Temperature Effect on Oxidation Behavior of Ni-Cr Alloys in CO₂ Gas Atmosphere

Oxidation kinetics determined by weight gain measurements are shown in Figure 1. In general, these weight gains can be approximated by parabolic kinetics, expressed by the equation:

where ΔW/A is the weight gain per unit surface area of the tested specimen, k_w is the parabolic rate constant, and t is the oxidation time. Table II summarizes the parabolic rate constants for the various alloys at different temperatures, found by linear regression on Eq. 1. Kinetic data at 800°C for Ni-20Cr was too irregular to treat this way. The slow kinetics observed for the 25 and 30 wt% Cr alloys were approximated as parabolic, with rate constants found by regression.

Zoom In Zoom Out Reset image size

It is obvious that temperature exerted a significant effect on the oxidation kinetics of these alloys in Ar-CO₂ gas mixtures. Generally, weight gain rates increased with increasing temperature for alloys with Cr concentrations ranging from 5 to 15 wt%. At 650°C, the weight gain rates showed no significant change when Cr concentrations increased from 5 to 25 wt% (Fig. 1 and Table II). Further increasing Cr to 30 wt% led to an apparent rate decrease. Similar effects are observed also at 700 and 800°C, but the critical chromium content for the occurrence of this significant rate decrease changed to lower values: 25 wt% at 700°C and 20 wt% at 800°C (Fig. 1 and Table II). Further increasing alloy chromium content above the critical chromium concentration led to a further rate decrease (Table II).

Figure 2 shows the surface XRD patterns obtained from samples exposed for 500 h at 650°C in Ar-20%CO₂. The oxide formed on the surface of Ni-5Cr alloy was mainly NiO, and Cr₂O₃ peaks were not identified in this case. For the alloys containing 10 to 25 wt% chromium, Cr₂O₃ peaks were detectable but NiO was still the predominant phase, as shown by the peak intensities. As the chromium content of the alloys increased to 30 wt%, the Cr₂O₃ peak intensities increased, although NiO was still present. It is also noteworthy that the alloy substrate was clearly detected for Ni-30Cr, indicating the formation of a much thinner oxide scale. This decrease in the thickness of the oxide scale corresponds with the slowest oxidation rate observed for this alloy at 650°C, as reflected in Fig. 1 and Table II.

Zoom In Zoom Out Reset image size

BSE-SEM cross-section images of the tested alloys after reaction for 500 h in Ar-20%CO₂ at 650°C are shown in Figs. 3a–3f. It can be seen that all the alloys developed multi-layered structures on their surfaces during oxidation. For alloys containing 15 wt% Cr or less, an inner oxide layer was observed between an external NiO layer and an internal oxidation zone (IOZ). The inner oxide layer on Ni-5Cr was relatively thin, and developed a nonplanar interface with the IOZ. Its IOZ was characterized by many dendritic shaped Cr-rich precipitates penetrating into the alloy matrix to a certain depth. This morphology has been reported previously for Ni-Cr alloys reacted in air and oxygen.^20,21 In contrast, the IOZ of Ni-10Cr and Ni-15Cr alloys contained very fine oxide precipitates. The inner layer formed on Ni-10Cr was continuous and comparable to the outer layer in thickness, but that on Ni-15Cr was discontinuous and disrupted by IOZ regions. Quantitative EDS analysis results for the scales formed on Ni-15Cr are shown in Table III. It is apparent that the outer layer contained essentially Ni and O, the inner layer a high O concentration, and the IOZ a high Ni concentration.

Zoom In Zoom Out Reset image size

In order to identify the phases present in the respective layers, the sample surface was subjected to successive grinding and XRD analysis until the alloy substrate was reached. The XRD patterns presented in Fig. 4 were collected during this process, and their corresponding locations in the scale are marked approximately in Fig. 3c. Clearly the scale consisted of an outer NiO layer, an inner NiO and Cr₂O₃ mixed layer, and an internally oxidized zone containing Cr₂O₃ precipitates located below the scale. No NiCr₂O₄ spinel was found at this temperature. Because the alloys of Ni-5Cr and Ni-10Cr developed Cr-rich inner layers similar to that of Ni-15Cr, it is supposed that their inner layers were of the same phase constitution.

Zoom In Zoom Out Reset image size

Unlike the cases of the alloys described above, the inner oxide layer nearly disappeared when the Cr content of the alloy was increased to 20 wt%, and only very small scattered islands are occasionally seen to connect with pores close to the original alloy surface (Fig. 3d). When the Cr concentration was further increased to 25 wt%, the inner oxide layer disappeared completely, and a thin, semi-continuous band was formed at the IOZ precipitation front. Figure 3g reveals a higher Cr concentration in this band, implying that it consisted of Cr-rich oxides. The IOZ depth was less at some locations where the band had formed than at other places where it was absent (Fig. 3e). This Cr-rich oxide band became continuous, uniform and of an appreciable thickness when the alloy contained 30 wt% Cr. The EDS line profiles shown in Fig. 3h confirmed this band to be Cr enriched, and the corresponding EDS quantitative analysis results indicated that this band had an atomic ratio of Cr to O of 1:1.42. Therefore, it is reasonably inferred that the oxide band formed at the reaction zone base was Cr₂O₃.

To study the detailed process of chromia band formation and development, BSE-SEM cross-sections of Ni-25Cr and Ni-30Cr alloys were produced after different exposure times (Figs. 5 and 6, respectively). After 50 h reaction, two layers were formed on the surface of Ni-25Cr, with a thin external NiO layer surmounting an internal oxidation zone (Fig. 5a). With increased reaction time, this two-layered structure remained unchanged, but the thickness of both layers increased. In addition, small regions of oxide band had formed locally at the IOZ-alloy interface after 150 h reaction (Fig. 5b). The coverage of these bands increased with the reaction time and reached about 57% of the whole interface after 500 h reaction, as shown in Fig. 5d and Table IV. The thickness of the two major layers as a function of the square root of time is shown in Fig. 5e. The thickening kinetics of both layers approximately obeyed the parabolic law:

where X represents layer thickness.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

For Ni-30Cr alloy, chromia band regions of appreciable length were formed at the reaction zone base, even when the oxidation time was as short as 50 h (Fig. 6a). The band became essentially continuous and covered the whole IOZ-alloy interface when the exposure time increased to 150 h (Fig. 6b), and the average IOZ thickness increased during this period. With further increases in experimental time, the continuous chromia band thickened (Figs. 6c and 6d). With the development of this continuous band at the reaction front, the IOZ on Ni-30Cr ceased to grow (Fig. 6e). It is noted that the IOZ shown in Figs. 6c and 6d appears darker than those shown in Figs. 6a and 6b, indicating a lower average atomic number under BSE image mode. Corresponding EDS analysis did reveal that the oxygen concentration of IOZ increased from about 15 to 25 wt% as the reaction time increased from 150 to 310 h. Although the measurement is only semi-quantitative, the large increase in oxygen uptake is attributed to the reaction of Ni in the former IOZ to NiO. The higher oxygen potential required for this reaction comes about because the underlying continuous chromia band acts as a barrier to inward oxygen diffusion.

The surface XRD patterns obtained at 700°C (not shown here) confirmed that the oxide phases on different alloys were similar to their counterparts at 650°C. BSE-SEM cross-sections of the alloys reacted for 500 h in Ar-20%CO₂ at 700°C are shown in Fig. 7. Compared with those formed at 650°C, the morphologies and microstructures of scales grown on alloys of Ni-5Cr and Ni-10Cr at 700°C were essentially the same, except that the density of acicular precipitates in the IOZ of Ni-5Cr increased and many globular and short rod-shaped precipitates appeared at the precipitation front. The inner oxide layer on Ni-15Cr became more continuous at 700°C, and also the corresponding layer on Ni-20Cr accounted for a larger volume fraction near the outer-inner layer interface than that at 650°C. At 700°C, the innermost Cr-rich oxide band formed by Ni-25Cr, which was discontinuous at 650°C, became continuous and uniform. In the case of Ni-30Cr, an obvious chromia subscale was developed underneath the outer NiO layer. These observations are consistent with earlier results obtained at 700°C.¹⁸

Zoom In Zoom Out Reset image size

The evolution of oxidation products grown on Ni-25Cr and Ni-30Cr alloys is shown in Figs. 8 and 9, respectively. The same process of chromia band growth on Ni-25Cr observed at 650°C (Fig. 5) was found to take place at an accelerated rate at 700°C. Discrete chromia patches were formed at the base of the IOZ after 50 h exposure (Fig. 8a), then thin, presumably continuous and uniform, completely continuous chromia bands were developed after 150 and 310 h, respectively (Figs. 8b and 8c). The process of forming a continuous chromia band was much faster at 700°C, as shown by the area fraction measurements in Table IV. As with the IOZ of Ni-30Cr formed at 650°C, the IOZ developed by Ni-25Cr after exposure for 500 h (Fig. 8d) was darker than that after 150 h exposure (Fig. 8b), with a roughly 10 wt% increase in oxygen concentration detected. The growth kinetics of the NiO layer and IOZ are shown in Fig. 8e. Their thickening kinetics were approximately parabolic up to 310 h, but slowed greatly afterwards.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

In the case of Ni-30Cr, a continuous chromia subscale had formed by the earliest stage observed at 700°C (Fig. 9a) and no extensive IOZ developed. This subscale thickened gradually with increasing reaction time. It should be noted that some isolated internal oxidation regions were found enclosed by the chromia subscale (Figs. 9a–9c). The growth kinetics of the NiO layer and chromia subscale are shown in Fig. 9e, both roughly following the parabolic law, but much slower than those shown in Figs. 5e, 6e and 8e.

BSE-SEM cross-sections of the model alloys after reaction for 500 h in Ar-20%CO₂ at 800°C are shown in Fig. 10. Scale morphologies developed by alloys containing 5 to 15 wt% Cr were somewhat different from those formed at 700°C. For Ni-5Cr, the interface between inner oxide layer and IOZ became more flat. The precipitates in its IOZ failed to develop the initial dendritic shape seen at lower temperatures, instead forming short rods. In the case of both Ni-10Cr and Ni-15Cr alloys, the oxide particles precipitated in the IOZ were larger in size than those formed at 650 and 700°C. A similar trend of increasing precipitate size with increasing depth beneath the inner oxide layer-IOZ interface was observed for these two alloys. In contrast, the microstructure of the oxide scales on Ni-20Cr changed substantially. No apparent IOZ was visible, but a continuous chromia scale formed instead, with some discontinuous NiO regions above and small scattered Cr₂O₃ particles underneath. Owing to the non-uniformity of the scale, the weight gain curve for Ni-20Cr in Fig. 1c was irregular, deviating from the parabolic law to some extent.

Zoom In Zoom Out Reset image size

The evolution with time of scale morphologies developed by the Ni-20Cr alloy at 800°C is shown in Fig. 11. As shown in Figs. 11a–11c, Ni-20Cr grew a thin chromia scale over most of its surface. This was occasionally interrupted by the growth of localized nodules consisting of outer NiO and underlying internal oxidation islands, surrounded by the chromia scale.

Zoom In Zoom Out Reset image size

At higher Cr levels, uniform Cr₂O₃ layers, identified by EDS analysis and XRD results (not shown here), were formed on both Ni-25Cr and Ni-30Cr after all the reaction times observed. The thickness of their chromia scales increased very slowly with reaction time.

Intergranular carbides were revealed by etching of high chromium content alloys, Ni-25Cr and Ni-30Cr reacted in CO₂ gas for 500 h at all test temperatures. Typical microstructures in Fig. 12 show the carbides to be scattered along alloy grain boundaries. Increasing the Cr content seems to increase the size and amount of carbides (Fig. 12). Analysis by SEM-EDS confirmed these precipitates to be rich in chromium and carbon. Carbides were also detected for Ni-20Cr after reaction at 800°C for 500 h but at a much lower volume fraction and in a shallow subsurface zone. No apparent carbides were observed in the lower Cr concentration alloys reacted under the same condition.

Zoom In Zoom Out Reset image size

As shown in Fig. 3, all alloys studied failed to form an exclusive chromia scale on their surfaces at 650°C. However, alloys of Ni-25Cr and Ni-30Cr did form protective scales, exclusively of Cr₂O₃, after exposure in the same atmosphere at 800°C, showing that the critical chromium concentration for external chromia scale formation was influenced by temperature. According to Wagner's theory, this critical concentration of chromium, N⁽¹⁾_Cr, required to form external rather than internal oxide can be estimated from the equation:²⁴

where g is a critical volume fraction of oxide precipitates, generally approximated as about 0.3,²⁵ v the stoichiometric coefficient of CrO_v (1.5), V_m the molar volume of alloy (6.8 cm³/mol), the molar volume of the oxide (14.6 cm³/mol), D_O the diffusion coefficient of oxygen in the alloy, and the alloy inter-diffusion coefficient. Sample surface treatment by electropolishing, as employed in this work, allows the evaluation of Eq. 3, using well characterized bulk alloy diffusion coefficients. Without this treatment, which removes the cold worked surface region, subsurface diffusion is predominantly due to the high dislocation density in the deformed metal and grain boundary diffusion in any subsequently recrystallized zone. In addition to being difficult to quantify, this enhanced diffusion is not characteristic of an alloy after long term service at high temperature, as the damage is by then annealed out.

Eq. 3 was derived on the assumption that no external scale of, in this case NiO, forms. This is dealt with here by calculating N^(S)_O for the Ni/NiO equilibrium (see below). Further inaccuracy may result from neglect of the volume change induced by internal oxidation, any contribution to diffusion by precipitate-matrix interfaces, and any interaction with dissolved carbon. The volume change effect is relatively small, and the dissolved carbon effect has been shown to be negligible at high temperatures.²⁶ The contribution of interfacial diffusion cannot be assessed without high resolution characterization of precipitate morphology. As this information is to date unavailable, it is noted that the predictions of Eq. 3 are imprecise, but nonetheless of semi-quantitative use.

The quantityN^(S)_O in Eq. 3 is the dissolved oxygen concentration at the alloy surface, determined by the partial pressure of oxygen at the scale-alloy interface, as calculated from Sievert's equation:

where K₅ is the equilibrium constant for the dissolution of oxygen in the alloy:

The value of K₅ can be calculated from data provided in the literature.²⁷ Using a value for determined by the equilibrium:

the value of N^(S)_O can be obtained. Diffusivity data of oxygen and chromium are unavailable at such low temperatures, so values are extrapolated from high temperature data in literature.^27,28

Once a chromia scale is formed, the flux of Cr diffusing from the bulk alloy to the scale-alloy interface must be sufficient to maintain its growth. The critical concentration of chromium, N⁽²⁾_Cr, associated with this criterion can be expressed as:²⁹

where k_p is the parabolic rate constant for chromia layer thickening, which can be derived from Eq. 2. Unfortunately, no value of k_p at 650°C can be deduced from the scale thickening kinetic data, as no exclusive chromia scales ever formed on any alloys throughout the oxidation times in the present study. As a result, N⁽²⁾_Cr values were not calculated at 650°C. At 700°C, in addition to the predominant scale morphology shown in Fig. 7f, some localized areas of exclusive chromia scale, of about 1.5 μm thickness were also observed on the Ni-30Cr alloy surface after oxidation for 500 h. This corresponds to k_p≈ 6.3 × 10⁻¹⁵ cm²/s. Similarly, for k_pcalculation at 800°C, the roughly 2.5 μm thick chromia scale shown in Fig. 10f leads to the estimate k_p≈ 1.7 × 10⁻¹⁴ cm²/s.

Calculated values of N⁽¹⁾_Cr and N⁽²⁾_Cr are summarized in Table V, along with the underlying data. It is seen that both N⁽¹⁾_Cr and N⁽²⁾_Cr are predicted to vary inversely with oxidation temperature. According to the predictions, at 650°C all alloys except Ni-30Cr in the present study should be unable to form an external chromia scale. The chromium content (32.6 at % Cr) of Ni-30Cr alloy is slightly higher than N⁽¹⁾_Cr, so marginal behavior could be expected. This was in fact evidenced by a continuous Cr₂O₃ band formed at the oxide scale base (Fig. 3f). Similarly, at 700°C, the chromium concentration of Ni-25Cr (27.3 at % Cr) is marginal for chromia formation but not sufficient to maintain it. As a result, only a thin but continuous Cr₂O₃ band is formed at the base of the reaction zone (Fig. 7e). At this temperature, Ni-30Cr alloy is predicted to develop a chromia layer on the surface, and not internally, but still fail to form an exclusive layer. This is in agreement with the experimental observation (Fig. 7f). Since Ni-25Cr and Ni-30Cr alloys are able to form continuous chromia layers, lower mass uptake rates as manifested by the parabolic rate constants shown in Table II are explained.

Table V. Critical concentrations of N⁽¹⁾_Cr and N⁽²⁾_Cr.

T (°C)	N(S)O	DO (cm2/s)	(cm2/s)	kp (cm2/s)	N(1)Cr	N(2)Cr
650	6.6E-05	2.6E-11	2.6E-15	–	0.311	–
700	9.4E-05	7.7E-11	1.4E-14	6.3E-15	0.275	0.392
800	1.7E-04	5.1E-10	2.3E-13	1.7E-14	0.235	0.159

At 800°C, the chromium content (22 at % Cr) of Ni-20Cr alloy is close to N⁽¹⁾_Cr and higher than N⁽²⁾_Cr. As a result, this alloy should exhibit marginal behavior in terms of chromia formation but succeed in maintaining the chromia layer once it emerges. These inferences are supported by the finding of predominant chromia layer formation, accompanied with localized NiO islands (Fig. 10d). The significantly reduced weight gain rates shown in Fig. 1c also reflected this success. For alloys of Ni-25Cr and Ni-30Cr, the observed exclusive chromia layers are in good agreement with the predicted results (Figs. 10e and 10f). It is concluded that predictions based on Wagner's diffusion theory agree rather well with the experimental observations.

It is to be noted that this agreement is found only for electropolished alloy specimens, with no deformed subsurface regions. In the more common case of alloys with cold worked surfaces, accelerated alloy diffusion in the deformed subsurface region leads via Eq. 3 and Eq. 7 to lower critical chromium concentrations, and the more ready formation of protective chromia scales.

Careful etching experiments revealed carbide formation for high Cr content alloys. A preliminary investigation³⁰ failed to detect the carbides, and is in that respect incorrect. For carbide precipitation in Ni-Cr alloys, the reaction is written:

the free energy change can be obtained from the reactions:

where underlining indicates a solute species. Thus, the equilibrium condition for carbide precipitation is written as:

in which the activity of carbide, a_CrCv, is assumed to be unity, and ΔG°, and the standard free energy changes for the formation of CrC_v and the partial molar free energies of dissolution of Cr and C in nickel, respectively.^31,32 The minimum carbon activity for Cr₂₃C₆ precipitation is calculated for Ni-30Cr, with a_Cr = 0.46 extrapolated from higher temperature data,³³ to be a_C = 1.2 × 10⁻⁴ at 800°C. Therefore, to form this carbide, a_C should be higher than this value.

As can be seen from Table I, carbon activities of the equilibrium gas phase are very low, far lower than the carbon activities required for carbide formation. However, as reported for iron-based chromium-containing alloys,^1,2 the CO₂ gas can produce much higher carbon activity at the interface between iron oxide and metal matrix, because of the significantly lower oxygen partial pressure at this location which raises the carbon activity. The correlation between the oxygen partial pressure and carbon activity is seen in the reaction equilibria:

Combination of Eqs. 16 and 14 yields:

The value of is unknown, but has a maximum equal to that of the gas phase.

At the scale/metal interface, the equilibrium is much lower than that in the gas, and as a result, a high carbon activity can be achieved, according to Eq. 17. In the case of an external chromia scale formed on the Ni-30Cr alloy oxidized at 800°C, the at the scale-alloy interface is equal to 1.2 × 10⁻²⁷ atm. For = 0.2 atm, a_C underneath the chromia scale is calculated to be about 9.7 × 10⁶ which is more than enough for carbide formation. Even at much reduced values, carbon activities will be higher than the calculated threshold a_C value for Cr₂₃C₆ formation in Ni-30Cr. The observation of carbide formation only at grain boundaries indicates that a_C does not in fact reach any such value within grains. The reason is the external oxide scale which greatly reduces inward carbon diffusion, resulting in a much lower and a_C values in the underlying alloy. As a result, no intragranular carbide is formed inside the matrix. Only at grain boundaries, where carbon diffusion is fast and carbide nucleation is facilitated, does carbide precipitation occur.

If an internal oxidation zone beneath an external NiO scale prevails, e.g. the oxidation products grown on Ni-25Cr alloy oxidized at 650 and 700°C, the at the scale-alloy interface should be governed by the dissociation pressure of NiO which is calculated to be 4.5 × 10⁻¹⁷ atm at 700°C. Correspondingly, the maximum value of interfacial a_C = 2.8 × 10⁻⁶ is found from Eq. 17. On the other hand, the critical a_C value required for Cr₂₃C₆ precipitation is 1.1 × 10⁻³ at 700°C according to Eq. 12, and carbide formation is not predicted for Ni-25Cr on this basis. However, the above calculation is based on the assumption that the NiO scale is pure. Actually, it has been reported³⁴ that there is an appreciable solubility of Cr in NiO, about 1 wt% at 950°C. This stabilizes the oxide scale, lowering the equilibrium value. It follows from Eq. 17 that the corresponding a_C should be much higher than 2.8 × 10⁻⁶. On this basis it is inferred that carbon activity beneath the scale is enhanced by the dissolution of Cr into NiO, leading to carbide formation.

As discussed above, a protective chromia scale can be formed when the alloy Cr content is higher than both N⁽¹⁾_CrandN⁽²⁾_Cr. When the concentration is far below these critical values, multi-layered oxide scales are formed, typically in the form of external NiO, inner oxide layer, and IOZ (Figs. 3, 7 and 10). As the alloy Cr concentration is increased to these critical values, a transition in reaction product morphology from a multi-layered structure to an exclusive chromia scale is expected. This is confirmed by the observation of chromia band formation at the reaction front (Figs. 5, 6 and 8). The evolution of this chromia band is shown schematically in Fig. 13. In a short reaction time, an external NiO layer together with an IOZ is expected to be formed (Fig. 13a), since the Cr concentration is slightly lower than the critical value for avoiding internal oxidation. With the gradual thickening of IOZ and external NiO layer, the chromium concentration at the reaction front reaches the critical value in some regions and as a result Cr₂O₃ band nucleation occurs (Fig. 13b). The local variations in microstructure leading to chromia band nucleation in some places but not in others are not revealed by the present results.

Zoom In Zoom Out Reset image size

As soon as a continuous Cr₂O₃ band is formed, both outward diffusion of Cr into the IOZ and inward oxygen diffusion through the Cr₂O₃ band are strongly retarded. As a result, oxygen potentials in the IOZ increase and become high enough to further oxidize Ni in the IOZ, converting it to an inner mixed oxide layer (Fig. 13c). This transformation is completed when the Ni oxidation front reaches the chromia band (Fig. 13d). The phenomenon of Cr₂O₃ band formation is apparent at temperatures of 650°C and 700°C, but not at 800°C where, in accord with Wagner's prediction, a protective Cr₂O₃ is the sole reaction product.

The formation of a continuous chromia band coincides with the reduction in weight gain rate shown in Figs. 1a and 1b, indicating the transition from non-protective to protective oxidation. This change is associated with a reduction in thickening rates of the external NiO layer and IOZ, as seen by comparing Figs. 5e, 6e and 8e. When the chromia band formed at the IOZ-alloy interface was discontinuous, it had an insignificant effect on the scale thickening kinetics (Fig. 5). However, when a continuous chromia band was formed (Figs. 6 and 8), the growth rates of both outer NiO layer and IOZ were reduced, but not fully stopped, with the extent of reduction more pronounced for the IOZ (Figs. 6e and 8e). This observation implies that the Cr₂O₃ is permeable to nickel, which is somewhat surprising as the diffusion coefficient of nickel in the Cr₂O₃ is very small, 3.3 × 10⁻¹⁸ cm²/s at 700°C,³⁵ unable to maintain the growth rate shown in Fig. 8e. Other possible sources of nickel for external NiO layer thickening are the sparsely distributed Ni particles at the NiO layer-IOZ interface (Fig. 8), and diffusion of Ni from the matrix of the IOZ as it is oxidized.