On flux effects in a low alloy steel from a Swedish reactor pressure vessel

Journal of Nuclear Materials 484 (2017) 110e119

Contents lists available at ScienceDirect

Journal of Nuclear Materials journal homepage: www.elsevier.com/locate/jnucmat

On flux effects in a low alloy steel from a Swedish reactor pressure vessel
n b Magnus Boåsen a, *, Pål Efsing a, Ulla Ehrnste a b

Department of Solid Mechanics, Royal Institute of Technology (KTH), SE-100 44 Stockholm, Sweden VTT Technical Research Centre of Finland Ltd, PO Box 1000, FI-02044 VTT, Finland

h i g h l i g h t s
Hardness testing is combined with post irradiation annealing at 330, 360 and 390  C.
Unstable matrix defects is studied in a reactor pressure vessel steel.
Comparison between surveillance material and accelerated irradiation.
No evidence of unstable matrix defects, i.e. not present in studied material.
Difference in hardness recovery between irradiation conditions found at 390  C.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 22 June 2016 Received in revised form 7 November 2016 Accepted 27 November 2016 Available online 29 November 2016

This study aims to investigate the presence of Unstable Matrix Defects in irradiated pressure vessel steel from weldments of the Swedish PWR Ringhals 4 (R4). Hardness tests have been performed on low flux (surveillance material) and high flux (Halden reactor) irradiated material samples in combination with heat treatments at temperatures of 330, 360 and 390  C in order to reveal eventual recovery of any hardening features induced by irradiation. The experiments carried out in this study could not reveal any hardness recovery related to Unstable Matrix Defects at relevant temperatures. However, a difference in hardness recovery was found between the low and the high flux samples at heat treatments at higher temperatures than expected for the annihilation of Unstable Matrix Defectsethe observed recovery is here attributed to differences of the solute clusters formed by the high and low flux irradiations. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Mechanical properties of structural materials have been shown to degrade due to microstructural changes induced by irradiation; these phenomena are specifically evident in the components of nuclear power plants subjected to irradiation by neutrons from the nuclear core. The main effect on the structural materials that make up the Reactor Pressure Vessel (RPV), i.e. both the base material and especially the weld metal, is an evident hardening and embrittlement, implying a weakening of the structural integrity of the RPV. It has been realized for many years that small clusters, commonly called solute clusters or agglomerates are formed within the microstructure due to neutron irradiation. These agglomerates

* Corresponding author. KTH Solid Mechanics, Teknikringen 8D, SE-100 44 Stockholm, Sweden. E-mail address: [email protected] (M. Boåsen). http://dx.doi.org/10.1016/j.jnucmat.2016.11.026 0022-3115/© 2016 Elsevier B.V. All rights reserved.

impede on the dislocation motion during plastic deformation of the material, which manifests as an increase of the tensile yield strength, but also as an increase of the ductile-to-brittle transition temperature (DBTT), and a lowering of the other fracture mechanical properties such as the fracture toughness and the upper shelf energy (USE) [1], [2], [3]. The weld metal investigated in this study has a characteristic high Mn-Ni e low Cu composition, leading to the formation of agglomerates mainly composed of Mn and Ni during irradiation. Among the agglomerates of alloying elements that emerge during irradiation, it has been proposed that there may emerge a new type of microstructural feature called Unstable Matrix Defects. Unlike the agglomerates, these defects are proposed to be unstable at the irradiation temperature, making the number density heavily dependent on the incident flux of neutrons from the core [4]. The purpose of this study has been to investigate the presence of Unstable Matrix Defects in irradiated pressure vessel steel from weld metal of the Swedish PWR Ringhals 4 (R4) by combining hardness testing and Post-Irradiation Annealing (PIA).

M. Boåsen et al. / Journal of Nuclear Materials 484 (2017) 110e119

The features that emerge following irradiation of low alloyed steels are commonly characterized into groups of agglomerates and matrix features. The group of matrix features has been proposed by Odette et al. [4] to be further divided into Stable Matrix Features (e.g. dislocation loops, nano-voids, etc.) and Unstable Matrix Defects (UMD). Both the agglomerates and the stable matrix features has been characterized to develop in rough proportion to the accumulated neutron fluence [2], while the latter has been found to develop as a function of several factors such as temperature, flux and fluence [4], [5]. The UMD are deemed unstable due to how these features are prone to annealing in-situ at RPV operating temperatures unlike other identified embrittlement mechanisms which appear thermally stable at reactor operating temperatures. Odette et al. proposed that the presence of UMD in the material would have dual properties. Firstly, by acting as a hardening feature, i.e. as a dislocation obstacle, impeding the dislocation motion, but not to the extent as of the agglomerates [2], [4], and secondly, as a sink to the radiation enhanced diffusion that is driving the formation of other features during irradiation such as the agglomerates, stable matrix features and a general diffusion of alloying elements towards e.g. grain boundaries and dislocations [2], [4], [5]. The precise constitution of the proposed UMD is not yet known, it is however theorized that the main population of UMD could exist as both interstitial and vacancy clusters complexed with segregated solutes such as C, N and/or Ni. Of the two proposed configurations, the vacancy clusters are characterized by being the more unstable e dissolving at a faster rate, thus being more probable of the two in relation to the experimental results of Odette et al. [2], [4]. The annihilation of proposed vacancy clusters has been theorized to occur by vacancy emission with recovery times strongly dependent on temperature, cluster size and free surface energy. Odette et al. suggested equation (1) as a simplified description of the number density of the UMD resulting from neutron irradiation


  ft Nðf; ftÞ ¼ ftsUMD Na 1  exp ft

(1)

where ft is the neutron fluence, f is the neutron flux, t is the characteristic annihilation time of the UMD, sUMD is the formation cross-section of the UMD, and Na is the atom density of the lattice (8.55$1028 1/m3 for iron). This implies that the number density of the UMD will form with a build-up phase and then saturate at a steady-state level, where the steady-state number density is linearly proportional to the neutron flux; for clarity in the conceptual formation of the UMD, see Fig. 1. In PIA studies presented by Odette et al. [4], [5], heat treatments are combined with micro-hardness testing in which a significant decrease of the hardness could be observed in a timeframe of 1.8$104 s (5 h) at a temperature of 343  C, and 3.25$105 s (90 h) at 290  C, proposedly connected to the annihilation of UMD within the material. These studies led to the theory that the UMD would mainly consist of small, sub-nm vacancy clusters that annihilate by vacancy emission at reactor relevant temperatures. Experimental results from Odette et al. [4], [5], on the contribution from UMD hardening can be found in Table 1 for different flux and fluence levels. Another study of the UMD, or the neutron flux effect on irradiation hardening in general, is presented in two papers by
rard in 2011 [6] and 2013 [7]. The study was Chaouadi and Ge conducted on a large set of materials where in the 2013 paper, eight different RPV materials were covered in the experimental effort, i.e. a large combination of different RPV-relevant chemical compositions was explored. The experiments consisted of samples irradiated at fluxes between 0.4 and 5.6$1014 n/s cm2 and to fluences of

111

Fig. 1. Conceptual development of the number density of Unstable Matrix Defects at high and low flux levels.

0.7e9.4$1019 n/cm2, which later was exposed to heat treatments at temperatures of 345  C and 355  C in combination with tensile testing. The studies concluded that no flux effect on the irradiation hardening of the tested materials could be proved to be present by the experiments, which according to the UMD theory proposed by Odette et al. should have been present [4]. Although surveillance specimens are mainly used for assessment of the structural integrity of the RPV, materials are also irradiated in test reactors to increase the amount of available material. A further understanding of the possible role of UMD is important, as the presence of these defects could affect results from mechanical test results on materials irradiated using different flux. In the extensive study collaborated by IAEA [8], it was concluded that the sole presence of Ni in RPV-steels does not necessarily make the changes of mechanical properties of the material especially radiation sensitive. However, in synergy with alloying elements such as Mn and Cu, the changes of mechanical properties by the conditions in a nuclear reactor can be quite severe with increasing aging. Prior studies of the low Cu e high Mn-Ni surveillance material of R4 have shown vast changes in the mechanical properties of the material due to irradiation, DBTT-changes (T41J) of 162  C, and increases in tensile yield strength of up to 215 MPa at a neutron fluence of 6.0$1019 n/cm2 [9]. It has also been shown that large number densities of solute clusters consisting of Mn, Ni, Cu, and Si form under irradiation along with segregation of P to dislocations and grain boundaries [10], [11], both being responsible for irradiation induced changes of material properties. In the study by Styman et al. in Ref. [11] it is shown how post-irradiation annealing treatments of the irradiated surveillance material from Ringhals

Table 1 Experimental results of UMD hardening from Odette et al. Reactor

Flux f½n=s cm2   1012

Fluence ft½n=cm2   1019

DPHUMD ½kg=mm2 a

Pluto Pluto Pluto BR2 BR2 BR2

4 46 46 1.4 100 100

0.22 0.1 0.5 2.1 6.6 12.7

10 8 19 12 30 32

a

DPH e Diamond Pyramid Hardness.

112

M. Boåsen et al. / Journal of Nuclear Materials 484 (2017) 110e119

Table 2 Measured chemical composition of the reactor pressure vessel steel used in Ringhals unit 4. Material

C

Al

Si

P

S

Cr

Mn

Co

Ni

Cu

Mo

R4 (wt.%)

0.068

0.024

0.14

0.015

0.004

0.04

1.35

0.01

1.66

0.05

0.50

3 at 450  C gives the onset of annihilation of Mn-Ni-Si-Cu agglomerates, which also is correlated to the recovery of mechanical properties of the material. The number density and volume weighted mean size of the agglomerates in the R3-material irradiated to a fluence of 6.8$1019 n/cm2 (E > 1 MeV) reported in the study by Styman et al. are 4.2$1023 m3 and 3.5 nm, respectively. The formation of solute clusters of the type discussed above has been studied by the use of model alloys both in experimental and simulation efforts. In an experimental study by Meslin et al. [12] on a binary Fe-Mn alloy irradiated by Fe ions, it was concluded that solute clustering of Mn is radiation induced (opposed to radiation enhanced, where there exists a thermodynamic driving force for clustering of alloying solutes, accelerated by radiation) where the Mn atoms could be dragged by mobile point defects towards clustering sites. The results of this study was also supported by the conclusions of a study conducted by Ngayam-Happy et al. [13] where it was concluded by atomistic kinetic Monte Carlo simulations that radiation induced fluxes of point defects is a sufficient mechanism for driving solute clustering in Fe-CuMnNi model alloys. However, in the studies by Bonny et al. [14], [15] the formation of the above mentioned solute clusters is investigated via a thermodynamic driving force in association with radiation induced defects. The studies were conducted by Monte Carlo simulations where solute clustering on irradiation induced defects was investigated in FeNiMnCu alloys, e.g. thermodynamic formation of clusters on dislocation loops, which themselves are formed by irradiation. The results of these studies suggest a formation mechanism of the agglomerates which is not entirely radiation induced or radiation enhanced, but rather something in between. 2. Material and methods 2.1. Material The material investigated in this study has been irradiated in the Halden Materials Testing Reactor situated in Norway and is representative for the weldments of the RPV in Ringhals unit R4. In addition to the material samples irradiated in Halden, one sample from the ordinary surveillance program of R4 has been included in the study. The material from R4 is a low Cu e high Mn-Ni content material, and the full chemical composition can be found in Table 2 [9]. Ringhals Unit 4 was taken into operation in 1983, and was designed by Westinghouse Electric Company [9]. The reactor

pressure vessel was fabricated by Uddcomb using SA 508 class 2 € ckner Werke, the same company also supring forgings from Klo plied the surveillance blocks. The surveillance material was fabricated by joining of two ring-shaped forgings using the same welding procedures as that used in the fabrication of the pressure vessels, all in accordance with ASTM E185-73 [16]. Chemical analysis of the weld material in Unit 4 showed that the nickel content was 1.66 wt % (1.35 wt% Mn and 0.05 wt% Cu), the highest recorded nickel content of any Westinghouse PWR. The microstructure of the weld metal is mainly acicular ferrite with some amount of grain boundary- and polygonal ferrite. The weld microstructure consist of several different macroscopic zones, i.e., the as-welded microstructure, the reheated microstructure between two weld passes and the twice re-heated microstructure between two weld beads, Fig. 2. The microstructure will affect both the as-manufactured hardness as well as the effect of irradiation. Typically, the hardness of the as-manufactured re-heated microstructure is slightly lower than that of the as-welded microstructure. The irradiation in Halden was conducted with an acceleration factor of 20 for the neutron spectrum with respect to the surveillance program conditions in the R4-reactor and corresponds to approximately 20, 30 and 60 years of reactor operation. For corresponding flux and fluence levels see Table 3 (E > 1 MeV). The accumulated irradiation dose in Halden was measured by neutron fluence monitor wires, which prior to irradiation were subjected to full-length gamma scans, and after irradiation analyzed to determine the activation of the wires in order to determine the neutron fluence. The irradiation temperature was kept in the interval 290e295  C and the environment was chosen to simulate an averaged standardized PWR-beginning of cycle primary water chemistry, i.e. 1200 ppm B, 3 ppm Li and 35 ml H/kg H2O [17]. The shut-down cycle of the irradiation capsule at Halden in terms of moderator temperature, coolant temperature and reactor power can be seen in Fig. 3. It can be seen that the temperature does not drop instantly but in a rather slow manner (as expected). Regarding this temperature transient, one could believe that a portion of possible UMD would annihilate. However, it was estimated that the effect on the hardening contribution of any UMD present in the material would be decreased by ~5% as an effect of the shut-down cycle, i.e. the majority of the UMD would still remain within the material samples, probably detectable by hardness measurement. For clarity in this estimate, see Appendix A.

Fig. 2. RPV weld microstructure. Left: macrostructure at lower magnification. Right: microstructure at higher magnification.

M. Boåsen et al. / Journal of Nuclear Materials 484 (2017) 110e119 Table 3 Calculated flux and fluence for the material samples used in this investigation (E > 1 Mev). Specimen

Irradiated at

Flux f½n=s cm2   1012

Fluence ft½n=cm2   1019

mdpa

S0 S4.6 H2.0 H2.8 H6.4

Baseline Ringhals Halden Halden Halden

0 0:15 2:31 3:25 3:81

0 4:56 2:00 2:81 6:35

0 68 30 42 95

The specimens used in the experiments were cut from broken Charpy specimens from archived surveillance material of the R4 weldments and the material irradiated in Halden. The specimens were sampled from the Charpy specimens in the form of crosssectional slices with a thickness of 2 mm using electrical discharge machining. Prior to all testing, the intended testing surfaces were prepared by polishing with a fineness of 3 mm and cleaned using ethanol in an ultrasonic bath to remove any residue from previous polishing and handling. The specimens used in the experiments are named after place of irradiation and fluence level, e.g. S4.6 corresponds to Surveillance, 4.56$1019 n/cm2, for clarity see Table 3. 2.2. Experimental The heat treatments of the specimens used in the PIA investigation were conducted in an inert helium environment to minimize the oxidation on the specimen surface. The furnace used was a Carbolite tube furnace in which a vacuum was imposed to ~1:5,102 mbar, prior to the introduction of helium gas at an overpressure relative to the atmospheric pressure. The heat treatments were conducted at 330, 360 and 390  C for accumulated times up to 30 h (1.08$105 s) per temperature. For each heat treatment temperature, a new set of specimens were used, i.e. different as-irradiated specimens from the same Charpy specimens were used for the different temperatures in the experiment. The specimens were exposed to the elevated temperatures in time increments where hardness measurements were carried out at each increment, i.e. heat treatment for a time increment, followed by a hardness measurement, to be followed by another heat treatment etc. To monitor the temperature of the specimens in the furnace, a thermocouple was used, where the estimated error of the thermocouple was given to be ±1 %; Fig. 4 illustrates a heat treatment increment used in the experiments. The specimens were kept in a fixture that would allow the heat transfer to occur easily, and that would have a low thermal mass for a fast temperature response.

113

Prior to the hardness testing, the specimens were polished to remove possible surface oxides that may have formed during the heat treatment. Following the polishing, the specimens were cleaned in an ultrasonic bath containing ethanol to remove any residue left over from the polishing. Hardness testing was carried out using a Struers Duramin-A300 hardness tester at room temperature; the machine offers an automated placement and measurement of the indentation diagonals by optical means. Vickers was chosen as the testing method with an indentation load of five kilograms due to availability and experience. The testing was carried out according to ASTM E384-11, with the one exception of the placement of the indents. The distance was increased so that a conservative safety to influence among indents could be ensured, e.g. the standard recommends a center-tocenter-distance of 2.5 times the diagonals of the produced indents [18], while the placement in the experiment was chosen to give a center-to-center distance of up to 5e6 times the diagonals. All specimens showed signs of a gradient of mechanical properties over the surface intended for the hardness measurement, i.e. a spatially varying hardness over the specimens as a result of the varying microstructure in the weldments. This was accounted for by initially measuring the reference hardness of the specimens prior to the heat treatments. After each heat treatment, the following indentations were placed in close vicinity to the reference indentations. For clarity in the placement of the indentations made in the experiment, see Fig. 4. To determine whether a change had occurred or not within the material due to the heat treatments, the hardness after each heat treatment was normalized with respect to reference state of the specimens according as

HN ¼

HV H V0

(2)

where HN denotes the normalized hardness level, H V is the mean of the points measured after each heat treatment, and H V0 is the mean of the reference positions as irradiated of two adjacent rows in close vicinity of the indents made after each heat treatment. 3. Results Within the framework of the experiment discussed in Section 2, the testing was conducted in a stringent manner and resulted in good square shaped indentations with very few misshapen indentations, which were easily replaced with a new indentation placed in close vicinity. Fig. 5 illustrates the hardness change induced by irradiation, shown are the three sets of specimens used

Fig. 3. Halden reactor shut-down cycle Left: Power and moderator temperature, Right: Coolant temperature in the irradiation rig.

114

M. Boåsen et al. / Journal of Nuclear Materials 484 (2017) 110e119

Fig. 4. Left: Illustration of the placements of indentations on the specimens used in the experiment, Right: A heat treatment increment of two hours at 390  C.

in the study, i.e. one set per temperature, where different sets of asirradiated specimens were used for each experiment. The irradiation induced changes of the observed mechanical properties follows the same double-linear trend curve that has been observed in earlier studies of the same RPV material, i.e. a rapid build-up of changes in the mechanical properties followed by a phase with slower development of the mechanical properties with respect to the neutron fluence (E > 1 MeV) [17], [19]. Regarding the measurements carried out in this study, one should note that there is a significant scatter of the hardness associated with the spatial placement of the indentations, which have been taken into account by the measurement of the reference indentations. However, there still exist small variations in the hardness due to even more local scatter in material properties and measurement errors, e.g. measurement errors of the diagonal are probably the largest source of error. An erroneous measurement of the diagonal gives quite extensive transmitted errors in the calculated hardness. For example, a material that has a hardness of 300 kg/mm2 would have a diagonal of 175.8 mm, however, an erroneous measurement of the diagonal of 2.5 mm (1.4%) would give a transmitted error of 3% in the calculated hardness value [18], which is regarded here to be a reasonable uncertainty in all the hardness data. Therefore, the data in the 1:1-plot shown in Fig. 6 is

presented with dashed scatter lines representing ±3%. The circles in Fig. 6 can be seen to have varying sizes, the idea here is to, in a simple way, give an overview of all the specimens at a given temperature while at the same time illustrating the annealing time for the presented data, i.e. increasing radii e increasing annealing time. As can be seen, most of the data falls within the scatter lines at the two lower annealing temperatures, but for the higher temperature a decreasing trend can be observed for the data. The data in the 1:1plot shown in Fig. 6 is presented as mean values of the hardness data that is associated to each heat treatment. This refers to the mean value of the three hardness measurements that correspond to the as-irradiated condition and the mean value of the six measurements corresponding to the after heat treatment condition, for a visual representation of the placement of the as-irradiated/after heat treatment indents see Fig. 4. To display how the hardness is affected by the heat treatments, the normalizing approach laid out in Section 2.2. was chosen for its ability to show comparable hardness changes on a specimen-tospecimen level regardless of dose level or initial hardness. For the level of irradiation induced hardness at each specific dose level, the reader is referred to Fig. 5. The normalized hardness as function of annealing time is depicted in Fig. 7, some scatter is observed in the diagrams containing the data from the two lower temperatures.

Fig. 5. Irradiation induced hardness plotted against neutron fluence, E > 1 MeV.

M. Boåsen et al. / Journal of Nuclear Materials 484 (2017) 110e119

115

However, in the higher temperature anneal, a distinct lowering of hardness (8e10%) is shown as function of annealing time in the pieces irradiated with higher flux levels, i.e. specimens H6.4, H2.8 and H2.0, while the surveillance specimen S4.6(from R4) irradiated with one order of magnitude lower flux shows almost no hardness recovery at all. The uncertainties in the measurements presented in Fig. 7 are taken as the normalized 95%-confidence interval from the hardness measurements of each increment of heat treatment. The confidence interval is normalized according to the same procedure as the mean value, as presented in Section 2.2. 4. Discussion

Fig. 6. 1:1-diagrams of the hardness after and before heat treatments, increasing radii symbolizes longer accumulated heat treatment time. Top diagram: 330  C, middle diagram: 360  C, bottom diagram: 390  C.

As can be seen in the data presented in Section 3, the data exhibits a certain amount of scatter and for the heat treatments at the lower temperatures (330 and 360  C); the hardness does not change significantly from the reference level, indicating no annihilation of dislocation obstacles at those temperatures. However, the data from the higher temperature anneal (390  C) shows a decreasing trend of hardness with increasing annealing time for the specimens subjected to the higher flux irradiations. The surveillance specimen S4.6 from Ringhals does however not show the same hardness recovery behavior as the in Halden irradiated material. One could draw the conclusion that this distinction comes from a population of Unstable Matrix Defects being annihilated at the high flux specimens. However, if this was to be true, then the hardness recovery should also have occurred at the lower temperatures but in a slower manner, otherwise they could not be called unstable in the sense laid out by Odette et al. in Ref. [4]. The word “unstable” in UMD refers to the fact that these defects would be unstable at the irradiation temperature, this by both being generated and annihilated at the operating conditions of the reactor. The result from this type of evolution would be a steadystate UMD-population strongly dependent on the neutron flux as laid out by Odette in Ref. [4]. But the fact that no hardness recovery can be seen at the lower annealing temperatures for the material samples of this testing series, points towards the absence of Unstable Matrix Defects within the weld material of the RPV material of Ringhals reactor 4, both after irradiation in the Halden MTR (higher flux) and in the surveillance positions in R4 (lower flux), and towards another explanation of the observed change in hardness at the higher temperature anneal. The different recovery of hardness of the specimens irradiated in Halden and Ringhals seen in Fig. 7 is speculated to come from the constitution of the agglomerates that are formed during the irradiation, i.e. Mn-Ni-Si-Cu rich solute clusters. It is theorized that a discrepancy can be found in the number density and size of the agglomerates in the different material samples depending on how the material samples have been irradiated, i.e. different number density and size distributions for different neutron flux but similar dose levels and irradiation temperature. Assuming that irradiation is mainly responsible for forming the agglomerates, and another driving force being responsible for driving the clusters to grow by diffusion of alloying elements from the bulk of the material. It is here speculated that a thermal aging component is biasing the growth of the (by irradiation) formed agglomerates at the irradiation temperature. The result from this would be an agglomerate population with a number density and size distribution largely dependent on dose, but via the thermal aging component also being dependent on flux. I.e. a high flux allows for a high dose to build up with a small thermal component (short time) e resulting in a high number density of smaller clusters compared to lower flux, taking longer time to build up the same dose, implying a larger thermal component e and thus a smaller number density of larger clusters. However, the hardening in the material samples seems to

116

M. Boåsen et al. / Journal of Nuclear Materials 484 (2017) 110e119

Fig. 7. Diagrams of normalized hardness as function of annealing time. Top: 330  C, bottom left: 360  C, bottom right: 390  C.

develop similarly with dose regardless of flux, as seen in Fig. 5. This implies that the square root of the product of number density and size of the agglomerates in the high and low flux samples attains similar values at the same dose level, if assuming a dispersed barrier strengthening mechanism, as also noted in a study by Efsing et al. [17]. The result from this reasoning is that the resulting resistance to dislocation glide is similar if the aging creates small clusters at a high number density (high flux and small thermal

component) or larger clusters at a lower number density (low flux and larger thermal component). However, if the difference in the aging conditions is very large, the hardening may eventually no longer become the same, as could be the case between a boiling water reactor (even lower flux) and material test reactor (high flux). The reasoning laid out in the paragraph above is based on the idea that larger clusters will have a higher thermal stability, dissolving at higher temperatures than smaller clusters, implying that

M. Boåsen et al. / Journal of Nuclear Materials 484 (2017) 110e119

the specimens irradiated in Halden contains smaller clusters while the surveillance specimen from Ringhals contains larger clusters. This is supported by the hardness recovery seen in Fig. 7 where the specimens named H2.0, H2.8 and H6.4 (high flux) shows a hardness recovery of 8e10% relative to the as-irradiated condition, a behavior not observed in the specimen S4.6 (lower flux) from Ringhals 4 surveillance program. If the experiments were to be repeated at a higher temperature, e.g. 450  C, then all agglomerates would most probably completely dissolve in the samples irradiated in the Halden reactor (which would result in a substantial recovery in mechanical properties), while the surveillance sample would most probably be at the onset of recovery, as seen in the study of surveillance material from R4s sister reactor R3 (with similar chemical composition) using Atom Probe Tomography combined with hardness testing conducted by Styman et al. [11]. Similar results as implied in this study were also recently published by Wagner et al. [20], where results from a SANS-study was compared with hardness testing on materials irradiated by different neutron flux, covering a wide range of chemical compositions relevant to RPV steels. In their study an A508 steel is irradiated to a fluence of 6.4$1019 n/cm2 using two fluxes separated by two orders of magnitude. It is concluded that in the high flux irradiation, a higher number density of smaller clusters appear, while in the low flux irradiation a lower number density of larger clusters appears e with no significant resultant difference in hardening between the two conditions. In the study by Dohi et al. [21], an A533 plate steel was irradiated in a “commercial LWR” and in the Halden reactor at flux of 1011 and 5$1012 n/s cm2, respectively. The irradiated specimens were in the study subjected to Vickers hardness testing and microstructural characterization by atom probe tomography. The conclusions in the study by Dohi et al., was that solute clusters had formed due to irradiation and that in the high flux irradiation, the number density was higher and the size of the clusters smaller than in the low flux irradiations. The flux effect on the measured hardness was thus concluded to be non-significant between the irradiations. Furthermore, in the study by Soneda et al. [22]. high and low flux irradiations of A508 forging and A533 plate material has been studied using Charpy impact testing and atom probe tomography. It is found that the high flux irradiation creates small clusters of high number density, while the low flux irradiation creates larger clusters of lower number density. The results from the Charpy testing indicate that the shift in T41J of the low-Cu material falls on the same trend curve for the high and low flux irradiations. These results are similar to what is implied in this study. In another study by Soneda et al. [23]. defect accumulation in neutron irradiated bcc-FE was studied using Kinetic Monte Carlo simulations. The effect of environmental variables such as irradiation temperature, neutron flux and spectrum was explored in the study, while also including the effect of grain boundaries as sinks where point defects could annihilate. The main conclusions of this study was that there exists a flux-effect on the “number of vacancy jumps”, which correlates to the possibility of diffusion of solute elements and the formation of solute clusters. The number of vacancy jumps is a sum of the irradiation induced vacancy jumps and the thermal vacancy jumps in the Fe-lattice. This results in a flux effect at low and high flux with a plateau in between, where at low flux the number of thermal vacancy jumps is high and results in a large total number of vacancy jumps (high level of diffusion, implying cluster growth). While at high flux, the number of irradiation induced vacancy jumps is lowered (lower diffusion) at the same fluence. This couples the irradiation dose and flux effect through the irradiation temperature, as implied in this study. The most significant uncertainties in this study come, as discussed in Section 4, most probably from the measurements of the diagonals of the

117

indentations. The source of error would be setting of the focus before measuring each set of indentations, which is done manually. The reason that the focus needs to be reset, is that the specimens are all measured in sequence and since the thickness of the specimens are not exactly the same; the same focus settings cannot be kept for all the specimens. 5. Conclusions
The observed hardening in the samples irradiated at two different fluxes seem to fall on the same trend curve e implying that the results of accelerated aging of materials mechanical properties in MTR conditions are valid for comparison with the mechanical properties of surveillance specimens. I.e. the effect from neutron flux on overall hardening is small or negligible.
No hardening Unstable Matrix Defects detectable by hardness testing could be found to be present in the material, from either surveillance irradiation in Ringhals reactor 4 or from accelerated irradiation in the Halden Reactor.
A distinct difference in the recovery behavior at the highest temperature (390  C) was found in the samples irradiated at different flux. These findings supports results from previous studies where a difference in cluster size and number density has been found in material irradiated at high and low flux. The current hypothesis is that the solute clusters would attain a number density and size distribution that is largely dependent on dose, but via a thermal aging component also dependent on flux. High resolution microstructural analysis is needed and will be performed to support or dismiss the hypothesis. Acknowledgements This investigation has been funded by a grant from The Swedish Radiation Safety Authority (SSM). An appreciation is directed to NUGENIA þ for the young scientist short training grant that made travelling to VTT for experiments possible, grant agreement no: 604965. The authors are also deeply grateful for the assistance in the experimental effort by the laboratory staff M.Sc. Petteri Lappalainen, Mr. Arto Kukkonen, Mrs. Marketta Mattila and Mr. Jari Lydman at VTT in Espoo, Finland. Appendix A. Estimate of annihilation due to reactor shutdown At the shut-down of the reactor at the end of the irradiation cycle, the neutron flux decreases faster than that of the temperature of the water, thus potentially annihilating UMD theorized to be present in the material. Below follows an estimation of the annihilation of the proposed UMD due to the cooling of the reactor after shut-down. Odette et al. presented characteristic annihilation times of the UMD at two temperatures as, ta343 C ¼ 1.8$104 s and ta290 C ¼ 3.25$105 s. Using these times in an Arrhenius-type equation

 
Q 1 1 ; t ¼ ta ,exp   R Ta T

(A1)

makes it possible to solve for the activation energy Q as

RTTa ln ðt=ta Þ Q ¼ z1:574,105 J=mol T  Ta

(A2)

Assuming that the annihilation rate of the UMD can be described by an Arrhenius relation, one finds

118

M. Boåsen et al. / Journal of Nuclear Materials 484 (2017) 110e119

  dN Q ¼ k,exp ; dt RT

(A3)

where N is taken as a fraction of the UMD number density in the material and k as a pre-exponential multiplier. Integrating (A3) one finds

  Q þ Nð0Þ N ¼ kt,exp RT

NC 1 z  4:462,1012  Q h t,exp RT

(A5)

The reactor shut-down temperature as a function of time is taken from Fig. 3 (Section 2.1.) and is simplified into the diagram shown in Fig. A1, where t ¼ 0 in the diagram is representative for the decrease in reactor power related to the shut down and thus decrease in neutron flux, i.e. end of UMD-production.

Fig. A1. Simplified temperature as a function of time in the irradiation rig after the core shut-down has been initialized.

The decrease of the UMD number density due to the temperature transient in Fig. A1 is given by

DN ¼

  1:574,105 dt 4:462,1012 ,exp 8:314,564:6

0 20:6 Z

þ

 4:462,1012 ,exp

6:5

 1:574,105 dt 8:314,ð651:9  13:43tÞ (A6)

which evaluates into

DNz  0:0939;

(A7)

and gives the total remaining UMD population as

N þ DN ¼ 1  0:0939 ¼ 0:906

(A8)

By assuming that the UMD would give a hardening contribution

(A9)

where the change in mechanical properties such as hardness due to the shut-down cycle can be found by taking the relative difference as

pffiffiffiffiffiffiffi

DUMD ¼

C ¼ 1 ½  ;

Z6:5

pffiffiffiffiffiffiffi

tUMD ¼ aGb Nd;

(A4)

where N(t ¼ 0) ¼ 1 and N(t ¼ 5 h, T ¼ 616 K) ¼ 0 is assumed. Following this, the constants can be determined as

k¼

according to a dispersed barrier strengthening mechanism, the increased resistance to plastic flow can then be expressed as

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

pffiffiffiffiffiffiffiffiffiffiffiffiffi aGb Nd  aGb 0:906Nd pffiffiffiffiffiffiffi ¼ 1  0:906z0:0482 aGb Nd (A10)

The conclusion drawn from this analysis is that the potential hardness contribution from the proposed Unstable Matrix Defects would decrease by 4.8% during the shut-down cycle, after the neutron flux is taken away. References [1] W.J. Phytian, C.A. English, Microstructural evolution in reactor pressure vessel steels, J. Nucl. Mater. 205 (1993) 162e177. [2] G.R. Odette, G.E. Lucas, Recent progress in understanding reactor pressure vessel steel embrittlement, Radiat. Eff. Defects Solids 144 (1998) 189e231. [3] G.R. Odette, G.E. Lucas, Embrittlement of nuclear reactor pressure vessels, JOM 53 (2001) 18e22. [4] G.R. Odette, E.V. Mader, G.E. Lucas, W.J. Phythian, A.C. English, STP1175, in: “The Effect of Flux on the Irradiation Hardening of Pressure Vessel Steels,” in 16th International Symposium, ASTM, Philadelphia, 1993. [5] G.R. Odette, Advanced models of LWR pressure vessel embrittlement for low flux-high fluence conditions, Santa Barbara, U. S. DOE Nucl. Energy uni. Program (2013) 6e23.
rard, Neutron flux and annealing effects on irradiation [6] R. Chaouadi, R. Ge hardening of RPV materials, J. Nucl. Mater. 418 (2011) 137e142.
rard, Confirmatory investigations on the flux effect and [7] R. Chaouadi, R. Ge associated unstable matrix damage in RPV material exposed to high neutron fluence, J. Nucl. Mater. 437 (2013) 267e274. [8] International Atomic Energy Agency (IAEA), Effects of Nickel on Irradiation Embrittlement of Light Water Reactor Pressure Vessel Steels IAEA-TECDOC1441, IAEA, Vienna, 2005. [9] P. Efsing, C. Jansson, T. Mager, G. Embring, Analysis of the ductile-to-brittle transition temperature shift in a commercial power plant with high nickel containg weld material, J. ASTM Int. 4 (7) (2007). [10] M.K. Miller, K.A. Powers, R.K. Nanstad, P. Efsing, Atom probe tomography characterizations of high nickel, low copper surveillance RPV welds irradiated to high fluences, J. Nucl. Mater. 437 (1-3) (2013) 107e114. [11] P.D. Styman, J.M. Hyde, D. Parfitt, K. Wilford, M.G. Burke, C.A. English, P. Efsing, Post-irradiation annealing of Ni-Mn-Si-enriched clusters in a neutron irradiated RPV steel weld using atom probe Tomography, J. Nucl. Mater. 459 (2015) 127e134. [12] E. Meslin, B. Radiguet, M. Loyer-Prost, Radiation-induced precipitation in a ferritic model alloy: an experimental and theoretical study, Acta Mater. 61 (16) (2013) 6246e6254. [13] R. Ngayam-Happy, C.S. Becquart, C. Domain, L. Malerba, Formation and evolution of MnNi clusters in neutron irradiated dilute Fe alloys modelled by a first principle-based AKMC method, J. Nucl. Mater. 426 (1e3) (2012) 198e207. [14] G. Bonny, D. Terentyev, A. Bakaev, E.E. Zhurkin, M. Hou, D. Van Neck, L. Malerba, On the thermal stability of late blooming phases in reactor pressure vessel steels: an atomistic study, J. Nucl. Mater. 442 (1e3) (2013) 282e291. [15] G. Bonny, D. Terentyev, E.E. Zhurkin, L. Malerba, Monte Carlo study of decorated dislocation loops in FeNiMnCu model alloys, J. Nucl. Mater. 452 (1e3) (2014) 486e492. [16] American Society of Testing and Materials (ASTM), E185e73 Surveillance tests for nuclear reactor vessels, ASTM, West Conshohocken (1973). [17] P. Efsing, J. Rouden and P. Nilsson, “Flux Effects on Radiation Induced Aging Behaviour of Low Alloy Steel Weld Material with High Nickel and Manganese Content,” in ASTM 26th Symposium on Effects of Radiation on Nuclear Materials, Indianapolis, 2013. [18] American Society of Testing and Materials (ASTM), E384e11 standard test method for knoop and Vickers hardness of materials, ASTM, West Conshohocken (2011). [19] M.K. Miller, A.A. Chernobaeva, Y.I. Shtrombakh, K.F. Russel, R.K. Nanstad, D.Y. Erak, O.O. Zabusov, Evolution of the nanostructure of VVER-1000 RPV materials under neutron irradiation and post irradiation annealing, J. Nucl. Mater. 385 (3) (2009) 615e622.
ndez-Mayoral, [20] A. Wagner, B. Bergner, R. Chaouadi, H. Hein, M. Herna M. Serrano, A. Ulbricht, E. Altstadt, Effect of neutron flux on the characteristics

M. Boåsen et al. / Journal of Nuclear Materials 484 (2017) 110e119 of irradiation-induced nanofeatures and hardening in pressure vessel steels, Acta Mater. 104 (2016) 131e142. [21] K. Dohi, K. Nishida, A. Nomoto, N. Soneda, H. Matsuzawa, M. Tomimatsu, “Effect of Neutron Flux at High Fluences on Microstructural and Hardness Changes of RPV Steels,” in ASME 2010 Pressure Vessel & Piping, Bellevue, 2010.

119

[22] N. Soneda, K. Nishida, A. Nomoto, K. Dohi, Avignon, Flux Effect on Embrittlement of Reactor Pressure Vessel Steels Irradiated to High Fluences,” in Fontevraud, vol. 8, 2014. [23] N. Soneda, S. Ishino, A. Takahashi, K. Dohi, Modeling the microstructural evolution in bcc-Fe during irradiation using kinetic Monte Carlo computer simulation, J. Nucl. Mater. 323 (2003) 169e180.