Microstructure Study of NiCrAlY and FeCrAlY Exposed to Superheated Steam at 800 °C

Concerns with greenhouse gas emissions and the uncertainty of long-term supply of fossil fuels have resulted in renewed interest in nuclear energy as an essential part of the energy mix for the future. To improve the current Gen-II and Gen-III nuclear plant efficiency, Gen-IV International Forum [1] was formed to develop Generation-IV reactor technologies that will meet the needs of the present efficiency demand and also be sustainable for the future [1]. The reactors being considered in the Gen-IV International Forum program include the six most promising designs that the International Scientific and Engineering Community has selected from over 100 candidates. Canada is currently taking the lead role in the design of Gen-IV supercritical water-cooled reactor (SCWR). The SCWR design utilizes single-phase supercritical water (SCW) as coolant, which will lead to higher thermodynamic efficiency (i.e., about 45% versus 33% for current reactors) and considerable plant simplification. The estimated operating condition for the SCWR calls for an operating pressure of 25 MPa, a core inlet temperature of 280 °C, and a core outlet temperature up to 620 °C. Peak fuel cladding temperatures may reach as high as 850 °C in certain designs [2,3]. The selection of suitable base materials and development of compatible corrosion resistant coatings for in-core and out-of-core components are an integral part for the successful design of SCWR [4–11]. As no single alloy has been identified so far which can meet all of the properties required for critical SCWR application, one of the proposed solutions is to apply a surface coating to the base material. The utilization of corrosion resistant coatings on high temperature steels (in core) or nickel-based alloys (out-of-core) has emerged as an answer to offer possible combination of mechanical and corrosion resistance for the construction of the SCWR. Coating development for gas and steam turbines can be divided into two categories: MCrAlY (M = Ni, Fe, NiCo, or CoNi) overlay coating and aluminide (β-Ni(Fe)Al) diffusional coating. These coatings rely on the formation of dense, adherent, and renewable alumina and/or chromia (via oxidation of aluminum and chromium) to provide high temperature oxidation and corrosion resistance. Aluminide coating, applied by slurry method, such as FeAl has shown to offer 4 magnitude orders of surface protection to P91 under 80% steam condition at 650 °C [12]. Stress relief cracks were observed on the coating surface due to mismatch in thermal expansion coefficients of aluminide and substrate alloys [13]. The MCrAlY coatings, comprised of β aluminide embedded in more ductile γ-Ni(Fe,Co) solid solution, are less brittle than aluminide coatings. They can be deposited by either thermal spray or electron beam physical vapor deposition techniques. Aluminum contents in commercial MCrAlY coatings typically vary between 5 wt % and 12 wt % and when Al is combined with Cr, a synergistic effect can result in less Al is required to form a dense Al₂O₃ layer. Thermal sprayed FeAlCr(MoSi) showed improved corrosion resistance than P92 when being tested under flowing steam at 650 °C [14]. Furthermore, thermal sprayed coating suffers less Al depletion, due to inward Al diffusion, than slurry iron aluminide coating under similar test condition.

Our past research showed that the corrosion rate of iron-based FeCrAlY and nickel-based NiCrAl coating samples fell below 0.05 mg/cm² after 6000 h in SCW at 500 °C [15–18]. These results were promising when being compared to a corrosion rate of 0.17 mg/cm² for 316 austenitic stainless (at 500 °C) [3] and 0.2 mg/cm² for 304 NG (at 550 °C) [19] after 1000 h testing. However, the cladding temperature of Canadian SCW has been estimated to reach a peak temperature as high as 800 °C. The behavior of these two coating alloys at higher temperatures is unknown. Studies carried out by others suggest that MCrAlY type alloys may offer oxidation protection to temperatures as high as 1200 °C [20–22]. The purpose of this study is to investigate the behavior of FeCrAlY and NiCrAl alloys at the Canadian SCWR peak clad temperature.

Because of no SCW autoclave or loop that can reach 800 °C, superheated steam (SHS) was used in this study as surrogate to SCW. The use of superheated steam to simulate the performance of SCW has been evaluated at 625 °C where high Cr alloys have demonstrated similar performance [23]. Furthermore, according to Guzonas and Cook at temperatures above about 500 °C at 25 MPa, the corrosion of alloys in SCW is not unlike corrosion in high-temperature superheated steam [24]. The corrosion behavior in superheated steam will be analyzed through weight change, scanning electron microscopy (SEM), energy dispersive X-ray spectrometry (EDS), and X-ray diffractometer (XRD) techniques and reported in this study.

FeCrAlY samples (1 cm × 1 cm) and NiCrAl (1 cm diameter) were cut from cast bars purchased from sophisticated alloys. To hang samples in the steam rig, a press was used to drill a 3.175 mm diameter hole in each sample. Samples were ground using 240, 320, 400, and 600 grit SiC abrasive papers. After grinding, the samples were cleaned in an ultrasonic bath (Branson 2510) using soap and water for 15 min followed by 15 min in acetone. To remove all moisture, the samples were then placed in a furnace (Cole-Parmer Stable Temp) at 200 °C for 2 h. Prior to testing, sample dimensions were measured and samples were weighed (Mettler Toledo AG285, ±0.1 mg precision) both before and after testing in SHS to calculate the weight change per unit surface area. Sample compositions are provided in Table 1. An SHS rig was used to simulate SCW conditions. A photograph of the superheated system is shown in Fig. 1. The water distiller supplies the gear pump (Cole-Parmer) with water, which is then pumped into the steam generator (MHI). The MHI discharges superheated steam into the superheater where the superheated steam reaches the final testing temperature. The steam flows into an alumina tube, where the samples are hung on ceramic rods. Exhaust steam from the superheater is discharged to a water reservoir. The steam is maintained at absolute pressure of 0.1 MPa throughout the test.

A total of six samples were tested at SHS at 800 °C and 0.1 MPa. The surface microstructure after SHS testing was characterized using SEM/EDS. Phase composition of the surface oxides formed was also analyzed using XRD. A Co Kα radiation X-ray source was used with an applied voltage and current of 35 kV and 40 mA, respectively. All the tests were carried out within a 2θ range of 25–80 deg (θ is the angle between the incident X-ray beam and the horizontal surface of the sample).

Three samples of each material type were inspected after being tested for 300 h. And two of the samples were further tested to 600 h. The visual appearances of both FeCrAlY and NiCrAl were recorded and shown in Figs. 2 and 3. FeCrAlY sample changed from a typical polished metallic appearance to more a speckled surface of mixed colors (Fig. 2(a)), suggesting different oxide scale types, structures, or thicknesses. After 600 h in SHS, the surface became more uniform in its appearance (Fig. 2(b)) and less reflective. The difference in surface appearance of the samples indicates the progression of oxidation. Changes were also observed on NiCrAl samples from 300 to 600 h, however, the oxidation occurred to a lesser extent. The NiCrAl samples after 300 h in SHS (Fig. 3(a)) still maintained much of the metallic appearance and only slight discoloration was observed. After 600 h (Fig. 3(b)), the NiCrAl sample surface started to display a mixture of colors. Again, the oxidation progressed from 300 h to 600 h based on the color change.

Detailed SEM analysis is presented in Sec. 3.2.

Sample weight changes were measured after 300 and 600 h in SHS. These changes were normalized by sample surface areas, giving the weight change in mg/cm². The average weight changes for FeCrAlY and NiCrAl are shown in Fig. 4. Also included in this figure are weight changes of several fuel clad candidate materials (800 H, 310 S, and IN625) tested under the same condition. FeCrAlY has the most weight gain in the first 600 h, while IN625 the least. NiCrAlY has the second lowest weight gain among the five materials exhibited here.

Although the 600 grit grinding lines are still visible on the surface (Fig. 5(a)), oxidation has taken place on FeCrAlY sample surface after 300 h in SHS. Two noticeable changes are: grain boundaries are heavily decorated by precipitates (A) (Fig. 5(b)) and discrete particles (B) reside within the grains (Fig. 5(c)).

In order to identify the precipitates formed on the grain boundaries and within the grains, EDS analysis was carried out at various locations, with the results summarized in Table 2. A surface area analysis revealed the presence of substantial (35.6 at %) of oxygen, an indication of oxide formation on the surface. Alumina (Al₂O₃) was formed in particular along the grain boundaries, based on the EDS spot analysis results show in Table 2. The discrete particles within the grain (Fig. 5(c)) contain similar composition as grain boundary alumina, however, without the presence of Y. The region C inside the grain is enriched in O, Al, Fe, and Cr, perhaps a mixture of different oxides.

More oxidation took place on FeCrAlY after 600 h in SHS, as the percentage of oxygen detected on a surface area (Table 3) was greater than that measured on FeCrAlY after 300 h. It still contained grain boundary precipitates (Fig. 6); however, they became much wider and exhibited visible crystalline features. There were two distinct types of grain boundary precipitates, linear ones forming along the grain boundaries (B) and also spherical ones (A in Fig. 6). The linear precipitates were identified by EDS as Al₂O₃ with Y addition, while the spherical ones contained more Fe, likely with a magnetite type oxide. The matrix within the grain developed well defined, needle-shaped alumina (Fig. 6(c)), as confirmed by EDS analysis. XRD analysis further confirmed the presence of both Al₂O₃ and magnetite (Fig. 9(a)).

The SHS exposed NiCrAl sample did not show the decorated grain boundary features seen on the FeCrAlY samples. The surface instead exhibits light and gray areas (Fig. 7(a)). Under a large magnification, the gray area is seen to contain spherical and needle shaped particles. The composition measurement (Table 4) of the bright area in Fig. 7(a) and A in Fig. 7(c) revealed the presence of mostly Ni-based oxide (the low oxygen count suggests X-ray signals from the metallic substrate under the thin oxide layer). The spherical particles shown in Fig. 7(b) and C in Fig. 7(c) are mixed oxide (spinel) with possible composite of NiO·(Cr,Al)₂O₃. The fine needles shown in Fig. 7(b) and D in Fig. 7(c) are chromia in nature with some incorporated Al. There are occasional dark, discrete particles, labeled B in Fig. 7(c). They are identified as Al₂O₃. However, it is not clear whether these particles were formed from the base metal reaction with SHS or condensed from SHS during cooling as the test chamber was made of Al₂O₃ tube.

After further exposure to SHS, NiCrAl sample surface became more uniformly gray (Fig. 8(a)). A more cuboidal phase (A in Fig. 8(b)) is now well developed and is composed of primarily Al₂O₃ (Table 5). The coverage of alumina on the surface increased after 600 h. The spinel particle size (oblong shaped phase marked with C in Fig. 8(c)) has enlarged in size but reduced in coverage and Al content in spinel also increased (C—Table 5). The fine needles (Cr₂O₃) formed on NiCrAl after 300 h maintained the same morphology (C in Fig. 8(b)) but less Al presence (Table 5) after 600 h. The presence of different phases after 600 h in SHS was further confirmed with XRD analysis (Fig. 9(b)).

X-ray diffractometer diffraction analysis was carried out on samples exposed to SHS for 600 h. As the oxides formed on the surface were still quite thin, the base metal was detected during analysis, shown as austenitic γ phase in Figs. 9(a) and 9(b). Overall both samples have alumina formation after 600 h in SHS in addition to other types of oxide. For FeCrAlY, the primary surface oxides formed are alumina along the grain boundaries and within the grain and also magnetite along grain boundaries. There is an unidentified peak at 2θ = 46.5 deg for both samples. The oxides formed on NiCrAl are more complex than that on FeCrAlY; three different types of oxides can be identified: Al₂O₃, Cr₂O₃, and also spinel.

To reduce oxidation in a SCW or superheated steam, a stable and adherent surface oxide layer is required. The known oxide formers are chromium, aluminum, silicon, zirconium, yttrium, niobium, tantalum, and combination of the above [25]. To be classified as an oxide former, the element must meet several requirements: (1) a driving force to form oxide under given test condition (oxide is stable) (Ellingham diagram) [26], (2) the oxide formed has suitable adhesion to the base alloy, (3) oxide is dense, i.e., forming barrier to diffusing species (inward or outward), and (4) the volume of the oxide should be equivalent to the metal so that no significant stress is induced during oxide formation (Pilling–Bedworth ratio) [27]. Elements such as Fe, Ni, and Cu are not considered oxide formers since Fe-oxide expanses excessively, while NiO and Cu₂O are not stable at elevated temperature based on Ellingham diagram. In addition, based on Pourbaix diagram for chromium [7,28], chromium(III) oxide can form volatile Cr⁶⁺ under high temperature oxidative steam. In fact, previous research has shown the potential of Cr oxide dissolution in SCWR core due to water radiolysis [24,29]. Because of the potential for Cr oxide dissolution in the SCWR core, the use of other oxide-forming species such as Al is currently examined. The incorporation of Al in the FeCrAlY and NiCrAl alloys in this study was to examine the ability of these alloys to form alumina on the surface during SCW and SHS exposure. Based on the results obtained from SEM analysis, it is evident that both alloys were able to form Al₂O₃ on the surface.

In terms of weight change, it only provides a rudimentary indication as to the extent of oxide formation or scale spallation. While weight gain indicates oxide formation and weight loss signals scale spallation or dissolution, smaller weight gain does not necessarily rank an alloy's ability to resist oxidation as oxidation formation and subsequently dissolution can lead to either weight loss and reduced weight gain. Corrosion studies of Ni-based alloys often show less weight gain than that seen in Fe-based alloy [29,30]; however, Ni-based alloys may be susceptible to intergranular attack and pitting in SCW [31], leading to weight loss. The same NiCrAl and FeCrAlY alloys were previously tested under SCW condition at 500 °C. The weight changes per unit surface area were approximately: 0.01 mg/cm³ (FeCrAlY) and −0.010 mg/cm³ (equivalent to 600 h) [17,18], suggesting weight loss on NiCrAlY. Furthermore, for IN 625 tested in SHS at 800 °C, the average weight change after 600 h was found to be −0.044 mg/cm³ [32]. In this study, the weight change for FeCrAlY was 0.3 mg/cm³ after 600 in SHS at 800 °C, a magnitude greater that from the test in SCW at 500 °C. Instead of initial weight loss for nickel-based alloys such as IN 625 in SHS, however, there was a weight gain observed for NiCrAl at 0.15 mg/cm³ (NiCrAlY) after SHS test. This suggest that the inclusion of Al in NiCrAl (and perhaps the high temperature of SHS) leads to no or reduced weight loss, compared to that observed on IN 625 tested in SHS and the same NiCrAl tested in SCW at 500 °C. Also, the Ni-based NiCrAl showed more alumina formation and less weight gain that Fe-based FeCrAlY.

Ongoing Auger study has shown that Ni-based faced centered cubic structure may result in excessive O diffusion into the base metal. This should be a concern as the fuel clad is <1 mm and the presence of O may embrittle the base material. According to Persaud et al. [33], the internal oxidation of Cr in nickel base alloy can produce stresses that are relieved through the expulsion of metallic Ni nodules to the surface. These processes can modify the diffusion of oxygen. Further study is underway and the results will be reported in the future communications.

Finally, although the use of nickel or nickel-based alloys in a nuclear reactor core is not ideal due to its large neutron absorption cross section and susceptibility to irradiation damage, a thin coating of a nickel-based material may be acceptable if it provides superior corrosion resistance. Longer terms test in SHS or SCW at 800 °C will be needed to obtain conclusive results.

In this study, FeCrAlY and NiCrAl coating samples were tested in SHS at 800 °C for up to 600 h. The FeCrAlY was covered with surface scale more rapidly than that on NiCrAl. The weight change results also suggest that more oxide formation took place on FeCrAlY than NiCrAl. For FeCrAlY, grain boundary oxide (Al₂O₃) formed rapidly upon exposure to SHS for 300 h. Further exposure caused more intragranular Al₂O₃ to form, in addition to magnetite formation on the grain boundaries. For NiCrAl, NiO seems to have formed initially upon SHS exposure due to the high Ni content in the alloy. Spinel and (Cr,Al)₂O₃ also formed after 300 h with limited amount of Al₂O₃. After 600 h, Al₂O₃ became well developed and the coverage of spinel and Cr₂O₃ on the surface reduces.

A =

area, m²

I =

current, A

L =

length, m

m =

mass, kg

p =

pressure, Pa

t =

temperature,  °C

U =

voltage, V

Greek Symbols

γ =

austenitic phase

θ =

angle, deg

Kα =

Co electronic transition

Acronyms and Abbreviations

EDS =

energy dispersive X-ray spectrometry

MHI =

denomination of the steam generator device

SCW =

supercritical water

SCWR =

supercritical water reactor

SEM =

scanning electron microscopy

SHS =

super-heated steam

Wt =

weight

XRD =

X-ray diffractometer