Enhanced Oxidation Resistance of Mo–Si–B–Ti Alloys by Pack Cementation
As new high-temperature structural materials, Mo–Si–B alloys satisfy several requirements such as oxidation and creep resistance. Recently, novel Ti-rich Mo–Si–B alloys have shown an increased creep resistance compared to Ti-free alloys. However, due to the formation of a duplex SiO2–TiO2 oxide layer, which allows for fast ingress of oxygen, the oxidation resistance is poor. To improve the oxidation resistance, a borosilicate-based coating was applied to a Mo–12.5Si–8.5B–27.5Ti (in at.%) alloy. After co-deposition of Si and B by pack cementation at 1000 °C in Ar, a conditioning anneal at 1400 °C is used to develop an outer borosilicate layer followed by an inner MoSi2 and Mo5Si3 layer. During both isothermal and cyclic oxidation after an initial mass loss during the first hours of exposure, a steady state is reached for times up to 1000 h at temperatures ranging from 800 to 1200 °C, demonstrating a significantly enhanced oxidation resistance.
KeywordsHigh-temperature materials Mo–Si–B–Ti alloys Pack cementation Oxidation-resistant coating
New high-temperature structural materials are required to deal with the challenges in power generation and air transportation. To increase the efficiency of power generation and aircraft engines, a higher combustion temperature is one possible solution. To withstand the additional oxidation and creep demands, materials with higher melting points like Nb- or Mo silicide alloys are promising candidates [1, 2, 3]. Because of their good creep and oxidation behavior at 1200 °C, multiphase Moss + Mo5SiB2 + Mo3Si alloys have attracted a growing interest for the last two decades [4, 5, 6, 7]. Unfortunately, such alloys exhibit a relatively high density of 9.6 g/cm3 compared to densities below 9 g/cm3 for common Ni-based superalloys  and fail in a brittle manner at lower temperatures [9, 10, 11, 12]. Furthermore, at temperatures at around 600–900 °C, the oxidation resistance of Mo–Si–B alloys is dominated by MoO3 vaporization and results in a catastrophic oxidation behavior .
Recently, novel Ti-rich Moss + Mo5SiB2 + Mo3Si alloys have shown an increased creep resistance and lowered density of 7.8 g/cm3 compared to Ti-free Mo–Si–B alloys [14, 15]. The enhanced creep resistance is due to the formation of Ti silicide precipitates within the Mo solid solution phase (Moss) during processing. However, these alloys exhibit poor oxidation resistance, due to the formation of a duplex SiO2–TiO2 surface layer that allows for fast inward diffusion of oxygen . This leads to a catastrophic oxidation at 820 and 1000 °C and to an almost linear mass loss at 1100 °C and above. Furthermore, Yang et al.  found that alloying with Ti can stabilize the Mo5Si3 phase in exchange for Mo3Si. As first reported by Burk et al. , the Ti-rich Mo5Si3 phase shows an increased oxidation resistance compared to the Ti-free phase and Mo3Si. Since the Mo5Si3 phase is difficult to stabilize under the conditions described by Yang et al. , Fe as further minor alloying element was added by Azim et al.  to enlarge the Moss + Mo5SiB2 + Mo5Si3 phase field. Nevertheless, even by stabilizing a multiphase Moss + Mo5SiB2 + Mo5Si3 microstructure, the oxidation resistance is still small compared to Ni-based superalloys .
For structural applications, the requirements for high strength at high temperature along with environmental attack resistance are not usually satisfied concurrently for most materials. One effective strategy to address an enhanced resistance to environmental attack in high-strength materials is the application of oxidation-resistant coatings. While this strategy is attractive, the successful implementation of protective coatings over the entire operating temperature range needs the satisfaction of other requirements involving thermodynamic and mechanical compatibility between the coating and the alloy substrate and the robust stability of the coating in order to maintain the integrity of the environmental protection. For further improvement in oxidation resistance, a pre-treatment of the surface is inevitable to preserve the excellent creep behavior and low density of Ti-rich Mo–Si–B alloys. In the past, different coating systems were applied on Mo–Si–B alloys. To deposit an Al-containing coating, Sakidja et al.  used the Al pack cementation process which was already established for Ni-based superalloys. After subsequent annealing in air, the coated samples were protected against further oxidation by an outer Al2O3 layer at all tested temperatures ranging from 732 to 1372 °C. Additionally, the coating consists of two more inner layers: (1) a mixture of Mo3Al8 and Mo3Si and of (2) Mo3Si and Mo5SiB2. Park et al.  covered Mo–Si–B alloys by a SiO2/ethanol slurry using an air gun followed by annealing at 1000 or 1200 °C in air. After annealing, the coating consists of three separate layers: (1) an exterior borosilicate glass layer, (2) a mixed glass and MoO2 layer and (3) the Moss phase. Sakidja et al.  designed silicide-based coatings by (Si + B) pack cementation, which also resulted in a three-layer arrangement after exposure in air: (1) an outer borosilicate layer, (2) a B-saturated Mo5Si3 layer and (3) a Mo5SiB2 layer. Both Park and Sakidja found an enhanced oxidation resistance of their coated alloys compared to uncoated samples at temperatures ranging from 1000 to 1300 °C.
To enhance the oxidation resistance of Ti-rich Mo–Si–B alloys, the present work examined the effectiveness of (Si + B) pack cementation on a model Mo–12.5Si–8.5B–27.5Ti alloy. The reaction of the coating with the substrate during processing and with the oxidizing atmosphere was determined as a function of temperature and time. Both isothermal and cyclic oxidation exposures have been carried out at temperatures ranging from 800 to 1200 °C.
The Mo–12.5Si–8.5B–27.5Ti (in at.%) substrate material was produced by repetitive arc melting of high-purity metals, Si and B, in a Zr-gettered high-purity Ar atmosphere. After homogenization treatment at 1600 °C for 100 h, slices of this alloy were prepared for pack cementation. The substrate slices with a thickness of 4.5 mm were embedded in a powder mixture consisting of 62.5 wt% Al2O3, 35 wt% Si + B and 2.5 wt% NaF. Subsequently, the pack cementation was carried out in an Ar atmosphere at 1000 °C for 40 h to co-deposit the Si and B. Finally, the conditioning takes place at 1400 °C for 10 h in air. A detailed description of the pack cementation process can be found elsewhere [22, 23].
List of oxidation tests
Temperature in °C
Time Sample A in h
Time Sample B in h
Time Sample C in h
SEM examination was carried out in secondary electron (SE) and backscattered electron (BSE) mode using a Zeiss EVO50 microscope. The relevant elements as well as their distribution were determined qualitatively by energy-dispersive X-ray spectroscopy (EDS). To confirm thermal stability of the substrate microstructure, the phase distribution was determined using an image processing program. A Zeiss Auriga 60 SEM equipped with an EDAX System for EDS and EBSD was used to determine the local crystal structure of the layer composition after pack cementation and the resulting oxidation products. Electron beam microanalysis (EPMA) with a Cameca SX51 instrument (8 kV, 20 na) was used for detection of boron in the coating.
Results and Discussion
Microstructure and Layer Arrangement After Conditioning
Phase distribution (vol %) in substrate for different exposure times and temperatures
Homogenization 1600 °C/100 h
Conditioned 1400 °C/10 h
Oxidation 800 °C/500 h
Oxidation 1100 °C/500 h
Oxidation 1200 °C/500 h
32.09 (± 1.1)
39.70 (± 2.2)
43.77 (± 2.1)
45.30 (± 3.0)
44.35 (± 1.6)
33.45 (± 0.8)
34.14 (± 3.1)
37.15 (± 2.4)
33.27 (± 3.1)
37.48 (± 1.6)
4.57 (± 0.5)
19.13 (± 1.9)
17.18 (± 3.7)
20.67 (± 3.7)
18.17 (± 1.3)
29.88 (± 0.9)
6.77 (± 2.3)
1.66 (± 1.6)
1.04 (± 0.78)
Composition measurements for the coating and Mo5Si3 (T1) phase
Coating cross section
The long-term oxidation behavior of the applied borosilicate layer was studied under thermal cycling conditions, which are known to be more crucial for materials in technical applications compared to isothermal oxidation, because spallation is promoted. Spallation can be caused by either poor adhesion between the substrate and the coating or different thermal expansion coefficients of the substrate and the coating, resulting in interfacial stress during cooling and heating .
The weight change curves of the cyclic oxidation tests for each test temperature are summarized in Fig. 3b. For the sake of clarity, a scale break from 550 to 850 h is applied to the time axis. At 800 °C, all samples show an initial weight loss around 3–4 mg/cm2 followed by a steady state with no further distinct weight change. With increasing temperature, the weight loss becomes less for 1100 °C and turns into a weight gain at 1200 °C. However, at higher temperatures, the weight change is more variable compared to 800 °C. One of the samples oxidized at 1100 °C has a larger weight loss than the others, and at 1200 °C, one of the three samples reached a steady state, while the other two show a mass gain over time. A better overview of the initial states of oxidation is represented in Fig. 3c, where the specific weight change for the first 25 h of cyclic oxidation is plotted. The majority of mass loss at 800 °C occurs within the first 10 h of oxidation. At the same time, the oxidation behavior at 1100 °C turns from mass loss to mass gain, whereas this point is reached after approximately 5 h at 1200 °C. As for the isothermal oxidation tests, the initial weight loss is assumed to be a result of an incomplete conditioning by the formation of cracks within the borosilicate layer during cooling or by incomplete covering of the samples surface. The initial weight loss at 800 °C during thermal cycling (Fig. 3b, c) for the coated sample is rather minimal compared to the uncoated sample (Fig. 3d) as indicated by the weight change scales in Fig. 3. The initial loss is related to the increased time during heating to a low temperature (i.e., 800 °C) for the coating to flow and cover defects (Fig. 2c) and cracks that develop in the MoSi2 layer (Fig. 2b) during cooling. The faster turnover from mass loss to mass gain at 1200 °C in contrast to 1100 °C is caused by the lower viscosity of the borosilicate glass at higher temperatures [13, 26].
Figure 3d summarizes the weight change curves for the thermal cycling and provides a comparison with the weight change behavior of uncoated Mo–12.5Si–8.5B–27.5Ti investigated by Azim et al.  at the same oxidation temperatures. While the weight change at 1100 and 1200 °C is negligible for the coating, the weight change for the uncoated alloy is 22 and 43 mg/cm2 after 100 h of oxidation without reaching steady state. However, at 800 °C, the uncoated substrate is fully consumed by the oxidation process within the first 20 h of oxidation, while the weight loss for the coated substrate is about 2 mg/cm2 at the same time. The main reason for the improved oxidation resistance is the suppressed formation of TiO2 during oxidation by the MoSi2 layer, which acts as a diffusion barrier for Ti due to negligible solubility of Ti in MoSi2 . The determined oxidation behavior of the coated Mo–Si–B–Ti alloy is consistent with the observed behavior of Ti-free Mo–Si–B alloys coated by Al pack cementation [20, 23].
By co-deposition of (Si + B) during pack cementation followed by a high-temperature conditioning treatment, a borosilicate-based coating has been established on a model Mo–Si–B–Ti alloy. In addition to the representative coating layers of an outer borosilicate glass and inner MoSi2 and Ti5Si3 layers, some TiO2 develops either as surface plate-shaped particles or as isolated particles within the glass. The substrate microstructure is unaltered by the coating process. During both isothermal and cyclic oxidation after an initial mass loss of about 4 mg/cm2 during the first hours of exposure, a steady state is reached for times up to 1000 h at temperatures ranging from 800 to 1200 °C, demonstrating a significantly enhanced oxidation resistance. The effectiveness of the coating is not diminished by the presence of isolated TiO2 particles. Furthermore, the coating appears to exhibit a self-healing capability to damage such as cracking in the MoSi2 layer due to the formation of the borosilicate glass.
The financial support of Deutsche Forschungsgemeinschaft (DFG) and by Karlsruhe House of Young Scientists (KYHS) is gratefully acknowledged. AK thanks the Carl Zeiss Foundation for funding. The support of ONR (N00014-10-1-0913) for JHP is gratefully acknowledged. This work was partly carried out with the support of the Karlsruhe Nano Micro Facility (KNMF, www.knmf.kit.edu), a Helmholtz Research Infrastructure at Karlsruhe Institute of Technology (KIT, http://www.kit.edu).
- 14.D. Schliephake, M. Azim, K. von Klinski-Wetzel, B. Gorr, H.-J. Christ, H. Bei, E. P. George and M. Heilmaier, Metallurgical and Materials Transactions A: Physical Metallurgy and Materials Science 45, 2013 (1).Google Scholar
- 24.K. Meier, H. Mehrer and R. Gerhard, Zeitschrift für Metallkunde 70, 1979 (271).Google Scholar
- 27.P. S. Frankwicz and J. H. Perepezko, in Symposium Q—High-Temperature Ordered Intermetallic Alloys IV, MRS, Vol. 213, p. 169 (1990).Google Scholar
- 30.K. Grente, F. Rebillat and F. Langlais, High Temperature Corrosion and Materials Chemistry IV Book Series: Electrochemical Society Series, Volume 2003, issue 16, 545 (2003).Google Scholar
- 31.R. H. Doremus, American Ceramic Society Bulletin 82, (3), 2003 (59).Google Scholar
- 32.N. P. Bansal and R. H. Doremus (eds.), Handbook of Glass Properties, (Academic Press, Orlando, 1919).Google Scholar