Materials and Design 89 (2016) 759–769

Contents lists available at ScienceDirect

Materials and Design

journal homepage: www.elsevier.com/locate/jmad

Influence of thermo-mechanical embrittlement processing on

microstructure and mechanical behavior of a pressure vessel steel
Xue Bai a,b, Sujun Wu a,⁎, Peter K. Liaw b
a
School of Materials Science and Engineering, Beihang University, Beijing 100191, P.R. China
b
Department of Materials Science and Engineering, The University of Tennessee, Knoxville, TN 37996, USA

a r t i c l e i n f o a b s t r a c t

Article history: A thermo-mechanical embrittlement processing (TMEP) consisting of thermal aging and cold strain could cause
Received 11 June 2015 the deterioration of reactor pressure vessel (RPV) steels in the form of an increase in the ductile to brittle transi-
Received in revised form 30 September 2015 tion temperature (DBTT) and a decrease in the upper-shelf energy (USE). In this study, the TMEP was employed
Accepted 5 October 2015
to investigate the microstructure and the evolution of mechanical behavior of the SA508-IV pressure vessel steel.
Available online 9 October 2015
In the microstructure of the as-received state, Cr and Mn atoms replace Fe atoms and form alloying cementites,
Keywords:
(Fe,Cr)3C and (Fe,Mn)3C, through in-situ nucleation. Due to the slower diffusion coefficient, Cr precipitates in the
SA508-IV outer layer of the Mn clusters. In the subsequent embrittlement process, needle-shaped Mo2C, fine copper-rich
Thermo-mechanical embrittlement processing precipitates (CRPs) and P-rich precipitates are formed, which play a great role in the mechanical behavior evolu-
(TMEP) tion. Mechanical test results show that a series of changes in mechanical behavior occurred. It has been found that
Precipitates DBTT fits a linear function of the square root of embrittling time at 520 °C (t 1/2) from 10 h to 90 h degradation and
Reactor pressure vessel steel the degree of embrittlement reaches saturation after 90 h.
Ductile to brittle transition temperature (DBTT) © 2015 Elsevier Ltd. All rights reserved.

1. Introduction
A main objective for the nuclear industry is to enhance the safety
and the reliability of nuclear power plants and to extend their operation
lifetime. The reactor pressure vessel (RPV) steel plays the most impor-
tant role in safety rankings of the whole nuclear power plant system.
The SA508-IV Ni–Cr–Mo low alloy steel has excellent mechanical be-
havior [1], which possesses higher strength and fracture toughness by
adding Ni and Cr elements [2]. Neutron irradiation has been considered
as the main factor responsible for the embrittlement of RPVs [3]. As the
experiments under neutron irradiation are not easy due to the radioac-
tivity, the thermo-mechanical embrittlement processing (TMEP) meth-
od has been chosen to study the performance of RPVs.
Neutron irradiation is generally known as a major reason for the em-
brittlement of RPV steels because of the damage produced in collision
cascades consisting of self-interstitial atoms and vacancies. The neutron
irradiation embrittlement of RPV steels is known to be related to the
presence of residual elements [4], such as Phosphorus, which can also
be obtained from long-term tempering [5,6]. Copper-rich precipitates
(CRPs) are also considered as the reason for irradiation embrittlement
[7]. In the temperature range of 300–500 °C, temper embrittlement
has been considered as the main factor responsible for the loss of
⁎ Corresponding author.
E-mail address: wusj@buaa.edu.cn (S. Wu).
toughness and the premature failure of steels [8,9], which is a non-
hardening embrittlement caused by the grain-boundary segregation of
impurity elements, mainly phosphorus [2].
Currently, the nature and mechanisms of irradiation embrittlement,
especially the changes in the microstructure and mechanical behavior of
materials, are the subjects of major studies. However, as neutron irradi-
ation is harmful, the thermo-mechanical process, thermal aging and
cold strain were employed to deteriorate the RPVs to simulate the effect
of neutron irradiation. To be specific, the temper-embrittlement process
could be combined with cold strain, which is called the TMEP procedure.
The heat treatment can produce a larger concentration of vacancies and
vacancy clusters that are supposed to accelerate the formation of CRPs
based on a solute-vacancy exchange diffusion mechanism [10]. After
that, the cold strain may cause the interaction of dislocations and pre-
cipitates, and then the steel could be embrittled by the large numbers
of dislocation loops forming behind them.
In the previous study, the pre-strain followed by aging process was
applied to study the microstructure evolution of RPV steels. Ghosh
et al. [11] have found that the 50% pre-strain prior to aging could trigger
the precipitation by decreasing the activation energy values for Cu pre-
cipitation from 182 to 126 kJ/mol. Kamada et al. have conducted a series
of experiments to simulate the irradiation embrittlement of nuclear RPV
steels [12]. The Fe-1 weight percent (wt.%) Cu alloys with and without
pre-deformation in the solid-state solution were thermally aged at
773 K (500 °C) for various times. After that, the evolution of hardness,

http://dx.doi.org/10.1016/j.matdes.2015.10.024
0264-1275/© 2015 Elsevier Ltd. All rights reserved.
760 X. Bai et al. / Materials and Design 89 (2016) 759–769

conductivity, and microstructure were investigated. It has been found

that the pre-deformation enhanced Cu precipitation and caused precip-
itation at dislocations.
Although the microstructure evolution of the above methods was
close to the results of the neutron irradiation state, the mechanical
tests could not obviously show the embrittlement of RPVs. Wu et al.
[13] studied the TMEP process, applying the 8% uniform cold strain fol-
lowing the heat treatment, to simulate the effect of neutron irradiation
on the mechanical properties of a pressure vessel weld metal. The frac-
ture tests were carried out using blunt notch four-point-bend speci-
mens in slow bend over a range of temperatures. The ductile–brittle
transition temperature (DBTT) was shown to increase by approximately
100 °C as a result of the degradation.
In this study, the TMEP was applied to the new type of pressure ves- Fig. 1. Schematic of the degradation process.
sel steel A508-IV. It has been stated by Fatehi et al. [14] that the precip-
itation of Cu in steels is usually observed in the temperature range of
400–650 °C. Thus, the current work was conducted at 520 °C using the 150 kg at 9 positions, and the results are averaged to assure the
TMEP deterioration process. It was focused on studying the embrittle- repeatability.
ment effect of the hardening and grain-boundary P segregation, to com-
pare with the in-service SA508-III and the neutron irradiation 2.3. Microstructure analysis
embrittlement. The microstructure evolution, which leads to the em-
brittlement, will be studied in details and the resultant degradation of After standard grinding and polishing, the samples were etched by 5
the mechanical behavior will also be discussed. volume percent (vol.%) nital. The microstructures were examined using
the optical microscopy (OM) and scanning electron microscopy (SEM).
Transmission electron microscope (TEM) was employed to examine
2. Experimental procedure the fine details of the microstructure and to obtain the diffraction pat-
tern of the selected phases. The TEM specimens were prepared using a
2.1. Testing materials and specimens double-jet electrolytic thinning technique. The images were taken
under both TEM and scanning transmission electron microscope
The chemical composition of the SA508-IV Ni–Cr–Mo low-alloy steel (STEM) mode.
is listed in Table 1. As shown in Fig. 1, the specimens were subjected to a
heat treatment of 1.5 h at 920 °C, followed by water quenching (WQ) to 3. Results and discussion
room temperature (RT), re-heating to 650 °C for 30 h and furnace
cooling to RT. Specimens without the further TMEP treatment are 3.1. Microstructural observation
termed as “the as-received state”. The degradation procedure has been
worked out with aging, followed by the 10% uniform cold strain (Fig. 3.1.1. Undegraded state
1). The cold strain is applied by the compressive stress at the speed of Figs. 3(a) and (b) are SEM micrographs of the undegraded (as-re-
0.5 mm/min using the 100 t computer-controlled electro-hydraulic ceived) samples of SA508-IV. Fig. 3(a) shows that it has an overall mar-
servo universal testing machine. The corresponding specimens are tensitic structure. In Fig. 3(b), it could be found that there are some
termed as “the degraded state”. coarse carbides precipitated along grain boundaries. According to the
The above two states of materials are all fabricated into tensile research of Lee et al. [15], they should be the M3C-type carbides. Then
specimens and Charpy-impact test specimens. The tensile specimens based on the study of Hashimoto et al. [16], the carbides distributed
are bars with a gage length of 25 mm and gage diameter of 5 mm. The di- along grain and/or lath boundaries could be defined as M23C6 and Cr-,
mension of Charpy-impact test specimens is 10 mm × 10 mm × 55 mm Ni-rich M6C. The fine M2C-type carbides are also dispersed inside
with the notch direction transverse to acting surfaces applied by the laths, similar to the results mentioned in Refs. [13] and [14]. In Fig. 4,
compressive stress (Fig. 2). it can be seen that the plate-shaped Cr-rich M23C6 carbides have a diam-
eter of about 142 nm [Fig. 4(a)], and the rod-shaped M6C is about
129 nm in length [Fig. 4(b)]. Besides, a few MC precipitates mainly dis-
2.2. Mechanical tests tribute along dislocations (see Fig. 5).
In the tempering process, the alloying elements diffuse quickly and
Tensile tests at room temperature were carried out using a re-locate inside α-Fe and Fe3C phases. It can be obviously seen from
computer-controlled 50 kN testing machine with a loading rate of 0.5 the STEM distribution diagram of elements (Fig. 6) of the undegraded
mm/min.
Charpy-impact tests were conducted using a 300 J pendulum impact
test machine in a temperature range from −160 °C to RT. The test re-
sults were used to plot ductile to brittle transition temperature (DBTT)
curves.
In order to assess the age-hardening effect, Rockwell C Hardness
(HRC) values were measured for each state of samples under a load of

Table 1
Chemical composition of SA508-IV Ni–Cr–Mo alloy steel in wt.%.

C Si Mn P S Ni Cr Cu Mo Al Fe

0.15 0.36 0.34 0.011 0.008 3.26 1.66 0.041 0.46 0.005 Balance
Fig. 2. Position relationship of the compressive stress and notch location.
 X. Bai et al. / Materials and Design 89 (2016) 759–769
Fig. 3. (a) SEM micrographs of the undegraded state of the SA508-IV RPV steel, showing a martensitic structure; (b) amplified photo of (a). Nital etched.
RPV steel that the precipitates are full of Cr and Mn elements but lack of
Fe and contain a mild quantity of Mo atoms.
In Fig. 6, the diameter of Cr element is measured as 148.24 nm and
that of Mn is 91.67 nm (as arrows indicated in Fig. 6). Thus it could be
considered that Cr elements form the outer layer surrounding the Mn
clusters, which means that Mn atoms move faster than Cr atoms. The re-
sults are consistent with the study of Zhang et al. [17], the diffusion co-
efficient of Mn is greater than that of Cr, DMn N DCr (the element Cr is one
order of magnitude slower than Mn). After Cr and Mn gather near Fe3C,
they will replace Fe atoms to alloy cementites (Fe,Cr)3C and (Fe,Mn)3C
in the manner of the in-situ nucleation. Therefore, there are no Fe
atoms in the positions of Cr and Mn atoms. Furthermore, it can be
seen from Fig. 6 that Mo atoms disperse in the area and do not form
clusters.
3.1.2. Degraded state
The microstructure of the degraded state samples could be found in
Fig. 7. They correspond to the 10 h, 30 h, 90 h, and 180 h-degraded state
samples, respectively. Fig. 7(b), (d), (f), and (h) are the enlarged ones of
(a), (c), (e), and (g), respectively. Comparing Fig. 7(a), (c), (e), and (g),
the grain size of these degraded state samples do not change significant-
ly. However, the size of carbides increases as the thermal aging in-
creases from 10 h to 180 h. Unlike Fig. 3, there are more tiny M2C
precipitates inside the martensite laths in the 10 h-degraded state sam-
ple. In the 30 h-degraded state sample [Fig. 7(d)], the needle-like car-
bides began to precipitate inside the martensitic laths, which is
supposed to be Mo2C [18]. In 90 h-degraded state sample, the size of
the coarse precipitates on grain boundaries and lath boundaries become
Fig. 4. M23C6 precipitates and M6C precipitates on grain and/or lath boundaries of the undegraded state, (a) M23C6 precipitates; (b) M6C precipitates.
761
smaller, the tiny M2C grows up a little [Fig. 7(e) and (f)]. After degraded
for 90 h [Fig. 7(f) and (h)], the coarse M3C carbides join together along
grain boundaries. Besides, the size of M23C6 and M6C carbides goes up,
and the density of M2C carbides increases. Obviously, the size of the
needle-like carbides Mo2C inside martensitic laths begins to grow up
[see in Fig. 7(h)].
Figs. 8, 9, 10, and 11 represent the STEM images of 10 h, 30 h, 90 h,
and 180 h-degraded state samples, respectively. In Fig. 8, the Cr-Mn
rich precipitates in the 10 h state could be found to be smaller compared
with Fig. 6, which represents the undegraded state. It can also be seen
from Fig. 8 that the elements Ni and Mo have precipitated together,
and the distributions of them are the same as phosphorus in 10 h-
degraded specimens. Kameda [18] holds the opinion that the presence
of Mo plays a key part in impeding intergranular P segregation and frac-
ture probably due to the attractive Mo-P interactions in the grain matrix
and a grain-boundary toughening effect of the segregated Mo. As the
aging time grows to 30 h, the attractive Mo-P interactions are dispersed
homogeneously (see Fig. 9). However, the beneficial effect of Mo is di-
minished in long-term aging where the dissolved Mo content decreases
as a result of the precipitation of fine Mo-rich carbides [18], Mo2C, which
are the tiny needle-like carbides precipitate inside the martensitic laths
[19], seen in Fig. 12. However, in Fig. 9, Cr atoms are found to segregate
and form clusters, so Mn atoms precipitate outside.
Mo2C is a member of the interstitial compounds where the transition
metal atoms form a close-packed-hexagonal arrangement, and the car-
bon atoms are found in the octahedral interstices, which has relatively
higher hardness, melting point, and dissolving temperature. Mo2C usu-
ally locates inside the martensite lath or segregates at the sub-
762 X. Bai et al. / Materials and Design 89 (2016) 759–769
Fig. 5. MC precipitates on dislocations of the undegraded state.
boundary. It forms through a nucleation and growth process in a charac-
teristic needle shape [19], which is one of the main strengthening phase
of steels in the temperature range of 500–650 °C.
The equilibrium solid solution of Cu in α-iron at 500 °C can be deter-
mined as 0.0328 wt.% extrapolated from the empirical relationship pro-
posed by Salje and Feller-Kniepmeier [20]. The bulk content of Cu in this
experimental steel is 0.041 wt.%. Therefore, the supersaturated Cu
would precipitate in the form of the CRPs. The calculation results obtain-
ed by Becquart [21] indicate that a strong vacancy affinity for copper
atoms, and slightly faster diffusion of the Cu impurities compared
with Fe. With the aging time goes by, the number of vacancies and
vacancy-clusters will increase rapidly. Diffusion of Cu will be accelerat-
ed by a large number of single vacancies and vacancy-clusters based on
the solute-vacancy exchange diffusion mechanism to form copper-rich
clusters [16]. It can be seen in Fig. 8 that there are plenty of CRPs
(CRPs with a size of 5–10 nm) in the 30 h-degraded RPV steel.
When the aging time rises up to 90 h (Fig. 10), CRPs grow larger, and
the density of CRPs decreases. This phenomenon is consistent with the
results observed by Valo et al. [22]. The reason is that during annealing,
copper precipitates do not massively re-dissolve back into the matrix.
However, they grow larger with the decrease of density.
Fig. 6. Elemental mapping of the undegraded RPV steel characterized by STEM.
The alloying elements Ni and Cr promote P segregation through the
mechanisms of co-segregation and a reduced carbon activity in alloys,
and Ni conversely improves the low temperature cleavage toughness
by decreasing the energy barrier of kink formation [18]. It has been
shown that the addition of Mn and/or Si influences the intergranular
embrittlement in iron, Cr-Mo and Ni–Cr–Mo–V steels with different
plasticity [18]. The content of P affects irradiation hardening more sig-
nificantly than that of Cu due to the distinct formation of P-rich precip-
itates arising from the stabilization of vacancies [23]. Therefore, it can be
expected that the TMEP hardening will be affected by the phosphorus
concentration on the grain boundary.
In Fig. 10, it could be found that there are no phosphorus precipi-
tates. Thus, it could be concluded that the phosphorus has completely
re-dissolved back into the matrix [22] after long-term degradation,
which is different from the evolution of copper precipitates. What's
more, there is no large difference between 180 h's aging and 90 h's
aging (Figs. 10 and 11).
3.2. Mechanical behavior
3.2.1. Tensile tests
Tensile test results are listed in Table 2, which has been plotted in
Fig. 13(a). It could be found that the yield strengths (σy) and tensile
strengths (σb) firstly increase and then decrease slightly with the degra-
dation time, showing peak values of 803 MPa and 866 MPa at 90 h's
aging, respectively. As for the elongation, the undegraded-state samples
possess good elongation of over 20%. However, with the degradation
time increases, elongation values fluctuate between 11.3%–14.7%. The
strain-hardening exponent, n, tends to decrease suddenly from the
undegraded state to the 10 h-degraded state, then rise up to a relatively
higher level at 30 h, and drop gradually from 30 h to 180 h.
As for the shift of strength, it can be seen in Fig. 13(a) that the yield
strength increases and the ductility decreases with the increase of deg-
radation time from 10 h to 90 h in the degradation process. This trend
can be attributed to the dislocation cells and distorted martensitic lath
in the degraded material, the aging, as well as the cold-worked material
[24]. From the degradation of 10 h to 30 h, there are large numbers of
nanoscale precipitates in the specimens, such as CRPs, phosphorus-
rich precipitates, and Mo2Cs. According to the formula,
τ ¼ Gb=L ð1Þ
where τ is the shear stress of particles placed upon dislocations, G is
shear modulus, b is Burgers vector, L is the distance between particles,
the material could be strengthened because of the close distance be-
tween precipitates.
It should be noticed that there is a turning point at 90 h in the tensile
strength versus degradation time curve, as seen in Fig. 13(a). The
 X. Bai et al. / Materials and Design 89 (2016) 759–769 763

Fig. 7. The microstructure of the 10 h, 30 h, 90 h and 180 h degraded state samples.

specimens degraded for 90 h possess the highest yield strength,
803 MPa, and relatively higher ductility, 14.7% elongation (Table 2).
After being degraded for 180 h, the precipitates become coarse, which
would weaken the strengthening effect. Otherwise, the softening be-
havior of the 180 h-degraded samples could also be attributed to the
partial annihilation of dislocations produced during quenching and/or
cold straining and tempering of martensite laths (the decomposition
of hard martensites), which over-balances the hardening feature due
to the Cu precipitation, as well as fine M2C carbides and the needle-
like Mo2C carbides that precipitate inside martensite laths. Besides,
764 X. Bai et al. / Materials and Design 89 (2016) 759–769

Fig. 8. Elemental mapping of the 10 h-degraded RPV steel characterized by STEM.

Fig. 9. Elemental mapping of the 30 h-degraded RPV steel characterized by STEM.

the coherent to incoherent transformation, the coarsening characteris-
tics of the nanoscale particles (mainly the CRPs) [14], M23C6 and M6C
carbides, which distributed along grain and/or lath boundaries, the coa-
lescence of the coarse M3C precipitates along grain boundaries, can all
lead to the slightly drop of the strength of the RPV steel. It has been re-
ported that [12] nano-size defect clusters can accumulate near
dislocations because of the mutual elastic interactions. The dislocations
could be seen in the 90 h-degraded state specimens (Fig. 14). After
being degraded for more than 90 h, there is enough time for the defects
to diffuse. In this case, the stress field of a dislocation that initially exists
in the RPV steel would be compensated for by the stress field of these
defects. Therefore, the strength would decrease after 90 h.

Fig. 10. Elemental mapping of the 90 h-degraded RPV steel characterized by STEM.
 X. Bai et al. / Materials and Design 89 (2016) 759–769 765

Fig. 11. Elemental mapping of the 180 h-degraded RPV steel characterized by STEM.

It has been found that the strengthening process is highly complex,
and there are various sources of strengthening combined to produce
the total flow stress, σy [25],
σ y ¼ σ 0 þ Δσ ss þ Δ σ disl þ Δσ gr þ Δ σ part
where the quantity, σ0, is the temperature and strain-rate-sensitive fric-
tional stress from the inherent lattice resistance to slip, Δσss is the solid-
solution-hardening effect increasing with the content of dissolved for-
eign atoms, Δσdisl represents the effect due to the long-range stress
fields of the dislocations and is proportional to the square root of the
dislocation density, Δσgr is the well-known Hall–Petch grain strength-
ening, and Δσpart is the contribution of particles, precipitates, or similar
structural defects. The last one mainly results from the carbides disper-
sion. The irradiation does not change the inherent lattice resistance, the
grain size, the dislocation density, and the carbide dispersion. It prefer-
entially produces a new dispersion of dislocation obstacles, copper-rich
precipitates, which are termed as irradiation defects. Therefore, the last
parameter Δσpart should be paid attention to in the neutron irradiation
embrittlement study.
Correspondingly, in the TMEP, the dislocation density and carbides
dispersion have changed. Specifically, the heat treatment produced
some precipitates, such as ε-Cu and Mo2C precipitates. The cold-strain
process forces the dislocations to move and then form many dislocation
loops behind, which leads to the embrittlement of the specimens. The
ð2Þ
TMEP in this study contributes to the changes of the parameters
Δσpart and Δσdisl in Eq. 3. Therefore, it can be used to forecast the perfor-
mance of RPV steels under real neutron irradiation and screen the ones
which possess the superior neutron irradiation embrittlement
resistance.
3.2.2. Hardness tests
The HRC hardness test results are listed in Table 3. The variation
trend can be seen in Fig. 13(b). It is evident that the hardness values
are in the direct proportion to the duration of the degradation time at
first, and then saturate gently after 30 h.
The first rapid increase stage of hardness values between 0 h-30 h
degraded states [see Fig. 13(b)] is caused by several reasons [24],
(a) The gradual precipitation of nano-scale CRPs and needle-like
Mo2Cs increases the hardness. As previously mentioned, Mo2C is one
of the mainly strengthening phases in steels in the temperature range
of 500–650 °C [19].
(b) It can be stated that the interaction of P with vacancies, and the
P-rich precipitation or clustering are responsible for the precipitation
hardening in the P-doped steels.
(c) The hardening behavior may arise from the increased density of
dislocation-pinning sites, which is the same as in the highly irradiated
steel.
(d) It is well known that Ni has a strong effect on hardening, arising
from the formation of Cu-rich clusters made up of a Ni shell and Cu core.
In addition, Ni increases the density of interstitial-type dislocation
loops, which increases the hardness.

Table 2
Tensile test results of the different degradation state of SA508-IV steel.

Sample Yield Tensile Elongation, n

condition strength, strength, δ (%)
σy (MPa) σb (MPa)

0 h (as-received) 509 622 22.2 0.161/0.165

10 h 743 777 12.9 0.142/0.148
30 h 769 812 11.3 0.159/0.154
90 h 803 866 14.7 0.144/0.133
180 h 797 854 13.0 0.084/0.100
Fig. 12. Mo2C precipitates in the 30 h-degraded RPV steel.
766 X. Bai et al. / Materials and Design 89 (2016) 759–769

Table 3
Mechanical test results of the different degradation state of the SA508-IV steel.

Samples σy/MPa Δσy/MPa DBTT41J/°C ΔDBTT41J/°C HRC ΔHRC

0 h (as-received) 509 – −81 – 30.93 –

10 h 743 234 −68 13 35.05 4.12
30 h 769 260 −53 28 40.67 9.74
90 h 803 294 −42.5 38.5 41.13 10.2
180 h 797 288 −42 39 41.71 10.78

where Δσy = σy (TMEP) − σy (as-received).

ΔDBTT 41J = DBTT 41J (TMEP) − DBTT 41J (as-received).
ΔHRC = HRC (TMEP) − HRC (as-received).

deformation and tempering of martensite laths (the decomposition of

hard martensites); (b) aging hardening due to Cu [26], clustering of
Cu atoms and coarsening of Cu precipitates, respectively; (c) P-rich pre-
cipitates or clusters re-dissolve back into the matrix after aging for a
long time. Therefore, the hardness values would not go up dramatically
at the initial stage.

3.2.3. Charpy-impact tests

In this study, the degraded specimens, which are supposed to simu-
late the irradiation effect, have been used to conduct impact testing. The
testing results are listed in Table 3. Fig. 15 reveals the variation of DBTT
under different degradation states. The DBTT curve of the as-received
state samples can be expressed as,

Impact Absorbed Energy ¼ 92:79 þ 83:09

 tanh½ðT þ 43:54Þ=52:10: ð3Þ

In this study, the temperature corresponding to the impact test ener-

Fig. 13. The variation of σy, HRC, DBTT and USE upon the degradation time.
The second gentle increase stage can be attributed to the combined ef-
fects of the concomitant softening (recovery) and age-hardening (pre-
cipitation) reactions taking place during the aging process. The factors
which cause this phenomenon are listed below: (a) the partial annihila-
tion of dislocations produced during quenching and/or cold
gy of 41 J has been used as the DBTT. From the intercepts of the DBTT
curves with the horizontal line, it can be concluded that the DBTT
keeps on increasing from the 0 h (the as-received state) to the 90 h-
degraded state. Fig. 16 shows the morphology of the fracture surfaces
of both the undegraded state [Fig. 16 (a)] and 90 h-degraded state
[Fig. 16 (b)] specimens. The impact absorbed energy at − 40 °C is
103.39 J for the undegraded specimen and 36.09 J for the 90 h-
degraded specimen. Compared with Fig. 16 (a), there are much more in-
tergranular facets in Fig. 16 (b), which makes it brittle. However, the
DBTT curves of 90 h and 180 h almost overlap. The variation tendency
of DBTT upon the degradation time could be concluded more clearly
from Fig. 13(c), which increases continuously at the beginning and
then reaches a plateau after 90 h, corresponding to the saturation of

Fig. 14. Dislocations of 90 h-degraded state specimen. Fig. 15. The variation of DBTT under different degradation states.
 X. Bai et al. / Materials and Design 89 (2016) 759–769 767

Fig. 16. Fracture surface of Charpy impact test specimens, (a) undegraded state, −40 °C, 103.39 J; (b) 90 h-degraded state, −40 °C, 36.09 J.

the embrittlement-hardening effect. The upper-shelf energy (USE) numbers of dislocation loops behind. Eventually, the specimens have
drops obviously after being degraded for 10 h, increases a little at 30 h been embrittled, resulting in the results of DBTT-upwards shift.
and decreases after 30 h. Then it keeps constant after 90 h [Fig. 13(d)]. Besides, in Ref. [24], it has been stated that the DBTT shift induced by
(1) Ductile–brittle transition temperature (DBTT). the real neutron irradiation (ΔT 41J) is linearly correlated with ΔCgbP,
Specifically, it can be seen from Fig. 15 that the shift of DBTT, which is that is,
termed as ΔDBTT 41J, is 39 °C after degradation for 180 h. The variation
tendency of DBTT [Fig. 13(c)] can be explained as follows. It is well ΔT 41 J ¼ 63 þ 530 ΔC gb P ð4Þ
known that the phosphorus segregation promotes the initiation of in-
tergranular cracking in the stress-intensified zone accompanied by the where ΔCgbP is the shift of the bulk content of phosphorus segregated at
cleavage and ductile crack propagation. According to the study by grain boundaries.
Druce [27], P segregation was shown to reduce the critical void- The segregation of P on prior austenite boundaries was proved to re-
nucleation strain giving rise to a higher proportion of the second- duce the fracture toughness of steels [28], and the enhanced segregation
phase particle distribution contributing to the fracture process and a of P was then found to be a linear function of the square root of the em-
consequential decrease in the average void diameter and toughness. brittling time at 520 °C, which means,
With the aging time goes from 0 h to 90 h, the dynamic interaction be-
tween vacancies and P near grain boundaries becomes more effective,
ΔC gb P ¼ C  t 1=2 ð5Þ
whereas their interaction in the grain interior is not greatly promoted.
Therefore, the DBTT rises gradually with the degree of the activation
of P. However, after 90 h, the grain boundary segregation of the solute where C is a constant, and t is the embrittling time at 520 °C.
elements reached maximum value and no more extra solute elements From Eqs. (4) and (5), it can be concluded that,
in the matrix can move to grain boundaries with longer tempering
time. Thus, there is no significant difference between the DBTT of 90 h ΔT 41 J ¼ 63 þ 530  C  t 1=2 ð6Þ
samples and that of 180 h. As for the effect of the cold strain of 10% ap-
plied on the specimens, the compressive stress forces the dislocations to Therefore, the DBTT shift induced by neutron irradiation (ΔT 41J) is
move and climb the carbides or precipitates and then leave large linearly correlated with the embrittling time at 520 °C.

Fig. 18. Relationship between ΔDBTT41J and Δσy of the RPV steel SA508-IV after different
Fig. 17. Relationship between DBTT 41J and the square root of embrittling time at 520 °C. degradation procedures.
768 X. Bai et al. / Materials and Design 89 (2016) 759–769
In the TMEP degraded samples, the DBTT41J values versus the square
root of embrittling time at 520 °C (t 1/2) have been plotted in Fig. 17,
from which it can be seen that DBTT fits a linear function of t 1/2 up to
90 h. The result above is consistent with the microstructure of 0 h to
180 h-degraded state, in which case the phosphorus segregation in-
creases gradually at the beginning and reaches saturation at 90 h. There-
fore, a direct relationship could be deduced between the TMEP up to
90 h and the real neutron irradiation embrittlement.
The ΔDBTT 0→180 h of SA508-IV, 39 °C, can also be compared with the
result of that of A533B RPV steel, which is 47.2 °C, studied by Wu and
Cao [5]. The tempering-degradation process in Ref. [5] is that holding
the specimens at 520 °C for 30, 90, and 180 h, respectively, and then
air cooling to room temperature. The fact that the ΔDBTT0→180h of
A533B is higher than that of SA508-IV implies that SA508-IV possesses
better embrittlement resistance than A533B.
(2) Upper-Shelf-Energy (USE)
The values of USE are supposed to keep on decreasing from 0 h
to 180 h degradation process. However, in Fig. 13(d) it can be
seen that the USE drops obviously after degradation for 10 h, in-
creases slightly at 30 h, and decreases after 30 h. Then it keeps con-
stant after 90 h.
The 30 h degradation samples possess the relatively higher USE. It
may be because of the nano-size precipitates and Mo2C strengthening
phase, unlike the coarse precipitates in 10 h, 90 h, and 180 h TMEP
states.
(3) The relationship between ΔDBTT 41J and Δσy.
It has been proved that the relationship between ΔDBTT 41J and Δσy
after the neutron-irradiation embrittlement follows the following equa-
tion [29],
ΔDBTT 41 J ¼ 0:65 Δσ y ð7Þ
However, in the present research, the test results of ΔDBTT 41J
against Δσy after TMEP have been plotted in Fig. 18, it fits the linear re-
lationship as expressed by Eq. (8),
ΔDBTT 41 J ¼ −87 þ 0:43 Δσ y ð8Þ
From which it can be seen that the ΔDBTT 41J of 10 h, 30 h, and 90 h-
degraded state specimens linearly increases with the Δσy, which is sim-
ilar to the neutron irradiated ones. Fig. 18 also shows that the difference
of the embrittlement results after 90 h and 180 h is insignificant which
indicates that the embrittlement effect is saturated after 90 h degrada-
tion and no longer time is necessary.
Based on the above analysis, it can be concluded that the mechanical
property change of the steel after TMEP from 10 h to 90 h is close to that
of the steel resulted from the real neutron irradiation process. Thus, the
neutron irradiation embrittlement resistance of RPV steels could be
studied by the TMEP from 10 h to 90 h instead of the real neutron irra-
diation experiments.
4. Conclusions
Mechanical behavior of RPV steels under neutron irradiation embrit-
tlement has been successfully simulated using this TMEP method. The
mutual effect of precipitates, vacancies, and dislocations in the long-
time degradation process leads to the variation of the mechanical prop-
erties of this steel. It can be concluded as follows:
(1) The microstructure of the undegraded state samples is tempered
martensite with alloy cementites (Fe,Cr)3C and (Fe,Mn)3C in the man-
ner of in-situ nucleation.
(2) Needle-shaped Mo2C, Cu-rich precipitates and P-rich precipi-
tates play a key role in the mechanical properties evolution after the
degradation process. The coalescence of M3C and the coarsening behav-
ior of M23C6 and M6C lead to the deterioration of the mechanical
properties.
(3) DBTT fits a linear function of the square root of embrittling time
at 520 °C (t 1/2) from 10 h to 90 h degradation.
(4) There is a linear relationship between ΔDBTT 41J and Δσy for
specimens after various degradation processes. Thus, the neutron irradi-
ation embrittlement resistance of RPV steels could be studied by the
TMEP instead of the real neutron irradiation experiments.
(5) Test results showed that SA508-IV possesses better embrittle-
ment resistance than the A533B RPV steel.
Acknowledgments
This work is financially supported by the National Science and Tech-
nology Key Specific Project: Life Management Technology of Nuclear
Power Plant (Grant No. 2011ZX06004-002).
References
[1] S.G. Park, M.C. Kim, B.S. Lee, D.M. Wee, Evaluation of temper embrittlement effect
and segregation behaviors on Ni–Mo–Cr high strength low alloy RPV steels with
changing P and Mn contents, Korean J. Met. Mater. 48 (2010) 122.
[2] S.G. Park, K.H. Lee, M.C. Kim, B.S. Lee, Effects of boundary characteristics on resis-
tance to temper embrittlement and segregation behavior of Ni–Cr–Mo low alloy
steel, Mater. Sci. Eng. A 561 (2013) 277–284.
[3] P. Bowen, S.G. Druce, J.F. Knott, Effects of microstructure on cleavage fracture in
pressure vessel steel, Acta Metall. 34 (1986) 1121–1131.
[4] E. Meslin, B. Radiguet, M. Loyer-Prost, Radiation-induced precipitation in a ferritic
model alloy: an experimental and theoretical study, Acta Mater. 61 (2013)
6246–6254.
[5] B. Yang, P.K. Liaw, H. Wang, L. Jiang, J.Y. Huang, R.C. Kuo, J.G. Huang, Thermographic
investigation of the fatigue behavior of reactor pressure vessel steels, Mater. Sci. Eng.
A 314 (2001) 131–139.
[6] S.J. Wu, L.W. Cao, Effect of intergranular failure on the critical fracture stress and the
fracture toughness of degraded reactor pressure vessel steel, Int. J. Press. Vessel. Pip.
101 (2013) 23–29.
[7] A.V. Barashev, S.I. Golubov, D.J. Bacon, P.E.J. Flewitt, T.A. Lewis, Copper precipitation
in Fe–Cu alloys under electron and neutron irradiation, Acta Mater. 52 (2004)
877–886.
[8] C. Lea, M.P. Seah, Site competition in surface segregation, Surf. Sci. 53 (1975)
272–285.
[9] P.K. Liaw, H. Wang, L. Jiang, B. Yang, J.Y. Huang, R.C. Kuo, J.G. Huang, Thermographic
detection of fatigue damage of pressure vessel steels at 1,000 Hz and 20 Hz, Scr.
Mater. 42 (2000) 389–395.
[10] L.W. Cao, S.J. Wu, B. Liu, On the Cu precipitation behavior in the thermos-
mechanically embrittlement processed low copper reactor pressure vessel model
steel, Mater. Des. 47 (2013) 551–556.
[11] S.K. Ghosh, A. Haldar, P.P. Chattopadhyay, Effect of pre-strain on the ageing behavior
of directly quenched copper containing micro-alloyed steel, Mater. Charact. 59
(2008) 1227–1233.
[12] Y. Kamada, S. Takahashi, H. Kikuchi, S. Kobayashi, K. Ara, J. Echigoya, Y. Tozawa, K.
Watanabe, Effect of pre-deformation on the precipitation process and magnetic
properties of Fe–Cu model alloys, J. Mater. Sci. 44 (2009) 949–953.
[13] S.J. Wu, J.F. Knott, Effects of degradation on the mechanical properties and fracture
toughness of a steel pressure-vessel weld metal, Int. J. Press. Vessel. Pip. 80 (2003)
807–815.
[14] A. Fatehi, J. Calvo, A.M. Elwazri, S. Yue, Strengthening of HSLA steels by cool defor-
mation, Mater. Sci. Eng. A 527 (2010) 4233–4240.
[15] S. Lee, S. Kim, B. Hwang, B.S. Lee, C.G. Lee, Effects of carbide distribution on the frac-
ture toughness in the transition temperature region of an SA 508 steel, Acta Mater.
50 (2002) 4755–4762.
[16] N. Hashimoto, R.L. Klueh, Microstructural evolution of nickel-doped 9Cr steels irra-
diated in HFIR, J. Nucl. Mater. 305 (2002) 153–158.
[17] H. Zhang, J.W. Wu, X.B. Liu, A. Baker, Studies on elements diffusion of Mn/Co coated
ferritic stainless steel for solid oxide fuel cell interconnects application, Int. J.
Hydrog. Energy 38 (2013) 5075–5083.
[18] J. Kameda, Y. Nishiyama, Combined effects of phosphorus segregation and partial in-
tergranular fracture on the ductile–brittle transition temperature in structural alloy
steels, Mater. Sci. Eng. A 528 (2011) 3705–3713.
[19] D.J. Dyson, S.R. Keown, D. Raynor, J.A. Whiteman, The orientation relationship and
growth direction of Mo2C in ferrite, Acta Metall. 14 (1966) 867–875.
[20] G. Salje, M. Feller-Kniepmeier, The diffusion and solubility of copper in iron, J. Appl.
Phys. 48 (1977) 1833–1839.
[21] C.S. Becquart, RPV steel microstructure evolution under irradiation: a multiscale ap-
proach, Nucl. Instrum. Meth. B. 228 (2005) 111–121.
[22] M. Valo, L. Debarberis, A. Kryukov, A. Chernobaeva, Copper and phosphorus effect on
residual embrittlement of irradiated model alloys and RPV steels after annealing, Int.
J. Press. Vessel. Pip. 85 (2008) 575–579.
[23] Y. Nishiyama, K. Onizawa, M. Suzuki, J.W. Anderegg, Y. Nagai, T. Toyama, M.
Hasegawa, J. Kameda, Effects of neutron-irradiation-induced intergranular phos-
phorus segregation and hardening on embrittlement in reactor pressure vessel
steels, Acta Mater. 56 (2008) 4510–4521.
 X. Bai et al. / Materials and Design 89 (2016) 759–769
[24] S.K. Ghosh, A. Haldar, P.P. Chattopadhyay, Effect of pre-strain on the ageing behavior
of directly quenched copper containing micro-alloyed steel, Mater. Charact. 59
(2008) 1227–1233.
[25] J. Bohmert, H.W. Viehrig, A. Ulbricht, Correlation between irradiation-induced
changes of microstructural parameters and mechanical properties of RPV steels, J.
Nucl. Mater. 334 (2004) 71–78.
[26] S.K. Ghosh, A. Haldar, P.P. Chattopadhyay, On the Cu precipitation behavior in
thermomechanically processed low carbon microalloyed steels, Mater. Sci. Eng. A
519 (2009) 88–93.
769
[27] S.G. Druce, Effects of austenitisation heat treatment on the fracture resistance and
temper embrittlement of MnMoNi steels, Acta Metall. 34 (1986) 219–232.
[28] P. Bowen, C.A. Hippsley, J.F. Knott, Effects of segregation on brittle fracture and fa-
tigue crack growth in coarse-grained, martensitic A533B pressure vessel steel,
Acta Metall. 32 (1984) 637–647.
[29] G.R. Odette, P.M. Lomborzo, R.A. Wullaert, In: Garner FA, Perrin JS (Eds). Effects of
Radiation on Materials, 12th International Symposium, ASTM STP 870. West
Conshohocken, PA: American Society for Testing and Materials; 1985.