- Email: [email protected]
Process development for 9Cr nanostructured ferritic alloy (NFA) with high fracture toughness
Journal of Nuclear Materials 449 (2014) 290–299
Contents lists available at ScienceDirect
Journal of Nuclear Materials journal homepage: www.elsevier.com/locate/jnucmat
Process development for 9Cr nanostructured ferritic alloy (NFA) with high fracture toughness Thak Sang Byun a,⇑, Ji Hyun Yoon b, David T. Hoelzer a, Yong Bok Lee b, Suk Hoon Kang b, Stuart A. Maloy c a
Oak Ridge National Laboratory, Oak Ridge, TN 37831, USA Korea Atomic Energy Research Institute, Daejeon 305-353, South Korea c Los Alamos National Laboratory, Los Alamos, NM 87545, USA b
a r t i c l e
i n f o
Article history: Available online 12 October 2013
a b s t r a c t This article is to summarize the process development and key characterization results for the newlydeveloped Fe–9Cr based nanostructured ferritic alloys (NFAs) with high fracture toughness. One of the major drawbacks from pursuing ultra-high strength in the past development of NFAs is poor fracture toughness at high temperatures although a high fracture toughness is essential to prevent cracking during manufacturing and to mitigate or delay irradiation-induced embrittlement in irradiation environments. A study on fracture mechanism using the NFA 14YWT found that the low-energy grain boundary decohesion in fracture process at a high temperature (>200 °C) resulted in low fracture toughness. Lately, efforts have been devoted to explore an integrated process to enhance grain bonding. Two base materials were produced through mechanical milling and hot extrusion and designated as 9YWTV-PM1 and 9YWTV-PM2. Isothermal annealing (IA) and controlled rolling (CR) treatments in two phase region were used to enhance diffusion across the interfaces and boundaries. The PM2 alloy after p CR treatments showed high fracture toughness (KJQ) at represented temperatures: 240–280 MPa m at p p room temperature and 160–220 MPa m at 500 °C, which indicates that the goal of 100 MPa m over possible nuclear application temperature range has been well achieved. Furthermore, it is also confirmed by comparison that the CR treatments on 9YWTV-PM2 result in high fracture toughness similar to or higher than those of the conventional ferritic–martensitic steels such as HT9 and NF616. Published by Elsevier B.V.
1. Introduction The core structures of advanced future nuclear reactors require high tolerance to extreme irradiation environments, and some critical components such as the fuel cladding in Sodium-cooled Fast Reactors (SFRs) and the first-wall structure in fusion reactors are required to maintain mechanical integrity to very high doses (>200 dpa) at high temperatures up to 700 °C [1–3]. Currently, the success in the development of future reactors is believed to largely depend on the capability of core materials. It is known, however, that no existing metallic materials can be applied to such harsh conditions for various reasons. Some known limitations are, for example, the high swelling and low strength in austenitic stainless steels [4–11], radiation-induced embrittlement in refractory metals [12–19], and phase instability, swelling and radioactivity buildup in irradiation in nickel-based superalloys [20–25]. In the past few decades, the Fe–(2–12%)Cr ferritic martensitic (FM) steels have been the primary materials for advanced reactor ⇑ Corresponding author. Address: 1 Bethel Valley Road, Oak Ridge National Laboratory, Oak Ridge, TN 37831, USA. Tel.: +1 865 576 7738; fax: +1 865 241 3650. E-mail address: [email protected] (T.S. Byun). 0022-3115/$ - see front matter Published by Elsevier B.V. http://dx.doi.org/10.1016/j.jnucmat.2013.10.007
core structures [26–40]. For example, the Fe–12%Cr alloy HT9 had been used for cladding and duct of the Fast Flux Testing Facility (FFTF) [32–40] and is still a primary candidate core material for future SFRs [32,33]. The major shortcoming is, however, that the FM steels have a maximum operating temperature at about 550 °C, which is well below the target operating temperature of core structures in the advanced reactors (650–700 °C). Decadeslong efforts to replace some high activation elements with low activation elements and to optimize the thermomechanical treatment (TMT) schedules with precipitation modifications have marginally improved high temperature capability [41–43]. It was shown that the high temperature strength of these highly refined FM steels was similar to or higher than that of traditional oxide dispersion strengthened (ODS) ferritic steels such as PM2000 [44–50], MA956 [48–50] and MA957 [50–53]. In the past few decades, more advanced ODS steels have been developed to further improve the high temperature capabilities and oxidation resistance, and those are often called the nanostructured ferritic alloys (NFAs) after their highly refined microstructures [54–69]. Although the matrix chemistry of NFAs varied widely, in particular, in chromium contents depending on the purpose of the materials, the vast majority of those were Fe-based 9Cr and 14Cr alloys. Only a few exceptions
T.S. Byun et al. / Journal of Nuclear Materials 449 (2014) 290–299
contained chromium contents higher than 14%, along with Al contents, as those aimed at high corrosion and oxidation resistance [44,61,62]. Currently, the relatively new NFAs with ultra-fine grains (<1 lm) and high particle density (>1023 m3) are considered to be the most promising iron-based materials that can provide sufficient strength at the core material service temperature in future fast neutron or fusion reactors (>650 °C) [1,2,4]. In definition, the NFAs are mechanically-alloyed particle-hardened materials of nano-grained Fe–Cr alloy matrix and Y–Ti–O–Cr–Fe enriched nanoclusters [54–57]. Some of the newest NFAs like 14YWT [54–57] usually have much more uniform and finer microstructures than the conventional oxide dispersion strengthened (ODS) steels such as PM2000 or earlier NFAs such as MA 956 and 957. Much of the improvement in high temperature strength and radiation resistance originates from the presence of uniformly dispersed nanoclusters which act as strong barriers to grain growth, dislocation glide, and grain boundary slip, as well as sinks for trapping of radiation-generated point defects and helium atoms [47,56]. Further, the remarkable thermal stability of nanoclusters provides excellent high temperature strength and creep resistance up to 800 °C [48,51,54–58] and highly delayed or muted radiation effects in mechanical properties [51,56,63,70-72]. Although the NFAs have admirable characteristics for high temperature applications, they show shortcomings, too. Their crack sensitivity is excessively high at high temperatures: the fracture p toughness is unacceptably low (<100 MPa m for the 14YWT) above 200 °C [72]. This latest temperature-dependence study also implied that the evaluation of materials in terms of the transition temperature parameters, such as the transition temperature (T0) determined by fracture toughness master curve method or the ductile-to-brittle transition temperature (DBTT) by impact tests, could not accurately describe the high temperature capability of NFAs as the drop of fracture toughness above 200 °C was profound. This can be a major drawback for a high temperature structural material not only because any high temperature reactor applications can be seriously jeopardized but also because many of hot manufacturing processes require proper toughness and ductility to prevent cracking. A series of recent studies on deformation and fracture mechanisms using 14YWT NFA indicate that the microcracks tend to propagate along grain boundaries without significant plastic deformation, resulting in low energy grain boundary decohesion and low fracture toughness [72,73]. It was also shown that the role of grain boundaries in strengthening was much larger than previously perceived; the grain boundary hardening was responsible for the majority of hardening in NFAs [74–77]. Lately, it was proposed that strengthening in grain bonding could be achieved through the thermo-mechanical treatments specially designed to enhance grain boundary inter-diffusion, which should also result in materials with improved fracture toughness. The main objective of this study was to develop a nanoparticle strengthened engineering grade steel for high performance reactor core structures. For the study, two base NFAs were produced through mechanical milling and hot extrusion and were designated as 9YWTV-PM1 and 9YWTV-PM2. As guided by earlier studies [72,73], improving high temperature grain boundary bonding has been attempted through various thermomechanical treatments (TMTs), i.e., isothermal annealing (IA) and controlled rolling (CR) treatments. Optimized TMTs resulted in fracture toughness similar to or higher than the values of the conventional ferritic–martensitic (FM) steels. This paper reports an integrated NFA processing route combined from the existing processes for 14YWT and new post-extrusion TMT schedules. It also provides a summary of characterization results for the newly developed nanostructured ferritic alloys (NFAs). Detailed characterization results will be presented in later publications.
291
2. Approach to enhance grain bonding Production process for a NFA generally consists of alloy powder production, high energy mechanical (or ball) milling, canning and degassing, consolidation, and post-consolidation TMTs. Since the NFAs have a relatively short history of development, no specification for production of NFAs has been developed. Among the existing NFAs, the 14YWT demonstrated uniquely fine and uniform microstructure of nanograin structure (grain size hundreds of nanometers) and high density (1024/m3) nanoclusters (2–3 nm). This should be due to its unique production process consisted of high energy mechanical milling for long time and relatively low-temperature consolidation, i.e., high power extrusion. As mentioned in introduction, however, our recent studies on the deformation and fracture behaviors of 14YWT have confirmed that the fracture mechanism of low energy grain boundary decohesion results in low fracture toughness at elevated and high temperatures (>200 °C) and that the grain boundary hardening is responsible for the majority of hardening in NFAs [74–77]. Naturally, these results have led to the conclusion that grain boundaries in NFAs need to be modified for stronger bonding. Since the unique nanostructures of 14YWT achieved through the current processes are required to be maintained, any remedy suggested to enhance grain boundary bonding should be limited to the development of post-consolidation TMTs at relatively low temperatures. The TMTs should enhance the joining of grains but should not incur excessive grain growth and coarsening of nanoclusters. In principle, therefore, only viable method to enhance grain boundary bonding and then fracture toughness is believed to be controlled TMTs that can maximize diffusion bonding among grains. The diffusion bonding is a process that produces solid-state coalescence between two materials usually at a temperature well below the melting point (Tm) of the materials [78,79]. It may be well known that diffusion occurs more rigorously through the interface boundaries since different elements are favored in different phases [80]. In this research, therefore, we propose to perform TMTs in two phase conditions to induce a high rate of elemental diffusion. Historically, the ferrite (a) to austenite (c) phase transformation has been used to produce variety of microstructures and properties in ferritic steels [80–82]. An important phenomenon that we should utilize in two phase treatment is that austenite phase is always formed along some preferred sites such as grain boundaries,
Fig. 1. A soft ferrite (gray)-hard martensite (light gray) dual phase structure of SA508 pressure vessel steel [82]. It is expected that in NFAs the austenite formed at high temperature does not transform to martensite during slow cooling after thermomechanical treatments; instead, it may transform back to ferrite as the ferrite–austenite interface is highly mobile at two phase temperature region.
292
T.S. Byun et al. / Journal of Nuclear Materials 449 (2014) 290–299
especially when driving force for phase transformation is not excessive. Fig. 1 shows a ferrite–martensite dual phase structure formed in an A508 cl.3 reactor pressure vessel steel, where a network of martensite phase covering grain boundaries and nearboundary areas is formed in ferritic matrix [82]. In this study, various thermal or thermomechanical treatments were attempted to utilize the partial phase transformation enhanced diffusion as seen in Fig. 1. It is well-known, however, that a steel that contains more than 12 wt.% chromium equivalent can have ferrite single phase up to melting point [83]. This indicates the chemistry of 14YWT, typically Fe(bal)–14Cr–2W–0.4Ti–0.3Y2O3 (in wt.%), cannot allow any TMTs involving the a-c phase transformation since only the aphase can be stable with such a high chromium amount (Cr-equivalent 17%) [84]. In 14YWT, therefore, the diffusion bonding (DB) between grains during powder consolidation and following TMTs should be limited because the diffusion of alloying elements between single phase grains should be much less active than among dissimilar material grains with large difference in chemistry, given that other conditions are the same. Considering the aforementioned conditions to enhance diffusion bonding, a few criteria have been set for the new NFAs and TMT schedules. First, the chemistries of the new alloys need to allow phase transformation and two phase heat treatments: the chromium equivalent should be well below 12%. Second, all the TMTs to be attempted to optimize mechanical properties (or maximize fracture toughness) are designed to involve partial phase transformation, which can be made by annealing in two-phase temperature region or annealing for a limit time at higher temperatures. Small amount of carbon ( 1%) is included in the alloy elements since it is a strong austenite former. Third, a high stress process is combined with the thermal annealing for having two phases. Controlled rolling after annealing is attempted to induce high diffusion bonding. Dislocation generation and annihilation occurring during the hot-rolling should also help to induce high diffusion rate. 3. An integrated process for high toughness NFA An integrated processing technology for production of high toughness NFAs has been developed by combining the existing process techniques for the NFA 14YWT and newly designed TMTs to induce enhanced grain bonding due to partial phase transformation with and without hot working. This section summarizes the key unit processes taken to produce the phase-transformable Fe– 9Cr NFAs. 3.1. Processes for base NFAs 3.1.1. Step-1. Alloy powder production To use the benefits of partial phase transformation, the target chemistries of two base alloys were decided for low chromium equivalents (in wt%): Fe(bal)–9Cr–2W–0.4Ti–0.2V–0.12C (Crequivalent 10%) and Fe(bal)–9Cr–2W–0.4Ti–0.2V–0.05C (Crequivalent 11.5%). Two different carbon contents were chosen for high variability in inducing phase transformation. The alloy powders for these two base compositions were produced in the ATI Powder Metals Co., in Pittsburgh, PA, using an integrated process of vacuum induction melting and argon gas atomization. The product chemistries of the two Fe–9Cr-alloy powders were Fe– 8.84Cr–2.04W–0.38Ti–0.21V–0.05C (heat L2476) and Fe–9.02Cr– 2.11W–0.38Ti–0.21V–0.11C (heat L2477). The impurity elements such as S, N, O are tightly controlled within 0.001–0.02 wt% for all particle size groups. For each heat, the total yield of powder was about 14 kg, including all batches with three different mesh
sizes: 35/+100 (500–149 l), 100/+325 (149–44 l), and 325 (<44 l). In the mechanical milling for this study the powders coarser than the mesh size 100 were not used to avoid incomplete alloying. As will be mentioned later, the as-consolidated NFA products with these chemistries are designated as 9YWTV-PM2 and 9YWTV-PM1, respectively. Meanwhile, the yttrium oxide (Y2O3) powder, which is mixed with these alloy powders for mechanical milling, was purchased from the Nanophase Technology Co., and has particle sizes ranging from 17 to 31 nm. 3.1.2. Step-2. High power mechanical alloying In powder metallurgy based materials, the ball milling process is always the most critical step for production of materials with desirable microstructures and other properties. A high energy collision among the flying balls in an attritor milling machine can generate repeated breaking and cold welding of powders, which eventually lead to a complete mixing of alloying elements and very high plasticity. In the production of NFAs, the milling process is designed to provide a pristine condition before creating nanoclusters in the microstructure. It usually means a completely-alloyed status of the material without any precipitation or evident elemental segregation. Since the highly deformed Fe–Cr steels can demonstrate ultra-high strength (>1 GPa), a high energy milling machine is needed for the production of NFAs. In addition, any impurity contamination should be avoided during the mechanical milling to prevent boundary embrittlement. In this study, the two alloy powders have been mechanicallymilled in Zoz GmbH CM08 attritor machine after being mixed with 0.3 wt.% yttrium oxide (Y2O3) powder. This machine can be operated up to 1000 rpm and generate a maximum relative ball velocity of >10 m/s. In the milling process, running the machine at too high speed was avoided to prevent the system from over-heating and the chamber was constantly cooled by water: over-heating temperature limit was set at 40 °C. Each of the loads for milling comprised 997 g of an alloy powder and 3 g of Y2O3 powder and was milled for 40 h in static argon environment in water cooled chamber. Six 1 kg batches were ball milled for a total of 6 kg milled powder of the heat L2476 (PM2) and three 1 kg batches for 3 kg powder of the heat L2477 (PM1). 3.1.3. Step-3. Canning and degassing This process is only a simple preparatory step for powder consolidation process. In the production of NFAs, however, this step needs to be done carefully. Degassing at an elevated temperature and sealing the can containing milled powder are the main processes. An important principle is that this warm degassing step should not alter the completely alloyed status of the powders since excessive oxidation at powder surfaces might affect the efficiency of consolidation process and eventually the properties of final product. In processing NFAs the mechanically-alloyed powder is filled into a mild steel can with 50 mm inner diameter and evacuated to a vacuum of 1 Pa at 6400 °C. The small hole used for evacuation is sealed at its connection to piping by welding equipment. 3.1.4. Step-4. Powder consolidation The consolidation process has two critical roles in production of NFAs: bonding of grains and powders and generation of nanoclusters. Both of these microstructural processes determine the properties of final products. At Oak Ridge National Laboratory this process normally has two steps of heating and extrusion. Prior to the extrusion, the sealed can containing milled powder is heated to a high temperature (6850 °C). Based on our experience, we have decided to set the extrusion temperature at 850 °C or lower to limit coarsening of nanoclusters and grains. In choosing the heating temperature prior to consolidation, it also needs to be considered that processing at too low temperatures might limit the strength of
T.S. Byun et al. / Journal of Nuclear Materials 449 (2014) 290–299
grain bonding and the ability of extrusion machine for high areareduction ratio. Based on high temperature test data, the required pressure for the consolidation was estimated to be 300–500 MPa to induce heavy plastic deformation in the milled ferritic powders. The extrusion operation for the Fe–9Cr NFAs was performed in the Watson–Stillman unit at ORNL materials procesing facility, which has a capacity of 1.25 kilotons. A force of 500–1000 tons was needed to extrude a 90 mm (outer) diameter can containing 1 kg mixed powder. The low-temperature (<850 °C) extrusion process is possible only with such a high power machine. The pressure applied to the powders being consolidated was estimated to be >700 MPa. It is believed that a strong diffusion bonding of grains and aggregates can be achieved by such an ultrahigh pressure and high plastic energy input. These two new NFAs in as-extruded condition (called as ‘base NFAs’ hereafter), were designated as 9YWTVPM1 and 9YWT-PM2, respectively, for the two nominal components: Fe–9Cr–2W–0.4Ti–0.2V–0.12C–0.3Y2O3 and Fe–9Cr–2W– 0.4Ti–0.2V–0.05C–0.3Y2O3. Fig. 2a and b shows the as-extruded bars and the coupons cut (nominally 18 45 90 mm with rounded cones) from the bars, respectively. In Fig. 2b the cut surfaces reveal two component materials: the 9Cr NFA in the meat portion and the mild steel can material in the rim, which is removed before or after the post-extrusion TMT prior to machining specimens. 3.2. Post-extrusion thermomechanical treatments in search of high toughness In search of optimized TMT condition(s), two series of treatments have been attempted for inducing diffusion bonding aided by partial ferrite–austenite (a–c) phase transformation and/or by high stress and plasticity. The first series of treatments were simple isothermal annealing (IA) treatments at various temperatures for at least 30 min followed by furnace cooling. In the second series, the base NFAs were controlled-rolled (CR) for 20% or 50% total reduction (i.e., isothermally annealed and rolled down on cooling in air). Conditioning these treatments associated with partial a–c transformation was guided by phase equilibrium calculation using Thermo-Calc program.
Fig. 2. Base NFAs produced by mechanical milling and consolidation: (a) asextruded bars cladded with mild steel and (b) blocks cut from the bars. Note that in (b) the meat portion in each block is NFA and that the first numbers 1 and 2 indicate 9YWTV-PM1 and -PM2, respectively.
293
The calculation results indicate that the two alloys behave differently in the phase transformation process, mainly due to the difference in carbon content (after-extrusion measurements: 0.155% versus 0.08%). With the chemistry of 9YWTV-PM1 the austenite phase (c) starts to form at 850 °C and its volume fraction reaches 100% at about 970 °C, which indicates that the intercritical temperature region is 850–970 °C for the higher carbon alloy (PM1). It is also shown that the austenitic transformation in the 9YWT-PM2 alloy is delayed by about 60 °C and much slower because of lower carbon content. The a–c transformation starts at 910 °C and reaches the maximum c volume fraction of about 53% at 1030 °C. For both alloys, the intercritical temperature region is about 120 °C wide. Although these calculation results show clearly-defined two phase regions for both alloys, it has to be considered in conditioning TMTs that the material in actual TMTs is always heated for a limited time, usually too short for an equilibrium. More importantly, no nano-features in NFAs, such as nanograins and nanoclusters which may alter any phase transformation kinetics as well as the thermodynamic state of the material significantly, could be factored in the commercial computer code. Therefore, it is admitted that the Thermo-Calc calculation results obtained above can be used only as approximate reference data for choosing annealing temperatures in TMTs. Other methodologies such as in situ X-ray detection and calorimetry have been attempted to measure the volume fractions of two phases in NFAs, and the results are to be included in a future publication. Such experimental attempts to obtain volume fraction data, however, have confirmed that the phase transformation behaviors in the nanostructured ferritic materials are significantly different from the knowledge we have or we can obtain with existing means. For example, the ferrite-to-austenite transformation detected by the in situ X-ray is much slower than the calculation predicts: an austenite fraction of 45% was measured for the PM2 alloy annealed at 1000 °C although the Thermo-Calc calculation predicted 100% transformation in the condition. Moreover, most of the austenite phase formed in nanograins transforms back to ferrite during cooling as the reverse transformation seems to be easily activated at mobile interfaces in nanograins. As an additional undesirable aspect, the X-ray data seemed to include a large error (some 5% in volume fraction), which seemed to originate from the nanograin features. These observations indicate that we currently do not have practical means to obtain accurate phase transformation information, and therefore, we could not evaluate some necessary transformation data such as the critical temperatures for the initiation and completion of austenitization (AC1 and AC3). We believe a long term scientific research is needed to understand the phase transformation behavior in the nanostructured steels. As we recognized that our information on phase deformation could not provide accurate guidance in process development, but knowing that the ultimate goal of this research was to develop a high toughness NFA, the final output, fracture toughness, was used as the main measure in selection of proper TMT conditions. Fig. 3a and b illustrate the TMT processes applied to the two base NFAs, and detailed schedules are summarized in Tables 1 and 2. As shown in Fig. 3a, the isothermal annealing was performed for the machined fracture specimens of both NFAs in the temperature range of 830–975 °C for 30 min to 20 h. According to the Thermo-Calc result, this temperature range overlaps the entire intercritical temperature region for 9YWTV-PM1 alloy and the majority of that for 9YWTV-PM2 alloy. At the highest temperature of 975 °C the volume fraction of austenite was estimated to be 100% for the PM1 alloy and about 45% for the PM2 alloy, which ensure to cover the whole grain boundary areas with austenite phase in both alloys. It is worthwhile to note that, in phase transforma-
294
T.S. Byun et al. / Journal of Nuclear Materials 449 (2014) 290–299
section, however, these isothermally annealed materials did not show significantly improved fracture toughness. Therefore, the later efforts were focused more on the controlled rolling treatments. The hot-rolling treatments have been applied to the coupons after heating to 900–1000 °C for 30 min. This temperature range was chosen considering the temperature range for a to c transformation (850–1000 °C), hot formability of NFAs (>900 °C), and prevention of abnormal grain growth (<1100 °C). It was also expected that high dislocation density after the first rolling can help to nucleate austenite phase in multi-step rolling process. As summarized in Table 2, the total thickness reduction ratio in rolling was 20% or 50%, in which the 20% rolling was made in one step while the 50% reduction was achieved by repetition of heating-rolling cycles for preventing the rolled NFA plates from cracking on further reduction. It was also considered in choosing treatment conditions that the overall strength of NFAs after these treatments falls within the strength range of the base NFAs or reference material 14YWT. Fig. 4 displays some examples of as-rolled plates, which are still cladded with mild steel.
4. Characterization of new 9Cr NFAs 4.1. Characteristics of base materials
Fig. 3. Schematics of the TMT processes applied to Fe–9Cr base NFAs: (a) isothermal annealing (IA) treatments and (b) controlled rolling (CR) treatments. Note that only the schedules for 30 and 200 min annealing treatments are illustrated in (a).
Table 1 Schedules for isothermal annealing treatments. Annealing temperature and period (°C)
30 or 60 min
200 min
830 850 875 900 950 975
O O O O O O
O O O O O
20 h O
Table 2 Schedules for controlled rolling (CR) treatments. Pre-heating temperature and thickness reduction (°C)
20%
50%
Hot-rolling process
900
O
O
Single pass for 20%, multi-pass for 50% Multi-pass Multi-pass Multi-pass Multi-pass
925 950 975 1000
O O O O
tion nucleating at grain boundaries, an austenite volume fraction of 20–30% is needed to form a well-connected c phase network covering all grain boundary areas [82]. As will be discussed in a later
The newly developed 9Cr base NFAs have been characterized to confirm the basic microstructural and mechanical properties before application of TMTs are comparable to those of our reference NFA, 14YWT. Specimens including tensile and fracture specimens and transmission electron microscopy (TEM) disks were machined form the meat portion of the coupons seen in Fig. 2b. Since creating a highly uniform ultra-fine grain structure is the key for a good combination of strength and ductility, grain structure was examined first for the base NFAs. SEM analysis was performed on the polished surfaces of both base NFAs. Back scattered electron (BSE) images shown in Fig. 5 indicates that the microstructures of both 9Cr NFA base materials containing uniform distributions of ultra-fine grains: 160 and 230 nm for PM1 and PM2 alloys, respectively. The past studies on 14YWT have shown that the observation of uniform ultra-small grain microstructures is always accompanied by formation of a homogeneous distribution of nanoclusters during extrusion. These highly uniform ultra-fine grain structures in Fig. 5a and b confirm that the ball milling process has effectively alloyed the elements of Y2O3 powder into the alloy powder, which lead to a high density distribution of nanoclusters at grain boundaries and within grains, and the relativelylow temperature consolidation has not caused abnormal or excessive grain growth. TEM analysis was also performed to confirm the grain and crystalline structures as well as the spatial and size distribution of nanoparticles in the new 9Cr NFAs. Fig. 6 illustrates the results of Field Emission TEM (FETEM) examination for the 9YWT-PM1 base material. The nano-grain structure is confirmed again in Fig. 6a, in which some dislocation lines and precipitations are observed. It is worth knowing that the grains in as-milled state are much finer (<50 nm) and further coarsening is expected in subsequent post-extrusion treatments. Diffraction pattern confirmed that those grains are ferrite or BCC phase. The Fe–M edge jump ratio EFTEM image in Fig. 6b displays the distribution of nanoparticles, from which the average size of the particles was measured to be about 2 nm. These results confirm that the new base NFAs retain nanostructure features. More detailed results of TEM analysis including those after TMTs optimized for high toughness will be presented in later publications. Figs. 7 and 8 present the tensile properties of the base materials compared to the reference NFA 14YWT-SM10. In Fig. 7, the yield
295
T.S. Byun et al. / Journal of Nuclear Materials 449 (2014) 290–299
Fig. 4. Examples of as-rolled plates. Each plate is still cladded with mild steel and contains a NFA meat portion of 5(t) 23(w) 150(l) mm.
(a)
(b) Nano-particle
(b) Fig. 6. Nanostructure of NFA 9YWTV-PM1: (a) bright field image of nano-grain structure, and (b) Fe–M edge jump ratio FETEM image showing nanoparticles. Fig. 5. SEM backscattered electron (BSE) micrographs showing the ultra-fine ferrite grain structure of the as-extruded 9Cr NFAs: (a) 9YWTV-PM1 (0.12%C) and (b) 9YWTV-PM2 (0.05%C).
stress of 9YWTV-PM1 is similar to or slightly higher than that of 14YWT over the entire test temperature. Also shown is that the strength, yield stress, of PM2 base NFA is slightly lower than that of 14YWT up to about 500 °C and then becomes similar to that of 14YWT at higher temperatures. It can be said that the overall
strength of the newly produced 9Cr base NFAs is in the same strength range of the reference material. Fig. 8 compares the total elongation data of the three NFAs. The overall ductility of PM1 alloy is slightly higher ductility compared to that of 14YWT, while the ductility of the PM2 alloy is the highest among the three NFAs over entire test temperature range. Meanwhile, the fracture toughness data for base NFAs have been obtained and compared with those after TMTs in the next session.
296
T.S. Byun et al. / Journal of Nuclear Materials 449 (2014) 290–299
Fig. 7. Yield stress versus temperature data for the new base NFAs compared to those of 14YWT-SM10.
Fig. 9. Fracture toughness at room temperature and 500 °C before and after isothermal annealing treatments in the 9YWTV-PM1 alloy.
Fig. 8. Total elongation versus temperature data for the new base NFAs compared to those of 14YWT-SM10.
4.2. Characteristics of Fe–9Cr NFAs after TMTs Since the ultimate goal of this study was to develop a high toughness NFA, the fracture toughness (KJQ) from static fracture resistance (J–R) testing was used to provide feedbacks in searching for optimized TMT condition. Small three-point bend (TPB) bar specimens (15 mm in length 2.5 mm thickness 5 mm width) were used for the J–R tests. Both precracking and static J–R testing were performed in a servohydraulic testing machine with vacuum furnace. Figs. 9–12 compare the fracture toughness data of 14YWT, base 9Cr NFAs, and 9Cr NFAs after TMTs. In these comparisons the two test temperatures, room temperature and 500 °C, were selected to represent fracture behaviors at low and high temperatures, respectively. Further, the fracture toughness data of these NFAs are compared with those of ferritic–martensitic (FM) steels, HT9 (12Cr–1MoWV) and NF616 (9Cr–2WVNb), since the ultimate goal of this work is to develop a 9Cr NFA with fracture toughness comparable to those of the melting-based FM steels. Both of the reference FM steels have fracture toughness higher p than 200 MPa m at room temperature and higher than p 150 MPa m at 500 °C. The fracture toughness of NF616, which is the toughest material among FM steels, is higher p than 250 MPa m at both temperatures, Fig. 9. The reference
Fig. 10. Fracture toughness at room temperature and 500 °C before and after isothermal annealing treatments in the 9YWTV-PM2 alloy.
NFA, 14YWT, shows significantly lower fracture toughness at p p both test temperatures: 143 MPa m at RT and 53 MPa m at 500 °C. At RT all of the 9YWTV-PM1 materials in as-extruded condition (base material) or after isothermal annealing (IA) in 850–975 °C display fracture toughness values lower than that of 14YWT. At 500 °C only the 975 °C annealed specimen has fracture toughness p higher than 100 MPa m, while all other NFAs show lower KJQ values. Overall, however, none of the IA treatments improve the fracp ture toughness of 9YWTV-PM1 to higher than 100 MPa m both at RT and at 500 °C. A conclusion from these observations is that none of the simple IA treatments is effective for the 9YWTV-PM1 alloy.
T.S. Byun et al. / Journal of Nuclear Materials 449 (2014) 290–299
Fig. 11. Fracture toughness at room temperature and 500 °C before and after hotrolling treatments in the 9YWTV-PM1 alloy.
297
The fracture toughness values of PM1 alloy before and after controlled rolling (CR) treatments are compared in Fig. 11. At both test temperatures the KJQ values after CR treatments, except for those after 975 °C CR treatment, are comparable to those of p 14YWT: the KJQ values are well above 100 MPa m at RT but are much lower than the bench mark level at 500 °C. For the PM1 alloy, p the KJQ values after 50% CR at 975 °C are higher than 150 MPa m; the improvement at 500 °C is particularly significant. It is, however, possible that this standing out case is an experimental error resulting from an inhomogeneity in material sampling. The improvement of fracture toughness by CR treatments is significantly more pronounced and consistent in the PM2 alloy. All the CR cases compared in Fig. 12 display KJQ values above p 150 MPa m at RT as well as at 500 °C. A noticeable behavior is that the typical decrease of fracture toughness at high temperatures (>200 °C) is not observed in the CR treated 9YWTV-PM2. In particular, the highest fracture toughness values were measured with the 20% and 50% CR treatments at 900 °C among all NFA cases compared in Figs. 9–12. It is also observed in this comparison that the multi-step 50% CR treatment at 900 °C yields slightly higher fracture toughness than that after 20% CR treatment. Therefore, we chose the multi-step 50% CR for the other CR treatment. Another positive point noticed in this comparison is that the improvement of fracture toughness to a similar degree is observed in the entire CR temperature range of 900–1000 °C, which can allow us to select processing temperature(s) without much restriction. In the meantime, the tensile strength and ductility of NFAs after TMTs were evaluated using subsize flat tensile specimens (gage section dimension: 7.62 mm in length 1.5 mm width 0.76 mm thickness) to confirm that the improvement in fracture toughness was achieved without significant change in strength. Tensile tests at RT and 500 °C showed that the tensile strength gradually decreased with increasing pre-rolling heating temperature, while ductility increased with the annealing temperature. Multiple rolling for the 50% total reduction reduced strength slightly more than the 20% reduction. At RT the tensile strengths after the hot-rolling at 900–1000 °C were in the range of 1450–1680 MPa for the PM1 alloy and in the range of 1150–1420 MPa for the PM2 alloy. At 500 °C the tensile strengths after the same TMTs became notably lower: 680–1000 MPa for the PM1 alloy and in the range of 660– 730 MPa for the PM2 alloy. Detailed temperature dependence of strength and ductility data is discussed in the next publication. For the highest fracture toughness case after the 50% hot-rolling at 900 °C, it is noticed that the tensile strengths at 500 °C for the PM2 alloys were reduced as much as 200 MPa from those of base (as-extruded) materials. This indicates the loss of strength by TMTs is much less significant compared to the 100% increase in fracture toughness in PM2. 4.3. Selection of an alloy and TMT routes for further studies
Fig. 12. Fracture toughness at room temperature and 500 °C before and after hotrolling treatments in the 9YWTV-PM2 alloy.
On the other hand, Fig. 10 indicates that fracture toughness improvement through IA is more effective in the 9YWTV-PM2. All of the IA treatments increased fracture toughness to about or p above 150 MPa m at RT. In the 500 °C test dataset, although the two cases of anneals at 850 °C and 950 °C resulted in KJQ values higher than 100 °C, fracture toughness appears to be much lower. A known, typical behavior for NFAs is that fracture toughness is relatively high at low temperatures and becomes significantly lower at elevated or high temperatures [72]. Such a typical behavior for NFAs, i.e., high-to-low toughness transition with temperature, is displayed in Fig. 10 for the PM2 alloy after various IA treatments.
Although the detailed toughening mechanisms by hot-rolling treatments require further studies, the comparisons in Figs. 9–12 confirm the belief that a well-conditioned TMT can lead to a high toughness NFA. To the same TMTs, however, the two base NFAs, 9YWTV-PM1 and -PM2, responded differently: no post-extrusion TMT could improve the fracture toughness of PM1 alloy to a large degree except for the 50% rolling at 975 °C while even some simple IA treatments resulted in meaningful increase of fracture toughness in the PM2 alloy. As discussed in the previous session, the improvement of fracture toughness in controlled rolled 9YWTVPM2 was particularly significant: the KJQ values after the hot rolling treatments were as high as those of non-ODS F/M steels. It is a speculation that the excessive carbide precipitation in grain boundary areas has resulted in little improvement of fracture toughness in the PM1 alloy. This can be considered as an expected
298
T.S. Byun et al. / Journal of Nuclear Materials 449 (2014) 290–299
result because in heating the carbide precipitation occurs before the formation of oxygen rich nanoclusters. Further, the higher carbon content in the PM1 alloy can induce higher carbon concentration in austenite phase during high temperature annealing, which leads to an excessive carbide precipitation around grain boundaries during cooling. Based on the understandings achieved so far, therefore, the 9YWTV-PM2 alloy is chosen for future applications and further studies. In addition, since the 9YWTV-PM2 controlled-rolled at 900 °C resulted in the best fracture toughness among the NFAs examined p in this study (>150 MPa m), the treatment is chosen for further applications. Although the initial intention of the present development for high toughness was to use the effect of highly effective diffusion bonding induced in two phase condition, it cannot confirmed yet if the postulated mechanism is highly responsible for the high toughness achieved by the CR treatments. Further, since the a–c interface formed in the intercritical region is mobile and thus the c-phase is expected to shrank during slow cooling [81,82], the final amounts of quenched martensite is not known. At the moment, therefore, the detailed behaviors of phase transformation, dislocation recovery, and diffusion during the repeated hot-rolling treatment are not understood well. A significant amount of effort might be required for understanding such high temperature phenomena occurring on TMT, especially in two phase temperature region. While further scientific understanding is left to future studies, achieving the main goal of developing engineering grade NFA is considered to be significant since the high toughness NFAs can be further processed for engineering purposes, such as tube drawing and forming.
5. Concluding remarks In the past, very high strength can be achieved in NFAs at the expense of fracture toughness and ductility. Low energy decohesion at boundaries causes poor fracture resistance at high temperatures. Developing a high toughness NFA was the ultimate goal of this study. To strengthen the powder-metallurgy produced weak boundaries by an enhanced diffusion bonding, the schedules of isothermal annealing (IA) and controlled rolling (CR) after annealing were attempted. Improvement of fracture toughness in the CR 9YWTV-PM2 was significant: its fracture toughness was as high as those of non-ODS F/M steels. In particular, the 9YWTV-PM2 controlled-rolled at 900 °C resulted in the best fracture toughness among NFAs p (>150 MPa m at both representative temperatures: RT and 500 °C. It can be claimed that the final goal of developing an engineering grade NFA has been achieved. More detailed microstructural and mechanical characterizations for selected materials, i.e., the 9YWTV-PM2 alloy after 50% CR at 900 °C are underway and the results are to be reported in separate articles. The topics in these companion articles include the detailed temperature dependence of mechanical properties, delayed phase transformation in nanostructure and stability of nanoclusters and grains in TMT processes, etc. Irradiation experiment is also being pursued.
Acknowledgements This research was sponsored by the International Nuclear Energy Research Initiative (I-NERI) Collaboration between United States and South Korea (I-NERI Project 2010-004-K). In the US, this research was also part of the Fuel Cycle R&D program sponsored by the Office of Nuclear Energy, US Department of Energy, under Contract DE-AC05-00OR22725 with UT-Battelle, LLC. The authors ex-
press special thanks to Drs. Chad M. Parish and Kurt. A. Terrani for their thorough reviews and thoughtful comments.
References [1] L.K. Mansur, A.F. Rowcliffe, R.K. Nanstad, S.J. Zinkle, W.R. Corwin, R.E. Stoller, J. Nucl. Mater. 329 (2004) 166–172. [2] S.J. Zinkle, J.T. Busby, Mater. Today 12 (2009) 12–19. [3] S.A. Maloy, M. Toloczko, J. Cole, T.E. Byun, J. Nucl. Mater. 415 (2011) 302–305. [4] S.J. Zinkle, G.S. Was, Acta Mater. 61 (2013) 735–758. [5] B.Z. Margolin, A.I. Murashova, V.S. Neustroev, Strength Mater. 44 (2012) 223– 240. [6] S.J. Zinkle, N.M. Ghoniem, J. Nucl. Mater. 417 (2011) 2–8. [7] E. Kenik, J.T. Busby, Mater. Sci. Eng. R 73 (2012) 67–83. [8] F.A. Garner, D.S. Gelles, ASTM STP 1046 (1999) 673–683. [9] T.R. Allen, J.I. Cole, C.L. Trybus, D.L. Porter, H. Tsai, F. Garner, E.A. Kenik, T. Yoshitake, J. Ohta, J. Nucl. Mater. 348 (2006) 148–164. [10] I.I. Balachov, E.N. Shcherbakov, A.V. Kozlov, I.A. Portnykh, F.A. Garner, J. Nucl. Mater. 329–333 (2004) 617–620. [11] V.S. Neustroev, F.A. Garner, J. Nucl. Mater. 386–388 (2009) 157–160. [12] S.J. Zinkle, N.M. Ghoniem, J. Nucl. Mater. 51–52 (2000) 55–71. [13] A.V. Babak, E.I. Uskov, Strength Mater. 16 (1984) 943–949. [14] Y. Mutoh, K. Ichikawa, K. Nagata, M. Takeuchi, J. Mater. Sci. 30 (1995) 770–775. [15] P. Gumbsch, J. Riedle, A. Hartmaier, H.F. Fischmeister, Science 282 (1998) 1293–1295. [16] B.V. Cockeram, R.W. Smith, L.L. Snead, J. Nucl. Mater. 346 (2005) 145–164. [17] B.V. Cockeram, E.K. Ohriner, T.S. Byun, M.K. Miller, L.L. Snead, J. Nucl. Mater. 382 (2008) 229–241. [18] B.V. Cockeram, R.W. Smith, K.J. Leonard, T.S. Byun, L.L. Snead, J. Nucl. Mater. 382 (2008) 1–23. [19] T.S. Byun, M. Li, B.V. Cockeram, L.L. Snead, J. Nucl. Mater. 376 (2008) 240–246. [20] B.H. Sencer, G.M. Bond, F.A. Garner, M.L. Hamilton, B.M. Oliver, L.E. Thomas, S.A. Maloy, W.F. Sommer, M.R. James, P.D. Ferguson, J. Nucl. Mater. 283–287 (2000) 324–328. [21] P. Yvon, F. Carre, J. Nucl. Mater. 385 (2009) 217–222. [22] T.S. Byun, K. Farrell, J. Nucl. Mater. 318 (2003) 292–299. [23] J.D. Hunn, E.H. Lee, T.S. Byun, L.K. Mansur, J. Nucl. Mater. 296 (2001) 203–209. [24] N. Hashimoto, J.D. Hunn, T.S. Byun, L.K. Mansur, J. Nucl. Mater. 318 (2003) 300–306. [25] A.F. Rowcliffe, L.K. Mansur, D.T. Hoelzer, R.K. Nanstad, J. Nucl. Mater. 392 (2009) 341–352. [26] R.L. Klueh, Int. Mater. Rev. 50 (2005) 287–310. [27] R.L. Klueh, D.R. Harris, High-chromium ferritic and martensitic steels for nuclear applications, ASTM (2001). ISBN 0-803102090-7. [28] T.R. Allen, J.T. Busby, R.L. Klueh, S.A. Maloy, M.B. Toloczko, JOM 60 (2008) 15– 23. [29] O. Anderoglu, T.S. Byun, M. Toloczko, S.A. Maloy, Met. Mater. Trans. A (2013) 70–83. [30] R. Chaouadi, G. Coen, L. Lucon, V. Massaut, J. Nucl. Mater. 403 (2010) 15–18. [31] T.S. Byun, M. Li, K. Farrell, Met. Mater. Trans. A 44 (2013) 84–93. [32] S.A. Maloy, M. Toloczko, J. Cole, T.S. Byun, J. Nucl. Mater. 415 (2011) 302–305. [33] J.S. Cheon, C.B. Lee, B.O. Lee, J.P. Raison, T. Mizuno, F. Delage, J. Carmack, J. Nucl. Mater. 392 (2009) 324–330. [34] B.H. Sencer, J.R. Kennedy, J.I. Cole, S.A. Maloy, F.A. Garner, J. Nucl. Mater. 393 (2009) 235–241. [35] F.H. Huang, M.L. Hamilton, J. Nucl. Mater. 187 (1992) 278–293. [36] F.H. Huang, J. Nucl. Mater. 43 (1992) 748–773. [37] F.H. Huang, D.E. Gelles, Eng. Fract. Mech. 19 (1984) 1–84. [38] T.S. Byun, J.H. Kim, J.H. Yoon, D.T. Hoelzer, J. Nucl. Mater. 407 (2010) 78–82. [39] T.S. Byun, W.D. Lewis, M.B. Toloczko, S.A. Maloy, J. Nucl. Mater. 421 (2012) 104–111. [40] T.S. Byun, M.B. Toloczko, T.A. Saleh, S.A. Maloy, J. Nucl. Mater. 432 (2013) 1–8. [41] R.L. Klueh, N. Hashimoto, P.J. Maziasz, Scripta Mater. 53 (2005) 275–280. [42] S. Hollner, B. Fournier, J. Nucl. Mater. 405 (2010) 101–108. [43] L. Tan, D.T. Hoelzer, J.T. Busby, M.A. Sokolov, R.L. Klueh, J. Nucl. Mater. 422 (2012) 45–50. [44] J. Chen, M.A. Pouchon, A. Kimura, P. Jung, W. Hoffelner, J. Nucl. Mater. 386–388 (2009) 143–146. [45] P. Lours, S. Le Roux, G. Bernhart, J. Adv. Mater. 33 (2001) 11–20. [46] P. Krautwasser, A. Czyrska-Filemonowicz, M. Widera, F. Carsughi, Mater. Sci. Eng. A 177 (1994) 199–208. [47] J. Chen, P. Jung, W. Hoffelner, H. Ullmaier, Acta Mater. 56 (2008) 250–258. [48] A. Wasilkowska, M. Bartsch, U. Messerschmidt, R. Herzog, A. CzyrskaFilemonowicz, J. Mater. Process. Technol. (2003) 218–224. [49] A. Czyrska-Filemonowicz, B. Dubiel, J. Mater. Process. Technol. 64 (1997) 53– 64. [50] R.L. Klueh, J.P. Shingledecker, R.W. Swindeman, D.T. Hoelzer, J. Nucl. Mater. 341 (2005) 103–114. [51] M.B. Toloczko, D.S. Gelles, F.A. Garner, R.J. Kurtz, K. Abe, J. Nucl. Mater. 329– 333 (2004) 352–355. [52] P. Miao, G.R. Odette, T. Yamamoto, M. Alinger, D. Klingensmith, J. Nucl. Mater. 377 (2008) 59–64. [53] M.K. Miller, D.T. Hoelzer, E.A. Kenik, K.F. Russell, J. Nucl. Mater. 329–333 (2004) 338–341.
T.S. Byun et al. / Journal of Nuclear Materials 449 (2014) 290–299 [54] D.T. Hoelzer, J. Bentley, M.A. Sokolov, M.K. Miller, G.R. Odette, M.J. Alinger, J. Nucl. Mater. 367–370 (2007) 166–172. [55] M.J. Alinger, G.R. Odette, D.T. Hoelzer, Acta Mater. 57 (2009) 392–406. [56] G.R. Odette, M.J. Alinger, B.D. Wirth, Ann. Rev. Mater. Res. 38 (2008) 471–503. [57] M.K. Miller, D.T. Hoelzer, E.A. Kenik, K.F. Russell, Intermetallics 13 (2005) 387– 392. [58] S. Ukai, S. Ohtsuka, T. Kaito, H. Sakasegawa, N. Chikata, S. Hayashi, S. Ohnuki, Mater. Sci. Eng. A 510–511 (2009) 115–120. [59] H. Sakasegawa, S. Ukai, M. Tamura, S. Ohtsuka, H. Tanigawa, H. Ogiwara, A. Kohyama, M. Fujiwara, J. Nucl. Mater. 373 (2008) 82–89. [60] P. Olier, J. Malaplate, M.H. Mathon, D. Nunes, D. Hamona, L. Toualbi, Y. de Carlan, L. Chaffron, J. Nucl. Mater. 428 (2012) 40–46. [61] H.S. Cho, R. Kasada, A. Kimura, J. Nucl. Mater. 367 (2007) 239–243. [62] H. Kishimoto, R. Kasada, O. Hashitomi, A. Kimura, J. Nucl. Mater. 386 (2009) 533–536. [63] N. Akasaka, S. Yamashita, T. Yoshitake, S. Ukai, A. Kimura, J. Nucl. Mater. 329 (2004) 1053–1056. [64] S. Ukai, M. Harada, H. Okada, M. Inoue, S. Nomura, S. Shikakura, K. Asabe, T. Nishida, M. Fujiwara, J. Nucl. Mater. 204 (1993) 65–71. [65] S. Ukai, T. Nishida, T. Okuda, T. Yoshitake, J. Nucl. Mater. 258 (1998) 1745– 1749. [66] A.-L. Rouffié, P. Wident, L. Ziolek, F. Delabrouille, B. Tanguy, J. Crépin, A. Pineau, V. Garat, B. Fournier, J. Nucl. Mater. 433 (2013) 108–115. [67] A. Steckmeyer, M. Praud, B. Fournier, J. Malaplate, J. Garnier, J.L. Béchade, I. Tournié, A. Tancray, A. Bougault, P. Bonnaillie, J. Nucl. Mater. 405 (2010) 95– 100.
299
[68] J.H. Kim, J.H. Lee, J.Y. Min, S.W. Kim, C.H. Park, J.T. Yeom, T.S. Byun, J. Alloys Compd. 580 (2013) 125–130. [69] D.K. Mukhopadhyay, F.H. Froes, D.S. Gelles, J. Nucl. Mater. 258–263 (1998) 1209–1215. [70] D.A. McClintock, D.T. Hoelzer, M.A. Sokolov, R.K. Nanstad, J. Nucl. Mater. 386 (2009) 307–311. [71] D.A. McClintock, M.A. Sokolov, D.T. Hoelzer, R.K. Nanstad, J. Nucl. Mater. 392 (2009) 353–359. [72] T.S. Byun, J.H. Kim, J.H. Yoon, D.T. Hoelzer, J. Nucl. Mater. 407 (2010) 78–82. [73] J.H. Kim, T.S. Byun, D.T. Hoelzer, J. Nucl. Mater. 407 (2010) 143–150. [74] J.H. Kim, T.S. Byun, D.T. Hoelzer, S.W. Kim, B.H. Lee, Mater. Sci. Eng. A 559 (2013) 101–110. [75] J.H. Kim, T.S. Byun, D.T. Hoelzer, C.H. Park, J.T. Yeom, J.K. Hong, Mater. Sci. Eng. A 559 (2013) 111–118. [76] J.H. Kim, T.S. Byun, D.T. Hoelzer, J. Nucl. Mater. 425 (2012) 147–155. [77] J.H. Kim, T.S. Byun, D.T. Hoelzer, J. Nucl. Mater. in press. [78] N. Orhan, M. Aksoy, M. Eroglu, Mater. Sci. Eng. A 271 (1999) 458–468. [79] B. Kurt, N. Orhan, E. Evin, A. Çalik, Mater. Lett. 61 (2007) 1747–1750. [80] S. Sun, M. Pugh, Mater. Sci. Eng. A 276 (2000) 167–174. [81] G.R. Speich, V.A. Demarest, R.L. Miller, Metall. Trans. A 12 (1981) 1419–1428. [82] Y.S. Ahn, H.D. Kim, T.S. Byun, Y.J. Oh, G.M. Kim, J.H. Hong, Nucl. Eng. Des. 194 (1999) 161–177. [83] http://www.calphad.com/iron–chromium.html. [84] S.H. Ryu, J. Yu, Met. Trans. 29A (1998) 1573–1578.
Contents lists available at ScienceDirect
Journal of Nuclear Materials journal homepage: www.elsevier.com/locate/jnucmat
Process development for 9Cr nanostructured ferritic alloy (NFA) with high fracture toughness Thak Sang Byun a,⇑, Ji Hyun Yoon b, David T. Hoelzer a, Yong Bok Lee b, Suk Hoon Kang b, Stuart A. Maloy c a
Oak Ridge National Laboratory, Oak Ridge, TN 37831, USA Korea Atomic Energy Research Institute, Daejeon 305-353, South Korea c Los Alamos National Laboratory, Los Alamos, NM 87545, USA b
a r t i c l e
i n f o
Article history: Available online 12 October 2013
a b s t r a c t This article is to summarize the process development and key characterization results for the newlydeveloped Fe–9Cr based nanostructured ferritic alloys (NFAs) with high fracture toughness. One of the major drawbacks from pursuing ultra-high strength in the past development of NFAs is poor fracture toughness at high temperatures although a high fracture toughness is essential to prevent cracking during manufacturing and to mitigate or delay irradiation-induced embrittlement in irradiation environments. A study on fracture mechanism using the NFA 14YWT found that the low-energy grain boundary decohesion in fracture process at a high temperature (>200 °C) resulted in low fracture toughness. Lately, efforts have been devoted to explore an integrated process to enhance grain bonding. Two base materials were produced through mechanical milling and hot extrusion and designated as 9YWTV-PM1 and 9YWTV-PM2. Isothermal annealing (IA) and controlled rolling (CR) treatments in two phase region were used to enhance diffusion across the interfaces and boundaries. The PM2 alloy after p CR treatments showed high fracture toughness (KJQ) at represented temperatures: 240–280 MPa m at p p room temperature and 160–220 MPa m at 500 °C, which indicates that the goal of 100 MPa m over possible nuclear application temperature range has been well achieved. Furthermore, it is also confirmed by comparison that the CR treatments on 9YWTV-PM2 result in high fracture toughness similar to or higher than those of the conventional ferritic–martensitic steels such as HT9 and NF616. Published by Elsevier B.V.
1. Introduction The core structures of advanced future nuclear reactors require high tolerance to extreme irradiation environments, and some critical components such as the fuel cladding in Sodium-cooled Fast Reactors (SFRs) and the first-wall structure in fusion reactors are required to maintain mechanical integrity to very high doses (>200 dpa) at high temperatures up to 700 °C [1–3]. Currently, the success in the development of future reactors is believed to largely depend on the capability of core materials. It is known, however, that no existing metallic materials can be applied to such harsh conditions for various reasons. Some known limitations are, for example, the high swelling and low strength in austenitic stainless steels [4–11], radiation-induced embrittlement in refractory metals [12–19], and phase instability, swelling and radioactivity buildup in irradiation in nickel-based superalloys [20–25]. In the past few decades, the Fe–(2–12%)Cr ferritic martensitic (FM) steels have been the primary materials for advanced reactor ⇑ Corresponding author. Address: 1 Bethel Valley Road, Oak Ridge National Laboratory, Oak Ridge, TN 37831, USA. Tel.: +1 865 576 7738; fax: +1 865 241 3650. E-mail address: [email protected] (T.S. Byun). 0022-3115/$ - see front matter Published by Elsevier B.V. http://dx.doi.org/10.1016/j.jnucmat.2013.10.007
core structures [26–40]. For example, the Fe–12%Cr alloy HT9 had been used for cladding and duct of the Fast Flux Testing Facility (FFTF) [32–40] and is still a primary candidate core material for future SFRs [32,33]. The major shortcoming is, however, that the FM steels have a maximum operating temperature at about 550 °C, which is well below the target operating temperature of core structures in the advanced reactors (650–700 °C). Decadeslong efforts to replace some high activation elements with low activation elements and to optimize the thermomechanical treatment (TMT) schedules with precipitation modifications have marginally improved high temperature capability [41–43]. It was shown that the high temperature strength of these highly refined FM steels was similar to or higher than that of traditional oxide dispersion strengthened (ODS) ferritic steels such as PM2000 [44–50], MA956 [48–50] and MA957 [50–53]. In the past few decades, more advanced ODS steels have been developed to further improve the high temperature capabilities and oxidation resistance, and those are often called the nanostructured ferritic alloys (NFAs) after their highly refined microstructures [54–69]. Although the matrix chemistry of NFAs varied widely, in particular, in chromium contents depending on the purpose of the materials, the vast majority of those were Fe-based 9Cr and 14Cr alloys. Only a few exceptions
T.S. Byun et al. / Journal of Nuclear Materials 449 (2014) 290–299
contained chromium contents higher than 14%, along with Al contents, as those aimed at high corrosion and oxidation resistance [44,61,62]. Currently, the relatively new NFAs with ultra-fine grains (<1 lm) and high particle density (>1023 m3) are considered to be the most promising iron-based materials that can provide sufficient strength at the core material service temperature in future fast neutron or fusion reactors (>650 °C) [1,2,4]. In definition, the NFAs are mechanically-alloyed particle-hardened materials of nano-grained Fe–Cr alloy matrix and Y–Ti–O–Cr–Fe enriched nanoclusters [54–57]. Some of the newest NFAs like 14YWT [54–57] usually have much more uniform and finer microstructures than the conventional oxide dispersion strengthened (ODS) steels such as PM2000 or earlier NFAs such as MA 956 and 957. Much of the improvement in high temperature strength and radiation resistance originates from the presence of uniformly dispersed nanoclusters which act as strong barriers to grain growth, dislocation glide, and grain boundary slip, as well as sinks for trapping of radiation-generated point defects and helium atoms [47,56]. Further, the remarkable thermal stability of nanoclusters provides excellent high temperature strength and creep resistance up to 800 °C [48,51,54–58] and highly delayed or muted radiation effects in mechanical properties [51,56,63,70-72]. Although the NFAs have admirable characteristics for high temperature applications, they show shortcomings, too. Their crack sensitivity is excessively high at high temperatures: the fracture p toughness is unacceptably low (<100 MPa m for the 14YWT) above 200 °C [72]. This latest temperature-dependence study also implied that the evaluation of materials in terms of the transition temperature parameters, such as the transition temperature (T0) determined by fracture toughness master curve method or the ductile-to-brittle transition temperature (DBTT) by impact tests, could not accurately describe the high temperature capability of NFAs as the drop of fracture toughness above 200 °C was profound. This can be a major drawback for a high temperature structural material not only because any high temperature reactor applications can be seriously jeopardized but also because many of hot manufacturing processes require proper toughness and ductility to prevent cracking. A series of recent studies on deformation and fracture mechanisms using 14YWT NFA indicate that the microcracks tend to propagate along grain boundaries without significant plastic deformation, resulting in low energy grain boundary decohesion and low fracture toughness [72,73]. It was also shown that the role of grain boundaries in strengthening was much larger than previously perceived; the grain boundary hardening was responsible for the majority of hardening in NFAs [74–77]. Lately, it was proposed that strengthening in grain bonding could be achieved through the thermo-mechanical treatments specially designed to enhance grain boundary inter-diffusion, which should also result in materials with improved fracture toughness. The main objective of this study was to develop a nanoparticle strengthened engineering grade steel for high performance reactor core structures. For the study, two base NFAs were produced through mechanical milling and hot extrusion and were designated as 9YWTV-PM1 and 9YWTV-PM2. As guided by earlier studies [72,73], improving high temperature grain boundary bonding has been attempted through various thermomechanical treatments (TMTs), i.e., isothermal annealing (IA) and controlled rolling (CR) treatments. Optimized TMTs resulted in fracture toughness similar to or higher than the values of the conventional ferritic–martensitic (FM) steels. This paper reports an integrated NFA processing route combined from the existing processes for 14YWT and new post-extrusion TMT schedules. It also provides a summary of characterization results for the newly developed nanostructured ferritic alloys (NFAs). Detailed characterization results will be presented in later publications.
291
2. Approach to enhance grain bonding Production process for a NFA generally consists of alloy powder production, high energy mechanical (or ball) milling, canning and degassing, consolidation, and post-consolidation TMTs. Since the NFAs have a relatively short history of development, no specification for production of NFAs has been developed. Among the existing NFAs, the 14YWT demonstrated uniquely fine and uniform microstructure of nanograin structure (grain size hundreds of nanometers) and high density (1024/m3) nanoclusters (2–3 nm). This should be due to its unique production process consisted of high energy mechanical milling for long time and relatively low-temperature consolidation, i.e., high power extrusion. As mentioned in introduction, however, our recent studies on the deformation and fracture behaviors of 14YWT have confirmed that the fracture mechanism of low energy grain boundary decohesion results in low fracture toughness at elevated and high temperatures (>200 °C) and that the grain boundary hardening is responsible for the majority of hardening in NFAs [74–77]. Naturally, these results have led to the conclusion that grain boundaries in NFAs need to be modified for stronger bonding. Since the unique nanostructures of 14YWT achieved through the current processes are required to be maintained, any remedy suggested to enhance grain boundary bonding should be limited to the development of post-consolidation TMTs at relatively low temperatures. The TMTs should enhance the joining of grains but should not incur excessive grain growth and coarsening of nanoclusters. In principle, therefore, only viable method to enhance grain boundary bonding and then fracture toughness is believed to be controlled TMTs that can maximize diffusion bonding among grains. The diffusion bonding is a process that produces solid-state coalescence between two materials usually at a temperature well below the melting point (Tm) of the materials [78,79]. It may be well known that diffusion occurs more rigorously through the interface boundaries since different elements are favored in different phases [80]. In this research, therefore, we propose to perform TMTs in two phase conditions to induce a high rate of elemental diffusion. Historically, the ferrite (a) to austenite (c) phase transformation has been used to produce variety of microstructures and properties in ferritic steels [80–82]. An important phenomenon that we should utilize in two phase treatment is that austenite phase is always formed along some preferred sites such as grain boundaries,
Fig. 1. A soft ferrite (gray)-hard martensite (light gray) dual phase structure of SA508 pressure vessel steel [82]. It is expected that in NFAs the austenite formed at high temperature does not transform to martensite during slow cooling after thermomechanical treatments; instead, it may transform back to ferrite as the ferrite–austenite interface is highly mobile at two phase temperature region.
292
T.S. Byun et al. / Journal of Nuclear Materials 449 (2014) 290–299
especially when driving force for phase transformation is not excessive. Fig. 1 shows a ferrite–martensite dual phase structure formed in an A508 cl.3 reactor pressure vessel steel, where a network of martensite phase covering grain boundaries and nearboundary areas is formed in ferritic matrix [82]. In this study, various thermal or thermomechanical treatments were attempted to utilize the partial phase transformation enhanced diffusion as seen in Fig. 1. It is well-known, however, that a steel that contains more than 12 wt.% chromium equivalent can have ferrite single phase up to melting point [83]. This indicates the chemistry of 14YWT, typically Fe(bal)–14Cr–2W–0.4Ti–0.3Y2O3 (in wt.%), cannot allow any TMTs involving the a-c phase transformation since only the aphase can be stable with such a high chromium amount (Cr-equivalent 17%) [84]. In 14YWT, therefore, the diffusion bonding (DB) between grains during powder consolidation and following TMTs should be limited because the diffusion of alloying elements between single phase grains should be much less active than among dissimilar material grains with large difference in chemistry, given that other conditions are the same. Considering the aforementioned conditions to enhance diffusion bonding, a few criteria have been set for the new NFAs and TMT schedules. First, the chemistries of the new alloys need to allow phase transformation and two phase heat treatments: the chromium equivalent should be well below 12%. Second, all the TMTs to be attempted to optimize mechanical properties (or maximize fracture toughness) are designed to involve partial phase transformation, which can be made by annealing in two-phase temperature region or annealing for a limit time at higher temperatures. Small amount of carbon ( 1%) is included in the alloy elements since it is a strong austenite former. Third, a high stress process is combined with the thermal annealing for having two phases. Controlled rolling after annealing is attempted to induce high diffusion bonding. Dislocation generation and annihilation occurring during the hot-rolling should also help to induce high diffusion rate. 3. An integrated process for high toughness NFA An integrated processing technology for production of high toughness NFAs has been developed by combining the existing process techniques for the NFA 14YWT and newly designed TMTs to induce enhanced grain bonding due to partial phase transformation with and without hot working. This section summarizes the key unit processes taken to produce the phase-transformable Fe– 9Cr NFAs. 3.1. Processes for base NFAs 3.1.1. Step-1. Alloy powder production To use the benefits of partial phase transformation, the target chemistries of two base alloys were decided for low chromium equivalents (in wt%): Fe(bal)–9Cr–2W–0.4Ti–0.2V–0.12C (Crequivalent 10%) and Fe(bal)–9Cr–2W–0.4Ti–0.2V–0.05C (Crequivalent 11.5%). Two different carbon contents were chosen for high variability in inducing phase transformation. The alloy powders for these two base compositions were produced in the ATI Powder Metals Co., in Pittsburgh, PA, using an integrated process of vacuum induction melting and argon gas atomization. The product chemistries of the two Fe–9Cr-alloy powders were Fe– 8.84Cr–2.04W–0.38Ti–0.21V–0.05C (heat L2476) and Fe–9.02Cr– 2.11W–0.38Ti–0.21V–0.11C (heat L2477). The impurity elements such as S, N, O are tightly controlled within 0.001–0.02 wt% for all particle size groups. For each heat, the total yield of powder was about 14 kg, including all batches with three different mesh
sizes: 35/+100 (500–149 l), 100/+325 (149–44 l), and 325 (<44 l). In the mechanical milling for this study the powders coarser than the mesh size 100 were not used to avoid incomplete alloying. As will be mentioned later, the as-consolidated NFA products with these chemistries are designated as 9YWTV-PM2 and 9YWTV-PM1, respectively. Meanwhile, the yttrium oxide (Y2O3) powder, which is mixed with these alloy powders for mechanical milling, was purchased from the Nanophase Technology Co., and has particle sizes ranging from 17 to 31 nm. 3.1.2. Step-2. High power mechanical alloying In powder metallurgy based materials, the ball milling process is always the most critical step for production of materials with desirable microstructures and other properties. A high energy collision among the flying balls in an attritor milling machine can generate repeated breaking and cold welding of powders, which eventually lead to a complete mixing of alloying elements and very high plasticity. In the production of NFAs, the milling process is designed to provide a pristine condition before creating nanoclusters in the microstructure. It usually means a completely-alloyed status of the material without any precipitation or evident elemental segregation. Since the highly deformed Fe–Cr steels can demonstrate ultra-high strength (>1 GPa), a high energy milling machine is needed for the production of NFAs. In addition, any impurity contamination should be avoided during the mechanical milling to prevent boundary embrittlement. In this study, the two alloy powders have been mechanicallymilled in Zoz GmbH CM08 attritor machine after being mixed with 0.3 wt.% yttrium oxide (Y2O3) powder. This machine can be operated up to 1000 rpm and generate a maximum relative ball velocity of >10 m/s. In the milling process, running the machine at too high speed was avoided to prevent the system from over-heating and the chamber was constantly cooled by water: over-heating temperature limit was set at 40 °C. Each of the loads for milling comprised 997 g of an alloy powder and 3 g of Y2O3 powder and was milled for 40 h in static argon environment in water cooled chamber. Six 1 kg batches were ball milled for a total of 6 kg milled powder of the heat L2476 (PM2) and three 1 kg batches for 3 kg powder of the heat L2477 (PM1). 3.1.3. Step-3. Canning and degassing This process is only a simple preparatory step for powder consolidation process. In the production of NFAs, however, this step needs to be done carefully. Degassing at an elevated temperature and sealing the can containing milled powder are the main processes. An important principle is that this warm degassing step should not alter the completely alloyed status of the powders since excessive oxidation at powder surfaces might affect the efficiency of consolidation process and eventually the properties of final product. In processing NFAs the mechanically-alloyed powder is filled into a mild steel can with 50 mm inner diameter and evacuated to a vacuum of 1 Pa at 6400 °C. The small hole used for evacuation is sealed at its connection to piping by welding equipment. 3.1.4. Step-4. Powder consolidation The consolidation process has two critical roles in production of NFAs: bonding of grains and powders and generation of nanoclusters. Both of these microstructural processes determine the properties of final products. At Oak Ridge National Laboratory this process normally has two steps of heating and extrusion. Prior to the extrusion, the sealed can containing milled powder is heated to a high temperature (6850 °C). Based on our experience, we have decided to set the extrusion temperature at 850 °C or lower to limit coarsening of nanoclusters and grains. In choosing the heating temperature prior to consolidation, it also needs to be considered that processing at too low temperatures might limit the strength of
T.S. Byun et al. / Journal of Nuclear Materials 449 (2014) 290–299
grain bonding and the ability of extrusion machine for high areareduction ratio. Based on high temperature test data, the required pressure for the consolidation was estimated to be 300–500 MPa to induce heavy plastic deformation in the milled ferritic powders. The extrusion operation for the Fe–9Cr NFAs was performed in the Watson–Stillman unit at ORNL materials procesing facility, which has a capacity of 1.25 kilotons. A force of 500–1000 tons was needed to extrude a 90 mm (outer) diameter can containing 1 kg mixed powder. The low-temperature (<850 °C) extrusion process is possible only with such a high power machine. The pressure applied to the powders being consolidated was estimated to be >700 MPa. It is believed that a strong diffusion bonding of grains and aggregates can be achieved by such an ultrahigh pressure and high plastic energy input. These two new NFAs in as-extruded condition (called as ‘base NFAs’ hereafter), were designated as 9YWTVPM1 and 9YWT-PM2, respectively, for the two nominal components: Fe–9Cr–2W–0.4Ti–0.2V–0.12C–0.3Y2O3 and Fe–9Cr–2W– 0.4Ti–0.2V–0.05C–0.3Y2O3. Fig. 2a and b shows the as-extruded bars and the coupons cut (nominally 18 45 90 mm with rounded cones) from the bars, respectively. In Fig. 2b the cut surfaces reveal two component materials: the 9Cr NFA in the meat portion and the mild steel can material in the rim, which is removed before or after the post-extrusion TMT prior to machining specimens. 3.2. Post-extrusion thermomechanical treatments in search of high toughness In search of optimized TMT condition(s), two series of treatments have been attempted for inducing diffusion bonding aided by partial ferrite–austenite (a–c) phase transformation and/or by high stress and plasticity. The first series of treatments were simple isothermal annealing (IA) treatments at various temperatures for at least 30 min followed by furnace cooling. In the second series, the base NFAs were controlled-rolled (CR) for 20% or 50% total reduction (i.e., isothermally annealed and rolled down on cooling in air). Conditioning these treatments associated with partial a–c transformation was guided by phase equilibrium calculation using Thermo-Calc program.
Fig. 2. Base NFAs produced by mechanical milling and consolidation: (a) asextruded bars cladded with mild steel and (b) blocks cut from the bars. Note that in (b) the meat portion in each block is NFA and that the first numbers 1 and 2 indicate 9YWTV-PM1 and -PM2, respectively.
293
The calculation results indicate that the two alloys behave differently in the phase transformation process, mainly due to the difference in carbon content (after-extrusion measurements: 0.155% versus 0.08%). With the chemistry of 9YWTV-PM1 the austenite phase (c) starts to form at 850 °C and its volume fraction reaches 100% at about 970 °C, which indicates that the intercritical temperature region is 850–970 °C for the higher carbon alloy (PM1). It is also shown that the austenitic transformation in the 9YWT-PM2 alloy is delayed by about 60 °C and much slower because of lower carbon content. The a–c transformation starts at 910 °C and reaches the maximum c volume fraction of about 53% at 1030 °C. For both alloys, the intercritical temperature region is about 120 °C wide. Although these calculation results show clearly-defined two phase regions for both alloys, it has to be considered in conditioning TMTs that the material in actual TMTs is always heated for a limited time, usually too short for an equilibrium. More importantly, no nano-features in NFAs, such as nanograins and nanoclusters which may alter any phase transformation kinetics as well as the thermodynamic state of the material significantly, could be factored in the commercial computer code. Therefore, it is admitted that the Thermo-Calc calculation results obtained above can be used only as approximate reference data for choosing annealing temperatures in TMTs. Other methodologies such as in situ X-ray detection and calorimetry have been attempted to measure the volume fractions of two phases in NFAs, and the results are to be included in a future publication. Such experimental attempts to obtain volume fraction data, however, have confirmed that the phase transformation behaviors in the nanostructured ferritic materials are significantly different from the knowledge we have or we can obtain with existing means. For example, the ferrite-to-austenite transformation detected by the in situ X-ray is much slower than the calculation predicts: an austenite fraction of 45% was measured for the PM2 alloy annealed at 1000 °C although the Thermo-Calc calculation predicted 100% transformation in the condition. Moreover, most of the austenite phase formed in nanograins transforms back to ferrite during cooling as the reverse transformation seems to be easily activated at mobile interfaces in nanograins. As an additional undesirable aspect, the X-ray data seemed to include a large error (some 5% in volume fraction), which seemed to originate from the nanograin features. These observations indicate that we currently do not have practical means to obtain accurate phase transformation information, and therefore, we could not evaluate some necessary transformation data such as the critical temperatures for the initiation and completion of austenitization (AC1 and AC3). We believe a long term scientific research is needed to understand the phase transformation behavior in the nanostructured steels. As we recognized that our information on phase deformation could not provide accurate guidance in process development, but knowing that the ultimate goal of this research was to develop a high toughness NFA, the final output, fracture toughness, was used as the main measure in selection of proper TMT conditions. Fig. 3a and b illustrate the TMT processes applied to the two base NFAs, and detailed schedules are summarized in Tables 1 and 2. As shown in Fig. 3a, the isothermal annealing was performed for the machined fracture specimens of both NFAs in the temperature range of 830–975 °C for 30 min to 20 h. According to the Thermo-Calc result, this temperature range overlaps the entire intercritical temperature region for 9YWTV-PM1 alloy and the majority of that for 9YWTV-PM2 alloy. At the highest temperature of 975 °C the volume fraction of austenite was estimated to be 100% for the PM1 alloy and about 45% for the PM2 alloy, which ensure to cover the whole grain boundary areas with austenite phase in both alloys. It is worthwhile to note that, in phase transforma-
294
T.S. Byun et al. / Journal of Nuclear Materials 449 (2014) 290–299
section, however, these isothermally annealed materials did not show significantly improved fracture toughness. Therefore, the later efforts were focused more on the controlled rolling treatments. The hot-rolling treatments have been applied to the coupons after heating to 900–1000 °C for 30 min. This temperature range was chosen considering the temperature range for a to c transformation (850–1000 °C), hot formability of NFAs (>900 °C), and prevention of abnormal grain growth (<1100 °C). It was also expected that high dislocation density after the first rolling can help to nucleate austenite phase in multi-step rolling process. As summarized in Table 2, the total thickness reduction ratio in rolling was 20% or 50%, in which the 20% rolling was made in one step while the 50% reduction was achieved by repetition of heating-rolling cycles for preventing the rolled NFA plates from cracking on further reduction. It was also considered in choosing treatment conditions that the overall strength of NFAs after these treatments falls within the strength range of the base NFAs or reference material 14YWT. Fig. 4 displays some examples of as-rolled plates, which are still cladded with mild steel.
4. Characterization of new 9Cr NFAs 4.1. Characteristics of base materials
Fig. 3. Schematics of the TMT processes applied to Fe–9Cr base NFAs: (a) isothermal annealing (IA) treatments and (b) controlled rolling (CR) treatments. Note that only the schedules for 30 and 200 min annealing treatments are illustrated in (a).
Table 1 Schedules for isothermal annealing treatments. Annealing temperature and period (°C)
30 or 60 min
200 min
830 850 875 900 950 975
O O O O O O
O O O O O
20 h O
Table 2 Schedules for controlled rolling (CR) treatments. Pre-heating temperature and thickness reduction (°C)
20%
50%
Hot-rolling process
900
O
O
Single pass for 20%, multi-pass for 50% Multi-pass Multi-pass Multi-pass Multi-pass
925 950 975 1000
O O O O
tion nucleating at grain boundaries, an austenite volume fraction of 20–30% is needed to form a well-connected c phase network covering all grain boundary areas [82]. As will be discussed in a later
The newly developed 9Cr base NFAs have been characterized to confirm the basic microstructural and mechanical properties before application of TMTs are comparable to those of our reference NFA, 14YWT. Specimens including tensile and fracture specimens and transmission electron microscopy (TEM) disks were machined form the meat portion of the coupons seen in Fig. 2b. Since creating a highly uniform ultra-fine grain structure is the key for a good combination of strength and ductility, grain structure was examined first for the base NFAs. SEM analysis was performed on the polished surfaces of both base NFAs. Back scattered electron (BSE) images shown in Fig. 5 indicates that the microstructures of both 9Cr NFA base materials containing uniform distributions of ultra-fine grains: 160 and 230 nm for PM1 and PM2 alloys, respectively. The past studies on 14YWT have shown that the observation of uniform ultra-small grain microstructures is always accompanied by formation of a homogeneous distribution of nanoclusters during extrusion. These highly uniform ultra-fine grain structures in Fig. 5a and b confirm that the ball milling process has effectively alloyed the elements of Y2O3 powder into the alloy powder, which lead to a high density distribution of nanoclusters at grain boundaries and within grains, and the relativelylow temperature consolidation has not caused abnormal or excessive grain growth. TEM analysis was also performed to confirm the grain and crystalline structures as well as the spatial and size distribution of nanoparticles in the new 9Cr NFAs. Fig. 6 illustrates the results of Field Emission TEM (FETEM) examination for the 9YWT-PM1 base material. The nano-grain structure is confirmed again in Fig. 6a, in which some dislocation lines and precipitations are observed. It is worth knowing that the grains in as-milled state are much finer (<50 nm) and further coarsening is expected in subsequent post-extrusion treatments. Diffraction pattern confirmed that those grains are ferrite or BCC phase. The Fe–M edge jump ratio EFTEM image in Fig. 6b displays the distribution of nanoparticles, from which the average size of the particles was measured to be about 2 nm. These results confirm that the new base NFAs retain nanostructure features. More detailed results of TEM analysis including those after TMTs optimized for high toughness will be presented in later publications. Figs. 7 and 8 present the tensile properties of the base materials compared to the reference NFA 14YWT-SM10. In Fig. 7, the yield
295
T.S. Byun et al. / Journal of Nuclear Materials 449 (2014) 290–299
Fig. 4. Examples of as-rolled plates. Each plate is still cladded with mild steel and contains a NFA meat portion of 5(t) 23(w) 150(l) mm.
(a)
(b) Nano-particle
(b) Fig. 6. Nanostructure of NFA 9YWTV-PM1: (a) bright field image of nano-grain structure, and (b) Fe–M edge jump ratio FETEM image showing nanoparticles. Fig. 5. SEM backscattered electron (BSE) micrographs showing the ultra-fine ferrite grain structure of the as-extruded 9Cr NFAs: (a) 9YWTV-PM1 (0.12%C) and (b) 9YWTV-PM2 (0.05%C).
stress of 9YWTV-PM1 is similar to or slightly higher than that of 14YWT over the entire test temperature. Also shown is that the strength, yield stress, of PM2 base NFA is slightly lower than that of 14YWT up to about 500 °C and then becomes similar to that of 14YWT at higher temperatures. It can be said that the overall
strength of the newly produced 9Cr base NFAs is in the same strength range of the reference material. Fig. 8 compares the total elongation data of the three NFAs. The overall ductility of PM1 alloy is slightly higher ductility compared to that of 14YWT, while the ductility of the PM2 alloy is the highest among the three NFAs over entire test temperature range. Meanwhile, the fracture toughness data for base NFAs have been obtained and compared with those after TMTs in the next session.
296
T.S. Byun et al. / Journal of Nuclear Materials 449 (2014) 290–299
Fig. 7. Yield stress versus temperature data for the new base NFAs compared to those of 14YWT-SM10.
Fig. 9. Fracture toughness at room temperature and 500 °C before and after isothermal annealing treatments in the 9YWTV-PM1 alloy.
Fig. 8. Total elongation versus temperature data for the new base NFAs compared to those of 14YWT-SM10.
4.2. Characteristics of Fe–9Cr NFAs after TMTs Since the ultimate goal of this study was to develop a high toughness NFA, the fracture toughness (KJQ) from static fracture resistance (J–R) testing was used to provide feedbacks in searching for optimized TMT condition. Small three-point bend (TPB) bar specimens (15 mm in length 2.5 mm thickness 5 mm width) were used for the J–R tests. Both precracking and static J–R testing were performed in a servohydraulic testing machine with vacuum furnace. Figs. 9–12 compare the fracture toughness data of 14YWT, base 9Cr NFAs, and 9Cr NFAs after TMTs. In these comparisons the two test temperatures, room temperature and 500 °C, were selected to represent fracture behaviors at low and high temperatures, respectively. Further, the fracture toughness data of these NFAs are compared with those of ferritic–martensitic (FM) steels, HT9 (12Cr–1MoWV) and NF616 (9Cr–2WVNb), since the ultimate goal of this work is to develop a 9Cr NFA with fracture toughness comparable to those of the melting-based FM steels. Both of the reference FM steels have fracture toughness higher p than 200 MPa m at room temperature and higher than p 150 MPa m at 500 °C. The fracture toughness of NF616, which is the toughest material among FM steels, is higher p than 250 MPa m at both temperatures, Fig. 9. The reference
Fig. 10. Fracture toughness at room temperature and 500 °C before and after isothermal annealing treatments in the 9YWTV-PM2 alloy.
NFA, 14YWT, shows significantly lower fracture toughness at p p both test temperatures: 143 MPa m at RT and 53 MPa m at 500 °C. At RT all of the 9YWTV-PM1 materials in as-extruded condition (base material) or after isothermal annealing (IA) in 850–975 °C display fracture toughness values lower than that of 14YWT. At 500 °C only the 975 °C annealed specimen has fracture toughness p higher than 100 MPa m, while all other NFAs show lower KJQ values. Overall, however, none of the IA treatments improve the fracp ture toughness of 9YWTV-PM1 to higher than 100 MPa m both at RT and at 500 °C. A conclusion from these observations is that none of the simple IA treatments is effective for the 9YWTV-PM1 alloy.
T.S. Byun et al. / Journal of Nuclear Materials 449 (2014) 290–299
Fig. 11. Fracture toughness at room temperature and 500 °C before and after hotrolling treatments in the 9YWTV-PM1 alloy.
297
The fracture toughness values of PM1 alloy before and after controlled rolling (CR) treatments are compared in Fig. 11. At both test temperatures the KJQ values after CR treatments, except for those after 975 °C CR treatment, are comparable to those of p 14YWT: the KJQ values are well above 100 MPa m at RT but are much lower than the bench mark level at 500 °C. For the PM1 alloy, p the KJQ values after 50% CR at 975 °C are higher than 150 MPa m; the improvement at 500 °C is particularly significant. It is, however, possible that this standing out case is an experimental error resulting from an inhomogeneity in material sampling. The improvement of fracture toughness by CR treatments is significantly more pronounced and consistent in the PM2 alloy. All the CR cases compared in Fig. 12 display KJQ values above p 150 MPa m at RT as well as at 500 °C. A noticeable behavior is that the typical decrease of fracture toughness at high temperatures (>200 °C) is not observed in the CR treated 9YWTV-PM2. In particular, the highest fracture toughness values were measured with the 20% and 50% CR treatments at 900 °C among all NFA cases compared in Figs. 9–12. It is also observed in this comparison that the multi-step 50% CR treatment at 900 °C yields slightly higher fracture toughness than that after 20% CR treatment. Therefore, we chose the multi-step 50% CR for the other CR treatment. Another positive point noticed in this comparison is that the improvement of fracture toughness to a similar degree is observed in the entire CR temperature range of 900–1000 °C, which can allow us to select processing temperature(s) without much restriction. In the meantime, the tensile strength and ductility of NFAs after TMTs were evaluated using subsize flat tensile specimens (gage section dimension: 7.62 mm in length 1.5 mm width 0.76 mm thickness) to confirm that the improvement in fracture toughness was achieved without significant change in strength. Tensile tests at RT and 500 °C showed that the tensile strength gradually decreased with increasing pre-rolling heating temperature, while ductility increased with the annealing temperature. Multiple rolling for the 50% total reduction reduced strength slightly more than the 20% reduction. At RT the tensile strengths after the hot-rolling at 900–1000 °C were in the range of 1450–1680 MPa for the PM1 alloy and in the range of 1150–1420 MPa for the PM2 alloy. At 500 °C the tensile strengths after the same TMTs became notably lower: 680–1000 MPa for the PM1 alloy and in the range of 660– 730 MPa for the PM2 alloy. Detailed temperature dependence of strength and ductility data is discussed in the next publication. For the highest fracture toughness case after the 50% hot-rolling at 900 °C, it is noticed that the tensile strengths at 500 °C for the PM2 alloys were reduced as much as 200 MPa from those of base (as-extruded) materials. This indicates the loss of strength by TMTs is much less significant compared to the 100% increase in fracture toughness in PM2. 4.3. Selection of an alloy and TMT routes for further studies
Fig. 12. Fracture toughness at room temperature and 500 °C before and after hotrolling treatments in the 9YWTV-PM2 alloy.
On the other hand, Fig. 10 indicates that fracture toughness improvement through IA is more effective in the 9YWTV-PM2. All of the IA treatments increased fracture toughness to about or p above 150 MPa m at RT. In the 500 °C test dataset, although the two cases of anneals at 850 °C and 950 °C resulted in KJQ values higher than 100 °C, fracture toughness appears to be much lower. A known, typical behavior for NFAs is that fracture toughness is relatively high at low temperatures and becomes significantly lower at elevated or high temperatures [72]. Such a typical behavior for NFAs, i.e., high-to-low toughness transition with temperature, is displayed in Fig. 10 for the PM2 alloy after various IA treatments.
Although the detailed toughening mechanisms by hot-rolling treatments require further studies, the comparisons in Figs. 9–12 confirm the belief that a well-conditioned TMT can lead to a high toughness NFA. To the same TMTs, however, the two base NFAs, 9YWTV-PM1 and -PM2, responded differently: no post-extrusion TMT could improve the fracture toughness of PM1 alloy to a large degree except for the 50% rolling at 975 °C while even some simple IA treatments resulted in meaningful increase of fracture toughness in the PM2 alloy. As discussed in the previous session, the improvement of fracture toughness in controlled rolled 9YWTVPM2 was particularly significant: the KJQ values after the hot rolling treatments were as high as those of non-ODS F/M steels. It is a speculation that the excessive carbide precipitation in grain boundary areas has resulted in little improvement of fracture toughness in the PM1 alloy. This can be considered as an expected
298
T.S. Byun et al. / Journal of Nuclear Materials 449 (2014) 290–299
result because in heating the carbide precipitation occurs before the formation of oxygen rich nanoclusters. Further, the higher carbon content in the PM1 alloy can induce higher carbon concentration in austenite phase during high temperature annealing, which leads to an excessive carbide precipitation around grain boundaries during cooling. Based on the understandings achieved so far, therefore, the 9YWTV-PM2 alloy is chosen for future applications and further studies. In addition, since the 9YWTV-PM2 controlled-rolled at 900 °C resulted in the best fracture toughness among the NFAs examined p in this study (>150 MPa m), the treatment is chosen for further applications. Although the initial intention of the present development for high toughness was to use the effect of highly effective diffusion bonding induced in two phase condition, it cannot confirmed yet if the postulated mechanism is highly responsible for the high toughness achieved by the CR treatments. Further, since the a–c interface formed in the intercritical region is mobile and thus the c-phase is expected to shrank during slow cooling [81,82], the final amounts of quenched martensite is not known. At the moment, therefore, the detailed behaviors of phase transformation, dislocation recovery, and diffusion during the repeated hot-rolling treatment are not understood well. A significant amount of effort might be required for understanding such high temperature phenomena occurring on TMT, especially in two phase temperature region. While further scientific understanding is left to future studies, achieving the main goal of developing engineering grade NFA is considered to be significant since the high toughness NFAs can be further processed for engineering purposes, such as tube drawing and forming.
5. Concluding remarks In the past, very high strength can be achieved in NFAs at the expense of fracture toughness and ductility. Low energy decohesion at boundaries causes poor fracture resistance at high temperatures. Developing a high toughness NFA was the ultimate goal of this study. To strengthen the powder-metallurgy produced weak boundaries by an enhanced diffusion bonding, the schedules of isothermal annealing (IA) and controlled rolling (CR) after annealing were attempted. Improvement of fracture toughness in the CR 9YWTV-PM2 was significant: its fracture toughness was as high as those of non-ODS F/M steels. In particular, the 9YWTV-PM2 controlled-rolled at 900 °C resulted in the best fracture toughness among NFAs p (>150 MPa m at both representative temperatures: RT and 500 °C. It can be claimed that the final goal of developing an engineering grade NFA has been achieved. More detailed microstructural and mechanical characterizations for selected materials, i.e., the 9YWTV-PM2 alloy after 50% CR at 900 °C are underway and the results are to be reported in separate articles. The topics in these companion articles include the detailed temperature dependence of mechanical properties, delayed phase transformation in nanostructure and stability of nanoclusters and grains in TMT processes, etc. Irradiation experiment is also being pursued.
Acknowledgements This research was sponsored by the International Nuclear Energy Research Initiative (I-NERI) Collaboration between United States and South Korea (I-NERI Project 2010-004-K). In the US, this research was also part of the Fuel Cycle R&D program sponsored by the Office of Nuclear Energy, US Department of Energy, under Contract DE-AC05-00OR22725 with UT-Battelle, LLC. The authors ex-
press special thanks to Drs. Chad M. Parish and Kurt. A. Terrani for their thorough reviews and thoughtful comments.
References [1] L.K. Mansur, A.F. Rowcliffe, R.K. Nanstad, S.J. Zinkle, W.R. Corwin, R.E. Stoller, J. Nucl. Mater. 329 (2004) 166–172. [2] S.J. Zinkle, J.T. Busby, Mater. Today 12 (2009) 12–19. [3] S.A. Maloy, M. Toloczko, J. Cole, T.E. Byun, J. Nucl. Mater. 415 (2011) 302–305. [4] S.J. Zinkle, G.S. Was, Acta Mater. 61 (2013) 735–758. [5] B.Z. Margolin, A.I. Murashova, V.S. Neustroev, Strength Mater. 44 (2012) 223– 240. [6] S.J. Zinkle, N.M. Ghoniem, J. Nucl. Mater. 417 (2011) 2–8. [7] E. Kenik, J.T. Busby, Mater. Sci. Eng. R 73 (2012) 67–83. [8] F.A. Garner, D.S. Gelles, ASTM STP 1046 (1999) 673–683. [9] T.R. Allen, J.I. Cole, C.L. Trybus, D.L. Porter, H. Tsai, F. Garner, E.A. Kenik, T. Yoshitake, J. Ohta, J. Nucl. Mater. 348 (2006) 148–164. [10] I.I. Balachov, E.N. Shcherbakov, A.V. Kozlov, I.A. Portnykh, F.A. Garner, J. Nucl. Mater. 329–333 (2004) 617–620. [11] V.S. Neustroev, F.A. Garner, J. Nucl. Mater. 386–388 (2009) 157–160. [12] S.J. Zinkle, N.M. Ghoniem, J. Nucl. Mater. 51–52 (2000) 55–71. [13] A.V. Babak, E.I. Uskov, Strength Mater. 16 (1984) 943–949. [14] Y. Mutoh, K. Ichikawa, K. Nagata, M. Takeuchi, J. Mater. Sci. 30 (1995) 770–775. [15] P. Gumbsch, J. Riedle, A. Hartmaier, H.F. Fischmeister, Science 282 (1998) 1293–1295. [16] B.V. Cockeram, R.W. Smith, L.L. Snead, J. Nucl. Mater. 346 (2005) 145–164. [17] B.V. Cockeram, E.K. Ohriner, T.S. Byun, M.K. Miller, L.L. Snead, J. Nucl. Mater. 382 (2008) 229–241. [18] B.V. Cockeram, R.W. Smith, K.J. Leonard, T.S. Byun, L.L. Snead, J. Nucl. Mater. 382 (2008) 1–23. [19] T.S. Byun, M. Li, B.V. Cockeram, L.L. Snead, J. Nucl. Mater. 376 (2008) 240–246. [20] B.H. Sencer, G.M. Bond, F.A. Garner, M.L. Hamilton, B.M. Oliver, L.E. Thomas, S.A. Maloy, W.F. Sommer, M.R. James, P.D. Ferguson, J. Nucl. Mater. 283–287 (2000) 324–328. [21] P. Yvon, F. Carre, J. Nucl. Mater. 385 (2009) 217–222. [22] T.S. Byun, K. Farrell, J. Nucl. Mater. 318 (2003) 292–299. [23] J.D. Hunn, E.H. Lee, T.S. Byun, L.K. Mansur, J. Nucl. Mater. 296 (2001) 203–209. [24] N. Hashimoto, J.D. Hunn, T.S. Byun, L.K. Mansur, J. Nucl. Mater. 318 (2003) 300–306. [25] A.F. Rowcliffe, L.K. Mansur, D.T. Hoelzer, R.K. Nanstad, J. Nucl. Mater. 392 (2009) 341–352. [26] R.L. Klueh, Int. Mater. Rev. 50 (2005) 287–310. [27] R.L. Klueh, D.R. Harris, High-chromium ferritic and martensitic steels for nuclear applications, ASTM (2001). ISBN 0-803102090-7. [28] T.R. Allen, J.T. Busby, R.L. Klueh, S.A. Maloy, M.B. Toloczko, JOM 60 (2008) 15– 23. [29] O. Anderoglu, T.S. Byun, M. Toloczko, S.A. Maloy, Met. Mater. Trans. A (2013) 70–83. [30] R. Chaouadi, G. Coen, L. Lucon, V. Massaut, J. Nucl. Mater. 403 (2010) 15–18. [31] T.S. Byun, M. Li, K. Farrell, Met. Mater. Trans. A 44 (2013) 84–93. [32] S.A. Maloy, M. Toloczko, J. Cole, T.S. Byun, J. Nucl. Mater. 415 (2011) 302–305. [33] J.S. Cheon, C.B. Lee, B.O. Lee, J.P. Raison, T. Mizuno, F. Delage, J. Carmack, J. Nucl. Mater. 392 (2009) 324–330. [34] B.H. Sencer, J.R. Kennedy, J.I. Cole, S.A. Maloy, F.A. Garner, J. Nucl. Mater. 393 (2009) 235–241. [35] F.H. Huang, M.L. Hamilton, J. Nucl. Mater. 187 (1992) 278–293. [36] F.H. Huang, J. Nucl. Mater. 43 (1992) 748–773. [37] F.H. Huang, D.E. Gelles, Eng. Fract. Mech. 19 (1984) 1–84. [38] T.S. Byun, J.H. Kim, J.H. Yoon, D.T. Hoelzer, J. Nucl. Mater. 407 (2010) 78–82. [39] T.S. Byun, W.D. Lewis, M.B. Toloczko, S.A. Maloy, J. Nucl. Mater. 421 (2012) 104–111. [40] T.S. Byun, M.B. Toloczko, T.A. Saleh, S.A. Maloy, J. Nucl. Mater. 432 (2013) 1–8. [41] R.L. Klueh, N. Hashimoto, P.J. Maziasz, Scripta Mater. 53 (2005) 275–280. [42] S. Hollner, B. Fournier, J. Nucl. Mater. 405 (2010) 101–108. [43] L. Tan, D.T. Hoelzer, J.T. Busby, M.A. Sokolov, R.L. Klueh, J. Nucl. Mater. 422 (2012) 45–50. [44] J. Chen, M.A. Pouchon, A. Kimura, P. Jung, W. Hoffelner, J. Nucl. Mater. 386–388 (2009) 143–146. [45] P. Lours, S. Le Roux, G. Bernhart, J. Adv. Mater. 33 (2001) 11–20. [46] P. Krautwasser, A. Czyrska-Filemonowicz, M. Widera, F. Carsughi, Mater. Sci. Eng. A 177 (1994) 199–208. [47] J. Chen, P. Jung, W. Hoffelner, H. Ullmaier, Acta Mater. 56 (2008) 250–258. [48] A. Wasilkowska, M. Bartsch, U. Messerschmidt, R. Herzog, A. CzyrskaFilemonowicz, J. Mater. Process. Technol. (2003) 218–224. [49] A. Czyrska-Filemonowicz, B. Dubiel, J. Mater. Process. Technol. 64 (1997) 53– 64. [50] R.L. Klueh, J.P. Shingledecker, R.W. Swindeman, D.T. Hoelzer, J. Nucl. Mater. 341 (2005) 103–114. [51] M.B. Toloczko, D.S. Gelles, F.A. Garner, R.J. Kurtz, K. Abe, J. Nucl. Mater. 329– 333 (2004) 352–355. [52] P. Miao, G.R. Odette, T. Yamamoto, M. Alinger, D. Klingensmith, J. Nucl. Mater. 377 (2008) 59–64. [53] M.K. Miller, D.T. Hoelzer, E.A. Kenik, K.F. Russell, J. Nucl. Mater. 329–333 (2004) 338–341.
T.S. Byun et al. / Journal of Nuclear Materials 449 (2014) 290–299 [54] D.T. Hoelzer, J. Bentley, M.A. Sokolov, M.K. Miller, G.R. Odette, M.J. Alinger, J. Nucl. Mater. 367–370 (2007) 166–172. [55] M.J. Alinger, G.R. Odette, D.T. Hoelzer, Acta Mater. 57 (2009) 392–406. [56] G.R. Odette, M.J. Alinger, B.D. Wirth, Ann. Rev. Mater. Res. 38 (2008) 471–503. [57] M.K. Miller, D.T. Hoelzer, E.A. Kenik, K.F. Russell, Intermetallics 13 (2005) 387– 392. [58] S. Ukai, S. Ohtsuka, T. Kaito, H. Sakasegawa, N. Chikata, S. Hayashi, S. Ohnuki, Mater. Sci. Eng. A 510–511 (2009) 115–120. [59] H. Sakasegawa, S. Ukai, M. Tamura, S. Ohtsuka, H. Tanigawa, H. Ogiwara, A. Kohyama, M. Fujiwara, J. Nucl. Mater. 373 (2008) 82–89. [60] P. Olier, J. Malaplate, M.H. Mathon, D. Nunes, D. Hamona, L. Toualbi, Y. de Carlan, L. Chaffron, J. Nucl. Mater. 428 (2012) 40–46. [61] H.S. Cho, R. Kasada, A. Kimura, J. Nucl. Mater. 367 (2007) 239–243. [62] H. Kishimoto, R. Kasada, O. Hashitomi, A. Kimura, J. Nucl. Mater. 386 (2009) 533–536. [63] N. Akasaka, S. Yamashita, T. Yoshitake, S. Ukai, A. Kimura, J. Nucl. Mater. 329 (2004) 1053–1056. [64] S. Ukai, M. Harada, H. Okada, M. Inoue, S. Nomura, S. Shikakura, K. Asabe, T. Nishida, M. Fujiwara, J. Nucl. Mater. 204 (1993) 65–71. [65] S. Ukai, T. Nishida, T. Okuda, T. Yoshitake, J. Nucl. Mater. 258 (1998) 1745– 1749. [66] A.-L. Rouffié, P. Wident, L. Ziolek, F. Delabrouille, B. Tanguy, J. Crépin, A. Pineau, V. Garat, B. Fournier, J. Nucl. Mater. 433 (2013) 108–115. [67] A. Steckmeyer, M. Praud, B. Fournier, J. Malaplate, J. Garnier, J.L. Béchade, I. Tournié, A. Tancray, A. Bougault, P. Bonnaillie, J. Nucl. Mater. 405 (2010) 95– 100.
299
[68] J.H. Kim, J.H. Lee, J.Y. Min, S.W. Kim, C.H. Park, J.T. Yeom, T.S. Byun, J. Alloys Compd. 580 (2013) 125–130. [69] D.K. Mukhopadhyay, F.H. Froes, D.S. Gelles, J. Nucl. Mater. 258–263 (1998) 1209–1215. [70] D.A. McClintock, D.T. Hoelzer, M.A. Sokolov, R.K. Nanstad, J. Nucl. Mater. 386 (2009) 307–311. [71] D.A. McClintock, M.A. Sokolov, D.T. Hoelzer, R.K. Nanstad, J. Nucl. Mater. 392 (2009) 353–359. [72] T.S. Byun, J.H. Kim, J.H. Yoon, D.T. Hoelzer, J. Nucl. Mater. 407 (2010) 78–82. [73] J.H. Kim, T.S. Byun, D.T. Hoelzer, J. Nucl. Mater. 407 (2010) 143–150. [74] J.H. Kim, T.S. Byun, D.T. Hoelzer, S.W. Kim, B.H. Lee, Mater. Sci. Eng. A 559 (2013) 101–110. [75] J.H. Kim, T.S. Byun, D.T. Hoelzer, C.H. Park, J.T. Yeom, J.K. Hong, Mater. Sci. Eng. A 559 (2013) 111–118. [76] J.H. Kim, T.S. Byun, D.T. Hoelzer, J. Nucl. Mater. 425 (2012) 147–155. [77] J.H. Kim, T.S. Byun, D.T. Hoelzer, J. Nucl. Mater. in press. [78] N. Orhan, M. Aksoy, M. Eroglu, Mater. Sci. Eng. A 271 (1999) 458–468. [79] B. Kurt, N. Orhan, E. Evin, A. Çalik, Mater. Lett. 61 (2007) 1747–1750. [80] S. Sun, M. Pugh, Mater. Sci. Eng. A 276 (2000) 167–174. [81] G.R. Speich, V.A. Demarest, R.L. Miller, Metall. Trans. A 12 (1981) 1419–1428. [82] Y.S. Ahn, H.D. Kim, T.S. Byun, Y.J. Oh, G.M. Kim, J.H. Hong, Nucl. Eng. Des. 194 (1999) 161–177. [83] http://www.calphad.com/iron–chromium.html. [84] S.H. Ryu, J. Yu, Met. Trans. 29A (1998) 1573–1578.