Introduction

A supercritical-CO2 (S-CO2) Brayton cycle is considered as a promising power conversion system for nuclear [1,2,3], solar thermal [4], geothermal [5] and fossil fuel [6] power generation systems. A few advantages of S-CO2 Brayton cycle over both the steam and gas turbine systems include the reduced compressive power requirement, higher thermal efficiency, much smaller components size and comparatively less material issues at high temperature [1, 2].

Recently, the corrosion and carburization behaviour of several Fe- and Ni-based alloys has been extensively investigated in high-temperature CO2 and S-CO2 environments [7,8,9,10,11,12]. Overall, Ni-based alloys showed better corrosion resistance than stainless steels because of the formation of stable and continuous chromia (Cr2O3) layer on the surface [10]. For austenitic stainless steels, discontinuous islands of carburized M23C6 regions were found below the oxide layer, which appeared to promote spallation of oxide layer after long-term exposure [9, 10]. Whereas, an amorphous C layer was observed at the interface between the chromia layer and underlying matrix for the Ni-based alloys exposed up to 1000 h at 550–650 °C (20 MPa) [11, 12]. Despite the presence of the C layer, carbide precipitates were not found in the matrix, and spallation of oxide layer was not observed. Nonetheless, the development of the amorphous C layer under the inner chromia was rather surprising, considering the presence of the continuous and seemingly protective chromia layer. Meanwhile, no such C layer was observed when pre-formed alumina layer was present on the surface [13].

In this regard, the role of oxide layers, especially a chromia layer, on the characteristics of the oxide layer and its subsequent impact on the carburization behaviour of Ni-based alloys need to be understood. Therefore, the objective of this study is to investigate the effect of Cr content in Ni-based binary alloys on the corrosion and carburization behaviour in high-temperature S-CO2 condition, emphasizing the role of chromia layer on the carburization behaviour. For this, pure Ni and Ni-xCr binary alloys were prepared and exposed to S-CO2 environment at 600 °C and 20 MPa for 200 h. The corrosion and carburization resistance were characterized by weight gain measurements and several analytical techniques. Finally, carburization resistance was discussed in view of the equilibrium oxygen partial pressure (pO2)eq of the oxides formed on the Ni-xCr binary alloys.

Materials and Experimental

The pure Ni and Ni-xCr (x = 7, 14, 22 and 27 wt%) binary alloys were prepared by vacuum arc remelting using pure Ni and Cr metal scrap pieces with 99.995% purity. The chemical composition of the as-cast alloys was confirmed by inductively coupled plasma (ICP) analysis. The as-cast pure Ni showed a fully austenite microstructure with an ASTM grain size number of about 8.5. For the Ni-xCr binary alloys containing Cr ≥ 7%, grains were much coarser with an ASTM grain size number of about 2.5. All the materials were cut into coupon-type specimens with 12 mm in diameter and 1 mm in thickness. A hole of 1.5 mm in diameter was drilled at the upper part of the specimens for a hanging wire. The specimens were ground to 1200 grit finish and ultrasonically cleaned with ethanol before the corrosion test. The specimens were hung with a Pt wire of 0.5 mm diameter, separated with alumina spacers and kept in an alumina boat placed in the autoclave. An isothermal corrosion test was conducted at 600 °C and 20 MPa for 200 h in an S-CO2 corrosion test facility by feeding a research-grade CO2 gas (99.999% purity) to a high-pressure CO2 pump. The oxygen partial pressure pO2 was calculated as 1.23 × 10−7 atm [12]. The detailed description of experimental set-up and procedure could be found elsewhere [11].

After the tests, the weight gains of the specimens were measured using a microbalance (Mettler Toledo AT21 Comparator) with 0.001 mg resolution. To characterize the oxide layer and underlying microstructure in the Ni-xCr binary alloys, X-ray diffraction (XRD, Rigaku D/MAX-2500), transmission electron microscope (TEM, Titan cubed G2 60–300) equipped with energy dispersive spectroscope (EDS) and secondary ion mass spectroscope (SIMS, CAMECA IMS 7f) analyses were performed. The TEM specimens were taken from the corroded specimens using a focused ion beam (FIB, Helios Nanolab 450 F1). The equilibrium oxygen partial pressure (pO2)eq of various oxides in contact with matrix of Ni-xCr binary alloys and carbide stability diagram were calculated using Thermo-Calc®, a commercial thermodynamic software, to understand the carburization mechanism.

Results and Discussion

Corrosion Behaviour in S-CO2 Environment

Figure 1 shows the measured weight gain and oxide thickness of pure Ni and Ni-xCr binary alloys exposed to S-CO2 environment. Compared to pure Ni and Ni-7Cr, binary alloys containing at least 14% Cr showed better corrosion resistance with smaller weight gain and thinner oxide thickness. In addition, the corrosion resistance was about the same for the alloys with at least 14% Cr. On the other hand, the corrosion resistance of Ni-7Cr binary alloy was the worst among the tested alloys in terms of weight gain and oxide thickness. The large increase in weight gain for Ni-7Cr alloy compared to pure Ni could have been attributed to the oxidation of Cr present in the matrix. XRD results of the alloys exposed to S-CO2 environment are shown in Fig. 2. For all tested alloys, the predominant peaks observed from XRD analysis can be attributed to the matrix, consistent with the thin oxide thickness shown in Fig. 1. Other than matrix peaks, only NiO peaks were observed for pure Ni, whereas minor chromia peaks as well as strong NiO peaks were shown for Ni-7Cr binary alloy. Meanwhile, only chromia peaks were observed for the alloys with at least 14% Cr. In addition, alloys with 14–27% Cr exhibited similar XRD results, suggesting the presence of similar oxide characteristics for these alloys. Overall, XRD results shown in Fig. 2 combined with weight gain results shown in Fig. 1 confirmed that the formation of the predominantly chromia layer showed better corrosion resistance in high-temperature S-CO2 environment.

Fig. 1
figure 1

Weight gains and oxide thickness of Ni-xCr binary alloys corroded in S-CO2 at 600 °C (20 MPa) for 200 h

Full size image
Fig. 2
figure 2

XRD analysis results of Ni-xCr binary alloys corroded in S-CO2 at 600 °C (20 MPa) for 200 h

Full size image

The detailed cross-sectional STEM-EDS analyses results of oxide layer and underlying matrix of the tested alloys are shown in Figs. 3, 4, 5 and 6. For pure Ni exposed to S-CO2 environment, a rather uniform oxide layer was observed on the surface, which was identified as NiO based on EDS mapping and line scanning results (Fig. 3). It should be noted that the slight presence of Cr and C in the oxide layer can be attributed to the electronic noise evolved during the STEM-EDS line scan process since it had been confirmed that no Cr had been present in the pure Ni specimens by ICP analysis. Whereas, Ni-7Cr binary alloy showed a complex oxide structure consisted of outer uniform NiO layer and inner layer of a rather irregular Cr2O3 mixed with NiO (Fig. 4b, c). The formation of the rather complex oxide structure could be explained as follows. Initially, NiO would be formed on the surface due to small amount of Cr in the Ni-matrix insufficient to form the protective chromia. It was reported that for dilute Ni–Cr binary alloys oxidized at high temperature, the internal Cr2O3 and NiO would grow by either oxygen transport from NiO dissociation or through the short-circuit transport of oxidants from the atmosphere [14]. Meanwhile, the presence of amorphous C region around the internal chromia (Fig. 4d) suggests that carbonaceous species would have transported into the matrix to grow Cr2O3 and NiO internally.

Fig. 3
figure 3

TEM analyses results of oxide layer on pure Ni corroded in S-CO2 at 600 °C (20 MPa) for 200 h. a Cross-sectional STEM micrograph and EDS mapping images and b line scan at the matrix–oxide interface

Full size image
Fig. 4
figure 4figure 4

TEM analyses results of oxide layer on Ni-7Cr binary alloy corroded in S-CO2 at 600 °C (20 MPa) for 200 h. a Cross-sectional STEM micrograph and EDS mapping images, b line scan I, c line scan II at the matrix–oxide interface and d STEM micrograph, HRTEM micrograph and EDS mapping of amorphous C region

Full size image
Fig. 5
figure 5

TEM analyses results of oxide layer on Ni-14Cr binary alloy corroded in S-CO2 at 600 °C (20 MPa) for 200 h. a Cross-sectional STEM micrograph and EDS mapping images, b line scan at the matrix–oxide interface and c HRTEM micrograph highlighting the amorphous C layer

Full size image
Fig. 6
figure 6

TEM analyses results of oxide layer on Ni-22Cr binary alloy corroded in S-CO2 at 600 °C (20 MPa) for 200 h. a Cross-sectional STEM micrograph and EDS mapping images and b line scan at the matrix–oxide interface

Full size image

Meanwhile, EDS mapping results (Fig. 4a) showed that NiO and Cr2O3 were separately present in the inner layer without forming mixed NiCr2O4 oxides, which is in agreement with the XRD results (Fig. 2). Similar results were previously reported for Ni–Cr binary alloys exposed in Ar-20CO2 environment, such that the NiCr2O4 spinel was absent at 650 °C [15], but present at slightly higher temperature of 700 °C [16]. EDS line scanning results showed the almost complete depletion of Cr in the matrix directly beneath the chromia layer on Ni-7Cr binary alloy (Fig. 4c), which is in agreement with the observation that the upper part of inner layer is predominantly Cr2O3 while lower part of inner layer is predominantly NiO. Interestingly, the presence of C layer was detected in the upper part of inner oxide layer where Cr2O3 was dominant (Fig. 4b), while C enrichment was not shown in the lower part of inner oxide layer where NiO was dominant (Fig. 4c). Further, the presence of dark regions throughout the internal oxide from STEM dark field mode revealed out to be a lump of non-diffracting, amorphous region within the regular oxide lattice from high-resolution bright field mode as shown in Fig. 4d. The STEM-EDS mapping in Fig. 4d showed that such regions were enriched with C encircled by inner chromia layer, similar to the amorphous C region previously observed in chromia-forming alloys [11]. The probable mechanism will be discussed in the following section.

For alloys with higher Cr content (greater than 14% Cr), a thin and continuous chromia layer was observed (Figs. 5, 6) with an average oxide thickness of around 130–150 nm. The reduced corrosion rate in chromia-forming alloys could be attributed to the formation of initial protective chromia layer, while Ni-7Cr binary alloy failed to form protective chromia due to the insufficient Cr content, resulting in the poorest corrosion resistance (Figs. 1, 2). Therefore, it could be said that the minimum Cr content for continuous chromia formation in 600 °C S-CO2 environment would be between 7 and 14%. Once Cr content in Ni-xCr binary alloys was enough for continuous chromia formation, the change in corrosion rate was rather small as in Fig. 2.

EDS line scanning results showed the depletion of Cr in the matrix directly beneath the chromia layer for Ni-14Cr binary alloy (Fig. 5b), while the degree of Cr depletion was not significant for Ni-22Cr binary alloy (Fig. 6b). The presence of Cr-depleted zone beneath the oxide layer for Ni-xCr binary alloys suggests that Cr would diffuse outwards from the matrix and be oxidized at the oxide–matrix interface [17]. Despite the severe Cr depletion below the chromia layer for Ni-14Cr binary alloy, NiO layer was not observed from STEM-EDS analyses (Fig. 5), which suggest the slow but stable growth of chromia layer could be supported by the outward diffusion of Cr from the underlying matrix. Previously, the stable growth of chromia layer despite the Cr depletion in the matrix was reported for Alloy 600 (16% Cr) exposed up to 1000 h at the same environment [11]. That is, the Ni and Cr oxides formed by the inward diffusion of oxygen, which was observed for the Ni-7Cr alloys, were not observed for alloys with more than 14% Cr. However, a recent similar study of Ni–Cr binary alloys in Ar-20CO2 environment at 650 °C showed continuous chromia was formed only when Cr content reached 30 wt%. In that case, the Ni–Cr binary alloys formed external NiO with an internal mixture of NiO and Cr2O3 layer, in which the internal layer was predominated by Cr2O3 on higher Cr content. Such distinct phenomena were observed because of their preceding electroplating procedure, which removed work hardened or deformed layer in the surface, thereby preventing enhanced diffusion of alloying elements towards the surface. Moreover, no such procedures were carried out in this study, which enabled rapid Cr-diffusion onto the surface to form continuous chromia even with the Cr content of 14 wt%.

Meanwhile, EDS line scanning results consistently showed the presence of C layer, despite the formation of continuous chromia layer for alloys with higher Cr content (greater than 14% Cr). On the other hand, C layer was not observed for pure Ni on which NiO was formed (Fig. 3). Previously, the carburized region at the interface between outer chromia layer and matrix was identified as amorphous C layer, which was confirmed by the lack of ordered atomic arrangement in high-resolution TEM image [11] as shown in Fig. 5c. The presence of amorphous C layer below the seemingly protective chromia layer has puzzled authors since the observation of such behaviours for several heat-resistant austenitic alloys [11, 12]. In the following section, authors tried to understand the mechanism behind the formation of C layer below chromia layer in S-CO2 environment.

Carburization in S-CO2 Environment

As mentioned in the previous section, carbon-enriched (10 wt%) region was observed at the oxide–matrix interface from the STEM-EDS line scanning of Ni-14Cr and Ni-22Cr binary alloys as shown in Figs. 5b and 6b. Meanwhile, a small amount C was detected for Ni-7Cr binary alloy directly beneath the chromia layer (Fig. 4b), while C was practically not observed beneath the NiO layer (Fig. 4c). SIMS analysis was performed to qualitatively check C concentration in the oxide layer and C penetration into the matrix, and the results are shown in Fig. 7. Based on SIMS results, C layer was absent beneath the oxide layer for pure Ni and very small amount of C layer was observed within the oxide layer for Ni-7Cr alloy. Meanwhile, the presence of C layer was clearly detected below the oxide layer for Ni-xCr binary alloys with more than 14% of Cr on which chromia layer was formed. Despite the presence of C layer at the oxide–matrix interface, no carbide precipitates were observed in the underlying matrix in all binary alloys, in agreement with the behaviour observed for the commercial chromia-forming Ni-based Alloy 600 and Alloy 690 [11].

Fig. 7
figure 7

SIMS depth profile of C and O from oxide layer in S-CO2 at 600 °C (20 MPa) for 200 h

Full size image

The carburization mechanism for the chromia-forming alloys could be described as follows. First, CO2 dissociates at high temperature into CO and O2 according to the following reaction:

$${\text{CO}}_{2} = {\text{ CO }} + {\text{ O}}_{2}$$
(1)

It has been reported that carbonaceous species would penetrate the oxide layer through the high-diffusion paths such as nano-channels, pores, cracks or grain boundaries [18,19,20] and dissociate to induce corrosion and carburization via the following reactions [21]:

$$x{\text{M }} + y{\text{CO}}_{2} = {\text{ M}}_{x} {\text{O}}_{y} + \, y{\text{CO,}}$$
(2)

where M is Cr or Ni; x, y = 1 for Ni; and x = 2, y = 3 for Cr. This is followed by C deposition at the oxide–matrix interface by Boudouard reaction [22]:

$$2{\text{CO }} = {\text{ CO}}_{2} + {\text{ C}}$$
(3)

However, the absence of amorphous C layer in pure Ni could not be explained by above mechanism. The probable difficulty in transport of carbonaceous species through NiO could not be the reason since a minute amount of C deposition has been observed for Ni-7Cr binary alloy having NiO in the outer layer (Fig. 4b). Therefore, other factors have to be considered in the C deposition process.

Gheno et al. [23] reported that the local C activity (a C) at the oxide–matrix interface depended directly on the local equilibrium oxygen partial pressure (pO2)eq when equilibrium condition was achieved at the oxide–matrix interface for CO2, CO and O2. The equilibrium condition for the reactions (1) and (3) can be expressed by the following equations:

$$\frac{{\text{pCO}}}{{\text{pCO}_{\text{2}} }} = \frac{{K_{\text{2}} }}{{\text{(pO}_{\text{2}} \text{)}_{{\text{eq}}}^{{\text{1/2}}} }}$$
(4)
$$a_{\text{C}} = K_{\text{4}} \frac{{\text{(pCO)}^{\text{2}} }}{{\text{pCO}_{\text{2}} }},$$
(5)

where \(K_{2}\) and \(K_{4}\) represent the equilibrium constant for reactions (1) and (3), respectively. According to above equations, the lower the (pO2)eq the higher the a C would be, which favours the deposition of C. The previous attempt to quantitatively determine the formation of C layer based on the estimation of a C using above equations was not successful for chromia as the supply of CO and CO2 to the interface was not enough to support the equilibrium [23, 24]. Nonetheless, above equations can be used to qualitatively compare the a C of different oxides such as Cr2O3 and NiO to explain the presence and absence of C layer at the interface.

The equilibrium reactions for Ni–Cr binary alloy system with O2 are given by:

$$\text{2Cr} + \frac{\text{3}}{\text{2}}\text{O}_{\text{2}} = \text{Cr}_{\text{2}} \text{O}_{\text{3}}$$
(6)
$$\text{Ni} + \frac{\text{1}}{\text{2}}\text{O}_{\text{2}} = \text{NiO}$$
(7)

The (pO2)eq at the oxide–matrix interface is equal to that of dissociation pressure of the corresponding oxide in oxide–metal equilibrium. The values were calculated from the oxide phase equilibrium diagram at 600 °C using Thermo-Calc® software [25]. It was calculated that (pO2)eq for Cr2O3 (corundum)–metal equilibrium was in the order of 10−36 atm, while for NiO (Halite)–metal equilibrium, it was in the order of 10−20 atm, significantly greater than that for chromia. Therefore, high (pO2)eq for NiO–metal equilibrium resulted in lower a C at the interface, which would make the accumulation of C difficult in pure Ni. Meanwhile, much higher a C for chromia could explain the formation of amorphous C layer in chromia-forming Ni-xCr binary alloys with more than 14% of Cr. In addition, the presence of small amount of C deposition within the oxide layer on Ni-7Cr binary alloy could be explained in view of the oxide structure. As shown in Fig. 4, the chromia was mostly embedded within the predominant NiO in the inner oxide layer and was located some distance away from the oxide–matrix interface. Therefore, due to the high (pO2)eq for the NiO + Cr2O3, C deposition was not expected. However, in some isolated locations where the chromia was in direct contact with matrix during its growth, the low enough (pO2)eq and high a C condition was achieved and C deposition happened locally. Then, due to the sparse Cr content in the matrix, discontinuous inner layer chromia was developed, which resulted in deposition of C along the irregular chromia–matrix interface during its growth. Further ingress of carbonaceous species would have oxidized adjacently situated Ni-rich matrix without any C deposition. As a result, C deposition was possible along the chromia interface for Ni-7Cr binary alloy as shown in Fig. 4b, d.

Also in a recent report, intergranular carbides observed in Ni–Cr binary alloy containing above 25 wt% Cr were attributed to higher a C arising from predominant chromia on the surface [15]. However in this study, despite the formation of amorphous C layer at the oxide–matrix interface for chromia-forming Ni-xCr binary alloys with more than 14% of Cr, further diffusion of C into the matrix and the formation of Cr-carbides were not observed by cross-sectional SEM images, which are not shown. Previously, it has been reported that the C at the interface would diffuse into the matrix and form carbides by combining with the available Cr for Fe-based alloys [9, 11, 23]. On the other hand, the underlying matrix was not affected by the presence of C layer for other Ni-based alloys like Alloy 600 and Alloy 690 [11, 12], which is in agreement with the current observation.

The carbon activity, a C, at the interface and the corresponding C diffusion into the matrix should be also considered regarding the matrix carbide precipitation. The a C required to precipitate carbides in Ni-xCr binary alloys at 600 °C was calculated by Thermo-Calc software, and the results are shown in Fig. 8. It could be said that carbide precipitation is thermodynamically possible in Ni-xCr binary alloys provided sufficient C and Cr are present in the matrix. The carbide formation is thermodynamically not favourable when the Cr content drops below the lower limit around 6% even when sufficient C is present as in the oxide–matrix interface. However, the lower C concentration in the matrix would require more Cr to form carbides, which was not the case when Cr was depleted in the matrix below the oxide layer as observed in Figs. 5 and 6. Therefore, for chromia-forming Ni-xCr binary alloys and commercial Ni-based alloys like Alloy 600, the a C at the interface of the inner chromia and matrix would be enough to deposit C there but insufficient to precipitate carbides in the underlying matrix in high-temperature S-CO2 environment.

Fig. 8
figure 8

Carbide stability diagram as the function of Cr content in Ni-xCr binary alloy at 600 °C as calculated by Thermo-Calc software

Full size image

Conclusions

Corrosion and carburization behaviour of pure Ni and Ni-xCr (x = 7, 14, 22 and 27 wt%) binary alloys in high-temperature S-CO2 environment was investigated. Specimens were exposed to S-CO2 environment at 600 °C and 20 MPa for 200 h. Based on the tests and subsequent analysis, the following conclusions were drawn:

  1. 1.

    For pure Ni, a continuous NiO layer was developed. Meanwhile, a relatively thin and continuous chromia (Cr2O3) was formed on alloys containing at least 14% Cr, resulting in better corrosion resistance. In addition, the corrosion resistance was about the same for the alloys with at least 14% Cr, whereas Ni-7Cr binary alloy showed the poorest corrosion resistance with a complex oxide structure consisted of outer uniform NiO layer and inner layer of a rather irregular Cr2O3 mixed with NiO.

  2. 2.

    Despite good corrosion resistance, the presence of C layer was detected below the chromia layer for Ni-xCr binary alloys containing at least 14% Cr, while C layer was not present for pure Ni with NiO layer. Meanwhile, for the Ni-7Cr binary alloy, amorphous C regions were present throughout the inner layer encircled with chromia.

  3. 3.

    Thermodynamic calculation was used to explain the formation of C layer below chromia layer. The low (pO2)eq for chromia–matrix interface resulted in higher carbon activity, which was in agreement with the presence of C layer below chromia layer. On the other hand, high (pO2)eq for NiO–matrix interface resulted in lower carbon activity, which explained the absence of C layer below NiO.