-rich a0 particles dominate hardening of Fe–15Cr alloys after aging beyond 1000 h at 475 °C.
(1) Thermal aging causes a significant increase in Vickers hardness, yield stress and ultimate tensile stress, with a concomitant decrease of total elongation. No remarkable effect of addition of the alloying elements, Mn, Si and Ni and impurity C was observed. (2) Twice larger hardening is required to cause the same amount of reduction in the total elongation for Fe–15Cr–Xs than the other alloys, suggesting that the alloy has a lower susceptibility to age-hardening embrittlement. (3) Based on the TEM observation, the age-hardening is mainly due to the formation of Cr-rich a0 precipitates. Although the addition of minor elements may affect the kinetics of a–a0 phase decomposition, it has a small effect on total agehardening. (4) The age-hardening in Fe–Cr alloy was interpreted in terms of the dispersed barrier models. Both Orowan type and SB models can be applied to explain the age-hardening caused by a–a0 phase decomposition after longer aging time. (5) It is considered that the strengthening factor may increase with increasing the size of a0 precipitates.
Acknowledgment The authors would like to extend gratitude to Dr. Chonghong Zhang, visiting professor of Institute of Advanced Energy, Kyoto University, for his assistance of the TEM observations. References [1] A. Kimura, R. Kasada, A. Kohyama, H. Tanigawa, J. Nucl. Mater. 367–370 (2007) 60–67. [2] S. Ukai, M. Fujiwara, J. Nucl. Mater. 307–311 (1) (2002) 749–757. [3] S. Ukai, T. Nishida, T. Okuda, T. Yoshitake, J. Nucl. Sci. Technol. 35 (4) (1998) 294–300. [4] T. Yoshitake, T. Ohmori, S. Miyakawa, J. Nucl. Mater. 307–311 (2002) 788. [5] G.R. Odette, M.J. Alinger, B.D. Wirth, Ann. Rev. Mater. Res. 38 (2008) 471–503. [6] H.S. Cho, A. Kimura, S. Ukai, J. Nucl. Mater. 329–333 (2004) 387. [7] P.J. Grobner, R.F. Steigerwald, J. Metals 28 (1977) 17–23. [8] F. Danoix, P. Auger, Mater. Charact. 44 (2000) 177–201. [9] H. Peter, B. Saeed, P. liu, O. Joakim, Mater. Sci. Eng. A 534 (2012) 552–556. [10] J.W. Cahn, Acta Metall. 9 (1961) 795–801. [11] R.O. Williams, Trans. TMS-AIME 212 (1958) 497–502. [12] G. Bonny, D. Terentyev, L. Malerba, J. Nucl. Mater. 385 (2009) 278–283. [13] M. Courtnall, F.B. Pickering, Metal Sci. 10 (1976) 273–276. [14] D. Chen, A. Kimura, C. Zhang, W. Han, in: Proc. 8th PRICM. Symp. C, 2013, pp. 529–536. [15] J. Ribis, S. Lozano, Mater Lett. 74 (2012) 143–146. [16] S.M. Dubiel, J. Zukrowski, Mater. Chem. Phys. 141 (2013) 18–21. [17] S.M. Dubiel, J. Zukrowski, Acta Mater. 61 (2013) 6207–6212. [18] G.E. Lucas, J. Nucl. Mater. 206 (1993) 287–305. [19] J. Friedel, Philos. Mag. 46 (1995) 1169. [20] K. Suganuma, H. Kayano, J. Nucl. Mater. 118 (1983) 234–241. [21] R.O. Scattergood, D.J. Bacon, Acta Metall. 30 (1982) 1665. [22] S.M. Hafez Haghighat, R. Schaubin, D. Raabe, Acta Mater. 64 (2014) 24–32.