Corrosion Behaviour of Ni-based Alloys 230, 617 and 601 in CO₂ Gas at 750 and 850 °C

Figure 1 compares the weight gains after 150 h and 500 h reaction of all three alloys at 750 (Fig. 1a) and 850 °C (Fig. 1b). In general, weight gains were small, less than 0.18 mg cm⁻² at 750 °C, but increased rapidly at 850 °C. No scale spallation was observed after cooling.

Zoom In Zoom Out Reset image size

Three Ni-based alloys after 150 h reaction were analyzed by XRD to identify phases of oxides and the results are shown in Fig. 2. For all three Ni-based alloys, corundum Cr₂O₃ and (Ni, Fe, Cr)₃O₄ spinel were detected, together with the Ni matrix. In addition, NiO was detected on the surface of alloys 230 and 617.

Zoom In Zoom Out Reset image size

Alloy 230

At 750 °C, the 230 alloy showed an excellent oxidation resistance, forming a thin uniform protective layer (Fig. 3). The SEM analysis in Fig. 3 showed some bright contrast nodules scattered on top of the thin oxide scale. There were also very fine dark precipitates underneath the oxide scale. Further TEM/EDS analysis was carried out and the results are shown in Fig. 4. As revealed by the EDS mapping and the point analysis, this oxide scale consisted of several layers. The top layer had complex compositions with the mixture of Ni-rich oxides (points 2 and 3), and Mn-rich oxides which contain also Cr and Fe (point 1). The second layer (point 4) was a continuous thin Cr-rich oxide layer covering the whole alloy surface. Underneath this thin chromia layer, there was another thin oxide layer (third layer) containing both W and Cr (point 5). Discrete Al and Si-rich oxide precipitates were observed mainly at the scale-alloy interface.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

At 850 °C, the metallographic and SEM-BSE cross-sections of 230 alloy are shown in Fig. 5. Similar to those at 750 °C, a thin uniform protective layer with some small bright contrast protrusions was observed after 500 h reaction. However, there were some internal oxides penetrating the matrix, probably along the grain boundaries (Figs. 5a and 5b). The SEM-EDS area mapping of the oxide scale showed that a discontinuous (Ni, Fe)-rich oxides formed above a thin uniform Cr-rich oxides with a thickness of 2.7 (± 0.7) μm, andapparent Mn dissolved in this layer. Meanwhile, some W was diffusing into Cr-rich oxide layer. Apparently, Al-rich oxide precipitates were found underneath the chromia scale and penetrated locally into the matrix via the grain boundary. No clear Si enrichment was found at this interface under the limited resolution of the SEM (Fig. 5j).

Zoom In Zoom Out Reset image size

Alloy 617

The SEM cross-section image of alloy 617 after 500 h at 750 °C showed a continuous protective oxide layer, together with some small particulate precipitates underneath this oxide layer (Fig. 6). The thickness of this protective oxide layer was not uniform. Further analysis of this oxide was carried out by TEM and the results are shown in Fig. 7. The oxide scale is rather complex, containing different oxides. There was a continuous Cr-rich oxide containing Ti on the alloy surface. On the top of this layer, there were Ni-rich oxides containing some iron, forming almost a continuous layer located on the surface of Cr-rich oxide layer. No Mo was detected inside the Cr-rich layer. Discrete Al oxide particles were formed underneath the chromia layer. It is worth noting that a large amount of Co was found inside the whole nickel oxide and top part of chromia layer.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

SEM-BSE cross-section of 617 alloy after 500 h reaction in Ar-20CO₂ at 850 °C is shown in Fig. 8. Compared with those formed at 750 °C, scale morphologies grown on 617 alloy at 850 °C remains the same. A thin, discontinuous Ni and Fe-rich oxide layer was formed on the top of Cr-rich oxide. Co was also found to be present in this layer. Meanwhile, an enrichment of manganese was observed in the chromium oxide layer. However, the thickness of chromium oxide layer was significantly increased, in comparison with that at 750 °C. There was also a thin aluminum enriched sub-layer beneath the Cr-rich oxide, and small, scattered aluminum oxide particles tended to penetrate deeper into the matrix.

Zoom In Zoom Out Reset image size

Alloy 601

Figure 9 shows the SEM cross-section of 601 alloy after 500 h at 750 °C. A thin protective oxide scale with some integrated metal particles was observed. There were some internal precipitates underneath this protective scale.

Zoom In Zoom Out Reset image size

Further analysis of this sample by TEM is shown in Fig. 10. A non-uniform Cr-rich oxide layer was formed on the alloy surface (Fig. 10a). Inside this oxide scale, Ti and Mn were also present (Figs. 10f and 10h). A thin Mn rich oxide scale was found to cover the whole Cr-rich oxide layer (Fig. 10f). Very small amounts of Fe and Ni in the form of particles were observed inside the Cr-rich oxide scale (Figs. 10d and 10e). Meanwhile a thin Al-rich oxide sub-layer formed underneath the Cr-rich oxide scale and tended to penetrate deeper into the bulk alloy.

Zoom In Zoom Out Reset image size

SEM-BSE cross-section of 601 alloy after 500 h reaction at 850 °C in Ar-20CO₂ and corresponding EDS mapping results are shown in Fig. 11. Microstructures of scales grown on 601 alloy at 750 °C and 850 °C were somewhat different. The (Cr, Mn, Ti)-rich scale was 5.4 ± 1.1 μm thick. The aluminum oxide particles were larger in size than those formed at 750 °C and penetrated deeper into the matrix. There were many fine metal particles (rich in Fe and Ni) beneath the Cr-rich oxide layer (Fig. 11a).

Zoom In Zoom Out Reset image size

Comparison of alloy carburization in Ar gas with that in CO₂ gas at 750 ^oC is shown in Fig. 12. There was slightly increased carbide density along the grain boundaries after reaction in CO₂ for alloy 230 (Figs. 12a, 12b). For alloys 617 and 601 almost no newly formed carbides were observed (Figs. 12d, 12e, 12g, 12h). At 850 °C, carburization was clearly observed along grain boundaries of alloys 230 and 617 (Figs. 12c, 12f), and some fine carbide precipitates were found in alloy 601 (Fig. 12i). Some large black precipitates seen in alloy 230 microstructure (Figs. 3a–3c) could come from etching pits of this particular alloy.

Zoom In Zoom Out Reset image size

The critical concentration of Cr for the transition from internal to external chromia can be calculated according to Wagner's theory: ¹⁴

$$\begin{eqnarray}&&{N}_{{Cr}}^{\left(1\right)}={\left(\frac{\pi g}{2v}\frac{{V}_{m}}{{V}_{{Cr}{O}_{v}}}\frac{{N}_{O}^{\left(S\right)}{D}_{O}}{{\widetilde{D}}_{{Cr}}}\right)}^{\frac{1}{2}}\end{eqnarray} \tag{ 2 }$$

Here g is a critical volume fraction of oxide precipitate, generally approximated as 0.3, ¹⁵ $\upsilon$ = 1.5 is the stoichiometric coefficient for CrO_1.5, V_m = 6.7 and ${V}_{{{\rm{CrO}}}_{{\rm{\upsilon }}}}$ = 14.6 cm³ mol⁻¹ are the molar volumes of alloy and oxide, respectively, ${N}_{{\rm{O}}}^{(S)}$is the oxygen solubility (mole fraction) in the alloy in equilibrium with the oxygen activity at the scale–alloy interface, D_O the diffusion coefficient of oxygen in the nickel alloys, ¹⁶ and ${\tilde{D}}_{{\rm{Cr}}}$ is the alloy interdiffusion coefficient of Cr. ¹⁷ ${\widetilde{D}}_{{Cr}}$ was calculated from the measured tracer diffusion coefficients for Cr + 65 wt% Ni and since the three Ni-based alloys have similar Cr and Ni content, the values of ${\widetilde{D}}_{{Cr}}$ are assumed to be the same.

For dissolution of oxygen in the alloy:

$$\begin{eqnarray}&&{\rm{\unicode{x000BD}}}{{\rm{O}}}_{2}\left(g\right)=\left[{\rm{O}}\right]\left({diss}\right)\end{eqnarray} \tag{ 3 }$$

and values of ${N}_{{\rm{O}}}^{\left(S\right)}$ could be calculated by Sievert's equation:

$$\begin{eqnarray}&&{N}_{O}^{\left(S\right)}=K\sqrt{P{o}_{2}}\end{eqnarray} \tag{ 4 }$$

Here K is the equilibrium constant for Eq. 4 and available from, ¹⁸ and $P{o}_{2}$ is the partial pressure of oxygen set by the NiO–metal interfacial equilibrium

$$\begin{eqnarray}&&{\rm{Ni}}+\frac{1}{2}{{\rm{O}}}_{2}={\rm{NiO}}\end{eqnarray} \tag{ 5 }$$

the equilibrium constant for which is obtained from standard thermodynamic data. ¹⁸

Calculated values of critical Cr contents at different temperatures are summarized in Table III. The results show that Ni-based alloys (${N}_{{Cr}}=0.26$ for alloy 230, ${N}_{{Cr}}=0.22$ for alloy 617, ${N}_{{Cr}}=0.25$ for alloy 601) are marginal to form a protective chromia oxide layer at 750 °C but all alloys would form the chromia at 850 °C. All three alloys developed a protective chromia oxide layer and small weight gain was observed after a long exposure even at 750 °C. This observation could be attributed to additional protection due to other element effects, e.g. the addition of Al, Mn and Si which will be discussed in the next section.

Table III. Calculated critical chromium contents at 750 °C and 850 °C by Wagner's theory and the parameters for the calculation.

T (°C)	${N}_{O}^{S}$	${D}_{o}$(cm2 s−1)	${\widetilde{D}}_{{Cr}}$ (cm2 s−1)	${N}_{{Cr}}^{\left(1\right)}$
$750$	$1.35\times {10}^{-4}$	$2.1\times {10}^{-10}$	$6.04\times {10}^{-14}$	0.26
850	$2.32\times {10}^{-4}$	$1.15\times {10}^{-9}$	$7.87\times {10}^{-13}$	0.22

It should be mentioned that above oxygen permeability calculation considers the equilibrium condition of Ni/NiO which is the dominant situation at the early stage of the reaction. However, this situation could be replaced later when NiCr-spinel is formed, leading to the reduced oxygen permeability and therefore a lowered ${N}_{{Cr}}^{\left(1\right)}.$ In practical condition, the formation of these Cr-rich oxides could cause the depletion of Cr underneath the Cr-rich oxide layer, offsetting the beneficial effect of the reduced ${N}_{{Cr}}^{\left(1\right)}.$ By this consideration, above ${N}_{{Cr}}^{\left(1\right)}$calculation based on Ni/NiO equilibrium is the up limit of Cr concentration required for the transition from the internal to external oxidation of chromia.

Furthermore, the second critical level of chromium ${N}_{{Cr}}^{\left(2\right)}$ was applied to determine whether there was sufficient chromium concentration to maintain chromia oxide layer growth: ¹⁹

$$\begin{eqnarray}&&{N}_{{Cr}}^{\left(2\right)}=\frac{{V}_{m}}{{V}_{{\left({CrO}\right)}_{v}}}{\left(\frac{\pi {k}_{p}}{2{\widetilde{D}}_{{Cr}}}\right)}^{\frac{1}{2}}\end{eqnarray} \tag{ 6 }$$

Here ${k}_{p}$ is the parabolic rate constant for chromia layer thickening, which could be shown as:

$$\begin{eqnarray}&&{X}^{2}=2{k}_{p}t\end{eqnarray} \tag{ 7 }$$

where $X$ represents the scale thickness of chromia layer and $t$ is time. The calculated values of ${N}_{{Cr}}^{\left(2\right)}$ in Table IV indicated that only 617 alloy satisfies the condition for maintaining the growth of chromia scale at 750 °C, but the other two are marginal. However, all alloys at 850 °C satisfy the condition for maintaining the growth of chromia scale, which is consistent with the experimental results. As the commercial alloys contain several other alloy elements, these alloy elements could provide additional protection.

Table IV. The second critical chromium contents at 750 °C and 850 °C.

T (°C)		230	617	601
750	${k}_{p},$ cm2 s−1	$1.5\times {10}^{-14}$	$8.6\times {10}^{-15}$	$1.3\times {10}^{-14}$
	${N}_{{Cr}}^{\left(2\right)}$	0.29	0.22	0.27
850	${k}_{p},$ cm2 s−1	$2.4\times {10}^{-14}$	$7.7\times {10}^{-14}$	$1.2\times {10}^{-13}$
	${{N}}_{{Cr}}^{\left(2\right)}$	0.10	0.18	0.22

Effect of Al

As discussed above, although the calculated results at 750 °C indicated that for all three alloys, only alloy 617 meets the conditions that can form a protective chromium oxide layer. However, the experimental results indicate the formation of a protective chromia layer with a small weight gain for all three alloys, showing a good oxidation resistance. This result could be related to the formation of Al oxides at the interface between the chromia and the matrix. According to Abbasi et al., ²⁰ the addition of Al up to 2.4 wt% leads to reducing weight gains, and the formation of Al oxides along the boundary under chromia layer is responsible for the improved oxidation behavior. The diffusion of Cr and O ions is restricted by Al₂O₃, which reduces the thickness of Cr₂O₃ layer.

Wagner's theory is applied to calculate the critical level of Al for the formation of the protective Al₂O₃ layer on Ni-base alloys. The calculated results are shown in Table V, where all parameters for the calculation were obtained from Gust et al. ²¹ The solubility of oxygen was calculated based on the $p{{\rm{O}}}_{2}\,$at the equilibrium of Cr/Cr₂O₃ at the reaction temperatures. The calculated results show that an Al₂O₃ layer could form for both 601 (${N}_{{Al}}=0.03$) and 617 (${N}_{{Al}}=0.02$) but could not form for 230 (${N}_{{Al}}=0.007$) at 750 °C, and all alloys could form an Al₂O₃ layer at 850 °C. This is basically consistent with the experimental results as shown by the formation of thin alumina layer, e.g. for alloy 617 (Fig. 8g). At 750 °C, the alloy 617 seemed to form densely packed Al₂O₃ precipitates (Fig. 7g), but relatively less densely packed Al₂O₃ particles were observed in the alloy 230 (Fig. 4h), confirming an increased tendency for alumina formation with the increasing temperature predicted by the calculation (Table V). For alloy 601 with relatively higher Al concentration, an obvious Al₂O₃ layer was formed beneath a chromia layer even at 750 °C (Fig. 10g). At 850 °C, large chromia precipitates penetrated deeply into the substrate, probably along the grain boundaries (Fig. 11g).

Table V. Calculated critical aluminum contents at 750 °C and 850 °C by Wagner's theory and the parameters for the calculation.

T (°C)	${V}_{{Al}{O}_{1.5}}$	${N}_{O}^{S}$	${D}_{o},$ cm2 s−1	$\widetilde{D}$ Al, cm2 s−1	${N}_{{Al}}^{\left(1\right)}$
750	12.9	$3.5\times {10}^{-7}$	$2.1\times {10}^{-10}$	$5.3\times {10}^{-14}$	$0.015$
850	12.9	$9.4\times {10}^{-7}$	$1.15\times {10}^{-9}$	$2.9\times {10}^{-12}$	$0.007$

Effect of Si

Effect of Mn

Douglass and Armijo ²⁷ reported that even though MnO is more stable than chromium oxide, manganese has a higher tendency to form a continuous and dense manganese-rich oxide spinel with lower oxygen activity by the following equation:

$$\begin{eqnarray}&&{\rm{MnO}}+{{\rm{Cr}}}_{2}{{\rm{O}}}_{3}={{\rm{MnCr}}}_{2}{{\rm{O}}}_{4}\end{eqnarray} \tag{ 8 }$$

Furthermore, Naoumidis ²⁸ proposed that diffusion coefficient of manganese in chromium oxide (${D}_{{Mn}}=1.6\times {10}^{-17}$ cm² s⁻¹ at 750 °C and ${D}_{{Mn}}=9.7\times {10}^{-17}$ cm² s⁻¹ at 850 °C) is faster than that in alloys (${D}_{{Mn}}=1.1\times {10}^{-18}$ at 750 °C and ${D}_{{Mn}}=1.2\times {10}^{-17}$ cm² s⁻¹ at 850 °C). ²⁹ This process can be expressed as follows:

$$\begin{eqnarray}&&2{\rm{Cr}}+\frac{3}{2}{{\rm{O}}}_{2}={{\rm{Cr}}}_{2}{{\rm{O}}}_{3}\end{eqnarray} \tag{ 9 }$$

$$\begin{eqnarray}&&{\rm{Mn}}+{{\rm{Cr}}}_{2}{{\rm{O}}}_{3}+\frac{1}{2}{{\rm{O}}}_{2}={{\rm{MnCr}}}_{2}{{\rm{O}}}_{4}\end{eqnarray} \tag{ 10 }$$

Due to the rapid diffusion of Mn through the chromia lattice, Mn-rich oxide formed in the outermost scale layer on the top of Cr-rich oxide layer for both alloys 230 and 601 (Figs. 4f, 5g, 10f, 11f) with relatively high Mn concentrations (0.5 and 0.61 wt%, respectively). The 617 alloy contains very little Mn content (0.04 wt%) which is not enough for Mn-rich oxide on the top of Cr-rich oxide layer. However, diffusion of manganese through the chromia layer was observed with increasing temperature to 850 °C (Fig. 8f). The low-valent Mn in the manganese-rich oxide layer will occupy the electron vacancies in the chromium oxide and play a role in reducing the concentration of defects, thereby improving the oxidation resistance.

In general, all three Ni-based alloys have a good oxidation resistance, forming a protective chromia scale. For alloys 230 and 617, Ni-rich oxide containing some iron was observed on top of the Cr-rich oxide layer (Figs. 4, 5, 7 and 8). This situation could be attributed to the rapid diffusion of Ni and Fe ions through the chromia layer. An explanation was proposed by Yi et al. ³⁰ that the outward migration of Ni is due to the release of stress by the addition of Al and Si during their internal oxidation process. Two different contrasts of oxide morphologies could be well observed by SEM in Figs. 5b and 8a. The bright contrast outermost layer was Ni-rich oxide and the dark contrast layer next to the matrix was Cr-rich oxide. From the XRD results in Fig. 2, Ni-rich oxide layer could be explained as consisting of a small amount of NiO particles on the outmost surface and (Fe, Ni, Cr)₃O₄ spinels underneath. It has been reported ³¹ that for this process, Cr₂O₃ forms first in the early stage and Ni and Fe ions diffuse along the grain boundary to form spinel and then NiO grows on the top of (Fe, Ni, Cr)₃O₄ spinels. For alloy 617, some Co diffuses outside, co-existing with the upper part of Cr-rich scale (Fig. 8h). This observation is explained by the fact that Co has much higher mobility than that of Ni ions in the oxide lattice which leads to the higher Co concentration near the scale-gas interface (${D}_{{Co}}\left(750\,\unicode{x02103}\right)=1.3\times {10}^{-14},$ ${D}_{{Ni}}\left(750\unicode{x02103}\right)=1\times {10}^{-17},$ ${D}_{{Co}}\left(850\,\unicode{x02103}\right)=1.2\times {10}^{-13},$ ${D}_{{Ni}}\left(850\,\unicode{x02103}\right)=6.5\times {10}^{-17}$ cm² s⁻¹). ^32–34

It is noted that alloy 230 contains up to 14% W. Some studies have reported that the addition of W reduces the oxidation resistance of Ni-base alloys. ^35,36 Because W is a high valance element, when these metal ions were mixed with Cr₂O₃, the metal vacancies increased and then the diffusion rate of other metals through the chromia layer increased. ³⁷ As a high valence element with a low diffusion rate, W could also retard the formation of Cr₂O₃ because of low oxygen activity. However, the large size of W could slow down the diffusion of other elements including the Al₂O₃ formation. ³⁸ Therefore, the effect of W on oxidation is not clear in this research. The experimental results (Fig. 4g) show (W, Cr)-rich oxides in the matrix because of the inward diffusion of oxygen and no outward diffusion of tungsten. Furthermore, because the 230 alloy contains low content of Al, it is not clear if W slowed the formation of Al₂O₃. However, it was observed that more Ni, Fe and Mn diffused into the outermost layer of the chromia layer.

Unlike 230 and 617, alloy 601 performs differently. Only a small amount of nickel and iron diffusion was observed, and no NiO formed. This is because the addition of a higher amount of chromium and aluminum increases its oxidation resistance. ³⁹ Furthermore, the outward diffusion of Ti and Mn was evident in the experiment results (Figs. 10h, 10f, 11h and 11f). The increase in temperature is conducive to the formation of titanium oxide at the position of the external interface but less inside the chromia layer. Lenglet et al. ⁴⁰ have explained that Ti oxides formed at the external interface proved that Ti diffuses through the chromia scale. Meanwhile, Naoumidis et al. ²⁸ explained that the high concentration of Ti dissolved in chromium oxide indicates that lattice diffusion could be responsible for the growth of Ti-rich oxides in the outer layer. Some authors ^41,42 also proposed that titanium promotes the formation of additional cation vacancies in the chromium oxide scale when Ti⁴⁺ ions replace Cr³⁺ ions. Since the matrix bulk cation diffusion and grain boundary control the growth of the chromium oxide layer, the increase of cation vacancies in the chromia scale will result in faster chromium diffusion and higher kinetic rates. ⁴³

Figure 13 schematically shows these element effects by classifying them into two categories, in general. The first category is able to form sublayered oxides (e.g. Al and Si). The second is able to integrate with chromia and to form even an additional spinel layer on the top of chromia, when the concentration is high. All these additional oxide formations increase the corrosion resistance of the chromia-forming alloys as discussed above.

Zoom In Zoom Out Reset image size

A comparison of oxide thickness at different temperatures in Table VI reveals that as the temperature increased, the scale thickness also increased. So, the increase in oxidation rate was observed with increasing temperature. Only the alloy 601 was observed with the formation of a thin Al₂O₃ layer in the IOZ at 750 °C, but a significant amount of aluminum oxide was observed in the internal oxidation zone at 850 °C. The huge volume expansion due to internal oxide formation generates high stress in the lattice of the alloy matrix. The stress relief process led to the outward transport of the alloy from the IOZ. ⁴⁴ Therefore, the alloy particles were observed at the interface between the IOZ and the external oxide layer at 850 °C for all Ni-base alloys (Figs. 3b, 5b, 6, 8a, 9, and 11a). The morphology of most of the oxides was similar at both temperatures. Even though the diffusion of some elements such as Mn, Ti, Al was not observed by SEM but can be revealed by TEM analysis at a higher magnification. It is worth noting that no outward diffusion of Mn was observed in alloy 617 at 750 °C even under TEM observation. However, as the oxidation rate increases, Mn diffuses outward through the chromia layer. Therefore, diffusion of Mn into the chromia layer was observed at 850 °C.

Except primary carbides from raw alloys, almost no newly formed carbides were found after 500 h exposure in Ar-20CO₂ at 750 °C for both 617 and 601, but some carbide formation was observed along the grain boundary of alloy 230. At 850 °C, the newly formed carbides were found in all three Ni-base alloys after the reaction. According to ThermoCalc calculation, the volume fraction of primary carbides is expected to be decreased from 750 °C to 850 °C. Therefore, the increased carburization observed at 850 °C should be contributed by the CO₂ gas.

In addition, other alloying elements also affect the diffusion of carbon. The beneficial effect of Al on decreasing carburization was reported by Searcy et al. ⁴⁵ The high binding energy of Al₂O₃ results in a low rate of atomic mobility within it. So Al₂O₃ could be accounted as a good barrier to hinder carbon diffusion into the matrix. In this work, the formation of Al₂O₃ was observed in all three alloys, resulting in a reduction in carburising kinetics and the density of carbide precipitates. Meanwhile, the additional formation of silicon and manganese oxides decreases the diffusion of carbon. ^46,47 In addition, the protective effect of alumina, silica, chromia and manganese oxide is attributed to their very low solubility of carbon, ⁴⁸ and therefore low carbon permeability in these oxides. As a result, they provide a barrier effect to resist carbon penetration.