Effect of the Noble Metals Addition on the Oxidation Behavior of Ni 3

Facultad de Quı́mica, Universidad Nacional Autónoma de Mexico, Edificio B, Cd. Universitaria, 04510 Ciudad de Mexico, Mexico CIICAp, Universidad Autónoma del Estado de Morelos, Avenida Universidad 1001, 62209 Cuernavaca, MOR, Mexico CIATEQ A.C., Eje 126 No. 225, Zona Industrial, 78395 San Luis Potośı, SLP, Mexico Instituto de Investigaciones en Materiales, Universidad Nacional Autónoma de Mexico, Cd. Universitaria, 04510 Ciudad de Mexico, Mexico Instituto Nacional de Electricidad y Enerǵıas Limpias, Avenida Reforma 113, Colonia Palmira, 62490 Cuernavaca, MOR, Mexico Instituto de Ciencias Fı́sicas, Universidad Nacional Autónoma de Mexico, Avenida Universidad s/n, 62210 Cuernavaca, MOR, Mexico Corrosion y Protección (CyP), Buffon 46, 11590 Mexico City, Mexico


Introduction
Ni 3 Al intermetallic compounds are excellent candidates for high-temperature structural applications due to their thermochemical and mechanical properties.ey can perform beyond conventional temperatures because of their excellent resistance to corrosion and oxidation [1].During solidification of an alloy with stoichiometry close to that of Ni 3 Al, there is in the liquid phase a competition for phase stability between β NiAl , γ′ Ni3Al , and γ Ni with stable eutectic (β-γ′), metastable (β-γ), and peritectic (γ′-γ) phases which have common borders in the Ni-Al binary diagram.Because the region of phase stability γ′ is very narrow, any deviation in the Ni/Al ratio favors the stability of any of the aforementioned phases; that is, an increase in the Ni content tends to favor the stability of γ′-γ phase.According to Verin [2], the microstructure of the Ni-23Al-0.001Ce-0.004Co(at %) is composed of a combination of dendritic phases γ′ and γ, where the matrix is the phase γ′ and the interdendritic phase is the phase γ; similar results were obtained by Chiba and Hanada [3] for the alloy Ni-23Al (at %) and Jozwik et al. [4] for the alloy Ni-22Al-0.002Zr-0.001B(at %).Moreover, Hunziker and Kurz [5] mentioned that the directional solidification of intermetallic Ni-25Al (at %) shows a microstructure where the phases β, γ, and γ′ are found.In this case, the γ′ phase is formed during cooling of metastable eutectic β-γ.It has been shown that there are subsequent transformations after the solidification process, which complicates the interpretation of the phases produced.
On the other hand, the solidification rate is a factor that influences the microstructure of the alloy.When the solidification is slow, the convection in the liquid phase favors formation of a single γ′-phase.However, in a rapid solidification process, there is little convection in the liquid phase and it promotes the formation of a structure with phases γ′' and β [6].e addition of a third element to the Ni-Al system may stabilize or destabilize a given phase.According to Inoue et al. [7], the addition of Cr, Mn, Fe, Co, or Si destabilizes the phase γ′, and according to Goetzinger et al. [8], the addition of Ti stabilizes the phase.Sun et al. [9] indicate that Pd, Ag, and Cu are elements forming the γ phase, while Pint et al. [10] show that Pt promotes the reduction of the temperature of the phase transformation from β to γ′ during the process of cooling and phase transformation from γ′ to β during heating.
During the oxidation process, β and γ′ phases play an important role due to their content of Al [11].According to Haerig and Hofmann [12], in the early stages of the oxidation process, NiO and Al 2 O 3 are formed simultaneously due to the diffusion of cations towards the outside of the alloy.Bardi et al. [13] show that the growth of the oxide is in the form of islands.Al consumption owing to the formation of Al 2 O 3 diminishes the aluminum content below 17% (at %), which is the content needed to form the outer protective layer [14].
is favors the growth of NiO onto the Al 2 O 3 layer.is pattern occurs because of the conducting properties of the formed oxides: Al 2 O 3 is an amphoteric oxide (movement of cations and anions is possible) allowing outward diffusion of Ni + cations.Al 2 O 3 forms internally due to inward diffusion of O 2− anions, while NiO is a p-type conductor, which preferably promotes conduction of cations [15][16][17].Al 2 O 3 growth onto alloys such as Ni 3 Al can adopt different crystallographic arrangements: between 427 and 927 °C, the γ-Al 2 O 3 phase is stable; until 1077 °C, the θ-Al 2 O 3 phase is stable; and above 1077 °C, the α-Al 2 O 3 phase is stable [18].e θ → α transformation is accompanied by a change in the transport properties of the scale.e θ-Al 2 O 3 scale mainly grows by outward diffusion of cations, forming cavities in the scale/alloy interface, resulting in the deterioration of the adhesion layer on the substrate.In contrast, the growth of α scale is dependent on the diffusion of the O 2− anion inwards, so that the oxide fills the preformed cavities.is improves the adhesion of the layer on the substrate.erefore, it is desirable for the growth of an α-Al 2 O 3 layer [1,14].It has been observed that the only growth of an outer scale of Al 2 O 3 is possible under the following conditions, by increasing both the temperature and concentration of Al [12], where the oxidation process is carried out at oxygen partial pressures below the equilibrium of NiO formation [13,[19][20][21] and also in alloys which were implanted with B + and Cr + ions [14].Increasing the oxygen partial pressure and temperature, it is possible to form the spinel NiAl 2 O 4 at the NiO/Al 2 O 3 interface [22].
Garza et al. [23] attributed this behavior to the difference between the free energies of formation of Al 2 O 3 and NiO which is four times larger than that between Al 2 O 3 and NiAl 2 O 4 .Velon and Olefjord [15] mentioned that the spinel could also be formed in the initial stages of oxidation at low temperatures (300 °C).
A key aspect to increase the oxidation resistance of an alloy is to improve the adherence of the protective oxide mainly because most applications implicate high-temperature thermal cycles [13,24].
e most effective method is the addition of elements or compounds such as Hf, Y, Zr, Cr, Pd, Pt, and SiC [25][26][27][28][29][30][31], which can change the localized concentration of aluminum and reduce its growth rate.Gleeson et al. [32] indicate that the addition of Pt has beneficial effects in promoting the formation of a stable Al 2 O 3 scale.Pt is an unreactive element whose addition decreases the chemical activity of Al in the γ′ phase, and it increases the Al/Ni ratio, which promotes the formation of Al 2 O 3 in certain crystallographic planes and increases the Al diffusion into the formed scale.It has been calculated that the preferential site of substitution of noble metals like Ag, Au, Pd, and Pt in the Ni 3 Al structure occurs in locations where there are normally nickel atoms [33][34][35][36][37]; therefore, the addition of these elements can modify the oxidation behavior of the alloy.
erefore, this paper evaluates the effect of the addition of these elements in low concentration (1 at.%) on the oxidation behavior of the Ni 3 Al intermetallic alloy.

Experimental Procedure
e intermetallic alloys with nominal composition Ni 75 Al 25 and Ni 75 Al 24 M 1 (where M � Ag, Au, Pd, and Pt) were manufactured in quartz crucibles in an induction furnace under inert atmosphere using elements with purity higher than 99.99%.e ingots were prepared by the fusion of stoichiometric amounts of the high-purity elements (solid rods), and the melt was homogenized by the convective currents generated by the induction furnace.After melting the metal elements, the ingots were allowed to cool to room temperature inside the induction furnace.From the ingots, samples were cut with dimensions of 20 × 5 × 2 mm to carry out the tests.
XRD analysis of alloys was carried with a Bruker D8 Advance diffractometer equipped with a Cu Kα (k � 1.5418 Å; 40 kV) monochromatic source and a LYNXEYE TM 1D positional detector.Scans were performed from 2θ � 5 °to 100 °with a 0.02 °step and an irradiation time of 0.2 seconds.In order to reveal the microstructure of the intermetallic alloys, they were etched by immersing in HF-HNO 3 solution (1 : 1 molar ratio) for 5 to 10 seconds at room temperature.
For the oxidation test, samples were grinded with a silicon carbide paper to 600-grit size, washed with water, and ultrasonically cleaned with acetone.In order to have a larger reaction area, two samples were employed in each test: one was used for surface analysis and the other for crosssectional analysis.
Oxidation tests were carried out in pure oxygen at 900, 1000, and 1100 °C.
is range was chosen since at these temperatures, there is a change in the stability of the different

XRD Analysis of Alloys.
e XRD profiles of the alloys with and without noble metal additions are shown in Figure 1.
e peaks observed in the XRD profiles of the doped alloys are wide, suggesting a small crystallite size.In addition, the intensities of some peaks increase and decrease depending on the added noble metal.
e Ni 75 Al 24 Ag 1 profile shows highest intensities in (100) and (200), and the intensity of the peak decreased in (111); this suggests the crystal growth in two preferential planes.is preferential growth was also observed in the Ni 75 Al 24 Pd 1 alloy, where the profile shows a slight increase in the (222) peak, indicating a preferential unidirectional growth [41].e XRD profile for the Ni 75 Al 24 Au 1 alloy shows a notable increase in the intensity of the (311) peak and a slight increase in the intensity of the (200) peak, and the characteristic peak (111) shows no change.On the other hand, the XRD profile of the Ni 75 Al 24 Pt 1 alloy shows an increase in the intensity of all its peaks, and an unidentified peak at 2θ � 29.56 is observed.Considering that the positions of the atoms in the unit cell affect the intensities but not the directions of the diffracted beams [42], the change in peak intensities can be completely attributable to the addition of noble metals.
Properties such as hardness show a significant dependence on crystallographic planes: for Ni 3 Al, the highest value occurs in (100), followed by ( 111) and (110) [43].However, several studies about the Ni 3 Al oxidation indicate that there are no significant differences in the oxidation behavior between planes (100), (110), and (111) [44][45][46][47][48]. Taking into account that the planes (100) and (110) consist of a series of two layers: the most stable having a composition of 50% Ni and 50% Al and the other with 100% Ni and the plane (111) having a composition of 75% Ni-25% Al [49,50]; then due to it, it is possible the presence of zones with different chemical compositions onto alloy surface (i.e., change in the Al/Ni ratio).In addition, substitution at preferential sites by the addition of noble metals can further change the surface chemical composition, thereby greatly influencing its oxidation behavior.e EDXA showed that the dendrite is a phase with high concentration of Al, so that the (γ-γ′) phase is in the matrix and the (β-γ′) phase is in the dendrite core as indicated by Hunziker and Kurz [51] and Rosas et al. [52], respectively.During the solidification process, there is no cooling rate sufficiently slow to maintain equilibrium in the melt; that is, there is insufficient time for the composition gradients to be diffusion removed and the dendrites retained [53].Its formation also may be  Advances in Materials Science and Engineering attributable to the difference between the melting points of the constituent phases.According to Lee et al. [11,17], when the solidification process begins, it is favored by the formation of the dendritic structures (β-γ′), and during the cooling process, the γ-phase decreases while the γ′-phase increases.

Microstructural Analysis.
Figure 3 shows the microstructure of intermetallic alloys evaluated, and it is observed that there are notable changes in 4 Advances in Materials Science and Engineering the microstructure of the doped alloys, which can be explained according to the preferential growth plane of each alloy (Figure 1) and an increase in the solidification rate.e dendrite structure of Ni 75 Al 25 (Figure 3(a)) has a spacing between secondary arms of 47 μm, and this value is typical in a conventional solidification process [54].e space between the secondary arms of the dendrites is inversely related to the cooling rate during the solidification process, which has been correlated according to the following equation [55]: where d 2 is the secondary dendrite arm spacing, v is the solidification cooling rate, and k is a constant of the alloy.Since the solidification cooling rate is inversely proportional to the space between the secondary arms of the dendrites, from Table 1, it can be noticed that the solidification cooling rate in the Ni 75 Al 25 alloy, so that there is a lower space between the secondary dendrite arms; moreover, several directions of dendrites in the Ni 75 Al 24 Pt 1 alloy are congruent with its preferential growth plane.
is same affirmation is consistent with the Ni 75 Al 24 Pd 1 alloy with almost only one preferential growth plane.
e solidification cooling rate in the Ni 75 Al 24 Au 1 alloy was slower than the undoped alloy such that the dendrites almost disappeared.
It is observed that the addition of Ag favors an increase in the ratio γ-γ′/β due to the reduction in the number and volume of the interdendritic phase (Figure 3(b)).Because of the low solubility of Ag in Ni [56], it is possible to observe the presence of Ag-rich phases.According to Zhou and Guo [57,58], the presence of Ag at grain boundaries begins during the solidification process since this element segregates at the solid-liquid interface, and at the same time, the Ag solubility decreases causing their precipitation into the grains.e microstructure of the Au-doped intermetallic alloy showed a honeycomb appearance (Figure 3(c)); that is, there is not a tendency to form precipitates as in the case of Ag-doped intermetallic alloy.Sheng et al. [59] found that the effect of addition of 1% Au on the microstructure of the NiAl intermetallic alloy causes a grain size reduction and the formation of α-AlAu 2 precipitates.It can be assumed that the absence of the α-AlAu 2 phase is due to the lower content of Al in Ni 3 Al compared to that in NiAl.Figures 3(d) and 3(e) show the microstructure of the intermetallic alloy doped with Pd and Pt, respectively.e Pd-doped alloy has dendritic structures of bigger area; however, the relationship γ-γ′/β is much greater than in the alloy doped with Ag and Au, indicating that, at room temperature, the γ-γ′ phase is stable.e Pt-doped alloy has very fine dendritic formations in which there is no enrichment of Al.In both cases, a uniform distribution of Pd and Pt is observed as in the case of the Au-doped intermetallic alloy.Meng et al. [60] indicate that the addition of Pd into Ni 3 Al intermetallic alloy generates an increase of delocalized bonding electrons, and this causes a strengthening of the grain boundaries and improves the ductility of the alloy.Furthermore, Li et al. [61,62], Oquab and Monceau [63], and Zhang et al. [64] mention that Pd plays an important role in stabilizing the β-phase in NiAl at high temperatures; the same effect is attributed to the addition of Pt [10].No precipitates were observed in the interdendritic region of the alloys, indicating that the phases found there are results of a eutectic reaction, contrasting with the work of Lee et al. [11], which presents cell interdendritic microstructures that Hunziker and Kurz [5] defined as peritectic structures.4 shows the curves (solid lines) of weight gain per unit area as a function of exposure time during the process of oxidation of the intermetallic alloys between 900 and 1100 °C.

Oxidation Kinetics. Figure
e curves were fitted (dashed lines) to an equation of the form [24,65]: where Δm is the weight gain per unit area, k is the oxidation rate constant, n is the exponent of velocity, and t is the time of exposure.e values of k and n to fit the experimental results are shown in Table 2. Adjustments were done using a nonlinear least-squares regression method (Gauss-Newton method) [66].According to the n values obtained, it can be observed that only at 900 °C, the oxidation kinetics of the Ni 3 Al alloy and the Ag-doped alloy follows a parabolic behavior, all other n values correspond to a subparabolic behavior.e oxidation mechanism of a protective oxide that grows with parabolic kinetics mainly consists of the diffusion of cations from the metal-oxide interface to the oxide-gas interface; however, when the oxidation mechanism approaches to cubic kinetics, some authors attribute this behavior to the oxide grain growth [67,68], although other authors claim that the formation of a protective layer like α-Al 2 O 3 is more a rule than an exception [69].An additional characteristic of the subparabolic kinetics is that at the beginning of the oxidation process, a great weight gain is observed with respect to that observed in a parabolic kinetic, being this a characteristic of the transient stage of the oxidation process.At 900 °C, Ag-and Au-doped alloys had the highest weight gain throughout the exposure period with respect to the undoped alloy, whereas Pd-and Pt-doped alloys had the lower weight gain.During first two hours, the Audoped alloy has a significantly greater weight gain, while Ag-, Pd-, and Pt-doped alloys have a similar and slightly higher weight gain than the undoped alloy.At 1000 °C, all doped alloys (except Pd-doped alloy) have the greater   6 Advances in Materials Science and Engineering terms of weight gain between the Pd-and Pt-doped alloys (lowest) and the Ag-and Au-doped alloys (higher) compared to undoped alloy, similar as at 900 °C.e Pt-doped alloy has the lower weight gain than the undoped alloy throughout the exposure period, while the Pd-doped alloy has a higher weight gain than the undoped alloy only in the first six hours.All alloys except Pd-doped alloy have a remarkably high weight gain in the first hour.e weight gain curves of Figure 4 give an impression about the effect of the addition of noble metals in the oxidation process, but it is also necessary to know the rate of oxidation.Considering strictly a parabolic behavior n � 0.5, for comparative purposes (2) can be rewritten as where k p is the parabolic rate constant.e linearity of (3) allows graphically displaying and estimating the k p value considering the slope of the Δm 2 − t relationship.In Figure 5, changes in the parabolic rate constant showing two k p values can be identified, the first k p value being greater than the second one; this is due to the fact that it is mathematically forcing a curve with a subparabolic behavior to be adjusted to a parabolic behavior [69], so the k p value that will be taken into account to characterize the oxidation process will be that of the second slope which is presented in Table 3.
At 900 °C, undoped and Ag-doped alloys show a constant k p value throughout the exposure period.e k p value of Pdand Pt-doped alloys is similar, and it is lower than that for the undoped alloy.At 1000 °C, the k p value of the Au-doped  Advances in Materials Science and Engineering          Au-doped alloy decreases at three hours after starting the test, and for Ag-, Pd-, and Pt-doped alloys, the k p value decreases at four hours after starting the test, while for the undoped alloy, its k p value remains constant.At 1100 °C, the k p value of all alloys (except Pt) decreases after 30 min, and for Pt-doped alloy, the k p value remains constant.When it is not possible to obtain accurate measurement from the slopes, it is necessary to consider the average of the instantaneous parabolic constants, which results in differentiating (3) versus time [65], e instantaneous parabolic constant (k p ) of the oxidation tests is shown in Figure 6. is type of representation of the experimental data provides more information regarding the evolution of oxidation rates.At 900 °C (Figure 6(a)), in the first two hours, the k p values of the Au-doped alloy decrease seven times its equilibrium value, while in the Pd-and Pt-doped alloys, the k p values decrease two times their equilibrium value; on the other hand, the Ag-doped alloy and undoped alloy maintain a constant rate.At 1000 °C (Figure 6(b)), it is observed that the values of the instantaneous parabolic constant k p of doped alloys tend to decrease in the first 6 hours, and then they remain constant, and for the undoped alloy, the k p values remain constant.For the Ag-doped alloy, its k p value decreases six times its equilibrium value, and for other doped-alloys, their k p values decrease two times their equilibrium value.Finally, at 1100 °C (Figure 6(c)), the Pt-doped alloy shows a stable behavior, and the other alloys show a fast decrease in their k p values in the first hour.
In order to visualize the effect of the temperature on the rate constants, it is necessary to determine the activation energy of the oxidation process from the Arrhenius equation.
where Q is the activation energy, R is the universal gas constant, T is the temperature, k p is the rate constant, and k 0 is the preexponential factor.Plotting ln k p versus −(RT) −1 from the slope of the lines, the Q value can be determined.Perez et al. [65] suggest assuming a parabolic kinetics in the whole temperature range; therefore, the average value of instantaneous k p from Figure 6 was calculated.Table 4 shows the average values of instantaneous k p .In order to compare the results of this work, Figure 7 shows the instantaneous k p values obtained along with those reported in other studies.At 900 and 1000 °C, the k p values of both doped and undoped alloys tend to k p values where the formation of the θ-Al 2 O 3 phase may be possible in NiAl as reported by Brumm and Grabke [72].At 1100 °C, the k p values are concentrated at a point between the formation of 12 Advances in Materials Science and Engineering θ-Al 2 O 3 and α-Al 2 O 3 ; an explanation for this behavior could be that, at this temperature, the stability of the α-Al 2 O 3 begins [18] and this tends to decrease the weight gain [70,71].At 900 °C and 1000 °C, the k p value of the undoped alloy is similar to results reported by Lee and Kim [26] and Choi et al. [73], and at 1100 °C, it is similar to results of Kuenzly et al. [74].e Pt-doped alloy has k p values similar to the undoped alloy, Ag-and Au-doped alloys have higher k p values, and the Pd-doped alloy has lower k p values.Activation energy is an important parameter that characterizes the predominant oxidation mechanism [75]; for example, the results presented by Ochiai et al. [75] and Kuenzly et al. [74] suggest that, in both cases, the oxidation mechanism is similar, although the experimental conditions were different, that is, cyclic oxidation tests and continuous oxidation tests.On the other hand, analyzing the results for the formation of α-Al 2 O 3 scale [72] and NiO scale [76], it is observed that the preexponential factor is inversely proportional to the protective capacity of the protective oxide.Table 5 shows the activation energies and preexponential factors obtained from the Arrhenius graph (Figure 7).e activation energy value of Ni 3 Al is similar to those values reported by Lee and Kim [26] and Cao et al. [77], but almost three times lower than that reported by other authors [73,74]. is difference is because the intermetallic alloy used in this study was not thermally treated.Activation energy values of the doped alloys show that the additions of the noble elements affect the Ni 3 Al oxidation mechanism in the following order: Pt > Au > Ag > Pd.On the other hand, the Pd-doped alloy showed the lowest value of the preexponential factor.
3.4.SEM Analysis.SEM analysis showed that the oxide scale formed on all alloys had nonuniform thicknesses.In the case of Ni 3 Al, the average thicknesses observed were between 2.44 and 2.86 μm in the temperature range studied.At 900 °C (Figure 8), a double-layer oxide was developed, an outer Ni-rich layer and an internal Al-rich layer, with similar thicknesses.Other authors have described similar results [78].However, at higher temperatures, only the presence of an Al-rich oxide was detected.
In general, the thickness of the oxide layer developed onto doped alloys was lower than that observed onto Ni 3 Al.At 900 °C, the average thicknesses were between 0.98 and 1.5 μm, at 1000 °C, between 2.13 and 2.20 μm (except the Pt-doped alloy, its average thickness was 4.61 μm), and at 1100 °C, between 1.43 and 2.55 μm (except the Ag-doped alloy, its average thickness was 4.52 μm). Figure 9 shows the thickness of the oxide formed on the intermetallic alloys oxidized to 900 °C; similar aspects were observed at higher temperatures.At all test temperatures, the Pd-doped alloy showed the oxide layer with the lowest thickness (Figure 10(a)).In general, all doped alloys developed an Alrich monolayer oxide (Figures 10(a)-10(c)), as has been reported in other studies [79].In the case of alloys doped with Pt (at 1000 °C) and with Ag (at 1100 °C), zones with multilayer oxides and thicknesses up to 20 μm were observed (Figures 10(d) and 10(g)).In all cases, an Al-depleted zone was observed at the interface alloy oxide (Figures 8 and 10).
is is due to the consumption of this element for the formation of protective oxide, Al 2 O 3 [80].Similar depletion processes have been reported in studies of corrosion by molten salts [81][82][83][84][85].
Figure 11 shows the surface morphology of the oxides formed at 900 °C.e formation of oxides in the form of needles can be observed only onto nondoped alloy and onto those alloys doped with Ag and Au. is type of morphology is associated with the presence of the metastable phases of alumina [86][87][88], whose transformation from γ-Al 2 O 3 to θ-Al 2 O 3 occurs between 800 and 950 °C [72].It is worth mentioning that the alloys that presented this type of morphology had a greater weight gain (Figure 4(a)).
Figure 12 shows the surface morphology of the oxide scale at 1000 °C.It is observed that, at this temperature, there is a greater presence of oxides in the form of needles (θ-Al 2 O 3 ) both in the nondoped alloy and in those alloys doped with Ag, Au, and Pt. e alloy doped with Pd showed a lower presence of this phase, possibly due to a decrease in the growth rate of the θ-Al 2 O 3 phase.Again, the surface morphology is correlated with the observed weight gain (Figure 4(b)); that is, the greater the presence of the θ-Al 2 O 3 phase, the greater the mass gain.
On the other hand, the morphology of the oxides formed at 1100 °C was completely different (Figure 13).In this case, the oxides formed show wrinkle shaped, which is characteristic of the α-Al 2 O 3 phase [72]. is morphology is more evident in the nondoped alloy and in those doped with Ag and Au, and in the alloys doped with Pd and Pt, a more compact and dense morphology was observed.ese morphological differences indicate a difference in the growth rates of the oxide layers.In all cases, the weight gain at 1100 °C was lower than that observed at 1000 °C (Figure 4). is suggests that oxide layers rich in α-Al 2 O 3 are more protective than those rich in θ-Al 2 O 3 .Figure 14 shows the typical morphology of the Ni 3 Al oxidized in oxygen for 24 hours.In general, a granular scale involving both heterogeneous Ni and Al oxides, mainly Ni    In some regions, the oxide layer has been cracked, probably during cooling, and it was possible to observe the presence of cavities into the scalealloy interface (Figure 15).ese sorts of cavities are formed by the consumption of Al and Ni in the course of formation of the oxide scale, and they have an impact on the adhesion of the scale [70].
e difference in the oxidation process between the dendritic and the interdendritic phases lies in the thermochemistry of alumina formation between the phases; that is, in the NiAl-rich dendritic phase, the following reaction may occur: 4NiAl + 3O 2 ⇆ 2Al 2 O 3 + 4Ni ΔG � −940.52 kJ @ 900 ∘ C ( ) On the other hand, in the interdendritic phase where the Ni3Al phase predominates, the following reaction occurs: ermochemically, the dendritic phase is more favored than the interdendritic phase.Considering also that in the doped alloys there is a considerable decrease in the dendritic phase (Section 3.2), it can be established that there exists a relation between the microstructural change and the thermochemical equilibrium conditions for the formation of an alumina layer.On the other hand, the absence of NiO in the surface of the doped alloys can also be attributed to the increase in the grain size of the Al 2 O 3 ; that is, to say, a decrease in the diffusion of Ni through the grain boundaries of the Al 2 O 3 layer, which is related to the kinetics of subparabolic growth.

Conclusions
e XRD analysis of the alloys showed that the addition of 1% (atomic percent) of noble metals to the Ni 3 Al intermetallic alloy promotes the preferential growth planes and increases the solidification rate by changing the microstructure of the base alloy.
e oxidation behavior of doped alloys may be attributable to the change in the surface composition, taking into account that every growth plane has different compositions and that the preferential substitution site of the noble metals in the structure of Ni

Figure 2
shows the microstructure of the intermetallic Ni 75Al 25  alloy, where the presence of two phases is observed.

Figure 1 :
Figure 1: e XRD patterns of Ni 3 Al and Ni 3 Al doped with 1 at.% of noble metals.

8
Advances in Materials Science and Engineering alloy decreases at three hours after starting the test, the k p value of Ag, Pd, and Pt doped alloys decreases at four hours after starting the test, while for the undoped alloy, its k p value remains constant.At 1100 °C, the k p value of the undoped and doped alloys decrease at the first half an hour after starting the test, except the Pt-doped alloy, for which the k p value remains constant.In most alloys, at all temperatures, it is observed at most only one change in k p , except the Pd-doped alloy at 1100 °C with two changes: half an hour and two hours after starting the test.At 1000 °C, the k p value of the 00011 −0.000105 −0.0001 −0.000095 −0.00009 −0.000085 −0.00008 −0.000075 −0.

Figure 7 :
Figure 7: Arrhenius graph showing the k p variation in function of temperature.

Figure 8 :
Figure 8: SEM micrograph (a) and mapping elements (b, c) of the scale formed onto Ni 3 Al at 900 °C.

3
Al increases the Al/Ni ratio.In general, the addition of noble metals promotes the formation of oxide layers rich in Al 2 O 3 and is evident that the diffusion of the Ni cations through Al 2 O 3 scale is decreased, since the formation of NiO during the test time was not evident.Ag and Au additions increase the value of the parabolic rate constant k p , while Pd and Pt decrease, the k p values tend to values where the formation of θ-Al 2 O 3 and α-Al 2 O 3 is possible.e noble metal addition increased the activation energy value of the oxidation process compared to the nondoped Ni 3 Al, except Pd. e low solubility of Ag and its low melting point caused microstructural localized failures into Ag-doped alloy.Financial support from Consejo Nacional de Ciencia y Tecnología (CONACYT, México) (Project no.159898 and Ph.D. scholarship to J. J. Ramos-Hernandez, Registration no.202898) is gratefully acknowledged.

Figure 15 :
Figure 15: Cavities formed due to diffusion of Al during the oxidation process of the Ni 3 Al in oxygen at (a) 900 °C, (b) 1000 °C, and (c) 1100 °C.

Table 1 :
Spaces between secondary arms and space ratio, taking as reference the undoped alloy.
[71]nces in Materials Science and Engineering weight gain with respect to the undoped alloy, in a decreasing order of Au-, Ag-, and Pt-doped alloys. is behavior may be because the doped alloys stabilize the θ-Al 2 O 3 phase and therefore reduce the α-Al 2 O 3 phase transformation as Pt does according to Cadoret et al.[70]and Hayashi and Gleeson[71].e θ-Al 2 O 3 phase generates more pathways for diffusion because of its smaller grain size compared to α-Al 2 O 3 phase.e k values at 1000 °C show magnitudes that explain the high increase in weight gain for doped alloys.At 1100 °C, there is a clear grouping in

Table 2 :
Values of k and n obtained from the curves of weight gain.

Table 3 :
Values of k p obtained from slopes of Figure5.

Table 4 :
Average values of instantaneous k p from Figure6.

Table 5 :
Values of k 0 and Q obtained from the Arrhenius graph.
Advances in Materials Science and Engineering