Effect of Chromium Addition on the Cyclic Oxidation Resistance of Pseudo-Binary (Mo,Cr)3 Si Silicide Alloy

I. Rosales; D. Ponce; MJ. Garcia-Ramirez; R. Guardian

doi:10.1515/htmp-2017-0099

Artikel Open Access

Effect of Chromium Addition on the Cyclic Oxidation Resistance of Pseudo-Binary (Mo,Cr)₃ Si Silicide Alloy

I. Rosales , D. Ponce , MJ. Garcia-Ramirez und R. Guardian

Veröffentlicht/Copyright: 13. Juni 2018

Veröffentlicht von

Veröffentlichen auch Sie bei De Gruyter Brill

Manuskript einreichen Informationen für Autor*innen

Aus der Zeitschrift High Temperature Materials and Processes Band 37 Heft 9-10

Abstract

This paper describes the performance under cyclic oxidation of (Mo,Cr)₃ Si compound with different Cr additions for possible coating application. Cyclic oxidation was carried out at 1,000 °C at different intervals during 250 h. Oxidized surface samples were analysed by scanning electron microscope where epitaxial oxide scales were observed mainly in samples with higher Cr content which may provide protection against surface oxidation. X-ray diffraction studies have shown the Cr₂O₃ and SiO₂ formation as the main oxide scale; after analyses, it was found that these oxides are responsible for the best oxidation protection, with 36 at.% Cr being the optimal chromium concentration. At lower chromium concentrations, pest reaction occurred in the oxidized samples at times less than 25 h as a result of the formation of the unstable molybdenum oxide.

Keywords: intermetallics compounds; SEMX-ray diffraction; oxidation

Introduction

High-temperature-resistant materials based on silicides are interesting due to their excellent behaviour in applications under extreme conditions such as high-temperature oxidation and corrosive environments [1, 2, 3, 4, 5, 6, 7, 8]. Several factors need to be considered to develop a material performance analysis when the metal is oxidized and the most important are the interface formed between the oxide scale and the metal base as well as the scale stability to prevent surface depletion. Shah et al. [9] investigated the Nb₃Al and Cr₃Si compounds, evaluating their mechanical properties at high temperature, reporting that Cr₃Si presents an excellent behaviour, but with limited possibilities of being alloyed. Raj [10] evaluated the mechanical properties at high temperature of Cr₃Si and Cr-Cr₃Si, produced by arc melting/casting and powder metallurgy, reporting that oxidized alloys developed Cr₂O₃ protective scales. An important requirement to be considered as high-temperature materials is its resistance to degradation in hot corrosive environments. There are different ways in which materials are deteriorated, such as oxidation, corrosion and hydrogen embrittlement, but the high-temperature oxidation plays the most important role in surface deterioration [10]. In our previous work, we have studied the effect of aluminium addition in Mo₃Si with interesting results [11] due to the alumina formation as protective oxide. Other authors [12, 13] have studied the oxidation performance in Cr₃Si single crystals and Mo-Si-B where the proposed mechanism was the formation of a protective oxide scale. Therefore, the purpose of the present work is to provide the knowledge concerning the optimal chromium concentration in (Mo,Cr)₃Si, evaluating the high-temperature oxidation resistance of the pseudo-binary compound and to elucidate the protection mechanisms for possible coating applications, specifically in engine burners and gas turbine blades.

Experimental procedure

Selected alloys with fixed nominal silicon concentration of 24 at.% and different Cr and Mo concentrations were prepared by using an arc-melting AMCO (AMCO furnace chamber) furnace with pure elements (99.98 %) under a partial argon pressure (99.999 % purity). The alloys were drop-cast into water-cooled copper molds with a diameter of 12.5 mm. Table 1 shows the alloy designation with the nominal compositions, considering the weight losses during melting and casting due to evaporation of molybdenum, silicon and chromium. Samples were annealed at 1,400 °C for 24 h under vacuum (10^–3 torrs).

Table 1:

Alloy designation and elemental composition for the alloys used in the oxidation experiments. Silicon was fixed in 24 at.% to keep the single-phase composition alloy.

Sample designation	Elemental content at.%
M1	Mo-76; Si-24 ———
M2	Mo-59; Si-24; Cr-17
M3	Mo-40; Si-24; Cr-36
M4	Mo-16; Si-24; Cr-60
M5	——– Si-24; Cr-76

Oxidation tests were carried out at 1,000 °C for a total time of 250 h. Samples with dimensions of 1×1×1 cm were placed inside a furnace chamber for specific times of 5, 25, 45, 80, 150 and 250 h. After been weighted, specimens were re-introduced inside the furnace in order to continue with the process, until the oxidation time was completed. Chemical and morphological analyses of the oxidized samples were performed in a Leo VP-1450 scanning electron microscope equipped with an energy-dispersive X-ray spectroscopy system (SEM/EDS), and line-scan analyses were carried out to characterize the oxide layers. X-ray diffraction (XRD) was carried out in a Siemens D5000 diffractometer with copper radiation CuKα (λ=1.5405 Å)

Results and discussion

Microstructural characterization

Figure 1 presents the Mo-Si-Cr ternary phase diagram [14] where the alloy selection was carried out, it can be Figure 2 shows the representative microstructure of the alloy with 36 at.% Cr after annealing treatment. A monophasic morphology is observed in this image, with the presence of large grains (>150 µm); similar microstructures were observed for the other compositions; therefore, transition from Mo₃Si to Cr₃Si was achieved successfully. In this case, the proposed mechanism is the substitutional type, where chromium takes partially the molybdenum places and gradually Mo atoms are totally replaced by Cr in the A15 crystal structure. The existence of Cr₃Si phase around 60 at.% Cr phase was identified by XRD, in good agreement with the ternary alloy phase diagram, corroborating that there are not significant changes due to the applied heat treatment (1,400 °C) and the diagram temperature (1,300 °C).

Figure 1:

Presents the Mo-Si-Cr ternary phase diagram [14] where the alloy selection was determinated, it can be observed that the alloys are positioned along the single-phase composition, considering the transition from Mo₃Si to Cr₃Si.

Figure 2:

Representative microstructure of the alloy with 36 at.% Cr after annealing for 24 h at 1,400 °C in vacuum. Large grains are developed after annealing.

Kinetic oxidation behaviour

Table 2 presents the numerical results of the weight variations obtained from the cyclic oxidation tests. It is observed from the samples that contain 0 and 17 at. % Cr (M1 and M2, respectively) that the weight losses are close to 1 g during the initial 5 h and similar behavior for both samples up to 25 hours; after that, the result was a catastrophic pest reaction. This effect occurs due to oxygen presence in the furnace chamber that reacts with molybdenum atoms existent on the alloy surface, generating predominantly molybdenum oxides; these oxides are volatile compounds at temperature of 700 °C, namely MoO₃ and MoO₂ which are not protective oxides [3, 4]. Samples M3, M4 and M5 showed lower weight losses, with the M4 sample exhibiting the best performance in weight losses (4.0 x10-4 g), showing also the best stability in comparison with M3 and M5 samples until the test was stopped.

Table 2:

Weight losses obtained from the continuous oxidation test after 250 h of exposure at 1,000 °C. Samples with low Cr concentration presented pest reaction.

Sample	Initial weight [g]	Final weight [g]	Weight losses [g]	Total time [h]	Average mass loss per surface area [g/cm²]	Observations
M1	1.3568	0.2694	1.0874	5	——–	(PESTING)
M2	0.8920	0.1336	0.7584	25	——–	(PESTING)
M3	0.6460	0.6446	0.0014	250	9.33×10⁻⁴	Stable
M4	0.6940	0.6936	0.0004	250	2.66×10⁻⁴	Stable
M5	1.0092	1.0072	0.0020	250	1.33×10⁻⁴	Stable

Figure 3 presents the oxidation kinetics for the alloys with different chromium contents, taking into consideration the weight losses respective to the initial weight for each sample. Pronounced weight losses are produced due to the unprotected surface in M1 and M2 samples in the initial stage; on the other hand, for samples M3, M4 and M5, it is observed that, after 50 h of exposure, a minimal weight variation was detected, describing a plateau in the curve during the following 200 h; this behaviour clearly indicates that a stable oxide scale has been developed and it remains on the sample surface with good adherence even when the tests were interrupted each time to take the weight loss measurement. Therefore, in order to explain the oxide formation mechanism, it can be established that the oxidation is not an oxygen diffusion-limited process; moreover, the partial pressure of oxygen at the interface is enough for oxidizing molybdenum, chromium and silicon. As such, the oxidation process takes place with the follow reactions:

(1)Mo+3/2O2g→MoO3g

(2)4/3Cr+O2g→2/3Cr2O3

(3)Si+O2g→SiO2

Figure 3:

Plot of the weight loss behaviour of the alloys with different chromium content. Only three samples are shown due to the pest reaction that occurred in the samples M1 and M2.

By analysing Equations (1) to (3), it is possible to establish that partial pressure of oxygen in equilibrium with Si and Cr depends directly on the metal content, while Mo reacts more easily with oxygen at lower temperatures.

Then, the oxidation process in the exposed alloys may occur according to the following chemical reactions:

(4)Mo3Si+6O2→3MoO3(volatile)+SiO2+1/2O2

and to oxidize Cr by

(5)2Cr3Si+7O2→3Cr2O3+2SiO2+1/2O2

These reactions explain the mass losses due to volatilization of molybdenum oxide (approximately ~750 °C), and therefore a silicon and chromium oxide interlayer can be developed, i. e. during the oxidizing process for the specific case of the samples M1 and M2, where initially the main oxide produced is MoO₃ and during the reaction a large amount of this oxide is volatilized, leaving a series of micro-voids that allow for a fast oxygen transport through the interface, causing molybdenum losses via oxidation reaction and as a consequence SiO₂ and Cr₂O₃ oxide layers are produced. Soleimani et al. [15] reported that Cr₂O₃ (chromia) forms the protective scale, while SiO₂ (cristobalite) is formed below the chromia scale, which is in agreement with our observations.

Surface oxidation analysis

XRD analysis

Figure 4 presents the XRD pattern of the oxidized surface. For sample M2, it was found that Cr₂O₃ phase is developed with a weak intensity and it grows up when the chromium content is increased; in the case of sample M4 which presents a combination of the two oxides on the surface sample, namely Cr₂O₃ and SiO₂, this behaviour takes place due to the chromium addition when it replaces molybdenum and also a silica peak is observed, which is produced due to fact that silicon reacts with oxygen. Therefore, the possibility that both oxide formation occurs simultaneously increases in agreement with the phase diagram. In general, it was found that crystalline phases were detected in the oxidized alloys, although in sample M4 an amorphous phase was detected. Due to the fact that the process was carried out in air, the probability of some nitridation product formation was expected; however, after a surface scanning, these types of compounds were not found.

Figure 4:

XRD pattern of the oxidized surface samples. In this series, M1 was not present due to the fact that sample was totally disintegrated.

SEM analyses

In Figure 5, the SEM/EDS analyses on the cross-section of the oxidized sample M3 at 1,000 °C are presented; it can be seen in both zones, namely in metal/oxide and in oxide interphase scale, that the molybdenum peak in matrix increases considerably, while in the scale/gas zone, the molybdenum peak presents a considerable reduction, indicating that Mo is almost absent in the outer part of the oxide scale. On the other hand, the oxygen increment in the oxide scale was observed. These observations permit us to formulate the assumptions that two oxides are present mainly in the oxidized scales, namely SiO₂, Cr₂O₃ and a small fraction of MoO₃. This element variation is observed in Figure 6, where the representative line-scan analyses in the cross-section of sample M3 are presented (across the matrix to the edge of the oxidized region). It can be observed that oxygen is present mainly in the oxide scale and almost absent in sample matrix, and the oxide scale width is approximately 200 µm, distributed homogeneously along the surface sample; silicon presence was detected in the oxide scale with high intensity, although this element also was detected with a moderated intensity in matrix; these results suggest that silicon oxide can be found in both zones. A similar behaviour was observed with chromium which is present in a major proportion in the oxide scale, although it was detected in a considerable proportion in the matrix of the sample also. Molybdenum was detected mainly in the matrix, while in the oxide scale it was almost absent, indicating that molybdenum oxide certainly was evaporated during the oxidation process. Therefore, because chromium and silicon are mainly present in the oxide scale, both elements contribute to the surface protection of the sample, which is in concordance with the previous thermodynamic and XRD analyses. Figures 7a–d present the images of the oxidized surface sample at different magnifications, where it is observed that some degree of superficial particles exists (≈ 100 µm) achieving a protection zone of approximately 95 % in area fraction. This effect can be explained in terms of an eventual evolution of sub-layers below the oxidized surface which plays an important role in the surface protection. Figures 7b–7d show an amplification of the surface layer, which is composed of small particles randomly distributed of approximately 1 µm. It is important to note that the spacing between the particles is considerably small and this array does not easily allow the constant oxygen migration from the atmosphere to the inner surface of the sample and vice versa generating protective oxidation products and that oxide layer possesses good stability and adhesion; this statement is demonstrated in Figure 8 that shows the SEM/EDS microanalysis with an image of the surface sample, where it is observed the evidence of an upper layer over an internal layer. This sub-layer (marked with a square) which is composed also of silicon and oxygen with lower chromium content produces layer-by-layer array (epitaxial growing). Therefore, it is established that the SiO₂ layer is developed above the Cr₂O₃ layer, in good agreement with the EDS line-scan observations. In this case, the SiO₂ scales are the most protective oxides, due to their high stability and the superior adhesion grade; therefore, the intrinsic protective characteristic of SiO₂ layers is attributed also to their high ability to react and to produce silicates as a result of the continuous chemical combination of elements present in the substrate and their self-restorative capacity at high temperature.

Figure 5:

SEM/EDS cross-section analyses of the oxidized sample M3 at 1,000 °C in matrix and in the oxide scale, showing the lower intensity of the Mo peak in the matrix.

Figure 6:

SEM/EDS line-scan analyses of the cross-section of the M3 sample, performed across the matrix and oxidized edge region. In the image, it is observed that silicon was present in the edge scale and chromium was present below the silicon region.

Figure 7:

SEM surface images of the oxidized surface sample at different magnifications, (a) and (c) a general view of the samples with 36 and 60 at.% Cr respectively, (b) and (d) a detailed view of the oxide layers of the same samples.

Figure 8:

SEM/EDS microanalysis with an image of the surface of the sample with 36 at.% Cr, where very small molybdenum traces were found in this zone.

Conclusions

Cyclic oxidation studies were carried out at 1,000 °C in pseudo-binary silicide samples with different Cr contents. Samples with cero and 17 at.% Cr showed a pest reaction at exposure times lower than 25 h, while the samples with higher chromium content, namely 36 and 60 at.%, presented an improved oxidation behavior. XRD results showed the formation of Cr₂O₃ and SiO₂ oxides which are responsible for this oxidation resistance improvement. No evidence of molybdenum oxides was detected after the test in the oxide scale. Evaluation at high temperature of thermal spray deposition (HVOF (High Velocity Oxygen Fuel) technique) of powders from selected alloys deposited on turbine blades is underway.

Funding statement: This work was partially supported by PRODEP and CONACyT under grant: S53120-Y.

Acknowledgements

The authors want to thank to Oak Ridge National Laboratory (ORNL), Oak Ridge, TN, for sample preparation. Also, thanks go to Professor Josue Delgado for reviewing this manuscript.

References

[1] J. Doychak, Oxidation Behavior of High-Temperature Intermetallics. Intermetallic Compunds: Vol. 1, Principles and Practice, New York, Edited by J.H. Westbrook and R.L. Fisher, John Wiley & Sons Ltd (1994).Suche in Google Scholar

[2] A.K. Vasudevan and J.J. Petrovic, Mater. Sci. Eng., A155 (1992) 1–17.10.1016/0921-5093(92)90308-NSuche in Google Scholar

[3] R. Tanaka, Mater. High Temp, 17(4) (2000) 457–464.10.1179/mht.2000.060Suche in Google Scholar

[4] P.V. Krakhmalev and J. Bergströn, Wear, 260 (2006) 450–457.10.1016/j.wear.2005.03.013Suche in Google Scholar

[5] M.K. Meyer and M. Akin, J. Am. Ceram. Soc., 79 (1996) 938–944.10.1111/j.1151-2916.1996.tb08528.xSuche in Google Scholar

[6] H. Inui, K. Ishikawa and M. Yamaguchi, Intermetallics, 8 (2000) 1131–1145.10.1016/S0966-9795(00)00049-2Suche in Google Scholar

[7] R. Sakidja, F. Rioult, J. Werner and J.H. Perepezko, Scr. Mater., 55 (2006) 903–906.10.1016/j.scriptamat.2006.07.044Suche in Google Scholar

[8] M. Akinc, M.K. Meyer, M.J. Kramer, J.A. Thom and J.J. Huebsch, Mater. Sci. Eng., A, 261 (1999) 16–23.10.1016/S0921-5093(98)01045-4Suche in Google Scholar

[9] D.M. Shah and D.L. Anton, Ternary Alloying of Refractory Intermetallics, MRS Proc., 213 (1990) 63.10.1557/PROC-213-63Suche in Google Scholar

[10] S.V. Raj, Mater. Sci. Eng. A, 201(1–2) (1995) 229–241.10.1016/0921-5093(95)09767-8Suche in Google Scholar

[11] I. Rosales, H. Martinez, D. Bahena, J.A. Ruiz, R. Guardian and J. Colin, Corros. Sci., 51 (2009) 534–538.10.1016/j.corsci.2008.12.004Suche in Google Scholar

[12] S. Okada, K. Toetsu and T. Mori, Pac. Sci. Rev., 13(1) (2011) 33–38.Suche in Google Scholar

[13] B. Gorr, L. Wang, S. Burk, M. Azim, S. Majumdar, H.-J. Christ, D. Mukherji, J. Rösler, D. Schliephake and M. Heilmaier, Intermetallics, 48 (2014) 34–43.10.1016/j.intermet.2013.10.008Suche in Google Scholar

[14] G. Petzow and G. Effenberg, Handbook of the Ternary Alloy Phase Diagrams, Vol. 8, MSI, Stuttgart, Germany, (1993).Suche in Google Scholar

[15] A. Soleimani-Dorcheh, W. Donner and M.C. Galetz, Corros, 65–12 (2014) 1143–1149.Suche in Google Scholar

Received: 2017-07-12

Accepted: 2018-02-04

Published Online: 2018-06-13

Published in Print: 2018-10-25

This article is distributed under the terms of the Creative Commons Attribution Non-Commercial License, which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Artikel in diesem Heft

https://doi.org/10.1515/htmp-2017-0099

Schlagwörter für diesen Artikel

intermetallics compounds; SEMX-ray diffraction; oxidation

Creative Commons

BY-NC-ND 3.0