- Email: [email protected]
Detection of intergranular embrittlement of reactor pressure vessel steel by electrochemical method
Materials Science & Engineering A 725 (2018) 88–97
Contents lists available at ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
Detection of intergranular embrittlement of reactor pressure vessel steel by electrochemical method Tenneti Sharmaa,b, Sunil Kumar Bonaganic, N. Naveen Kumarc, I. Samajdard, V. Kaina,c,
T
⁎
a
Homi Bhabha National Institute, Trombay, Mumbai 400085, India Indian Navy, India c Materials Group, Bhabha Atomic Research Centre, Trombay, Mumbai 400085, India d Department of Metallurgical Engineering & Materials Science, Indian Institute of Technology Bombay, Powai, Mumbai 400076, India b
A R T I C LE I N FO
A B S T R A C T
Keywords: Low Alloy Steel Aging Impact Toughness Carbide precipitates Segregation Electrochemical polarisation test
The reactor pressure vessel (RPV) steel, with manganese-nickel-molybdenum (Mn-Ni-Mo) alloying, was subjected to accelerated thermal aging treatment at 450 °C for durations upto 8400 h. Impact toughness decreased and the intergranular fracture (IGF) increased monotonically with increased aging time. This was accompanied by insignificant changes in average hardness, grain size, misorientation and carbide size. However, electrochemical polarisation tests in picric acid showed increasing width of attack at prior austenitic grain boundaries (PAGB). It was hypothesized that the accelerated aging led to phosphorus segregation at the grain boundary, a hypothesis supported by the post electrochemical attack on the PAGB, leading to a significant loss in impact properties.
1. Introduction Low Alloy Steels (LAS) are used as the structural material in fabrication of the nuclear reactor pressure vessel (RPV) due to the excellent high temperature mechanical properties it provides and its compatibility in the operating environment of the reactor. The nuclear RPV is exposed to neutron fluence in addition to the prolonged high operating temperature it sees all along the operation life of the reactor. These steels are also exceedingly sensitive to heat treatments that are a part of the fabrication processes [1] and the resulting microstructures have a wide scatter in the desirable properties. Since the life of the nuclear plant is decided by the RPV, understanding its microstructural evolution that takes place during fabrication and subsequently degradation during its operation is of paramount importance. Measuring the degradation and assessing the residual life of the material is essential to ensure safe plant operation and also to extend the life of the plant beyond the design life. The two major types of steels used worldwide for construction of RPVs are Mn-Ni-Mo steels (commonly referred to as western type RPVs) and Cr-Mo-V steels (commonly referred to as eastern type RPVs) [2–4]. Grades 15Kh2NMFAA and 15Kh2MFA [5] are examples of eastern grade RPV steels and 20MnMoNi55, SA508 grade 3, SA533, SA 302B are some of the grades of western RPVs [2–4,6,7]. The RPV materials having ferritic/ferritic-bainitic microstructure [4] (or body centered
⁎
cubic (bcc) crystal structure) show the typical ductile to brittle transition (DBT) phenomena (Fig. 1) [2]. At low temperature the bcc materials show low impact toughness while at room temperature or higher temperatures, these materials show high impact toughness (Fig. 1). The transition from low to high impact toughness with increase in temperature (typical S curve) is characterized by ductile to brittle transition temperature (DBTT) that is typically taken as the middle point between the high and the low plateau of toughness. Irradiation shifts the typical S curve to right side and lowers the upper shelf impact energy drastically (Fig. 1). Radiation embrittlement (RE) manifests as lowering of upper shelf energy and DBTT shift to higher temperature. Both western and eastern materials have distinguishable chemical compositions and thus it is expected that their response to the radiation embrittlement is also different. Though the radiation embrittlement of both types of the steels depends on the operating fluence, the major factor governing the irradiation sensitivity is the amount of the deleterious elements present. The main elements identified to contribute to RE are phosphorus (P) and copper. Nickel has also been reported to increase the irradiation damage sensitivity caused by copper [2]. The primary mode of degradation of the properties of the RPV steels is radiation embrittlement. Three basic mechanisms are considered to contribute to radiation embrittlement: (a) direct matrix damage by high energy neutrons, (b) irradiation-induced precipitation and (c) element segregation [8,9]. The effect of these three components to the overall
Corresponding author at: Homi Bhabha National Institute, Trombay, Mumbai 400085, India. E-mail address: [email protected] (V. Kain).
https://doi.org/10.1016/j.msea.2018.04.008 Received 21 January 2018; Accepted 3 April 2018 Available online 04 April 2018 0921-5093/ © 2018 Elsevier B.V. All rights reserved.
Materials Science & Engineering A 725 (2018) 88–97
T. Sharma et al.
boundaries results in temper embrittlement of these steels that does not affect the hardness and increases the DBTT but is reversible [13]. Cr-Mo steels are shown to be largely susceptible to degradation by carbide coarsening and Cr-Mo-V steels by segregation to grain boundaries [16]. Takahashi et al. [17] have shown P segregation in Cr-Mo (2.25 Cr-1Mo) and Cr-Mo-V steels. In the case of X20CrMoV12.1 alloy steel pipe, it was shown that the degradation of the material is mainly related to carbide coarsening [18]. The effects of thermal aging on the degradation of the properties have been studied by many. There is literature to suggest that the degradation effects of thermal embrittlement are negligible whereas a few researchers have suggested otherwise also. The effects of long-term aging at temperatures up to 350 °C (reactor operating temperature) on the ductile-to-brittle transition temperature of RPV steels have been studied by Corwin et al. [12]. Corwin et al. showed from literature data that there is no embrittlement at such reactor operation temperatures for operation periods of up to 100,000 h [12]. Similarly, DeVan et al. [19] reported that DBTT of RPV steel from the Arkansas 1 reactor shifted from − 1 to 9 °C after thermal ageing at 280 °C for 93,000 h. The material from the beltline region of the Oconee Unit 3 reactor showed an increase of about 1 °C after exposure to a temperature of 282 °C for 103,000 h [19]. Fukakura et al. [20] studied the effect of thermal aging on western grade RPV steel and concluded that after thermal aging for 10,000 h at temperatures of 350 °C, 400 °C and 450 °C, the increase in the DBTT was small. In contrast, P segregation due to thermal aging at higher temperatures has been confirmed by various studies. Gurovich [21], while investigating both eastern and western steels has attributed an essential part of the RE to P segregation to interfaces. The investigation by Oleg Zabusov [22] showed that continuous exposure of VVER-1000 RPV material at 320 °C caused an increase in P content at grain boundaries. It was observed that this effect could be due to the presence of carbides at the grain boundaries with increased P content at the carbide/matrix interface and lead to intergranular embrittlement of unirradiated RPV components by the end of the extended service life. Yaroslav I. Shtrombakh et al. [11] analysed VVER 1000 RPV steel (base material and welds) and found that found that during long term exposures at operating temperatures of 310–320 °C, thermal aging effects are observed only when the Ni content is high and are the result of the grain boundary segregation. Pierre et al. [23] has shown a shift in DBTT of 37 °C in 18MND5 (composition very close to 16MND5 grade used for RPVs) steels after thermal aging at 450 °C for 5000 h due to P segregation to grain boundaries. Antoine Andrieu et al. [24] using a mathematical model, has reconfirmed the DBTT shift of 40 °C in Mn-Ni-Mo steels due to thermal ageing at 450 °C for 5000 h. Broughton et al. [25] studied the changes in the grain boundary chemistry of D6ac (Ni-Cr-P) steels with heat treatments and found P concentrations at the grain boundary increased rapidly during the initial stages of ageing to reach a plateau level at 480 °C and a maximum value at both 520 and 560 °C. In a similar study of A533B plate [26] and A508 [27] forging materials, Druce et al. [28] studied the effects of thermal ageing in the temperature range 300–550 °C for durations up to 20,000 h and showed that ageing increased the DBTT by an amount dependent on the material, prior heat treatment, ageing temperature and time. The embrittling/ segregation potency of P was highly dependent on prior heat treatment; highest in the coarse-grained, higher hardness simulated HAZ condition, and the least in the conventionally quenched and tempered condition. Degradation in the material by formation and evolution of the carbides has also been widely reported. Nishizaka et al. [29] and Qu and Kuo [30] have shown that service-exposed Cr-Mo (reactor pressure vessel) and 2.25 Cr-Mo-V(steel bolts) embrittle due to precipitation or change in morphology of carbides during service. Williams and Wilshire [31] have further shown a progressive loss of the creep resistance of a 0.5Cr-0.5Mo-0.25V steel before and after service at elevated temperatures decreased with increasing intercarbide particle spacing and with
Fig. 1. Effect of irradiation on ductile to brittle transition temperature (DBTT) and upper shelf energy for ferritic steels.
Fig. 2. Three components of radiation embrittlement for RPV steels, as a function of neutron (energy > 0.5 MeV) fluence [8].
degradation is additive. Formation of radiation induced precipitates/ segregates (Cu-rich, may contain Ni also) and radiation defects in the steel matrix are hardening mechanisms resulting in increase in the yield strength and are primarily due to these features acting as obstacles for dislocation movement. The non-hardening mechanism of embrittlement comprises impurities segregation (primarily P) at grain boundaries and interfaces [10]. The presence of P at the grain boundary lowers the cohesion energy causing the material to fail through intergranular mode [3]. This is taken to be due to changes in local electron density and weakening of base material atomic bond energy [4]. The effects of precipitation and segregation are found to saturate with time whereas the direct matrix damage continues to increase all along its lifetime as shown in Fig. 2. [8]. Though most of the degradation in the properties is due to irradiation, thermal aging is known to contribute to these effects [11]. Corwin [12] brought out that a number of processes, viz., formation of hardening phases (such as copper-rich precipitates (CRP)) and segregation of P lead to the embrittlement of RPV steels subjected to longterm service at elevated temperatures. These processes are accelerated or enhanced due to irradiation. Pure thermal aging represents a zeroflux limit of damage rate-dependent effects [6]. Thermal ageing is a temperature, material state (microstructure) and time dependent degradation mechanism [6]. The degradation in the properties when LAS components are subjected to elevated operating temperatures is primarily of two types: (a) due to changes in microstructure and /or (b) due to diffusion of impurities such as P, Sn, As, and Sb to the grain boundaries/interfaces [13–15]. The changes in microstructure, such as carbide coarsening and precipitation of more stable carbides during service, can cause softening and irreversible embrittlement. Also carbide coarsening and precipitation of more stable carbides causes lowering of impact energy although there is reported to be no effect on DBTT [13]. In contrast, segregation of impurities at the grain 89
Materials Science & Engineering A 725 (2018) 88–97
T. Sharma et al.
decreasing alloying elements in the solid solution. Literature also show that coarsening of carbides, spheroidization of carbides, an increase in interparticle spacing, and precipitation of more stable carbides in the Cr-Mo-V and Cr-Mo type steels takes place when exposed to higher operating temperatures [13]. Yutaka et al. [16] has shown that degradation in properties of 2.25Cr-1Mo steels was due to presence of Mo6C carbides. He has shown that the finely dispersed acicular carbides of Mo2C at the beginning of the life provide effective barriers for dislocation movement. The effect of operation at high temperatures resulted in precipitation and growth of coarse Mo6C carbides which caused degradation in the mechanical properties. Yamashita et al. [32] has also shown the carbide coarsening of Mo6C in Cr-Mo-V steels.
Table 1 Elemental composition of Mn-Ni-Mo steel (wt%). C
Si
S
Mn
P
Cr
Ni
Fe
Mo
Cu
V
0.17
2.0
0.004
1.4
0.005
0.10
0.61
Rest
0.41
0.06
0.019
examined by electron back scattered diffraction (EBSD) technique. 2. Experimental 2.1. Material and heat treatment The material selected for the study is Mn-Ni-Mo steel, the composition of which was analysed using spark-optical emission spectroscopy (OES) and is presented in Table 1. The samples for the study were extracted from the cylindrical RPV forging of this material. The material was quenched and tempered (Q& T) as a part of the initial fabrication process. Samples of 12 × 12 × 60 mm were prepared and used for the heat treatments. Accelerated thermal aging of the above samples was carried out at 450 °C in a muffle furnace with a temperature control of ± 1 °C. The samples were periodically taken out of the furnace for testing after 700, 2100, 2800, 3500,4200, 4900, 5600, 7700 and 8400 h. The degradation of the RPV steels due to thermal aging has been accelerated by aging at 450 °C which is higher than 310–330 °C, the temperature range experienced during the operation of the reactor. The basis for this is the assumption that the degradation process remains the same at the operation and the aging temperatures. At higher temperatures, the microstructure may tend to change by grain growth and carbide spheroidization [18]. Accelerated thermal aging of the Cr-Mo-V steels was carried out at 450 °C by Yutaka watanabe [16]. Fukakura et al. [20]and they observed P segregation in SA508. K. J. Choi had also aged a dissimilar weld of alloy 152 and A533Gr B at 450 °C for 2750 h [37,38]. Thus, the temperature of 450 °C was chosen for accelerated isothermal aging in furnace and aging upto 8400 h duration was carried out in the present study.
1.1. Electrochemistry to measure the degradation The conventional methods for evaluation of the degradation are usually destructive and involve measurement of hardness or fracture toughness/impact energy, small punch tests or auger electron spectroscopy (AES) and are reliable [15,33]. But difficulty in extracting the samples from the industrial facilities in service and an involved testing could pose a serious restriction in the use of these methodologies. Presently, a host of non-destructive testing (NDT) [34] techniques are being explored to evaluate the degradation in the properties. Out of the various NDT methods, the chemical etching and electrochemical methods are gaining prominent attention due to the ease in their application. Electrochemical approach is relatively high-precision nondestructive detecting method in predicting the material degradation [34]. For 2.25Cr-lMo steels, potentiodynamic polarisation measurements in dilute sodium molybdate (Na2MoO4) solution was used for selective dissolution of coarse carbides of M6C type [16]. The peak value of current density, ΔIp, was shown as the measure of the coarse carbides M6C and was correlated to shifts in fracture appearance transition temperature (FATT) caused by carbide coarsening and with hardness change. Intergranular corrosion (IGC) occurred in temper embrittled samples of Cr-Mo-V samples due to polarisation in calcium nitrate (Ca (NO3)2) solution. The characteristic changes in the polarisation curves showed good correlation to the shifts in FATT due to temper embrittlement [16]. Selective dissolution of Mo-rich carbides in Cr-Mo-V steels by electrochemical polarisation in saturated picric acid showed correlation between the peak current of polarisation to the degradation [32]. Kwon et al. [35] studied the effect of thermal aging on mechanical and electrochemical behaviour of Cr-Mo-V alloys. The electrochemical corrosion characteristics were investigated by the potentiodynamic anodic polarisation and the reactivation methods in 50 wt% Ca(NO3)2 electrolyte, and the electrochemical characteristic values showed a good correlation with the rate of material degradation obtained from destructive testing [35]. The degree of temper embrittlement in 2.25Cr1Mo steels was evaluated by anodic and single loop electrochemical potentiokinetic reactivation (SL-EPR) polarisation curves of the embrittled specimens measured in 55 wt% Ca(NO3)2 solution at 30 °C. The difference in current density between active second peak and passive potential in an anodic polarisation test showed a linear correlation increase in FATT [36]. As shown above, the role of P segregation and carbide coarsening on the embrittlement due to thermal aging on certain Cr-Mo/Cr-Mo-V steels and use of electrochemistry for measurement of degradation has been reported. The aim of the present work is to investigate the degradation of the mechanical properties (indicated by measurement of hardness and impact energy) of Mn-Ni-Mo steels because of the long term thermal aging experienced during service. To accelerate the effects of the thermal aging, the temperature selected for aging was 450 °C. In this study aging was carried out for different durations. The effect of aging on the degradation of the mechanical properties as measured by impact tests were correlated to the characteristic aspects of the electrochemical polarisation and also to microstructural features as
2.2. Microhardness The microhardness measurements were made by a Vickers hardness tester using a 350gf load with a dwell time of 10 s and the results are reported in Vickers hardness (HV). An average of at least ten indentations was taken to establish the hardness value of the steels. The specimens were polished with successively finer grit of emery papers up to 1500 grit and then by diamond polish of 1 µm size prior to the measurement of hardness. 2.3. Microstructural examination Detailed microstructural examination of the samples was performed by EBSD using a FEI™ Quanta 3D- field emission gun (FEG) scanning electron microscope (SEM). Samples were prepared by standard metallographic preparation consisting of mechanical polishing upto 2000 grit followed by polishing using a diamond paste of 1 µm. These samples were then electro polished in 10% perchloricacid (HClO4) + 90% methanol (CH3OH) solution at − 20 °C using 20 V DC. A scan step size of 0.3 µm was used for EBSD measurements. For the EBSD analysis, data with more than 0.8 confidence index (CI) was used. The data analysis was done using TSL-OIM™ EBSD software. The nature of the precipitates was analysed by X-ray Diffraction (XRD) studies. For this, three samples with different aging (as-received, 3500 h and 7700 h) were chosen to represent the changes taking place due to accelerated thermal aging. Small specimens of 1 cm x 1 cm x 0.5 cm were cut, polished progressively up to 1200 grit emery papers and electro polished. The Mo Kα X-ray (wave length-0.70930 Å) was 90
Materials Science & Engineering A 725 (2018) 88–97
T. Sharma et al.
used and the scan was performed in the 2-theta range of 14–32° with step of 0.02° for all the runs. 2.4. Impact testing Toughness properties were measured using Charpy V-notch impact tests as per ASTM Standard E23-16b using standard sized specimens (10 mm x 10 mm × 55 mm), broken in an instrumented pendulum type impact machine [39]. The samples were prepared such that the notch was oriented in the radial direction of the original RPV forging block from the samples extracted. The isothermal aged samples of different duration were fractured at − 12 °C and three samples were tested for each condition and average values were reported. Fig. 3. Showing the measurement of Ecorr and Icorr from Tafel extrapolated lines.
2.5. Fractographic studies To establish the mode of the fracture and to assess the fraction of brittle fracture, the fracture surface from impact test (fractured halves after impact tests at − 12 °C) was subjected to a detailed microscopic examination using a FEI™ Quanta 3D-FEG SEM. To prevent oxidation of the fractured surfaces, the fractured samples were preserved in methanol immediately after the fracture and loaded into the SEM chamber with minimum exposure to atmosphere. To obtain an overall picture, the fracture surface analysis was carried out at macrolevel (at magnifications of 50–100×) and then the full fracture surface was recreated. Detailed analysis of the fracture topography and estimates of the regions with different modes of fracture were done by obtaining images at microlevel (at magnifications of 100 − 5000×). For each sample, the share of different types of fracture – brittle (intergranular (IG)/transgranular (TG)) fracture and ductile fracture, was determined using image processing software. The relative error of determining the share of brittle fracture in the fracture surface of fractured specimen by this method was 10–15% [40]. Grain boundary embrittlement was assessed by the share of brittle IG fracture. The loss in the ductility was also measured based on the lateral contraction analysis [41]. In this method, lateral contraction i.e., maximum contraction in specimen thickness direction ahead of the notch of the broken Charpy sample, was employed as the parameter to represent the absorbed energy. The measurement of the maximum contraction was done by taking the SEM image of the fractured surface and using an image analysis software. The lateral contraction strain (ratio of the maximum contracted length to the original dimension of 10 mm) was calculated for each aging condition.
and the remaining area was electrically insulated using a lacquer. To remove the oxide film formed on the exposed surface, the specimen was cathodically cleaned at − 1 VSCE for 1 min soon after specimen's immersion into the solution. The open circuit potential (OCP) was allowed to stabilize and then the potentiodynamic polarisation tests were performed at room temperature at a scan rate of 20 mV/min. Anodic polarisation tests were initiated at a potential 0.2 V cathodic to the measured OCP. Scans were continued up to commencement of the transpassive zone in each case but were not taken into the transpassive potential zone. The corrosion potential (Ecorr) and the corrosion current density (Icorr) for each aged sample was determined by the intersection of the extrapolated linear portions of the anodic and cathodic part of the polarisation curve (tafel analysis) as shown in Fig. 3. This was performed using the EC lab software. After the polarisation test, the tested specimen surface was dried and examined using optical microscopy (ZEISS make) as well as FESEM. The magnitude of attack after the polarisation test was estimated in two ways: (a) by measuring the width of the grain boundary attacked regions and (b) the area of the attacked grain boundary surface as a percentage of the total area. For this, the optical image of the attacked surface (at 1000 ×) was taken and the width and area of the attacked grain boundary were measured using ImageJ software. 3. Results and discussion 3.1. Characterization of as-received sample The microstructure of the sample in as-received (Q&T) condition is shown in Fig. 4. As can be seen from Fig. 4(a)&(b), the morphology is of a typical bainitic structure with lath length in the range of 20–30 µm and width in the range of 2–5 µm. Further the laths are seen to have several low angle grain boundaries inside them as shown in Fig. 4(c). This structure provides a unique combination of strength and toughness. The average hardness of the as-received material is 215HV.The ASTM grain size (ASTM) number measured by the analysis software (TSL-OIM™ EBSD software) is 15 and the average grain size as measured by standard linear intercept is 2.66 µm. The software considers a misorientation greater than 5° as a grain boundary. However, the ASTM grain number as measured from the optical micrographs (at 100× magnification) by drawing linear intercept was around 8.5. This indicated an average prior austenite grain size of 15–20 µm.
2.6. Electrochemical polarisation Electrochemical polarisation tests were performed in saturated picric acid solution at room temperature to obtain anodic polarisation curves (VSP 300, Biologic instruments). The solution was chosen based on the systematic investigation of the most sensitive solution for the selective attack at the prior austenitic grain boundary (PAGB). The other solutions considered were Na2MoO4 and 55 pct Ca(NO3)2 solution at 60 °C which did not give the desired results for Mn-Ni-Mo steel. All the electrochemical studies were carried out in a conventional threeelectrode cell using a Pt foil as the auxiliary (counter) electrode, saturated calomel electrode (SCE) as the reference electrode and the specimen as the working electrode. All the specimens for electrochemical tests were prepared by the standard procedure of grinding with emery papers with successively finer grit sizes up to 1200 grit and then mirror polished using 1 µm diamond paste. The specimen was prepared just before each electrochemical experiment. Electrical connection to the specimen was made by attaching an electrical wire to the top of the specimen and all these connections were covered with a teflon tape to avoid any contact with the solution. The lower part of the specimen was immersed in the solution. The surface area exposed for the test was approximately 1 cm2
3.2. Change in impact toughness and micro hardness by aging The degradation of the mechanical properties is most commonly ascertained by the loss of toughness as assessed by impact tests carried out at a specified temperature. In the present study, the variation in the impact energy with aging duration is shown in Fig. 5(a). The results show an asymptotic decrease in the impact energy with aging 91
Materials Science & Engineering A 725 (2018) 88–97
T. Sharma et al.
Fig. 4. EBSD analysis of Mn-Ni-Mo RPV material showing (a) inverse pole figure (IPF) map of the as-received material. In the same region, (b) high angle (> 15°blue) boundaries are shown. (c) Selected region of (b) is used to indicate low angle (2–5°-red, 5–15°-green) boundaries inside the martensite laths. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
and IG fracture (as a fraction of increasing brittle fracture). Out of the various modes of fracture, the presence of brittle IG fracture regions on the fracture surfaces containing brittle transgranular (TG - cleavage and quasicleavage) fracture and ductile regions, has been considered as an evidence for the reduction in the grain boundary cohesive strength attributed to P segregation [40]. Hence, in converse, the share of brittle IG fracture on fracture surfaces of Charpy specimens has been taken as the measure of the P segregation to grain boundary [11,40,43,44]. The results of the lateral contraction analysis of the fracture surface are shown in Fig. 8. As shown, the lateral contraction/compressive strain is found to be reducing with increasing aging duration indicating loss of ductility. This reducing lateral contraction/compressive strain is also indicative of the degradation and agrees with the overall trend shown in the impact energy variation with aging (Fig. 5(a)). The decrease in ductility with aging duration is accompanied by increase in the IG brittle fracture with aging duration, as shown in the Fig. 7. Thus, the increase in the brittle IGF with aging duration is indicative of the loss of cohesiveness at the grain boundary.
temperature. The decrease in impact energy in the initial stages of the aging is small but drops drastically at higher aging durations. On the other hand, the overall hardness did not show any reduction, as shown in Fig. 5(b). As reported, the non-hardening degradation is indicative of the segregation of elements to grain / interfaces boundaries whereas the reduction in the impact energy with increase in the hardness is attributed to changes in the phase composition due to formation of new phases and/or modification of existing density and size of hardening phases [11,42,43]. Thus, the changes in the mechanical properties, as shown in Fig. 5, are indicative of a non-hardening i.e. elemental segregation mechanism. 3.3. Fractographic studies The SEM micrographs of the fractured surface of the aged samples was taken to access the mode of fracture. Micrographs of three representative aging (as-received, 3500 h and 7700 h aged) is shown in Fig. 6. The as-received (unaged) specimen showed a fully ductile failure and no cleavage facets were observed. The typical as-received fractured surface is shown in Fig. 6(a)&(b). The aged samples showed a mixed mode of failure with an increasing fraction of brittle fracture. As shown in Fig. 7, the fraction of the IG fracture in the brittle fracture also increased with aging duration. The fraction of the IG failure reached nearly 65% of the total brittle fracture with aging duration. The typical fractographs of 3500 h and 7700 h aged samples are shown in Fig. 6(b) &(c) and Fig. 6(e)&(f). These indicate an increasing proportion of brittle
3.4. Microstructure of isothermally aged samples Degradation in the properties can also result from the changes occurring in the microstructure. Changes in the morphology and distribution of the precipitates can adversely affect the mechanical properties like creep resistance [16,29,31] and impact toughness [16].
Fig. 5. (a) Variation in the (a) impact energy and (b) micro-hardness with aging (450 °C) time. 92
Materials Science & Engineering A 725 (2018) 88–97
T. Sharma et al.
Fig. 6. SEM fractographs showing (a)-(b) typical ductile fracture in the as-received specimen, (c)-(d) Transgranular (TG) and (e)–(f) intergranular (IG) brittle fracture after respective aging (450 °C) treatments of 3500 h and 7700 h duration.
Fig. 8. Percentage strain, estimated from local lateral contraction [41], versus aging (450 °C) duration.
Fig. 7. Percentage ductility and intergranular (IG) fracture area (of brittle fracture) versus aging (450 °C) time. Experimental data points are shown with corresponding trend lines.
3.5. Size and distribution of carbides
Ultrafine alloying carbides are known to improve the creep resistance and strength while coarse carbides can become a source of crack [45]. Thus, to assess the changes in the microstructure with aging duration, microstructural examination of samples of three aged conditions of asreceived, 3500 h and 7700 h, was undertaken using EBSD technique (Fig. 9). As shown in the Fig. 9, the microstructure does not show any measurable changes with aging duration. The ASTM grain size number (as measured by the software) in all the three aged conditions is close to 15 (Fig. 10(a)). The distribution of different grain sizes (Fig. 9(a), (b)& (c)) and the average misorientation (Fig. 9(d), (e)&(f)) also did not change. This indicates that there are no major microstructural changes with aging heat treatment. This is further confirmed when the distributions of grain size and misorientation are plotted for the as-received, aged for 3500 h and for 7700 h samples in Fig. 10(a) and Fig. 10(b) respectively.
The SEM images of the samples aged at 450 °C for durations ranging up to 8400 h, after a hold for 400 s at OCP in saturated picric acid solution showed a high density of carbides along the interfaces (lath and GB) and in the matrix (Fig. 11). This result is significant as it indicates that the carbides did not change significantly in size or distribution even after different aging durations. To identify the nature of the carbides, the XRD of three specimens (as-received, aged for 3500 h and for 7700 h at 450 °C) was conducted. The results of the XRD studies, shown in Fig. 12. and it clearly indicates the presence of Mo2C type of carbides. A small peak corresponding to Fe3C was also observed. The intensity of the Mo2C/Fe3C peaks did not show any increase with aging duration indicating that there has been no growth in these carbides. Since, no new carbides have been identified, it is evident that no new carbides formed. However, XRD shows peaks corresponding to a phase/precipitate only if those phases/precipitates are above a certain minimum volume fraction. Therefore, it cannot be ruled out that small quantity of some other carbide was present. 93
Materials Science & Engineering A 725 (2018) 88–97
T. Sharma et al.
Fig. 9. EBSD maps showing IPF plus high angle (> 15°) grain boundaries for (a)-(b) as-received sample, and specimens subjected to (c)-(d) 3500 h and (e)-(f) 7700 h aging (450 °C) treatment.
Ca(NO3)2 solution at 60 °C, showed a very high level of attack on the entire sample and had no sensitivity to attack at specific features. On the other hand, attack was not observed on the specimens (i.e. no features could be observed either under optical microscope or under SEM) when the tests were carried out in Na2MoO4 and so were not very sensitive to the degradation. Based on the tests carried out in all the above electrolytes, results indicated that the picric acid solution was the only test that showed attack at microstructural features during the electrochemical tests for Mn-Ni-Mo steel. The potentiodynamic polarisation curves for various aging durations in saturated picric acid are shown in Fig. 13(a). All the specimens
3.6. Electrochemical polarisation Electrochemical potentiodynamic polarisation tests of the isothermally aged Mn-Ni-Mo specimens that showed severe degradation in the impact energy (Fig. 5(a)) were conducted. The electrochemical tests were done in three environments that were often reported to establish the degradation:(1) 55% Ca(NO3)2 solution at 60 °C, (2) 0.001 M Na2MoO4 at room temperature (25 °C), and (3) saturated picric acid solution at room temperature. However, the polarisation behaviour of the specimens in these electrolytes was very different. Electrochemical polarisation and etching behaviour of the specimens examined in 55%
Fig. 10. (a) ASTM grain size distribution (measured by standard linear intercept method) and (b) misorientation distribution for as-received sample and specimens subjected to 3500 h and 7700 h aging (450 °C) treatment. 94
Materials Science & Engineering A 725 (2018) 88–97
T. Sharma et al.
Fig. 11. The SEM micrographs of specimens after a hold for 400 s at OCP in saturated picric acid showing similar size and distribution of carbides for (a) as-received condition. (b) aged for 3500 h, (c) aged for 7700 h at 450 °C.
and the thickness of the grain boundary attacked regions increased with aging duration. The average thickness of the grain boundary attack is plotted with aging duration and is shown in Fig. 15(a). This clearly shows that with aging duration, as the segregation at PAGB increased, the electrochemical polarisation in picric acid caused more selective attack at the PAGB boundaries having segregation (Fig. 15(a)). Therefore, the width of attack at PAGB increased with aging duration at 450 °C (Fig. 15(a)). The area of the surface attacked as percentage of the total area is shown in Fig. 15(b) (attacked area shown in red). Similar to the grain boundary width, the area of the surface (at grain boundary) attacked also increased with increasing aging duration. The trend of increase in width of the grain boundary attack and in the attacked grain boundary surface area (Figs. 14 and 15) has clearly shown a correlation with the decrease in the impact energy (Fig. 5(a)). The increase in the share of the IG fracture with aging duration (Fig. 7) and the selective attack of the picric acid to the grain boundary (Fig. 15(b)) suggest that the degradation is localised to the changes taking place at the grain boundary. Picric acid has been shown to selectively attack P segregated at the grain boundary [14] and has been used as an etchant also [14,46–48]. The sensitivity of the picric acid to attack/etch regions rich in P has been used in the polarisation tests also. Thus, the increase in the grain boundary attack with polarisation in picric acid can be attributed to increased P levels at the grain boundary.
Fig. 12. X-ray Diffraction pattern of as-received sample and samples isothermally aged at 450 °C for 3500 and 7700 h.
showed similar overall trend in the polarisation. The observed change has been the trend of increase in Ecorr with aging duration. As shown in Fig. 13(b), Ecorr has shown a steady increase from − 112 mVSCE for the as-received sample to − 61 mVSCE of the 8400 h aged sample. Icorr has also shown an increasing trend with aging. Also, the increase in the Ecorr is seen to saturate with aging, while Icorr seems to increase linearly from 118 to 390 µm/cm2. Though, the current density in the passive potential regime is in a narrow band (Fig. 13(a)). The surface of the specimen after the complete polarisation was examined under an optical microscope and the results are shown in Fig. 14. As can be seen in the Fig. 14, picric acid has attacked the PAGB
3.7. Mechanism of IG fracture upon aging at 450 °C for Mn-Ni-Mo steel The degradation of LAS has been reported to be either due to carbide precipitation and growth or segregation of elements to interfaces (grain boundaries or laths). Though such mechanisms have been investigated by many in Cr-Mo(-V) steels, a detailed study of the same has not been reported for the Mn-Ni-Mo steels. Thermal aging of Mn-Ni-Mo steels, used as RPV material, has been
Fig. 13. (a) Polarisation behaviour of the isothermally aged samples. (b) variation in Ecorr and Icorr with aging (450 °C) duration. In (b), data points and corresponding trendlines are shown. 95
Materials Science & Engineering A 725 (2018) 88–97
T. Sharma et al.
Fig. 14. The developed microstructure, after the polarisation test in saturated picric acid solution showing increasing preferential attack at prior austenitic grain boundaries (PAGB) with increase in aging duration. (a)-(b) as-received condition, (c)-(d) 3500 h aged, (e)-(f) 7700 h aged conditions.
Fig. 15. Increasing (a) width of attack at PAGB and (b) attacked surface area after polarisation in saturated picric acid solution with increasing aging duration at 450 °C.
aging at 450 °C did not result in any changes in size or distribution of carbides (Fig. 11) and the nature of carbides too remained the same, i.e. Mo2C type (Fig. 12). Also, aging at 450 °C did not result in any changes in either grain size or misorientation of grains (Fig. 9(a)& (b)). Similarly, hardness too did not change upon aging upto 8400 h at 450 °C (Fig. 5(b)). The impact energy decreased monotonically as the aging duration was increased (Fig. 5(a)). The fracture surface analyses showed that the ductility (assessed from fracture surface analysis) reduced and IG fracture fraction increased with increased aging duration (Fig. 7). The electrochemical tests in saturated picric acid showed increased width of attack at PAGBs (Figs. 14 and 15) with increase in aging duration. The size of the PAGBs, as revealed in electrochemical test (Fig. 14) and the size of IG features revealed on fracture surfaces (Fig. 6(e)&(f)) matches quite well. This confirms that the segregation did take place pre-dominantly at PAGBs rather than at low or high
accelerated by isothermal heat treatment at 450 °C. The Impact test of these aged samples show steady drop in the impact energy with aging duration (Fig. 5(a)) indicating loss of ductility. The loss of ductility is also evident from the fractographic studies carried out (Figs. 7 and 8). In addition, the hardness of the specimens has not shown any appreciable change with aging duration even upto 8400 h (Fig. 5(b)), indicating the degradation to be of non-hardening (i.e. elemental segregation) type. Morphology, type of carbides, and the amount of alloying elements present in the ferrite are considered the main factors affecting the strengths of a low alloy ferritic steel. The carbide morphology and alloying elements in solid solution control movement of dislocations in the ferrite grains [13]. It has been suggested that the creep strength of these steels decreases as the carbides grow and lose coherency with the matrix, and is also affected by the level of molybdenum content in the matrix. However, the present study on Mn-Ni-Mo steel has shown that 96
Materials Science & Engineering A 725 (2018) 88–97
T. Sharma et al.
angle grain boundaries that exist inside the main features in the microstructure. Therefore, in the present study on Mn-Ni-Mo steel, a heat treatment at 450 °C could show thermal embrittlement in aging periods as low as 3500 h (and upto 7700 h). The embrittlement is also shown to be elemental segregation type (i.e. non-hardening type) and at PAGBs. This is also supported by the observations of no change in other possible modes of embrittlement e.g. size or distribution of carbides or nature of carbides (Figs. 11 and 12) nor any change in grain size or misorientation (Figs. 9 and 10). As shown by Yutaka et al. [16] and others et al. [11,21–23], [25,28]. RPV steels develop non-hardening type embrittlement by segregation of P at grain boundaries for Cr-Mo-V type steels. It has been reported that P does segregate to PAGB of Mn-Ni-Mo steels and its weldments [23,24,28]. Though, P has not been directly shown to be segregated at grain boundaries in this study, there are some studies reported [40] that used in-situ fracture and auger electron spectroscopy (AES) to establish P segregation. Also, there are some other studies that have shown [16,20,37] that 450 °C aging heat treatment results in P segregation in RPV steels. Therefore, in this study, where we have also used the same aging treatment at 450 °C, it is taken that P would segregate at grain boundaries.
2009. [5] B.A. Gurovich, E.A. Kuleshova, O.V. Lavrenchuk, K. Prikhodko, Y.I. Shtrombakh, J. Nucl. Mater. 264 (2007) 333–353. [6] IAEA-TECDOC-1556, Assessment and Management of Ageing of Major Nuclear Power Plant Components Important to Safety: PWR Pressure Vessels, 2007. [7] J. Kim, Nucl. Eng. Des. 174 (1997) 51–58. [8] L. Debarberis, A. Kryukov, F. Gillemot, B. Acosta, F. Sevini, Int. J. Press. Vessel. Pip. 82 (2005) 195–200. [9] G. Maussner, L. Scharf, R. Langer, B. Gurovich, Nucl. Eng. Des. 193 (1999) 359–376. [10] B. Gurovich, E. Kuleshova, Y. Shtrombakh, S. Fedotova, D. Maltsev, a. Frolov, O. Zabusov, D. Erak, D. Zhurko, J. Nucl. Mater. 456 (2015) 23–32. [11] Y.I. Shtrombakh, B.A. Gurovich, E.A. Kuleshova, D.A. Maltsev, S.V. Fedotova, A.A. Chernobaeva, J. Nucl. Mater. 452 (2014) 348–358. [12] W.R. Corwin, G.R. Odette, R.E. Stoller, J.A. Wang, in: Int. Conf. Struct. Mech. React. Technol. CONF-950804-–3, United States, 1995. [13] N.S. Cheruvu, Metall. Trans. A 20 (1989) 87–97. [14] R. Viswanathan, S.M. Bruemmer, J. Eng. Mater. Technol. Trans. ASME 107 (1985). [15] L. Debarberis, B. Acosta, A. Zeman, S. Pirfo, P. Moretto, P. Chernobaeva, Y. Nikoleav, Int. J. Microstruct. Mater. Prop. 2 (2007) 326–338. [16] Y. Watanabe, T. Shoji, Metall. Trans. A 22 (1991) 2097–2106. [17] F. Nogata, K. Seo, H. Takahashi, K. Tamagawa, S.H. Chung, Int. J. Press. Vessel. Pip. 45 (1991) 367–379. [18] C.-S. Seok, J.-M. Koo, Mater. Sci. Eng. A 395 (2005) 141–147. [19] M.J. DeVan, I.A.L. Lowe, C.S. Wade, A.S. Kumar, D.S. Gelles, R.K. Nanstad, E.A. Little, in: Eff. Radiat. Mater. 16th Int. Symp. ASTM STP 1175, Philedelphia, 1993. [20] J. Fukakura, M. Asano, M. Kikuchi, M. Ishikawa, Nucl. Eng. Des. 144 (1993) 423–429. [21] B.A. Gurovich, E.A. Kuleshova, Y.A. Nikolaev, Y.I. Shtrombakh, J. Nucl. Mater. 246 (1997) 91–120. [22] O.O. Zabusov, B.A. Gurovich, E.A. Kuleshova, M.A. Saltykov, S.V. Fedotova, A.P. Dementjev, Key Eng. Mater. 592–593 (2013) 577–581. [23] J. Pierre, R. Francois, P. Christophe, in: Proc. ASME 2013 Press. Vessel. Pip. Conf. PVP 2013, Paris, 2013, pp. 1–8. [24] A. Andrieu, A. Pineau, P. Joly, F. Roch, D. Ryckelynck, Procedia Mater. Sci. 3 (2014) 655–660. [25] W.R. Broughton, P.J.K. Paterson, Appl. Suface Sci. 15 (1983) 120–128. [26] ASTM A533/A533M-09 (reapproved2014), Asme Boil. Press. Vessel Code 9, 2004, pp. 9–12. [27] ASTM A508/A 508M, 1, 2015, pp. 1–5. [28] S. Druce, G. Gage, G. Jordan, Acta Metall. 34 (1986) 641–652. [29] Y. Nishizaka, Y. Hara, A. Hori, H. Tsukahara, K. Miyano, T. Wada, T.B. Cox, J. Press. Vessel Technol. 107 (1985) 285–294. [30] Z. Qu, K.H. Kuo, Metall. Trans. A Phys. Metall. Mater. Sci. 12 A (1981) 1333–1337. [31] K.R. Williams, B. Wilshire, Mater. Sci. Eng. 47 (1981) 151–160. [32] M. Yamashita, K. Nobuo, N. Sakuma, T. Watanabe, Method of Detecting Carbides in Alloy Steels by Electrochemical Polarisation, US Patent Number 5,505,827, 1996. [33] Z. Shenghan, L. Yaling, T. Yu, J. Spectrosc. 2015 (2015) 9–12. [34] L. Yaling, Z. Shenghan, T. Yu, J. Chem. Pharm. Res. 6 (2014) 1435–1439. [35] I.H. Kwon, S.S. Baek, H.S. Yu, Int. J. Press. Vessel. Pip. 80 (2003) 157–165. [36] T. Shibuya, T. Misawa, Zairo-to-Kankyo 154 (1999) 146–154. [37] K.J. Choi, J.J. Kim, B.H. Lee, C.B. Bahn, J.H. Kim, J. Nucl. Mater. 441 (2013) 493–502. [38] S.C. Yoo, K.J. Choi, C.B. Bahn, S.H. Kim, J.Y. Kim, J.H. Kim, J. Nucl. Mater. 459 (2015) 2–12. [39] ASTM E 23-12c, ASTM Int., 2013. [40] B. Gurovich, E. Kuleshova, O. Zabusov, S. Fedotova, A. Frolov, M. Saltykov, D. Maltsev, J. Nucl. Mater. 435 (2013) 25–31. [41] Y. Takashima, T. Nishi, H. Shoji, M. Ohata, F. Minami, Weld. World 58 (2014) 289–295. [42] B. Gurovich, E. Kuleshova, S. Fedotova, D. Maltsev, O. Zabusov, A. Frolov, D. Erak, D. Zhurko, in: Fontevraud 8 - Contrib. Mater. Investig. Oper. Exp. to LWRs' Safety, Perform. Reliab. Fr. Avignon, Sept. 14–18, 2014. [43] B. Gurovich, E. Kuleshova, Y. Shtrombakh, S. Fedotova, D. Erak, D. Zhurko, J. Nucl. Mater. 456 (2015) 373–381. [44] Y.I. Shtrombakh, B.A. Gurovich, E.A. Kuleshova, A.S. Frolov, S.V. Fedotova, D.A. Zhurko, E.V. Krikun, J. Nucl. Mater. 461 (2015) 292–300. [45] Z. Zhai, Y. Miyahara, H. Abe, Y. Watanabe, J. Nucl. Mater. 452 (2014) 133–140. [46] Z. Zhai, H. Abe, Y. Miyahara, Y. Watanabe, Corros. Sci. 92 (2015) 32–42. [47] Y. Kadoya, T. GOTO, M. Takei, Nirou Haruki, Takeshi Ikuno, JSME Int. J. 35 (1992) 226–233. [48] Y.M. Chen, Y. Long, Z.G. Han, Y.H. Wang, in:, Asia-Pacific Power Energy Eng. Conf. APPEEC, 2009, pp. 2–5.
4. Conclusions 1. Mn-Ni-Mo type RPV steel showed embrittlement after prolonged (8400 h) thermal aging (450 °C) treatment. Impact energy decreased and intergranular (IG) brittle fracture area fraction increased with aging time. However, there was no measurable change in average hardness, grain size, misorientation and carbide size. 2. Electrochemical polarisation tests, in picric acid, showed clear attack at the prior austenite grain boundaries (PAGB). Width of such boundaries increased monotonically with corresponding drop in impact energy. This indicated phosphorous segregation leading to grain boundary embrittlement. Acknowledgements The authors thank Dr A. B. Mukherjee, Distinguished Scientist, AD (RPG), BARC and Shri Vivek Shrivastav, SO(F), LWRD, BARC for their help in sourcing the material. The authors would also like to thank Dr Amit Verma, MMD, BARC for his assistance in XRD measurements and analysis. Data availability The raw/processed data required to reproduce these findings cannot be shared at this time as the data also forms part of an ongoing study. References [1] K.D. Haverkamp, K. Forch, K.H. Piehl, W. Witte, Nucl. Eng. Des. 81 (1984) 207–217. [2] L.M. Davies, Int. J. Press. Vessel. Pip. 76 (1999) 163–208. [3] Review of Phosphorus Segregation and Intergranular Embrittlement in Reactor Pressure Vessel Steels (PWRMRP-19) PWR Materials Reliability Project (PWRMRP), Palo Alto, 2000. [4] I.N.E.S. NP-T-3.11, Integrity of Reactor Pressure Vessels in Nuclear Power Plants: Assessement of Irradiation Embrittlement Effects in Reactor Presseure Vessel Steels,
97
Contents lists available at ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
Detection of intergranular embrittlement of reactor pressure vessel steel by electrochemical method Tenneti Sharmaa,b, Sunil Kumar Bonaganic, N. Naveen Kumarc, I. Samajdard, V. Kaina,c,
T
⁎
a
Homi Bhabha National Institute, Trombay, Mumbai 400085, India Indian Navy, India c Materials Group, Bhabha Atomic Research Centre, Trombay, Mumbai 400085, India d Department of Metallurgical Engineering & Materials Science, Indian Institute of Technology Bombay, Powai, Mumbai 400076, India b
A R T I C LE I N FO
A B S T R A C T
Keywords: Low Alloy Steel Aging Impact Toughness Carbide precipitates Segregation Electrochemical polarisation test
The reactor pressure vessel (RPV) steel, with manganese-nickel-molybdenum (Mn-Ni-Mo) alloying, was subjected to accelerated thermal aging treatment at 450 °C for durations upto 8400 h. Impact toughness decreased and the intergranular fracture (IGF) increased monotonically with increased aging time. This was accompanied by insignificant changes in average hardness, grain size, misorientation and carbide size. However, electrochemical polarisation tests in picric acid showed increasing width of attack at prior austenitic grain boundaries (PAGB). It was hypothesized that the accelerated aging led to phosphorus segregation at the grain boundary, a hypothesis supported by the post electrochemical attack on the PAGB, leading to a significant loss in impact properties.
1. Introduction Low Alloy Steels (LAS) are used as the structural material in fabrication of the nuclear reactor pressure vessel (RPV) due to the excellent high temperature mechanical properties it provides and its compatibility in the operating environment of the reactor. The nuclear RPV is exposed to neutron fluence in addition to the prolonged high operating temperature it sees all along the operation life of the reactor. These steels are also exceedingly sensitive to heat treatments that are a part of the fabrication processes [1] and the resulting microstructures have a wide scatter in the desirable properties. Since the life of the nuclear plant is decided by the RPV, understanding its microstructural evolution that takes place during fabrication and subsequently degradation during its operation is of paramount importance. Measuring the degradation and assessing the residual life of the material is essential to ensure safe plant operation and also to extend the life of the plant beyond the design life. The two major types of steels used worldwide for construction of RPVs are Mn-Ni-Mo steels (commonly referred to as western type RPVs) and Cr-Mo-V steels (commonly referred to as eastern type RPVs) [2–4]. Grades 15Kh2NMFAA and 15Kh2MFA [5] are examples of eastern grade RPV steels and 20MnMoNi55, SA508 grade 3, SA533, SA 302B are some of the grades of western RPVs [2–4,6,7]. The RPV materials having ferritic/ferritic-bainitic microstructure [4] (or body centered
⁎
cubic (bcc) crystal structure) show the typical ductile to brittle transition (DBT) phenomena (Fig. 1) [2]. At low temperature the bcc materials show low impact toughness while at room temperature or higher temperatures, these materials show high impact toughness (Fig. 1). The transition from low to high impact toughness with increase in temperature (typical S curve) is characterized by ductile to brittle transition temperature (DBTT) that is typically taken as the middle point between the high and the low plateau of toughness. Irradiation shifts the typical S curve to right side and lowers the upper shelf impact energy drastically (Fig. 1). Radiation embrittlement (RE) manifests as lowering of upper shelf energy and DBTT shift to higher temperature. Both western and eastern materials have distinguishable chemical compositions and thus it is expected that their response to the radiation embrittlement is also different. Though the radiation embrittlement of both types of the steels depends on the operating fluence, the major factor governing the irradiation sensitivity is the amount of the deleterious elements present. The main elements identified to contribute to RE are phosphorus (P) and copper. Nickel has also been reported to increase the irradiation damage sensitivity caused by copper [2]. The primary mode of degradation of the properties of the RPV steels is radiation embrittlement. Three basic mechanisms are considered to contribute to radiation embrittlement: (a) direct matrix damage by high energy neutrons, (b) irradiation-induced precipitation and (c) element segregation [8,9]. The effect of these three components to the overall
Corresponding author at: Homi Bhabha National Institute, Trombay, Mumbai 400085, India. E-mail address: [email protected] (V. Kain).
https://doi.org/10.1016/j.msea.2018.04.008 Received 21 January 2018; Accepted 3 April 2018 Available online 04 April 2018 0921-5093/ © 2018 Elsevier B.V. All rights reserved.
Materials Science & Engineering A 725 (2018) 88–97
T. Sharma et al.
boundaries results in temper embrittlement of these steels that does not affect the hardness and increases the DBTT but is reversible [13]. Cr-Mo steels are shown to be largely susceptible to degradation by carbide coarsening and Cr-Mo-V steels by segregation to grain boundaries [16]. Takahashi et al. [17] have shown P segregation in Cr-Mo (2.25 Cr-1Mo) and Cr-Mo-V steels. In the case of X20CrMoV12.1 alloy steel pipe, it was shown that the degradation of the material is mainly related to carbide coarsening [18]. The effects of thermal aging on the degradation of the properties have been studied by many. There is literature to suggest that the degradation effects of thermal embrittlement are negligible whereas a few researchers have suggested otherwise also. The effects of long-term aging at temperatures up to 350 °C (reactor operating temperature) on the ductile-to-brittle transition temperature of RPV steels have been studied by Corwin et al. [12]. Corwin et al. showed from literature data that there is no embrittlement at such reactor operation temperatures for operation periods of up to 100,000 h [12]. Similarly, DeVan et al. [19] reported that DBTT of RPV steel from the Arkansas 1 reactor shifted from − 1 to 9 °C after thermal ageing at 280 °C for 93,000 h. The material from the beltline region of the Oconee Unit 3 reactor showed an increase of about 1 °C after exposure to a temperature of 282 °C for 103,000 h [19]. Fukakura et al. [20] studied the effect of thermal aging on western grade RPV steel and concluded that after thermal aging for 10,000 h at temperatures of 350 °C, 400 °C and 450 °C, the increase in the DBTT was small. In contrast, P segregation due to thermal aging at higher temperatures has been confirmed by various studies. Gurovich [21], while investigating both eastern and western steels has attributed an essential part of the RE to P segregation to interfaces. The investigation by Oleg Zabusov [22] showed that continuous exposure of VVER-1000 RPV material at 320 °C caused an increase in P content at grain boundaries. It was observed that this effect could be due to the presence of carbides at the grain boundaries with increased P content at the carbide/matrix interface and lead to intergranular embrittlement of unirradiated RPV components by the end of the extended service life. Yaroslav I. Shtrombakh et al. [11] analysed VVER 1000 RPV steel (base material and welds) and found that found that during long term exposures at operating temperatures of 310–320 °C, thermal aging effects are observed only when the Ni content is high and are the result of the grain boundary segregation. Pierre et al. [23] has shown a shift in DBTT of 37 °C in 18MND5 (composition very close to 16MND5 grade used for RPVs) steels after thermal aging at 450 °C for 5000 h due to P segregation to grain boundaries. Antoine Andrieu et al. [24] using a mathematical model, has reconfirmed the DBTT shift of 40 °C in Mn-Ni-Mo steels due to thermal ageing at 450 °C for 5000 h. Broughton et al. [25] studied the changes in the grain boundary chemistry of D6ac (Ni-Cr-P) steels with heat treatments and found P concentrations at the grain boundary increased rapidly during the initial stages of ageing to reach a plateau level at 480 °C and a maximum value at both 520 and 560 °C. In a similar study of A533B plate [26] and A508 [27] forging materials, Druce et al. [28] studied the effects of thermal ageing in the temperature range 300–550 °C for durations up to 20,000 h and showed that ageing increased the DBTT by an amount dependent on the material, prior heat treatment, ageing temperature and time. The embrittling/ segregation potency of P was highly dependent on prior heat treatment; highest in the coarse-grained, higher hardness simulated HAZ condition, and the least in the conventionally quenched and tempered condition. Degradation in the material by formation and evolution of the carbides has also been widely reported. Nishizaka et al. [29] and Qu and Kuo [30] have shown that service-exposed Cr-Mo (reactor pressure vessel) and 2.25 Cr-Mo-V(steel bolts) embrittle due to precipitation or change in morphology of carbides during service. Williams and Wilshire [31] have further shown a progressive loss of the creep resistance of a 0.5Cr-0.5Mo-0.25V steel before and after service at elevated temperatures decreased with increasing intercarbide particle spacing and with
Fig. 1. Effect of irradiation on ductile to brittle transition temperature (DBTT) and upper shelf energy for ferritic steels.
Fig. 2. Three components of radiation embrittlement for RPV steels, as a function of neutron (energy > 0.5 MeV) fluence [8].
degradation is additive. Formation of radiation induced precipitates/ segregates (Cu-rich, may contain Ni also) and radiation defects in the steel matrix are hardening mechanisms resulting in increase in the yield strength and are primarily due to these features acting as obstacles for dislocation movement. The non-hardening mechanism of embrittlement comprises impurities segregation (primarily P) at grain boundaries and interfaces [10]. The presence of P at the grain boundary lowers the cohesion energy causing the material to fail through intergranular mode [3]. This is taken to be due to changes in local electron density and weakening of base material atomic bond energy [4]. The effects of precipitation and segregation are found to saturate with time whereas the direct matrix damage continues to increase all along its lifetime as shown in Fig. 2. [8]. Though most of the degradation in the properties is due to irradiation, thermal aging is known to contribute to these effects [11]. Corwin [12] brought out that a number of processes, viz., formation of hardening phases (such as copper-rich precipitates (CRP)) and segregation of P lead to the embrittlement of RPV steels subjected to longterm service at elevated temperatures. These processes are accelerated or enhanced due to irradiation. Pure thermal aging represents a zeroflux limit of damage rate-dependent effects [6]. Thermal ageing is a temperature, material state (microstructure) and time dependent degradation mechanism [6]. The degradation in the properties when LAS components are subjected to elevated operating temperatures is primarily of two types: (a) due to changes in microstructure and /or (b) due to diffusion of impurities such as P, Sn, As, and Sb to the grain boundaries/interfaces [13–15]. The changes in microstructure, such as carbide coarsening and precipitation of more stable carbides during service, can cause softening and irreversible embrittlement. Also carbide coarsening and precipitation of more stable carbides causes lowering of impact energy although there is reported to be no effect on DBTT [13]. In contrast, segregation of impurities at the grain 89
Materials Science & Engineering A 725 (2018) 88–97
T. Sharma et al.
decreasing alloying elements in the solid solution. Literature also show that coarsening of carbides, spheroidization of carbides, an increase in interparticle spacing, and precipitation of more stable carbides in the Cr-Mo-V and Cr-Mo type steels takes place when exposed to higher operating temperatures [13]. Yutaka et al. [16] has shown that degradation in properties of 2.25Cr-1Mo steels was due to presence of Mo6C carbides. He has shown that the finely dispersed acicular carbides of Mo2C at the beginning of the life provide effective barriers for dislocation movement. The effect of operation at high temperatures resulted in precipitation and growth of coarse Mo6C carbides which caused degradation in the mechanical properties. Yamashita et al. [32] has also shown the carbide coarsening of Mo6C in Cr-Mo-V steels.
Table 1 Elemental composition of Mn-Ni-Mo steel (wt%). C
Si
S
Mn
P
Cr
Ni
Fe
Mo
Cu
V
0.17
2.0
0.004
1.4
0.005
0.10
0.61
Rest
0.41
0.06
0.019
examined by electron back scattered diffraction (EBSD) technique. 2. Experimental 2.1. Material and heat treatment The material selected for the study is Mn-Ni-Mo steel, the composition of which was analysed using spark-optical emission spectroscopy (OES) and is presented in Table 1. The samples for the study were extracted from the cylindrical RPV forging of this material. The material was quenched and tempered (Q& T) as a part of the initial fabrication process. Samples of 12 × 12 × 60 mm were prepared and used for the heat treatments. Accelerated thermal aging of the above samples was carried out at 450 °C in a muffle furnace with a temperature control of ± 1 °C. The samples were periodically taken out of the furnace for testing after 700, 2100, 2800, 3500,4200, 4900, 5600, 7700 and 8400 h. The degradation of the RPV steels due to thermal aging has been accelerated by aging at 450 °C which is higher than 310–330 °C, the temperature range experienced during the operation of the reactor. The basis for this is the assumption that the degradation process remains the same at the operation and the aging temperatures. At higher temperatures, the microstructure may tend to change by grain growth and carbide spheroidization [18]. Accelerated thermal aging of the Cr-Mo-V steels was carried out at 450 °C by Yutaka watanabe [16]. Fukakura et al. [20]and they observed P segregation in SA508. K. J. Choi had also aged a dissimilar weld of alloy 152 and A533Gr B at 450 °C for 2750 h [37,38]. Thus, the temperature of 450 °C was chosen for accelerated isothermal aging in furnace and aging upto 8400 h duration was carried out in the present study.
1.1. Electrochemistry to measure the degradation The conventional methods for evaluation of the degradation are usually destructive and involve measurement of hardness or fracture toughness/impact energy, small punch tests or auger electron spectroscopy (AES) and are reliable [15,33]. But difficulty in extracting the samples from the industrial facilities in service and an involved testing could pose a serious restriction in the use of these methodologies. Presently, a host of non-destructive testing (NDT) [34] techniques are being explored to evaluate the degradation in the properties. Out of the various NDT methods, the chemical etching and electrochemical methods are gaining prominent attention due to the ease in their application. Electrochemical approach is relatively high-precision nondestructive detecting method in predicting the material degradation [34]. For 2.25Cr-lMo steels, potentiodynamic polarisation measurements in dilute sodium molybdate (Na2MoO4) solution was used for selective dissolution of coarse carbides of M6C type [16]. The peak value of current density, ΔIp, was shown as the measure of the coarse carbides M6C and was correlated to shifts in fracture appearance transition temperature (FATT) caused by carbide coarsening and with hardness change. Intergranular corrosion (IGC) occurred in temper embrittled samples of Cr-Mo-V samples due to polarisation in calcium nitrate (Ca (NO3)2) solution. The characteristic changes in the polarisation curves showed good correlation to the shifts in FATT due to temper embrittlement [16]. Selective dissolution of Mo-rich carbides in Cr-Mo-V steels by electrochemical polarisation in saturated picric acid showed correlation between the peak current of polarisation to the degradation [32]. Kwon et al. [35] studied the effect of thermal aging on mechanical and electrochemical behaviour of Cr-Mo-V alloys. The electrochemical corrosion characteristics were investigated by the potentiodynamic anodic polarisation and the reactivation methods in 50 wt% Ca(NO3)2 electrolyte, and the electrochemical characteristic values showed a good correlation with the rate of material degradation obtained from destructive testing [35]. The degree of temper embrittlement in 2.25Cr1Mo steels was evaluated by anodic and single loop electrochemical potentiokinetic reactivation (SL-EPR) polarisation curves of the embrittled specimens measured in 55 wt% Ca(NO3)2 solution at 30 °C. The difference in current density between active second peak and passive potential in an anodic polarisation test showed a linear correlation increase in FATT [36]. As shown above, the role of P segregation and carbide coarsening on the embrittlement due to thermal aging on certain Cr-Mo/Cr-Mo-V steels and use of electrochemistry for measurement of degradation has been reported. The aim of the present work is to investigate the degradation of the mechanical properties (indicated by measurement of hardness and impact energy) of Mn-Ni-Mo steels because of the long term thermal aging experienced during service. To accelerate the effects of the thermal aging, the temperature selected for aging was 450 °C. In this study aging was carried out for different durations. The effect of aging on the degradation of the mechanical properties as measured by impact tests were correlated to the characteristic aspects of the electrochemical polarisation and also to microstructural features as
2.2. Microhardness The microhardness measurements were made by a Vickers hardness tester using a 350gf load with a dwell time of 10 s and the results are reported in Vickers hardness (HV). An average of at least ten indentations was taken to establish the hardness value of the steels. The specimens were polished with successively finer grit of emery papers up to 1500 grit and then by diamond polish of 1 µm size prior to the measurement of hardness. 2.3. Microstructural examination Detailed microstructural examination of the samples was performed by EBSD using a FEI™ Quanta 3D- field emission gun (FEG) scanning electron microscope (SEM). Samples were prepared by standard metallographic preparation consisting of mechanical polishing upto 2000 grit followed by polishing using a diamond paste of 1 µm. These samples were then electro polished in 10% perchloricacid (HClO4) + 90% methanol (CH3OH) solution at − 20 °C using 20 V DC. A scan step size of 0.3 µm was used for EBSD measurements. For the EBSD analysis, data with more than 0.8 confidence index (CI) was used. The data analysis was done using TSL-OIM™ EBSD software. The nature of the precipitates was analysed by X-ray Diffraction (XRD) studies. For this, three samples with different aging (as-received, 3500 h and 7700 h) were chosen to represent the changes taking place due to accelerated thermal aging. Small specimens of 1 cm x 1 cm x 0.5 cm were cut, polished progressively up to 1200 grit emery papers and electro polished. The Mo Kα X-ray (wave length-0.70930 Å) was 90
Materials Science & Engineering A 725 (2018) 88–97
T. Sharma et al.
used and the scan was performed in the 2-theta range of 14–32° with step of 0.02° for all the runs. 2.4. Impact testing Toughness properties were measured using Charpy V-notch impact tests as per ASTM Standard E23-16b using standard sized specimens (10 mm x 10 mm × 55 mm), broken in an instrumented pendulum type impact machine [39]. The samples were prepared such that the notch was oriented in the radial direction of the original RPV forging block from the samples extracted. The isothermal aged samples of different duration were fractured at − 12 °C and three samples were tested for each condition and average values were reported. Fig. 3. Showing the measurement of Ecorr and Icorr from Tafel extrapolated lines.
2.5. Fractographic studies To establish the mode of the fracture and to assess the fraction of brittle fracture, the fracture surface from impact test (fractured halves after impact tests at − 12 °C) was subjected to a detailed microscopic examination using a FEI™ Quanta 3D-FEG SEM. To prevent oxidation of the fractured surfaces, the fractured samples were preserved in methanol immediately after the fracture and loaded into the SEM chamber with minimum exposure to atmosphere. To obtain an overall picture, the fracture surface analysis was carried out at macrolevel (at magnifications of 50–100×) and then the full fracture surface was recreated. Detailed analysis of the fracture topography and estimates of the regions with different modes of fracture were done by obtaining images at microlevel (at magnifications of 100 − 5000×). For each sample, the share of different types of fracture – brittle (intergranular (IG)/transgranular (TG)) fracture and ductile fracture, was determined using image processing software. The relative error of determining the share of brittle fracture in the fracture surface of fractured specimen by this method was 10–15% [40]. Grain boundary embrittlement was assessed by the share of brittle IG fracture. The loss in the ductility was also measured based on the lateral contraction analysis [41]. In this method, lateral contraction i.e., maximum contraction in specimen thickness direction ahead of the notch of the broken Charpy sample, was employed as the parameter to represent the absorbed energy. The measurement of the maximum contraction was done by taking the SEM image of the fractured surface and using an image analysis software. The lateral contraction strain (ratio of the maximum contracted length to the original dimension of 10 mm) was calculated for each aging condition.
and the remaining area was electrically insulated using a lacquer. To remove the oxide film formed on the exposed surface, the specimen was cathodically cleaned at − 1 VSCE for 1 min soon after specimen's immersion into the solution. The open circuit potential (OCP) was allowed to stabilize and then the potentiodynamic polarisation tests were performed at room temperature at a scan rate of 20 mV/min. Anodic polarisation tests were initiated at a potential 0.2 V cathodic to the measured OCP. Scans were continued up to commencement of the transpassive zone in each case but were not taken into the transpassive potential zone. The corrosion potential (Ecorr) and the corrosion current density (Icorr) for each aged sample was determined by the intersection of the extrapolated linear portions of the anodic and cathodic part of the polarisation curve (tafel analysis) as shown in Fig. 3. This was performed using the EC lab software. After the polarisation test, the tested specimen surface was dried and examined using optical microscopy (ZEISS make) as well as FESEM. The magnitude of attack after the polarisation test was estimated in two ways: (a) by measuring the width of the grain boundary attacked regions and (b) the area of the attacked grain boundary surface as a percentage of the total area. For this, the optical image of the attacked surface (at 1000 ×) was taken and the width and area of the attacked grain boundary were measured using ImageJ software. 3. Results and discussion 3.1. Characterization of as-received sample The microstructure of the sample in as-received (Q&T) condition is shown in Fig. 4. As can be seen from Fig. 4(a)&(b), the morphology is of a typical bainitic structure with lath length in the range of 20–30 µm and width in the range of 2–5 µm. Further the laths are seen to have several low angle grain boundaries inside them as shown in Fig. 4(c). This structure provides a unique combination of strength and toughness. The average hardness of the as-received material is 215HV.The ASTM grain size (ASTM) number measured by the analysis software (TSL-OIM™ EBSD software) is 15 and the average grain size as measured by standard linear intercept is 2.66 µm. The software considers a misorientation greater than 5° as a grain boundary. However, the ASTM grain number as measured from the optical micrographs (at 100× magnification) by drawing linear intercept was around 8.5. This indicated an average prior austenite grain size of 15–20 µm.
2.6. Electrochemical polarisation Electrochemical polarisation tests were performed in saturated picric acid solution at room temperature to obtain anodic polarisation curves (VSP 300, Biologic instruments). The solution was chosen based on the systematic investigation of the most sensitive solution for the selective attack at the prior austenitic grain boundary (PAGB). The other solutions considered were Na2MoO4 and 55 pct Ca(NO3)2 solution at 60 °C which did not give the desired results for Mn-Ni-Mo steel. All the electrochemical studies were carried out in a conventional threeelectrode cell using a Pt foil as the auxiliary (counter) electrode, saturated calomel electrode (SCE) as the reference electrode and the specimen as the working electrode. All the specimens for electrochemical tests were prepared by the standard procedure of grinding with emery papers with successively finer grit sizes up to 1200 grit and then mirror polished using 1 µm diamond paste. The specimen was prepared just before each electrochemical experiment. Electrical connection to the specimen was made by attaching an electrical wire to the top of the specimen and all these connections were covered with a teflon tape to avoid any contact with the solution. The lower part of the specimen was immersed in the solution. The surface area exposed for the test was approximately 1 cm2
3.2. Change in impact toughness and micro hardness by aging The degradation of the mechanical properties is most commonly ascertained by the loss of toughness as assessed by impact tests carried out at a specified temperature. In the present study, the variation in the impact energy with aging duration is shown in Fig. 5(a). The results show an asymptotic decrease in the impact energy with aging 91
Materials Science & Engineering A 725 (2018) 88–97
T. Sharma et al.
Fig. 4. EBSD analysis of Mn-Ni-Mo RPV material showing (a) inverse pole figure (IPF) map of the as-received material. In the same region, (b) high angle (> 15°blue) boundaries are shown. (c) Selected region of (b) is used to indicate low angle (2–5°-red, 5–15°-green) boundaries inside the martensite laths. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
and IG fracture (as a fraction of increasing brittle fracture). Out of the various modes of fracture, the presence of brittle IG fracture regions on the fracture surfaces containing brittle transgranular (TG - cleavage and quasicleavage) fracture and ductile regions, has been considered as an evidence for the reduction in the grain boundary cohesive strength attributed to P segregation [40]. Hence, in converse, the share of brittle IG fracture on fracture surfaces of Charpy specimens has been taken as the measure of the P segregation to grain boundary [11,40,43,44]. The results of the lateral contraction analysis of the fracture surface are shown in Fig. 8. As shown, the lateral contraction/compressive strain is found to be reducing with increasing aging duration indicating loss of ductility. This reducing lateral contraction/compressive strain is also indicative of the degradation and agrees with the overall trend shown in the impact energy variation with aging (Fig. 5(a)). The decrease in ductility with aging duration is accompanied by increase in the IG brittle fracture with aging duration, as shown in the Fig. 7. Thus, the increase in the brittle IGF with aging duration is indicative of the loss of cohesiveness at the grain boundary.
temperature. The decrease in impact energy in the initial stages of the aging is small but drops drastically at higher aging durations. On the other hand, the overall hardness did not show any reduction, as shown in Fig. 5(b). As reported, the non-hardening degradation is indicative of the segregation of elements to grain / interfaces boundaries whereas the reduction in the impact energy with increase in the hardness is attributed to changes in the phase composition due to formation of new phases and/or modification of existing density and size of hardening phases [11,42,43]. Thus, the changes in the mechanical properties, as shown in Fig. 5, are indicative of a non-hardening i.e. elemental segregation mechanism. 3.3. Fractographic studies The SEM micrographs of the fractured surface of the aged samples was taken to access the mode of fracture. Micrographs of three representative aging (as-received, 3500 h and 7700 h aged) is shown in Fig. 6. The as-received (unaged) specimen showed a fully ductile failure and no cleavage facets were observed. The typical as-received fractured surface is shown in Fig. 6(a)&(b). The aged samples showed a mixed mode of failure with an increasing fraction of brittle fracture. As shown in Fig. 7, the fraction of the IG fracture in the brittle fracture also increased with aging duration. The fraction of the IG failure reached nearly 65% of the total brittle fracture with aging duration. The typical fractographs of 3500 h and 7700 h aged samples are shown in Fig. 6(b) &(c) and Fig. 6(e)&(f). These indicate an increasing proportion of brittle
3.4. Microstructure of isothermally aged samples Degradation in the properties can also result from the changes occurring in the microstructure. Changes in the morphology and distribution of the precipitates can adversely affect the mechanical properties like creep resistance [16,29,31] and impact toughness [16].
Fig. 5. (a) Variation in the (a) impact energy and (b) micro-hardness with aging (450 °C) time. 92
Materials Science & Engineering A 725 (2018) 88–97
T. Sharma et al.
Fig. 6. SEM fractographs showing (a)-(b) typical ductile fracture in the as-received specimen, (c)-(d) Transgranular (TG) and (e)–(f) intergranular (IG) brittle fracture after respective aging (450 °C) treatments of 3500 h and 7700 h duration.
Fig. 8. Percentage strain, estimated from local lateral contraction [41], versus aging (450 °C) duration.
Fig. 7. Percentage ductility and intergranular (IG) fracture area (of brittle fracture) versus aging (450 °C) time. Experimental data points are shown with corresponding trend lines.
3.5. Size and distribution of carbides
Ultrafine alloying carbides are known to improve the creep resistance and strength while coarse carbides can become a source of crack [45]. Thus, to assess the changes in the microstructure with aging duration, microstructural examination of samples of three aged conditions of asreceived, 3500 h and 7700 h, was undertaken using EBSD technique (Fig. 9). As shown in the Fig. 9, the microstructure does not show any measurable changes with aging duration. The ASTM grain size number (as measured by the software) in all the three aged conditions is close to 15 (Fig. 10(a)). The distribution of different grain sizes (Fig. 9(a), (b)& (c)) and the average misorientation (Fig. 9(d), (e)&(f)) also did not change. This indicates that there are no major microstructural changes with aging heat treatment. This is further confirmed when the distributions of grain size and misorientation are plotted for the as-received, aged for 3500 h and for 7700 h samples in Fig. 10(a) and Fig. 10(b) respectively.
The SEM images of the samples aged at 450 °C for durations ranging up to 8400 h, after a hold for 400 s at OCP in saturated picric acid solution showed a high density of carbides along the interfaces (lath and GB) and in the matrix (Fig. 11). This result is significant as it indicates that the carbides did not change significantly in size or distribution even after different aging durations. To identify the nature of the carbides, the XRD of three specimens (as-received, aged for 3500 h and for 7700 h at 450 °C) was conducted. The results of the XRD studies, shown in Fig. 12. and it clearly indicates the presence of Mo2C type of carbides. A small peak corresponding to Fe3C was also observed. The intensity of the Mo2C/Fe3C peaks did not show any increase with aging duration indicating that there has been no growth in these carbides. Since, no new carbides have been identified, it is evident that no new carbides formed. However, XRD shows peaks corresponding to a phase/precipitate only if those phases/precipitates are above a certain minimum volume fraction. Therefore, it cannot be ruled out that small quantity of some other carbide was present. 93
Materials Science & Engineering A 725 (2018) 88–97
T. Sharma et al.
Fig. 9. EBSD maps showing IPF plus high angle (> 15°) grain boundaries for (a)-(b) as-received sample, and specimens subjected to (c)-(d) 3500 h and (e)-(f) 7700 h aging (450 °C) treatment.
Ca(NO3)2 solution at 60 °C, showed a very high level of attack on the entire sample and had no sensitivity to attack at specific features. On the other hand, attack was not observed on the specimens (i.e. no features could be observed either under optical microscope or under SEM) when the tests were carried out in Na2MoO4 and so were not very sensitive to the degradation. Based on the tests carried out in all the above electrolytes, results indicated that the picric acid solution was the only test that showed attack at microstructural features during the electrochemical tests for Mn-Ni-Mo steel. The potentiodynamic polarisation curves for various aging durations in saturated picric acid are shown in Fig. 13(a). All the specimens
3.6. Electrochemical polarisation Electrochemical potentiodynamic polarisation tests of the isothermally aged Mn-Ni-Mo specimens that showed severe degradation in the impact energy (Fig. 5(a)) were conducted. The electrochemical tests were done in three environments that were often reported to establish the degradation:(1) 55% Ca(NO3)2 solution at 60 °C, (2) 0.001 M Na2MoO4 at room temperature (25 °C), and (3) saturated picric acid solution at room temperature. However, the polarisation behaviour of the specimens in these electrolytes was very different. Electrochemical polarisation and etching behaviour of the specimens examined in 55%
Fig. 10. (a) ASTM grain size distribution (measured by standard linear intercept method) and (b) misorientation distribution for as-received sample and specimens subjected to 3500 h and 7700 h aging (450 °C) treatment. 94
Materials Science & Engineering A 725 (2018) 88–97
T. Sharma et al.
Fig. 11. The SEM micrographs of specimens after a hold for 400 s at OCP in saturated picric acid showing similar size and distribution of carbides for (a) as-received condition. (b) aged for 3500 h, (c) aged for 7700 h at 450 °C.
and the thickness of the grain boundary attacked regions increased with aging duration. The average thickness of the grain boundary attack is plotted with aging duration and is shown in Fig. 15(a). This clearly shows that with aging duration, as the segregation at PAGB increased, the electrochemical polarisation in picric acid caused more selective attack at the PAGB boundaries having segregation (Fig. 15(a)). Therefore, the width of attack at PAGB increased with aging duration at 450 °C (Fig. 15(a)). The area of the surface attacked as percentage of the total area is shown in Fig. 15(b) (attacked area shown in red). Similar to the grain boundary width, the area of the surface (at grain boundary) attacked also increased with increasing aging duration. The trend of increase in width of the grain boundary attack and in the attacked grain boundary surface area (Figs. 14 and 15) has clearly shown a correlation with the decrease in the impact energy (Fig. 5(a)). The increase in the share of the IG fracture with aging duration (Fig. 7) and the selective attack of the picric acid to the grain boundary (Fig. 15(b)) suggest that the degradation is localised to the changes taking place at the grain boundary. Picric acid has been shown to selectively attack P segregated at the grain boundary [14] and has been used as an etchant also [14,46–48]. The sensitivity of the picric acid to attack/etch regions rich in P has been used in the polarisation tests also. Thus, the increase in the grain boundary attack with polarisation in picric acid can be attributed to increased P levels at the grain boundary.
Fig. 12. X-ray Diffraction pattern of as-received sample and samples isothermally aged at 450 °C for 3500 and 7700 h.
showed similar overall trend in the polarisation. The observed change has been the trend of increase in Ecorr with aging duration. As shown in Fig. 13(b), Ecorr has shown a steady increase from − 112 mVSCE for the as-received sample to − 61 mVSCE of the 8400 h aged sample. Icorr has also shown an increasing trend with aging. Also, the increase in the Ecorr is seen to saturate with aging, while Icorr seems to increase linearly from 118 to 390 µm/cm2. Though, the current density in the passive potential regime is in a narrow band (Fig. 13(a)). The surface of the specimen after the complete polarisation was examined under an optical microscope and the results are shown in Fig. 14. As can be seen in the Fig. 14, picric acid has attacked the PAGB
3.7. Mechanism of IG fracture upon aging at 450 °C for Mn-Ni-Mo steel The degradation of LAS has been reported to be either due to carbide precipitation and growth or segregation of elements to interfaces (grain boundaries or laths). Though such mechanisms have been investigated by many in Cr-Mo(-V) steels, a detailed study of the same has not been reported for the Mn-Ni-Mo steels. Thermal aging of Mn-Ni-Mo steels, used as RPV material, has been
Fig. 13. (a) Polarisation behaviour of the isothermally aged samples. (b) variation in Ecorr and Icorr with aging (450 °C) duration. In (b), data points and corresponding trendlines are shown. 95
Materials Science & Engineering A 725 (2018) 88–97
T. Sharma et al.
Fig. 14. The developed microstructure, after the polarisation test in saturated picric acid solution showing increasing preferential attack at prior austenitic grain boundaries (PAGB) with increase in aging duration. (a)-(b) as-received condition, (c)-(d) 3500 h aged, (e)-(f) 7700 h aged conditions.
Fig. 15. Increasing (a) width of attack at PAGB and (b) attacked surface area after polarisation in saturated picric acid solution with increasing aging duration at 450 °C.
aging at 450 °C did not result in any changes in size or distribution of carbides (Fig. 11) and the nature of carbides too remained the same, i.e. Mo2C type (Fig. 12). Also, aging at 450 °C did not result in any changes in either grain size or misorientation of grains (Fig. 9(a)& (b)). Similarly, hardness too did not change upon aging upto 8400 h at 450 °C (Fig. 5(b)). The impact energy decreased monotonically as the aging duration was increased (Fig. 5(a)). The fracture surface analyses showed that the ductility (assessed from fracture surface analysis) reduced and IG fracture fraction increased with increased aging duration (Fig. 7). The electrochemical tests in saturated picric acid showed increased width of attack at PAGBs (Figs. 14 and 15) with increase in aging duration. The size of the PAGBs, as revealed in electrochemical test (Fig. 14) and the size of IG features revealed on fracture surfaces (Fig. 6(e)&(f)) matches quite well. This confirms that the segregation did take place pre-dominantly at PAGBs rather than at low or high
accelerated by isothermal heat treatment at 450 °C. The Impact test of these aged samples show steady drop in the impact energy with aging duration (Fig. 5(a)) indicating loss of ductility. The loss of ductility is also evident from the fractographic studies carried out (Figs. 7 and 8). In addition, the hardness of the specimens has not shown any appreciable change with aging duration even upto 8400 h (Fig. 5(b)), indicating the degradation to be of non-hardening (i.e. elemental segregation) type. Morphology, type of carbides, and the amount of alloying elements present in the ferrite are considered the main factors affecting the strengths of a low alloy ferritic steel. The carbide morphology and alloying elements in solid solution control movement of dislocations in the ferrite grains [13]. It has been suggested that the creep strength of these steels decreases as the carbides grow and lose coherency with the matrix, and is also affected by the level of molybdenum content in the matrix. However, the present study on Mn-Ni-Mo steel has shown that 96
Materials Science & Engineering A 725 (2018) 88–97
T. Sharma et al.
angle grain boundaries that exist inside the main features in the microstructure. Therefore, in the present study on Mn-Ni-Mo steel, a heat treatment at 450 °C could show thermal embrittlement in aging periods as low as 3500 h (and upto 7700 h). The embrittlement is also shown to be elemental segregation type (i.e. non-hardening type) and at PAGBs. This is also supported by the observations of no change in other possible modes of embrittlement e.g. size or distribution of carbides or nature of carbides (Figs. 11 and 12) nor any change in grain size or misorientation (Figs. 9 and 10). As shown by Yutaka et al. [16] and others et al. [11,21–23], [25,28]. RPV steels develop non-hardening type embrittlement by segregation of P at grain boundaries for Cr-Mo-V type steels. It has been reported that P does segregate to PAGB of Mn-Ni-Mo steels and its weldments [23,24,28]. Though, P has not been directly shown to be segregated at grain boundaries in this study, there are some studies reported [40] that used in-situ fracture and auger electron spectroscopy (AES) to establish P segregation. Also, there are some other studies that have shown [16,20,37] that 450 °C aging heat treatment results in P segregation in RPV steels. Therefore, in this study, where we have also used the same aging treatment at 450 °C, it is taken that P would segregate at grain boundaries.
2009. [5] B.A. Gurovich, E.A. Kuleshova, O.V. Lavrenchuk, K. Prikhodko, Y.I. Shtrombakh, J. Nucl. Mater. 264 (2007) 333–353. [6] IAEA-TECDOC-1556, Assessment and Management of Ageing of Major Nuclear Power Plant Components Important to Safety: PWR Pressure Vessels, 2007. [7] J. Kim, Nucl. Eng. Des. 174 (1997) 51–58. [8] L. Debarberis, A. Kryukov, F. Gillemot, B. Acosta, F. Sevini, Int. J. Press. Vessel. Pip. 82 (2005) 195–200. [9] G. Maussner, L. Scharf, R. Langer, B. Gurovich, Nucl. Eng. Des. 193 (1999) 359–376. [10] B. Gurovich, E. Kuleshova, Y. Shtrombakh, S. Fedotova, D. Maltsev, a. Frolov, O. Zabusov, D. Erak, D. Zhurko, J. Nucl. Mater. 456 (2015) 23–32. [11] Y.I. Shtrombakh, B.A. Gurovich, E.A. Kuleshova, D.A. Maltsev, S.V. Fedotova, A.A. Chernobaeva, J. Nucl. Mater. 452 (2014) 348–358. [12] W.R. Corwin, G.R. Odette, R.E. Stoller, J.A. Wang, in: Int. Conf. Struct. Mech. React. Technol. CONF-950804-–3, United States, 1995. [13] N.S. Cheruvu, Metall. Trans. A 20 (1989) 87–97. [14] R. Viswanathan, S.M. Bruemmer, J. Eng. Mater. Technol. Trans. ASME 107 (1985). [15] L. Debarberis, B. Acosta, A. Zeman, S. Pirfo, P. Moretto, P. Chernobaeva, Y. Nikoleav, Int. J. Microstruct. Mater. Prop. 2 (2007) 326–338. [16] Y. Watanabe, T. Shoji, Metall. Trans. A 22 (1991) 2097–2106. [17] F. Nogata, K. Seo, H. Takahashi, K. Tamagawa, S.H. Chung, Int. J. Press. Vessel. Pip. 45 (1991) 367–379. [18] C.-S. Seok, J.-M. Koo, Mater. Sci. Eng. A 395 (2005) 141–147. [19] M.J. DeVan, I.A.L. Lowe, C.S. Wade, A.S. Kumar, D.S. Gelles, R.K. Nanstad, E.A. Little, in: Eff. Radiat. Mater. 16th Int. Symp. ASTM STP 1175, Philedelphia, 1993. [20] J. Fukakura, M. Asano, M. Kikuchi, M. Ishikawa, Nucl. Eng. Des. 144 (1993) 423–429. [21] B.A. Gurovich, E.A. Kuleshova, Y.A. Nikolaev, Y.I. Shtrombakh, J. Nucl. Mater. 246 (1997) 91–120. [22] O.O. Zabusov, B.A. Gurovich, E.A. Kuleshova, M.A. Saltykov, S.V. Fedotova, A.P. Dementjev, Key Eng. Mater. 592–593 (2013) 577–581. [23] J. Pierre, R. Francois, P. Christophe, in: Proc. ASME 2013 Press. Vessel. Pip. Conf. PVP 2013, Paris, 2013, pp. 1–8. [24] A. Andrieu, A. Pineau, P. Joly, F. Roch, D. Ryckelynck, Procedia Mater. Sci. 3 (2014) 655–660. [25] W.R. Broughton, P.J.K. Paterson, Appl. Suface Sci. 15 (1983) 120–128. [26] ASTM A533/A533M-09 (reapproved2014), Asme Boil. Press. Vessel Code 9, 2004, pp. 9–12. [27] ASTM A508/A 508M, 1, 2015, pp. 1–5. [28] S. Druce, G. Gage, G. Jordan, Acta Metall. 34 (1986) 641–652. [29] Y. Nishizaka, Y. Hara, A. Hori, H. Tsukahara, K. Miyano, T. Wada, T.B. Cox, J. Press. Vessel Technol. 107 (1985) 285–294. [30] Z. Qu, K.H. Kuo, Metall. Trans. A Phys. Metall. Mater. Sci. 12 A (1981) 1333–1337. [31] K.R. Williams, B. Wilshire, Mater. Sci. Eng. 47 (1981) 151–160. [32] M. Yamashita, K. Nobuo, N. Sakuma, T. Watanabe, Method of Detecting Carbides in Alloy Steels by Electrochemical Polarisation, US Patent Number 5,505,827, 1996. [33] Z. Shenghan, L. Yaling, T. Yu, J. Spectrosc. 2015 (2015) 9–12. [34] L. Yaling, Z. Shenghan, T. Yu, J. Chem. Pharm. Res. 6 (2014) 1435–1439. [35] I.H. Kwon, S.S. Baek, H.S. Yu, Int. J. Press. Vessel. Pip. 80 (2003) 157–165. [36] T. Shibuya, T. Misawa, Zairo-to-Kankyo 154 (1999) 146–154. [37] K.J. Choi, J.J. Kim, B.H. Lee, C.B. Bahn, J.H. Kim, J. Nucl. Mater. 441 (2013) 493–502. [38] S.C. Yoo, K.J. Choi, C.B. Bahn, S.H. Kim, J.Y. Kim, J.H. Kim, J. Nucl. Mater. 459 (2015) 2–12. [39] ASTM E 23-12c, ASTM Int., 2013. [40] B. Gurovich, E. Kuleshova, O. Zabusov, S. Fedotova, A. Frolov, M. Saltykov, D. Maltsev, J. Nucl. Mater. 435 (2013) 25–31. [41] Y. Takashima, T. Nishi, H. Shoji, M. Ohata, F. Minami, Weld. World 58 (2014) 289–295. [42] B. Gurovich, E. Kuleshova, S. Fedotova, D. Maltsev, O. Zabusov, A. Frolov, D. Erak, D. Zhurko, in: Fontevraud 8 - Contrib. Mater. Investig. Oper. Exp. to LWRs' Safety, Perform. Reliab. Fr. Avignon, Sept. 14–18, 2014. [43] B. Gurovich, E. Kuleshova, Y. Shtrombakh, S. Fedotova, D. Erak, D. Zhurko, J. Nucl. Mater. 456 (2015) 373–381. [44] Y.I. Shtrombakh, B.A. Gurovich, E.A. Kuleshova, A.S. Frolov, S.V. Fedotova, D.A. Zhurko, E.V. Krikun, J. Nucl. Mater. 461 (2015) 292–300. [45] Z. Zhai, Y. Miyahara, H. Abe, Y. Watanabe, J. Nucl. Mater. 452 (2014) 133–140. [46] Z. Zhai, H. Abe, Y. Miyahara, Y. Watanabe, Corros. Sci. 92 (2015) 32–42. [47] Y. Kadoya, T. GOTO, M. Takei, Nirou Haruki, Takeshi Ikuno, JSME Int. J. 35 (1992) 226–233. [48] Y.M. Chen, Y. Long, Z.G. Han, Y.H. Wang, in:, Asia-Pacific Power Energy Eng. Conf. APPEEC, 2009, pp. 2–5.
4. Conclusions 1. Mn-Ni-Mo type RPV steel showed embrittlement after prolonged (8400 h) thermal aging (450 °C) treatment. Impact energy decreased and intergranular (IG) brittle fracture area fraction increased with aging time. However, there was no measurable change in average hardness, grain size, misorientation and carbide size. 2. Electrochemical polarisation tests, in picric acid, showed clear attack at the prior austenite grain boundaries (PAGB). Width of such boundaries increased monotonically with corresponding drop in impact energy. This indicated phosphorous segregation leading to grain boundary embrittlement. Acknowledgements The authors thank Dr A. B. Mukherjee, Distinguished Scientist, AD (RPG), BARC and Shri Vivek Shrivastav, SO(F), LWRD, BARC for their help in sourcing the material. The authors would also like to thank Dr Amit Verma, MMD, BARC for his assistance in XRD measurements and analysis. Data availability The raw/processed data required to reproduce these findings cannot be shared at this time as the data also forms part of an ongoing study. References [1] K.D. Haverkamp, K. Forch, K.H. Piehl, W. Witte, Nucl. Eng. Des. 81 (1984) 207–217. [2] L.M. Davies, Int. J. Press. Vessel. Pip. 76 (1999) 163–208. [3] Review of Phosphorus Segregation and Intergranular Embrittlement in Reactor Pressure Vessel Steels (PWRMRP-19) PWR Materials Reliability Project (PWRMRP), Palo Alto, 2000. [4] I.N.E.S. NP-T-3.11, Integrity of Reactor Pressure Vessels in Nuclear Power Plants: Assessement of Irradiation Embrittlement Effects in Reactor Presseure Vessel Steels,
97