- Email: [email protected]
Behaviour of a silicon-rich coating on Ti-46Al-8Ta (at.%) in hot-corrosion environments
Accepted Manuscript Title: Behaviour of a silicon-rich coating on Ti-46Al-8Ta (at. %) in hot-corrosion environments Authors: K. Rubacha, E. Godlewska, K. Mars PII: DOI: Reference:
S0010-938X(16)30845-9 http://dx.doi.org/doi:10.1016/j.corsci.2017.02.002 CS 6995
To appear in: Received date: Revised date: Accepted date:
27-9-2016 30-1-2017 2-2-2017
Please cite this article as: K.Rubacha, E.Godlewska, K.Mars, Behaviour of a siliconrich coating on Ti-46Al-8Ta (at.%) in hot-corrosion environments, Corrosion Science http://dx.doi.org/10.1016/j.corsci.2017.02.002 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Title: Behaviour of a silicon-rich coating on Ti-46Al-8Ta (at. %) in hot-corrosion environments Authors: K. Rubacha, E. Godlewska, K. Mars AGH University of Science and Technology, Faculty of Materials Science and Ceramics, Al. A. Mickiewicza 30, 30-059 Krakow, Poland
Corresponding author: Katarzyna Rubacha Address: AGH UST Al. Mickiewicza 30, 30-059 Krakow, Poland Phone number: +48 126174787 E-mail: [email protected]
Highlights
Adherent silicide coating was deposited on Ti-46Al-8Ta alloy in two-step process. Cyclic oxidation tests with different salt deposits were conducted. Silicon-rich coating had beneficial effect on oxidation resistance. Diffusion of reactive gases to substrate was prevented.
Abstract In this work silicon-rich coating was deposited on a Ti-46Al-8Ta (at. %) alloy in a two-step procedure comprising physical and chemical vapour deposition (magnetron sputtering and pack cementation, respectively). Protective properties of the coating were tested through cyclic oxidation in air at 800 oC with samples contaminated with NaCl, Na2SO4 or a mixture of these two salts. The coated samples exhibited better hot corrosion resistance compared with the uncoated reference. This was attributed to the formation of an oxide layer composed of amorphous silica with embedded rutile and α cristobalite crystals, constituting a barrier against oxygen penetration. Keywords: A. Intermetallics, B. SEM, Thermal cycling, C. Hot corrosion 1. Introduction Titanium alloys and titanium aluminides have been under investigation for at least twenty years owing to the advantageous combination of physical, mechanical and chemical properties. Intermetallic phases in the Ti-Al binary system attract attention because of relatively low density, very high specific strength and oxidation resistance at moderate temperatures [1-4]. A lot of work has been devoted to improve mechanical properties and oxidation resistance of titanium alloys at elevated temperatures. Particularly good results were obtained by alloying with Nb, Mo, Ta, Si or W [5]. Alloys with high percentage of aluminium (up to about 45 at. %) and ternary additions of Nb and/or Ta are among the most advanced intermetallic materials with good oxidation resistance [6-9]. This can be further improved by surface engineering. A great
deal of metallic and ceramic coatings and deposition techniques have been reported to provide some degree of protection against oxidation [10-14]. By far more complex problems appear, however, when mineral deposits such as NaCl and Na2SO4 are in contact with Ti-Al alloys exposed to oxidising atmospheres, as described previously [15, 16]. As far as hot corrosion protection of intermetallic alloys is concerned, very limited number of investigations have been undertaken and reported in the literature. For illustration, Ti50Al was coated with TiAlCr [17] or Al2O3 and enamel coating [18] which had beneficial effect on the oxidation resistance in a hot-salt environment but during long-term cyclic oxidation tests, the initially protective layer cracked and spalled off. As a result, the material corroded at an accelerated rate. Some studies were also performed with more advanced alloys [19, 20]. For illustration, the surface of Ti–48Al– 2Cr–2Nb (wt. %) was modified with gold by electrochemical deposition followed by vacuum annealing [19]. The layer of gold was thought an ideal barrier for hot corrosion because of chemical inertness of this noble metal and its very low solubility in molten salts. Indeed the layer provided the desired protection to the alloy but due to high prices of gold its common use is doubtful. The most recent studies concerning hot corrosion protection of the advanced titanium aluminide alloys are focused on composite coatings with alternate Y2O3-doped Al2O3 ceramic nano-layers and Pt–Au binary alloy nano-layers [20]. It is required that the proposed composite coating consist of at least seven layers to provide satisfactory protection which makes the deposition process quite complex and time consuming. Coatings containing silicon might be an interesting option to enhance high-temperature oxidation resistance by forming a protective SiO2 scale [21-23]. It has been reported [21] that a Si diffusion coating significantly improves scale adherence and a Si-Al coating has excellent resistance to oxidation. Interestingly, with increasing exposure time the protective properties of the silica scale become better [22]. It has been demonstrated also that silica is highly stable in molten salts such as chlorides and sulphates [24]. Titanium silicide layers were successfully produced on pseudo-α titanium alloys by pack cementation [25, 26]. Preliminary trials of using the same technology to develop silicide coatings on Ti46Al-8Ta (at. %) appeared not fully successful because of Kirkendall porosity around the coating/substrate interface. In the approach proposed in this work, alloy surface will be first enriched in titanium to prevent excessive loss of titanium from the substrate and next diffusion coated with silicon to yield a layer with the matrix based on titanium silicides modified by the components of the substrate.
2. Experimental procedure 2.1 Sample preparation The Ti–46Al–8Ta (at. %) ingots for hot corrosion studies were cut with a diamond saw into samples in a form of pellets, 13 mm in diameter and 1-2 mm in thickness, which were ground with SiC paper to 2000 grit and polished using water-based diamond suspensions (grain size 3 µm and 1 µm). Each sample was measured and weighed. Mirror-like surfaces were subsequently cleaned in acetone and dried prior to coating deposition. 2.2 Coating deposition
Silicon-rich layers were produced in an innovative two-stage procedure. The first stage involved modification of alloy surface with a titanium-silicon solid solution by magnetron sputtering using TEPRO NP 501A Koszalin equipment. Composition of the target used in this experiment was 90 at. % Ti, and 10 at. % Si. It was produced by sintering elemental powders at 1300 oC for 1 h under a pressure of 25 MPa. Sputtering conditions were the following: pressure 0.24 Pa, current 0.6A, temperature of sample holder 380 o
C. In the second stage, a silicide coating was grown by pack cementation. The samples were buried in a
powder mixture consisting of 54.58 mol. % silicon and 44.96 mol. % alumina (Al2O3) and halide activator, cryolite (Na3AlF6), in the amount of 0.46 mol. %. The whole assembly was maintained at 850 oC under argon for 6 h. After deposition the samples were weighed again. 2.3 Hot corrosion tests Samples with silicide coatings were immersed in an aqueous solution 10% by mass of Na2SO4 (0.77 mol/dm3) or NaCl (1.83 mol/dm3) or a mixture of two salts Na2SO4 (0.67 mol/dm3) + NaCl (0.22 mol/dm3) and dried in air. Average amount of the salt deposit was about 1 mg/cm2. The samples were placed in ceramic crucibles in order to collect spalled matter. Reference samples without the silicide coating but with salt deposits were prepared in the same way. The temperature of cyclic oxidation tests was 800 oC. Each cycle consisted of rapid heating (samples were introduced into hot furnace), exposure at constant temperature for 20 h and rapid cooling to room temperature in the laboratory air (samples were removed from the furnace). Total hot-dwell time was 80 h or 300 h. After every cycle the samples were weighed with an accuracy of 10-4 g to evaluate mass variations over time. 2.4 Post treatment examination The multilayer silicide coating was analysed by grazing incidence X-ray diffraction (GID) after removal of consecutive layers by grinding. Phase composition of samples surface after hot corrosion tests was determined by X-ray diffraction (PanalyticalX’Pert) in usual Bragg-Brentano configuration. Surfaces and transversal cross-sections of the samples were analysed by means of scanning electron microscopy (Nova Nano SEM 200 FEI Europe Company) combined with the energy dispersive X-ray spectrometry (EDS/EDAX).
3. Results 3.1 Coating characterization A cross-section of sample after the first stage of coating deposition is shown in Fig. 1. The layer is adherent and uniform with a thickness of about 2 µm. The role of titanium-silicon interlayer deposited by magnetron sputtering was to enrich alloy surface in titanium. Previous experiments without the Ti-Si interlayer ended up with poor coating adherence to the substrate, as a result of enhanced outward diffusion of titanium from the near-surface zone of the alloy. Fig. 2a presents a digital picture of the sample with a silicide coating. As can be seen in Fig. 2b (SEM image at a higher magnification), the surface is quite rough and consists of small equiaxed grains, average diameter of 1.5 µm. According to EDS the external surface is composed of titanium and silicon in an atomic ratio of about 1:2 which would correspond to TiSi2. Fig. 3
presents a cross-section of the silicide coating. The coating is well bonded with the substrate and its overall thickness is about 60 µm. It has clearly a layered structure and consecutive layers differ in composition. According to EDS (Table 1, Fig. 3) the outermost part of the coating consists mainly of silicon and titanium; the concentration of titanium increasing and that of silicon decreasing from the surface inwards. In points 3 and 4 the concentration of tantalum is higher than in other areas. The dark grey zone close to the alloy (point 5) is significantly enriched in aluminium and the Ti to Al ratio corresponds to TiAl 2. These results are roughly consistent with XRD measurements (Fig. 4). The occurrence of TiSi2 together with TiSi on the surface was confirmed. After removal of the outermost layers of the coating, the XRD patterns revealed Ti5Si4 and Ti5Si3. Even though EDS indicated tantalum enrichment at some distance from the surface, the tantalum-containing phases were not marked on the diffractogram because reflections corresponding to titanium silicides and tantalum silicides e.g. Ti5Si3 and Ta5Si3 are indistinguishable. It is possible therefore that either both titanium and tantalum silicides with the same crystallographic structure were present or a solid solution (Ti, Ta)5Si3. In the inner layers of the coating, reflections corresponding to TiAl2 were found along with titanium aluminides from the substrate: -TiAl and 2-Ti3Al. 3.2 Hot corrosion – thermogravimetric studies at 800 oC Gross mass changes of samples with and without silicide coatings during cyclic oxidation tests in air at 800 oC in the presence of different salt deposits are shown in Fig. 5. The most aggressive environment for the uncoated alloy was sodium chloride bringing about the largest mass gains and severe scale spallation. Protective properties of the silicide coating are unquestionable in that case: mass gain dropped from about 20 mg/cm2 (uncoated sample) to 5 mg/cm2 (coated sample) and scale adherence was much better. For the sample contaminated with sodium sulphate, the silicide coating appeared also effective: mass gain decreased from 1.5 to 0.7 mg/cm2. In the case of mixed salt deposit, no particular effect of the coating on gross mass changes was identified. After 300 h of cyclic oxidation tests mass gains of the two types of samples were almost the same. 3.3 Hot corrosion - surface morphology and chemical composition Surfaces of samples with silicide coatings after 80 h of oxidation at 800 oC in the presence of salt deposits differ significantly in morphology and average chemical composition (Table 1, Fig. 6a, c, e). The biggest amount of silicon was detected on samples with the silicide coating and the NaCl deposit. SEM images in Fig. 6a, b show some spherical particles sticking to the sample surface which is mostly covered with small, needle-shaped crystals. According to EDS (Table 1, point 1 in Fig. 6b) the outermost layer of the sample is composed of a mixture of silicon and titanium oxides. The concentration of aluminium is elevated in the round-shaped particles. Needle-shaped crystals covered with a smooth discontinuous top layer are found on the sample with a silicide coating contaminated with sodium sulphate (Fig. 6c, d). From Table 1, Fig. 6d (point 3 and 4) it
follows that the smooth areas have composition corresponding to the salt deposit and the crystals are a mixture of titanium and silicon oxides. Fig. 6e, f presents external surface of the silicide coating after 80 h oxidation in the presence of salt mixture, Na2SO4 + NaCl. The surface is covered with loosely bonded submicrometric crystals and some larger protrusions, which at a higher magnification appear petal-like crystal agglomerates, with an elevated concentration of aluminium (Table 1, Fig. 6f point 5).
3.4 Hot corrosion - surface phase composition Fig. 7 shows diffraction patterns from the surfaces of samples with a silicide coating and either NaCl or Na2SO4 deposit, tested for 80 h at 800 oC. These confirm that phases present on the surface are mainly SiO2 (α cristobalite structure) and TiO2 (rutile). A few peaks corresponding to TiSi2 (external layer of the coating) are also visible. For the sample contaminated with Na2SO4, some reflections from the salt residue are found in addition. XRD patterns from the surface of samples with a silicide coating and Na2SO4 + NaCl deposit after oxidation for 80 h, 160 h and 300 h are presented in Fig. 8. It can be seen that during oxidation the surface composition undergoes changes. After 80 h the phases present on the surface were Al2O3 (corundum), TiO2 (rutile), SiO2 (α cristobalite) and NaTaO3. After 160 h, NaTaO3 and Al2O3 were not identified. Al2O3, TiO2, SiO2 were found after 300 h of oxidation. 3.5 Hot corrosion – sample cross-sections Cross-sections of coated samples after 80 h of hot corrosion test can be seen in Fig. 9. In all cases the scale consisted of an amorphous silica matrix with dispersed TiO2 (rutile) and α cristobalite crystals. Its thickness was about 2 µm for Na2SO4 and 10 µm for Na2SO4 + NaCl. In the latter case, a smooth grey layer with a substantially higher concentration of aluminium could be seen under the scale (Table 1, Fig. 9d point 3). Generally, below the scale, there was a layer of unconsumed coating which mostly consisted of TiSi and TiSi2. Some cracks propagating parallel to the surface were probably formed during sample preparation for the analysis. At a bigger distance from the surface, there was a bright layer with an elevated concentration of tantalum (about 12 at. %) together with titanium (60 at. %) and silicon (25 at. %) (Fig. 9a, b and Table 1, Fig. 9c point 1, 2). The dark layer at the coating/substrate interface contained high amounts of aluminium and some tantalum-rich precipitates. Cross-section presented in Fig. 9d is somewhat different. After exposure, the coating with a mixed salt deposit appeared less compact, especially in the near-surface region. The bright layer visible in Fig. 9a, b is not present here. According to EDS (Table 1, Fig. 9d) the concentration of tantalum locally reaches the highest value of 12 at. %, as in point 6. 4. Discussion In a recent paper [15] it has been demonstrated that exposure of a Ti-46Al-8Ta (at. %) alloy contaminated with salt deposits to air at elevated temperatures leads to remarkable material losses.
Especially harmful was sodium chloride which brought about degradation by a self-sustaining mechanism involving formation of volatile species, porosity and poor adherence of the scale. The novel two-step procedure applied in this work resulted in an adherent multiphase coating which was formed in a complex diffusion process involving predominant inward diffusion of silicon and some outward diffusion of titanium. From the surface inwards, the coating consists of TiSi2/TiSi; Ti5Si4/Ti5Si3 probably with tantalum as a ternary addition; and aluminium-enriched zone TiAl2 (EDS Fig. 3 and XRD Fig. 4). This phase composition is in good agreement with the previously reported results [27, 28]. There are, however, some differences, e.g. Ti3Si on the pure Ti substrate in [27] or Ti(Al,Si)3 between Ti5Si4 and TiAl2 on a Ti-50Al alloy in [28]. The porous layer in the middle of the coating can be attributed to segregation of aluminium which is insoluble in all titanium silicides except Ti5Si3 [29]. The origin of aluminium in the amount of 20 at. % in the outer part of the coating (Table 1, Fig. 3 point 1 and 2) might be surprising. Noteworthy fact is that the porous region of the coating does not contain aluminium. As suggested earlier [28], at a high temperature aluminium can vaporize from TiAl3 at the TiAl3/Ti5Si4 interface and probably also from the zone where its segregation from the silicide phases initially occurred. The composition of salt deposit is an important factor when considering sample degradation in hot corrosion tests. Among salt deposits used in this work, sodium chloride appeared the most aggressive, as evidenced in Fig. 10 by net mass changes of the coated Ti–46Al–8Ta (at. %) samples. The silicide coating did not stop gradual mass losses due to scale exfoliation and possibly also evaporation of some volatile reaction products. However, compared with the uncoated alloy, these mass losses in a similar test were generally lower (e.g. 2 times for the samples with NaCl) [15]. In all experiments after about 100 h the mass of samples became essentially stable which is understandable in view of the fact that salt deposit was not reapplied during the test. Top views of samples after 300 h of hot corrosion tests at 800 oC with and without the silicide coatings are presented in Fig. 11. In the case of NaCl deposit, protective properties of the coating could be seen even with the naked eye. Degradation of the coated samples was significantly limited. The coated samples with the Na2SO4 deposit had much thinner scale compared with the uncoated reference and apparently no scale spallation occurred. After 300 h of the test with a mixed salt deposit, the scale on uncoated samples was thin and not uniform, the sample surface underneath was distinctly visible. On coated samples a relatively thick scale formed and only slight exfoliation of the scale was observed. Cross-sections of samples after 300 h of oxidation in the presence of NaCl with and without coating are shown on Fig. 12. The uncoated sample is seriously damaged. Many cracks caused decohesion of its external part. The extent of chemically changed material reaches over 100 µm. In contrast, the coated sample is in good condition after the same time of oxidation. The coating is adherent and continuous, without macroscopic defects. After 300 h of the test it still provides protection to the substrate. Longer oxidation tests would be needed, however, to assess its lifetime. The above-described results indicate that the silicide coatings had a beneficial effect on the Ti-46Al8Ta (at. %) alloy in the hot corrosion environments. The protective properties were attributed to the presence
of SiO2 (in a form of α cristobalite) and TiO2 (rutile) on the surface. The occurrence of cristobalite which is the high-temperature form of silicon dioxide might be surprising but already in 1980’s rutile and cristobalite were identified in the oxide scale on a Ti-Si alloy [30]. The tetragonal α cristobalite can exist in a metastable state at temperatures below 500 K [31]. There are also reports on the presence of amorphous silica on the surface of oxidized Ti5Si3 [32] and TiSi2 [33]. According to some unpublished results received at our lab, amorphous silica is also found in the scale. We suppose therefore that amorphous silica and α cristobalite, both, are present on the surface of coated samples after the test. The latter has been confirmed by XRD. The silica glass probably seals small discontinuities and porosity in the oxide layer, which contributes to its better gas tightness. Moreover, according to some reports, silicon can form solid solutions with rutile. It is claimed that small atoms of Si take interstitial positions in the crystal lattice of rutile, decrease the concentration of oxygen vacancies and thereby make diffusion of oxygen slower [30]. Worth noticing are also diffusion coefficients of oxygen in silica and rutile: 5∙10-16 cm2s-1 and 1∙10-11 cm2s-1, respectively [32]. Since the diffusion coefficient of oxygen in rutile is by five orders of magnitude bigger than in silica, it can be assumed that oxygen diffuses through the rutile phase mainly (along the grain boundaries and via oxygen vacancies in the crystal lattice). All the mentioned factors are beneficial for oxidation resistance of the coated samples. For comparison Fig. 13 presents XRD patterns taken from the surface of uncoated samples tested at 800 oC for 300 h with different salt deposits and from the scale collected in the crucible in the case of sample contaminated with NaCl. The only phase identified in the powdered scale was TiO 2. The only oxides identified on the alloy surface in NaCl environment were: TiO2 and Ta2O5. Phases found on the surface of samples oxidised with Na2SO4 or Na2SO4+ NaCl deposits were: TiO2, Ta2O5, Na2Ta4O11 and TiAl from the alloy. Additional scale component was Al2O3 when Na2SO4 alone was deposited. The assessment of degradation mechanism of the silicide coatings is beyond the scope of this study; however the results of hot-corrosion tests shed light on their overall performance and durability. Some of the thermodynamically possible reactions are presented in the following section. As the coating surface after the hot corrosion tests was covered with mixed oxides, the following reactions could take place [33]: TiSi2 + 3O2 → TiO2 + 2SiO2
(1)
TiSi + 2O2 → TiO2 + SiO2
(2)
or 2TiSi2 + 4O2 → TiO2 + 3SiO2 + TiSi
(3)
6TiSi2 + 10 O2 → TiO2 + 9SiO2 + Ti5 Si3
(4)
On the samples contaminated with mixed salt deposits, Na2SO4 + NaCl, additional phases detected by XRD were Al2O3 and NaTaO3 (Fig. 8). The latter could be formed in the following reactions involving decomposition of Na2SO4 at high temperature [15]: Na2 SO4 → Na2 O + SO3 Ta2 O5 + Na2 O → 2NaTaO3
(5) (6)
The scale can further react with salt deposits forming titanates or silicates which is accompanied by release of chlorine gas or sulphur (VI) oxide. 1
TiO2 + 2NaCl + 2 O2 → Na2 TiO3 + Cl2
(7)
TiO2 + Na2 SO4 → Na2 TiO3 + SO3
(8)
1
2SiO2 + 2NaCl + 2 O2 → Na2 Si2 O5 + Cl2
(9)
2SiO2 + Na2 SO4 → Na2 Si2 O5 + SO3
(10)
The gaseous species resulting from the above reactions (Eqs. 7-10) can penetrate inward via various defects and react with the coating or alloy components. From the point of view of thermodynamics, the most probable among chloride phases is AlCl3 as its standard Gibbs free energy of formation is the most negative (see Table 2). The TiO2/SiO2 barrier is not totally gas-tight. As already mentioned, oxygen may diffuse through rutile grain boundaries and oxygen vacancies in the crystal lattice. Also silicon dioxide may react with tantalum (V) chloride to form volatile species, Eq. (11) [34]. SiO2 (s) + 2TaCl5 (g) → SiCl4 (g) + 2TaOCl3 (g)
(11)
Furthermore oxygen and nitrogen may penetrate into the coating and react with its components, yielding multiple nitride, oxide or oxynitride phases. Table 3 shows standard Gibbs free energy of reactions (1-11) at 1100 K. For reactions (6) and (11) no thermodynamic data are available, however XRD analyses evidenced the presence of NaTaO3 in a scale with Na2SO4 + NaCl deposit. Almost all reactions presented in Table 3 have negative values of standard Gibbs free energy which means that they are thermodynamically favourable. Calculations of ΔGor provide positive value for decomposition of Na2SO4. Despite this fact, reaction (5) may occur which is explained in [35]. Upon oxidation, some compositional changes take place in the coating layer. Undoubtedly, titanium diffuses outwards leaving behind a depleted zone at the coating/substrate interface. The mobility of tantalum is very limited in titanium aluminides. This may explain why Ta-rich intermetallic appears between the layer of titanium silicide and the titanium depleted zone. To the best of our knowledge the studies on hot-corrosion behaviour of silicide coatings on advanced titanium alloys are almost non-existent in the literature. There is one paper published in 2015 [36], describing deposition of a silicide coating by halide activated pack cementation on the alloy with a nominal composition of Nb–20Ti–15Si–5Cr–3Hf–3Al (at. %). The researchers proposed a plausible mechanism of hot corrosion processes occurring beneath the layer of molten salt composed of Na2SO4 and NaCl. It has been stated that NaAlO2, AlNaSiO4, and Na2TiSi2O7 are formed along with TiO2 and SiO2. This scheme, however, is not comparable with the results obtained in this work. In particular, coating microstructure was different and neither of the aluminate or silicate phases was identified in the corrosion products.
5. Conclusions
1. The proposed two-step procedure enabled formation of an adherent silicide coating on the Ti46Al-8Ta (at. %) substrate. The coating was multiphase, consisting of TiSi 2/ TiSi in the outer part, lower silicides (Ti, Ta)xSiy and titanium aluminides TiAl2/TiAl3 in the inner part. 2. The silicide coating prevented inward diffusion of oxygen and chlorine to the titanium aluminide substrate. 3. Protective properties of the silicon-rich coating result from the formation of amorphous silicon oxide layer with embedded rutile and α cristobalite crystals, constituting an effective barrier against oxygen penetration. 4. After 300 h of the interrupted oxidation tests at 800 oC in the presence of mineral deposits, the silicide coating was preserved and did not show signs of any significant destruction, except some thickness loss due to oxidation on the surface. It was compact and adherent. Longer oxidation tests are needed to asses coating lifetime in hot corrosion conditions.
Acknowledgements The authors gratefully acknowledge financial support from National Science Centre project number 2013/11/N/ST8/01345. Alloy ingots were produced by ACCESS (Aachen).
References [1] X. Wu, Review of alloy and process development of TiAl alloys, Intermetallics 14 (2006) 1114-1122. [2] E.A. Loria, Gamma titanium aluminides as prospective structural materials, Intermetallics 8 (2000) 1339-1345. [3] M. Yamaguchi, H. Inui, K. Ito, High – temperature structural intermetallics, Acta Mater. 48 (2000) 307322. [4] S. Djanarthany J.-C. Viala, J. Bouix, An overview of monolithic titanium aluminides based on Ti3Al and TiAl, Mater. Chem. Phys. 72 (2001) 301–319. [5] Y. Shida, H. Anada, The effect of various ternary additives on the oxidation behavior of TiAl in high temperature air, Oxid. Met. 45 (1996) 197–218. [6] D. Vojtěch, T. Popela, J. Kubásek, J. Maixner, P. Novák, Comparison of Nb- and Ta-effectiveness for improvement of the cyclic oxidation resistance of TiAl-based intermetallics, Intermetallics 19 (2011) 493-501. [7] M. Mitoraj, E. Godlewska, O. Heintz, N. Geoffroy, S. Fontana, S. Chevalier, Scale composition and oxidation mechanism of the Ti-46Al-8Nb alloy in air at 700 and 800 oC, Intermetallics 19 (2011) 39-47. [8] M. Mitoraj, E. Godlewska, Oxidation of Ti–46Al–8Ta in air at 700 °C and 800 °C under thermal cycling conditions, Intermetallics 34 (2013) 112–121.
[9] E. Godlewska, M. Mitoraj, F. Devred, B. E. Nieuwenhuys, Reactivity of a Ti–45.9Al–8Nb alloy in air at 700–900°C, J. Therm. Anal. Calorim. 88 (2007) 1, 225–230. [10] E. Godlewska, M. Mitoraj, R. Mania, R. Gajerski., Oxidation of a Ti-46Al-8Nb alloy modified with some Ni-base coatings, Ann. Chim. - Sci. Mat. 33 (2008) 9–16. [11] D.B Wei, P.Z. Zhang, Z.J. Yao, J.T. Zhou, X.F. Wei, P. Zhou, Cyclic oxidation behavior of plasma surface chromising coating on titanium alloy Ti–6Al–4V, Appl. Surf. Sci. 261 (2012) 800 – 806. [12] J. Kong, C. Zhou, S. Gong, H. Xul., Low-pressure plasma-sprayed Al–Cu–Fe–Cr quasicrystalline coating for Ti-based alloy oxidation protection, Surf. Coat. Tech. 165 (2003) 281–285. [13] M. Góral, L. Swadzba, G. Moskal, M. Hetmanczyk, T. Tetsui, Si-modified aluminide coatings deposited on Ti46Al7Nb alloy by slurry method, Intermetallics 17 (2009) 965–967. [14] S. Teng, W. Liang, Z. Li, X. Ma, Improvement of high-temperature oxidation resistance of TiAl-based alloy by sol–gel method, J. Alloy. Compd. 464 (2008) 452-456. [15] E. Godlewska, M. Mitoraj, K. Leszczynska, Hot corrosion of Ti-46Al-8Ta (at. %) intermetallic alloy, Corros.Sci. 78 (2014) 63-70. [16] K. Zhang, Z. Li, W. Gao, Hot corrosion behaviour of Ti–Al based intermetallics, Mater. Lett. 57 (2002) 834–843. [17] Z. Tang, F. Wang, W. Wu, Effect of a sputtered TiAlCr coating on hot corrosion resistance of gammaTiAl, Intermetallics 7 (1999) 1271-1274. [18] Z. Tang, F. Wang, W. Wu, Effect of Al2O3 and enamel coatings on 900°C oxidation and hot corrosion behaviors of gamma-TiAl, Mat. Sci. Eng. A-Struct. 276 (2000) 70 - 75. [19] M. Bacos, M. Thomas, J.-L. Raviart, A. Morel, S. Mercier, P. Josso, Influence of an oxidation protective coating upon hot corrosion and mechanical behavior of Ti-48Al-2Cr-2Nb alloy, Intermetallics 19 (2011) 1120-1129. [20] X. Ma, Y. He, J. Lin, D. Wang, J. Zhang, Effect of a magnetron sputtered (Al 2O3– Y2O3)/(Pt– Au) laminated coating on hot corrosion resistance of 8Nb – TiAl alloy, Surf. Coat. Tech. 206 (2012) 2690– 2697. [21] H. Zheng, L. Xiong, Q. Luo, S. Lu, Microstructure and cyclic oxidation behavior of silicide coating and Si-Al coating on Cr-50Nb alloy, Corros.Sci. 89 (2014) 326-330. [22] N. Chaia, Y. Bouizi, S. Mathieu, M. Vilasi, Isothermal and cyclic oxidation behaviour of hot-pressed MSi2 compounds (with M = V, Ti, Cr), Intermetallics 65 (2015) 35 – 41. [23] S. Majumdar, J. Kishor, B. Paul, R.C. Hubli, J.K. Chakravartty, Isothermal oxidation behavior and growth kinetics of silicide coatings formed on Nb–1Zr–0.1C alloy, Corros. Sci. 95 (2015) 100–109. [24] T. Ishitsuka, K. Nose, Stability of protective oxide films in waste incineration environment – solubility measurement of oxides in molten chlorides, Corros. Sci. 44 (2002) 247-263. [25] E. Godlewska, M. Mitoraj, S. Kozinski, Influence of diffusion silicide coatings on the oxidation behavior of a Ti-6Al alloy in a cyclic oxidation test, Polish Ceramic Bulletin, Ceramics, 91/1 (2005) 551-558.
[26] E. Godlewska, M. Mitoraj, S. Kozinski, Deposition of silicide coatings on Ti-6Al alloy using pack cementation method, Polish Ceramic Bulletin, Ceramics, 91/1 (2005) 543-550. [27] B. Cockeram, R. Rapp, The kinetics of multilayered titanium-silicide coatings grown by the pack cementation method, Metall. Mater. Trans. A 26 (1995) 777- 791. [28] T. Munro, B. Gleeson, The deposition of aluminide and silicide coatings on γ-TiAl using the halideactivated pack cementation method, Metall. Mater. Trans. A 27 (1996) 3761- 3772. [29] Z. Zhang, H.M. Flower, Composition and lattice parameters of silicide and matrix in cast Ti–Si–Al–Zr alloys, Mater. Sci. Technol. 7 (1991) 812-817. [30] A.M. Chaze, C. Coddet, Influence of silicon on the oxidation of titanium between 550 and 700oC, Oxid Met. 27 (1987) 1-20. [31] R.T. Downs, D.C Palmer, The pressure behavior of α cristobalite, Am. Mineral. 79, (1994) 9-14. [32] A. Abba, A. Galerie, M. Caillet, High-temperature oxidation of titanium silicide coatings on titanium, Oxid. Met. 17 (1982) 43-54. [33] F.N. Schwettmann, R.A. Graft, M. Kolodney, Mechanism of the oxidation of titanium disilicide, J. Electrochem. Soc. 118 (1971) 1973-1977. [34] P. Schmidt, M. Binnewies, R. Glaum, M. Schmidt, Chemical Vapor transport reactions–methods, Materials, Modeling, in: S.O. Ferreira (Edt), Advanced Topics on Crystal Growth, InTech, Rijeka, Croatia, 2013, pp. 227- 305. [35] M. Mitoraj-Krolikowska, E. Godlewska, Hot corrosion behaviour of (γ + α2)-Ti-46Al-8Nb (at.%) and α-Ti-6Al-1Mn (at.%) alloys, Corros.Sci. 115 (2017) 18–29. [36] Y. Qiao, X. Guo, X. Li, Hot corrosion behavior of silicide coating on an Nb–Ti–Si based ultrahigh temperature alloy, Corros.Sci. 91 (2015) 75–85. [37] I. Barin, O. Knacke, Thermochemical properties of inorganic substances, Springer-Verlag, Berlin, 1973. [38] I. Barin, O. Knacke, Thermochemical properties of inorganic substances - Supplement, SpringerVerlag, Berlin, 1977. [39] M.W. Chase, Nist-Janaf thermochemical tables, Fourth Edition, Monograph 9 (Part I and Part II) (1998)
Figures
Fig. 1 Backscattered electron image of Ti-46Al-8Ta (at. %) with a Ti-10Si layer deposited by magnetron sputtering.
Fig. 2 Top view of as-received silicide coating a) digital picture, b) SEM image of the surface.
Fig. 3 Cross-section of a Ti–46Al–8Ta (at. %) with a silicide coating with EDS analysis in marked points (Table 1).
Fig. 4 XRD spectra from the surface of the as-coated sample and after removal of consecutive layers of the silicide coating.
Fig. 5 Gross mass changes (sample with a crucible) of the Ti–46Al–8Ta (at. %) samples with and without silicide coating after hot corrosion test in air at 800 oC with different salt deposits.
Fig. 6 Backscattered electron images of samples with a silicon-rich coating contaminated with NaCl (a, b); Na2SO4 (c, d) and Na2SO4 + NaCl (e, f) – at different magnifications - after 80 h of oxidation in air at 800 oC. EDS analysis in marked points (see Table 1)
Fig. 7 XRD spectra from the surface of coated sample after 80 h of exposure; experimental conditions: 800 oC, 20-h cycles, laboratory air, NaCl or Na2SO4 deposits.
Fig. 8 XRD spectra from the surface of coated sample after 80 h, 160 h and 300 h of oxidation tests at 800 oC with a Na2SO4 + NaCl deposit.
Fig. 9 Cross-sections of coated samples with a) NaCl, (b, c) Na2SO4 and d) Na2SO4 + NaCl deposits after 80 h of oxidation in air at 800 oC with EDS analysis in marked points (Table 1).
Fig. 10 Net mass changes of the Ti–46Al–8Ta (at. %) samples with silicide coating after hot corrosion test in air at 800 oC with different salt deposits.
NaCl
Na2SO4
Na2SO4+NaCl
a)
b)
Fig. 11 Digital pictures of samples contaminated with NaCl or Na2SO4 or Na2SO4+ NaCl after 300 h of oxidation in air at 800 oC without (a) and with (b) silicide coating.
Without coating
With coating
Fig. 12 Cross-sections of samples with and without silicide coating contaminated with a NaCl deposit after 300 h of oxidation in air at 800 oC (20 h cycles).
Fig. 13 XRD spectra from the surface of uncoated sample and spalled scale collected in alumina crucible (sample with NaCl) after 300 h of exposure at 800 oC, 20-h cycles, laboratory air, NaCl, Na2SO4 or Na2SO4+NaCl deposits.
Tables Table 1 Local or average chemical composition of samples shown in Figs. 3, 6, 9 (EDS). Fig. No
Type of analysis Point/ average
Ti
Fig. 3
1 2 3 4 5 6
37.6 35.2 59.9 59.2 28.5 41.8
14.0 0.4 48.0 20.7 7.2 36.9 3.1 13.1 23.9 5.4 12.1 23.3 59.2 6.8 5.5 49.6 8.6 0.0
Fig. 6a
average
8.3
3.8
Fig. 6b
1 2
Fig. 6c
average
5.1
1.4
-
10.7 18.5 55.6
Fig. 6d
3 4
6.1 12.7
0.8 0.8
-
5.2 22.5 47.2 18.2 10.8 8.8 62.8 4.1
Fig. 6e
average
23.7
2.7
1.0
5.8
3.5
63.4
-
Fig. 6f
5
2.5
31.8
0.0
1.2
1.2
63.3
-
Fig. 9c
1 2 3
57.6 1.9 12.7 27.8 62.9 3.9 12.7 20.5 30.0 65.5 4.5 0.0
-
-
-
Fig. 9d
1 2 3 4 5 6 7 8 9 10
7.7 17.0 21.6 55.2 52.7 64.7 26.8 28.1 35.3 42.1
2.0 1.3 0.2 1.1 1.2 2.3 0.0 0.0 0.0 0.0
64.3 58.6 42.4 19.1 16.5 -
-
Al
12.4 3.4 3.8 10.8
Ta
at. % Si
Na
O
S
-
-
-
-
20.2
5.5
62.2
-
0.0 -
11.1 19.6
7.5 8.2
65.6 57.6
-
2.0 2.4 23.1 2.5 1.8 9.5 2.7 4.8 13.9 12.0 67.5 5.7 67.4 4.5 55.6 9.1 49.6 8.3
24.0 20.7 10.2 13.3 22.1 7.1 0.0 0.0 0.0 0.0
8.7
Table 2 Standard Gibbs free energy of formation at 1100 K in kJ/mol Cl2 [39]. ΔGo [kJ/mol Cl2] AlCl3 TiCl4 TaCl5
-351.58 -315.32 -226.80
SiCl4
-259.04
Table 3 Standard Gibbs free energy of reactions at 1100 K [37, 38, 39]
Reaction number 1 2 3 4 5 6 7 8 9 10 11
Reaction ΔGor at 1100K [kJ] -495.9 TiSi2 + 3O2 → TiO2 + 2SiO2 -528.8 TiSi + 2O2 → TiO2 + SiO2 -463.0 2TiSi2 + 4O2 → TiO2 + 3SiO2 + TiSi -202.8 6TiSi2 + 10 O2 → TiO2 + 9SiO2 + Ti5 Si3 396.5 Na2 SO4 → Na2 O + SO3 N/A Ta2 O5 + Na2 O → 2NaTaO3 1 -439.2 TiO2 + 2NaCl + O2 → Na2 TiO3 + Cl2 2 -396.9 TiO2 + Na2 SO4 → Na2 TiO3 + SO3 1 -698.4 2SiO2 + 2NaCl + O2 → Na2 Si2 O5 + Cl2 2 -656.0 2SiO2 + Na2 SO4 → Na2 Si2 O5 + SO3 N/A SiO2 (s) + 2TaCl5 (g) → SiCl4 (g) + 2TaOCl3 (g)
S0010-938X(16)30845-9 http://dx.doi.org/doi:10.1016/j.corsci.2017.02.002 CS 6995
To appear in: Received date: Revised date: Accepted date:
27-9-2016 30-1-2017 2-2-2017
Please cite this article as: K.Rubacha, E.Godlewska, K.Mars, Behaviour of a siliconrich coating on Ti-46Al-8Ta (at.%) in hot-corrosion environments, Corrosion Science http://dx.doi.org/10.1016/j.corsci.2017.02.002 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Title: Behaviour of a silicon-rich coating on Ti-46Al-8Ta (at. %) in hot-corrosion environments Authors: K. Rubacha, E. Godlewska, K. Mars AGH University of Science and Technology, Faculty of Materials Science and Ceramics, Al. A. Mickiewicza 30, 30-059 Krakow, Poland
Corresponding author: Katarzyna Rubacha Address: AGH UST Al. Mickiewicza 30, 30-059 Krakow, Poland Phone number: +48 126174787 E-mail: [email protected]
Highlights
Adherent silicide coating was deposited on Ti-46Al-8Ta alloy in two-step process. Cyclic oxidation tests with different salt deposits were conducted. Silicon-rich coating had beneficial effect on oxidation resistance. Diffusion of reactive gases to substrate was prevented.
Abstract In this work silicon-rich coating was deposited on a Ti-46Al-8Ta (at. %) alloy in a two-step procedure comprising physical and chemical vapour deposition (magnetron sputtering and pack cementation, respectively). Protective properties of the coating were tested through cyclic oxidation in air at 800 oC with samples contaminated with NaCl, Na2SO4 or a mixture of these two salts. The coated samples exhibited better hot corrosion resistance compared with the uncoated reference. This was attributed to the formation of an oxide layer composed of amorphous silica with embedded rutile and α cristobalite crystals, constituting a barrier against oxygen penetration. Keywords: A. Intermetallics, B. SEM, Thermal cycling, C. Hot corrosion 1. Introduction Titanium alloys and titanium aluminides have been under investigation for at least twenty years owing to the advantageous combination of physical, mechanical and chemical properties. Intermetallic phases in the Ti-Al binary system attract attention because of relatively low density, very high specific strength and oxidation resistance at moderate temperatures [1-4]. A lot of work has been devoted to improve mechanical properties and oxidation resistance of titanium alloys at elevated temperatures. Particularly good results were obtained by alloying with Nb, Mo, Ta, Si or W [5]. Alloys with high percentage of aluminium (up to about 45 at. %) and ternary additions of Nb and/or Ta are among the most advanced intermetallic materials with good oxidation resistance [6-9]. This can be further improved by surface engineering. A great
deal of metallic and ceramic coatings and deposition techniques have been reported to provide some degree of protection against oxidation [10-14]. By far more complex problems appear, however, when mineral deposits such as NaCl and Na2SO4 are in contact with Ti-Al alloys exposed to oxidising atmospheres, as described previously [15, 16]. As far as hot corrosion protection of intermetallic alloys is concerned, very limited number of investigations have been undertaken and reported in the literature. For illustration, Ti50Al was coated with TiAlCr [17] or Al2O3 and enamel coating [18] which had beneficial effect on the oxidation resistance in a hot-salt environment but during long-term cyclic oxidation tests, the initially protective layer cracked and spalled off. As a result, the material corroded at an accelerated rate. Some studies were also performed with more advanced alloys [19, 20]. For illustration, the surface of Ti–48Al– 2Cr–2Nb (wt. %) was modified with gold by electrochemical deposition followed by vacuum annealing [19]. The layer of gold was thought an ideal barrier for hot corrosion because of chemical inertness of this noble metal and its very low solubility in molten salts. Indeed the layer provided the desired protection to the alloy but due to high prices of gold its common use is doubtful. The most recent studies concerning hot corrosion protection of the advanced titanium aluminide alloys are focused on composite coatings with alternate Y2O3-doped Al2O3 ceramic nano-layers and Pt–Au binary alloy nano-layers [20]. It is required that the proposed composite coating consist of at least seven layers to provide satisfactory protection which makes the deposition process quite complex and time consuming. Coatings containing silicon might be an interesting option to enhance high-temperature oxidation resistance by forming a protective SiO2 scale [21-23]. It has been reported [21] that a Si diffusion coating significantly improves scale adherence and a Si-Al coating has excellent resistance to oxidation. Interestingly, with increasing exposure time the protective properties of the silica scale become better [22]. It has been demonstrated also that silica is highly stable in molten salts such as chlorides and sulphates [24]. Titanium silicide layers were successfully produced on pseudo-α titanium alloys by pack cementation [25, 26]. Preliminary trials of using the same technology to develop silicide coatings on Ti46Al-8Ta (at. %) appeared not fully successful because of Kirkendall porosity around the coating/substrate interface. In the approach proposed in this work, alloy surface will be first enriched in titanium to prevent excessive loss of titanium from the substrate and next diffusion coated with silicon to yield a layer with the matrix based on titanium silicides modified by the components of the substrate.
2. Experimental procedure 2.1 Sample preparation The Ti–46Al–8Ta (at. %) ingots for hot corrosion studies were cut with a diamond saw into samples in a form of pellets, 13 mm in diameter and 1-2 mm in thickness, which were ground with SiC paper to 2000 grit and polished using water-based diamond suspensions (grain size 3 µm and 1 µm). Each sample was measured and weighed. Mirror-like surfaces were subsequently cleaned in acetone and dried prior to coating deposition. 2.2 Coating deposition
Silicon-rich layers were produced in an innovative two-stage procedure. The first stage involved modification of alloy surface with a titanium-silicon solid solution by magnetron sputtering using TEPRO NP 501A Koszalin equipment. Composition of the target used in this experiment was 90 at. % Ti, and 10 at. % Si. It was produced by sintering elemental powders at 1300 oC for 1 h under a pressure of 25 MPa. Sputtering conditions were the following: pressure 0.24 Pa, current 0.6A, temperature of sample holder 380 o
C. In the second stage, a silicide coating was grown by pack cementation. The samples were buried in a
powder mixture consisting of 54.58 mol. % silicon and 44.96 mol. % alumina (Al2O3) and halide activator, cryolite (Na3AlF6), in the amount of 0.46 mol. %. The whole assembly was maintained at 850 oC under argon for 6 h. After deposition the samples were weighed again. 2.3 Hot corrosion tests Samples with silicide coatings were immersed in an aqueous solution 10% by mass of Na2SO4 (0.77 mol/dm3) or NaCl (1.83 mol/dm3) or a mixture of two salts Na2SO4 (0.67 mol/dm3) + NaCl (0.22 mol/dm3) and dried in air. Average amount of the salt deposit was about 1 mg/cm2. The samples were placed in ceramic crucibles in order to collect spalled matter. Reference samples without the silicide coating but with salt deposits were prepared in the same way. The temperature of cyclic oxidation tests was 800 oC. Each cycle consisted of rapid heating (samples were introduced into hot furnace), exposure at constant temperature for 20 h and rapid cooling to room temperature in the laboratory air (samples were removed from the furnace). Total hot-dwell time was 80 h or 300 h. After every cycle the samples were weighed with an accuracy of 10-4 g to evaluate mass variations over time. 2.4 Post treatment examination The multilayer silicide coating was analysed by grazing incidence X-ray diffraction (GID) after removal of consecutive layers by grinding. Phase composition of samples surface after hot corrosion tests was determined by X-ray diffraction (PanalyticalX’Pert) in usual Bragg-Brentano configuration. Surfaces and transversal cross-sections of the samples were analysed by means of scanning electron microscopy (Nova Nano SEM 200 FEI Europe Company) combined with the energy dispersive X-ray spectrometry (EDS/EDAX).
3. Results 3.1 Coating characterization A cross-section of sample after the first stage of coating deposition is shown in Fig. 1. The layer is adherent and uniform with a thickness of about 2 µm. The role of titanium-silicon interlayer deposited by magnetron sputtering was to enrich alloy surface in titanium. Previous experiments without the Ti-Si interlayer ended up with poor coating adherence to the substrate, as a result of enhanced outward diffusion of titanium from the near-surface zone of the alloy. Fig. 2a presents a digital picture of the sample with a silicide coating. As can be seen in Fig. 2b (SEM image at a higher magnification), the surface is quite rough and consists of small equiaxed grains, average diameter of 1.5 µm. According to EDS the external surface is composed of titanium and silicon in an atomic ratio of about 1:2 which would correspond to TiSi2. Fig. 3
presents a cross-section of the silicide coating. The coating is well bonded with the substrate and its overall thickness is about 60 µm. It has clearly a layered structure and consecutive layers differ in composition. According to EDS (Table 1, Fig. 3) the outermost part of the coating consists mainly of silicon and titanium; the concentration of titanium increasing and that of silicon decreasing from the surface inwards. In points 3 and 4 the concentration of tantalum is higher than in other areas. The dark grey zone close to the alloy (point 5) is significantly enriched in aluminium and the Ti to Al ratio corresponds to TiAl 2. These results are roughly consistent with XRD measurements (Fig. 4). The occurrence of TiSi2 together with TiSi on the surface was confirmed. After removal of the outermost layers of the coating, the XRD patterns revealed Ti5Si4 and Ti5Si3. Even though EDS indicated tantalum enrichment at some distance from the surface, the tantalum-containing phases were not marked on the diffractogram because reflections corresponding to titanium silicides and tantalum silicides e.g. Ti5Si3 and Ta5Si3 are indistinguishable. It is possible therefore that either both titanium and tantalum silicides with the same crystallographic structure were present or a solid solution (Ti, Ta)5Si3. In the inner layers of the coating, reflections corresponding to TiAl2 were found along with titanium aluminides from the substrate: -TiAl and 2-Ti3Al. 3.2 Hot corrosion – thermogravimetric studies at 800 oC Gross mass changes of samples with and without silicide coatings during cyclic oxidation tests in air at 800 oC in the presence of different salt deposits are shown in Fig. 5. The most aggressive environment for the uncoated alloy was sodium chloride bringing about the largest mass gains and severe scale spallation. Protective properties of the silicide coating are unquestionable in that case: mass gain dropped from about 20 mg/cm2 (uncoated sample) to 5 mg/cm2 (coated sample) and scale adherence was much better. For the sample contaminated with sodium sulphate, the silicide coating appeared also effective: mass gain decreased from 1.5 to 0.7 mg/cm2. In the case of mixed salt deposit, no particular effect of the coating on gross mass changes was identified. After 300 h of cyclic oxidation tests mass gains of the two types of samples were almost the same. 3.3 Hot corrosion - surface morphology and chemical composition Surfaces of samples with silicide coatings after 80 h of oxidation at 800 oC in the presence of salt deposits differ significantly in morphology and average chemical composition (Table 1, Fig. 6a, c, e). The biggest amount of silicon was detected on samples with the silicide coating and the NaCl deposit. SEM images in Fig. 6a, b show some spherical particles sticking to the sample surface which is mostly covered with small, needle-shaped crystals. According to EDS (Table 1, point 1 in Fig. 6b) the outermost layer of the sample is composed of a mixture of silicon and titanium oxides. The concentration of aluminium is elevated in the round-shaped particles. Needle-shaped crystals covered with a smooth discontinuous top layer are found on the sample with a silicide coating contaminated with sodium sulphate (Fig. 6c, d). From Table 1, Fig. 6d (point 3 and 4) it
follows that the smooth areas have composition corresponding to the salt deposit and the crystals are a mixture of titanium and silicon oxides. Fig. 6e, f presents external surface of the silicide coating after 80 h oxidation in the presence of salt mixture, Na2SO4 + NaCl. The surface is covered with loosely bonded submicrometric crystals and some larger protrusions, which at a higher magnification appear petal-like crystal agglomerates, with an elevated concentration of aluminium (Table 1, Fig. 6f point 5).
3.4 Hot corrosion - surface phase composition Fig. 7 shows diffraction patterns from the surfaces of samples with a silicide coating and either NaCl or Na2SO4 deposit, tested for 80 h at 800 oC. These confirm that phases present on the surface are mainly SiO2 (α cristobalite structure) and TiO2 (rutile). A few peaks corresponding to TiSi2 (external layer of the coating) are also visible. For the sample contaminated with Na2SO4, some reflections from the salt residue are found in addition. XRD patterns from the surface of samples with a silicide coating and Na2SO4 + NaCl deposit after oxidation for 80 h, 160 h and 300 h are presented in Fig. 8. It can be seen that during oxidation the surface composition undergoes changes. After 80 h the phases present on the surface were Al2O3 (corundum), TiO2 (rutile), SiO2 (α cristobalite) and NaTaO3. After 160 h, NaTaO3 and Al2O3 were not identified. Al2O3, TiO2, SiO2 were found after 300 h of oxidation. 3.5 Hot corrosion – sample cross-sections Cross-sections of coated samples after 80 h of hot corrosion test can be seen in Fig. 9. In all cases the scale consisted of an amorphous silica matrix with dispersed TiO2 (rutile) and α cristobalite crystals. Its thickness was about 2 µm for Na2SO4 and 10 µm for Na2SO4 + NaCl. In the latter case, a smooth grey layer with a substantially higher concentration of aluminium could be seen under the scale (Table 1, Fig. 9d point 3). Generally, below the scale, there was a layer of unconsumed coating which mostly consisted of TiSi and TiSi2. Some cracks propagating parallel to the surface were probably formed during sample preparation for the analysis. At a bigger distance from the surface, there was a bright layer with an elevated concentration of tantalum (about 12 at. %) together with titanium (60 at. %) and silicon (25 at. %) (Fig. 9a, b and Table 1, Fig. 9c point 1, 2). The dark layer at the coating/substrate interface contained high amounts of aluminium and some tantalum-rich precipitates. Cross-section presented in Fig. 9d is somewhat different. After exposure, the coating with a mixed salt deposit appeared less compact, especially in the near-surface region. The bright layer visible in Fig. 9a, b is not present here. According to EDS (Table 1, Fig. 9d) the concentration of tantalum locally reaches the highest value of 12 at. %, as in point 6. 4. Discussion In a recent paper [15] it has been demonstrated that exposure of a Ti-46Al-8Ta (at. %) alloy contaminated with salt deposits to air at elevated temperatures leads to remarkable material losses.
Especially harmful was sodium chloride which brought about degradation by a self-sustaining mechanism involving formation of volatile species, porosity and poor adherence of the scale. The novel two-step procedure applied in this work resulted in an adherent multiphase coating which was formed in a complex diffusion process involving predominant inward diffusion of silicon and some outward diffusion of titanium. From the surface inwards, the coating consists of TiSi2/TiSi; Ti5Si4/Ti5Si3 probably with tantalum as a ternary addition; and aluminium-enriched zone TiAl2 (EDS Fig. 3 and XRD Fig. 4). This phase composition is in good agreement with the previously reported results [27, 28]. There are, however, some differences, e.g. Ti3Si on the pure Ti substrate in [27] or Ti(Al,Si)3 between Ti5Si4 and TiAl2 on a Ti-50Al alloy in [28]. The porous layer in the middle of the coating can be attributed to segregation of aluminium which is insoluble in all titanium silicides except Ti5Si3 [29]. The origin of aluminium in the amount of 20 at. % in the outer part of the coating (Table 1, Fig. 3 point 1 and 2) might be surprising. Noteworthy fact is that the porous region of the coating does not contain aluminium. As suggested earlier [28], at a high temperature aluminium can vaporize from TiAl3 at the TiAl3/Ti5Si4 interface and probably also from the zone where its segregation from the silicide phases initially occurred. The composition of salt deposit is an important factor when considering sample degradation in hot corrosion tests. Among salt deposits used in this work, sodium chloride appeared the most aggressive, as evidenced in Fig. 10 by net mass changes of the coated Ti–46Al–8Ta (at. %) samples. The silicide coating did not stop gradual mass losses due to scale exfoliation and possibly also evaporation of some volatile reaction products. However, compared with the uncoated alloy, these mass losses in a similar test were generally lower (e.g. 2 times for the samples with NaCl) [15]. In all experiments after about 100 h the mass of samples became essentially stable which is understandable in view of the fact that salt deposit was not reapplied during the test. Top views of samples after 300 h of hot corrosion tests at 800 oC with and without the silicide coatings are presented in Fig. 11. In the case of NaCl deposit, protective properties of the coating could be seen even with the naked eye. Degradation of the coated samples was significantly limited. The coated samples with the Na2SO4 deposit had much thinner scale compared with the uncoated reference and apparently no scale spallation occurred. After 300 h of the test with a mixed salt deposit, the scale on uncoated samples was thin and not uniform, the sample surface underneath was distinctly visible. On coated samples a relatively thick scale formed and only slight exfoliation of the scale was observed. Cross-sections of samples after 300 h of oxidation in the presence of NaCl with and without coating are shown on Fig. 12. The uncoated sample is seriously damaged. Many cracks caused decohesion of its external part. The extent of chemically changed material reaches over 100 µm. In contrast, the coated sample is in good condition after the same time of oxidation. The coating is adherent and continuous, without macroscopic defects. After 300 h of the test it still provides protection to the substrate. Longer oxidation tests would be needed, however, to assess its lifetime. The above-described results indicate that the silicide coatings had a beneficial effect on the Ti-46Al8Ta (at. %) alloy in the hot corrosion environments. The protective properties were attributed to the presence
of SiO2 (in a form of α cristobalite) and TiO2 (rutile) on the surface. The occurrence of cristobalite which is the high-temperature form of silicon dioxide might be surprising but already in 1980’s rutile and cristobalite were identified in the oxide scale on a Ti-Si alloy [30]. The tetragonal α cristobalite can exist in a metastable state at temperatures below 500 K [31]. There are also reports on the presence of amorphous silica on the surface of oxidized Ti5Si3 [32] and TiSi2 [33]. According to some unpublished results received at our lab, amorphous silica is also found in the scale. We suppose therefore that amorphous silica and α cristobalite, both, are present on the surface of coated samples after the test. The latter has been confirmed by XRD. The silica glass probably seals small discontinuities and porosity in the oxide layer, which contributes to its better gas tightness. Moreover, according to some reports, silicon can form solid solutions with rutile. It is claimed that small atoms of Si take interstitial positions in the crystal lattice of rutile, decrease the concentration of oxygen vacancies and thereby make diffusion of oxygen slower [30]. Worth noticing are also diffusion coefficients of oxygen in silica and rutile: 5∙10-16 cm2s-1 and 1∙10-11 cm2s-1, respectively [32]. Since the diffusion coefficient of oxygen in rutile is by five orders of magnitude bigger than in silica, it can be assumed that oxygen diffuses through the rutile phase mainly (along the grain boundaries and via oxygen vacancies in the crystal lattice). All the mentioned factors are beneficial for oxidation resistance of the coated samples. For comparison Fig. 13 presents XRD patterns taken from the surface of uncoated samples tested at 800 oC for 300 h with different salt deposits and from the scale collected in the crucible in the case of sample contaminated with NaCl. The only phase identified in the powdered scale was TiO 2. The only oxides identified on the alloy surface in NaCl environment were: TiO2 and Ta2O5. Phases found on the surface of samples oxidised with Na2SO4 or Na2SO4+ NaCl deposits were: TiO2, Ta2O5, Na2Ta4O11 and TiAl from the alloy. Additional scale component was Al2O3 when Na2SO4 alone was deposited. The assessment of degradation mechanism of the silicide coatings is beyond the scope of this study; however the results of hot-corrosion tests shed light on their overall performance and durability. Some of the thermodynamically possible reactions are presented in the following section. As the coating surface after the hot corrosion tests was covered with mixed oxides, the following reactions could take place [33]: TiSi2 + 3O2 → TiO2 + 2SiO2
(1)
TiSi + 2O2 → TiO2 + SiO2
(2)
or 2TiSi2 + 4O2 → TiO2 + 3SiO2 + TiSi
(3)
6TiSi2 + 10 O2 → TiO2 + 9SiO2 + Ti5 Si3
(4)
On the samples contaminated with mixed salt deposits, Na2SO4 + NaCl, additional phases detected by XRD were Al2O3 and NaTaO3 (Fig. 8). The latter could be formed in the following reactions involving decomposition of Na2SO4 at high temperature [15]: Na2 SO4 → Na2 O + SO3 Ta2 O5 + Na2 O → 2NaTaO3
(5) (6)
The scale can further react with salt deposits forming titanates or silicates which is accompanied by release of chlorine gas or sulphur (VI) oxide. 1
TiO2 + 2NaCl + 2 O2 → Na2 TiO3 + Cl2
(7)
TiO2 + Na2 SO4 → Na2 TiO3 + SO3
(8)
1
2SiO2 + 2NaCl + 2 O2 → Na2 Si2 O5 + Cl2
(9)
2SiO2 + Na2 SO4 → Na2 Si2 O5 + SO3
(10)
The gaseous species resulting from the above reactions (Eqs. 7-10) can penetrate inward via various defects and react with the coating or alloy components. From the point of view of thermodynamics, the most probable among chloride phases is AlCl3 as its standard Gibbs free energy of formation is the most negative (see Table 2). The TiO2/SiO2 barrier is not totally gas-tight. As already mentioned, oxygen may diffuse through rutile grain boundaries and oxygen vacancies in the crystal lattice. Also silicon dioxide may react with tantalum (V) chloride to form volatile species, Eq. (11) [34]. SiO2 (s) + 2TaCl5 (g) → SiCl4 (g) + 2TaOCl3 (g)
(11)
Furthermore oxygen and nitrogen may penetrate into the coating and react with its components, yielding multiple nitride, oxide or oxynitride phases. Table 3 shows standard Gibbs free energy of reactions (1-11) at 1100 K. For reactions (6) and (11) no thermodynamic data are available, however XRD analyses evidenced the presence of NaTaO3 in a scale with Na2SO4 + NaCl deposit. Almost all reactions presented in Table 3 have negative values of standard Gibbs free energy which means that they are thermodynamically favourable. Calculations of ΔGor provide positive value for decomposition of Na2SO4. Despite this fact, reaction (5) may occur which is explained in [35]. Upon oxidation, some compositional changes take place in the coating layer. Undoubtedly, titanium diffuses outwards leaving behind a depleted zone at the coating/substrate interface. The mobility of tantalum is very limited in titanium aluminides. This may explain why Ta-rich intermetallic appears between the layer of titanium silicide and the titanium depleted zone. To the best of our knowledge the studies on hot-corrosion behaviour of silicide coatings on advanced titanium alloys are almost non-existent in the literature. There is one paper published in 2015 [36], describing deposition of a silicide coating by halide activated pack cementation on the alloy with a nominal composition of Nb–20Ti–15Si–5Cr–3Hf–3Al (at. %). The researchers proposed a plausible mechanism of hot corrosion processes occurring beneath the layer of molten salt composed of Na2SO4 and NaCl. It has been stated that NaAlO2, AlNaSiO4, and Na2TiSi2O7 are formed along with TiO2 and SiO2. This scheme, however, is not comparable with the results obtained in this work. In particular, coating microstructure was different and neither of the aluminate or silicate phases was identified in the corrosion products.
5. Conclusions
1. The proposed two-step procedure enabled formation of an adherent silicide coating on the Ti46Al-8Ta (at. %) substrate. The coating was multiphase, consisting of TiSi 2/ TiSi in the outer part, lower silicides (Ti, Ta)xSiy and titanium aluminides TiAl2/TiAl3 in the inner part. 2. The silicide coating prevented inward diffusion of oxygen and chlorine to the titanium aluminide substrate. 3. Protective properties of the silicon-rich coating result from the formation of amorphous silicon oxide layer with embedded rutile and α cristobalite crystals, constituting an effective barrier against oxygen penetration. 4. After 300 h of the interrupted oxidation tests at 800 oC in the presence of mineral deposits, the silicide coating was preserved and did not show signs of any significant destruction, except some thickness loss due to oxidation on the surface. It was compact and adherent. Longer oxidation tests are needed to asses coating lifetime in hot corrosion conditions.
Acknowledgements The authors gratefully acknowledge financial support from National Science Centre project number 2013/11/N/ST8/01345. Alloy ingots were produced by ACCESS (Aachen).
References [1] X. Wu, Review of alloy and process development of TiAl alloys, Intermetallics 14 (2006) 1114-1122. [2] E.A. Loria, Gamma titanium aluminides as prospective structural materials, Intermetallics 8 (2000) 1339-1345. [3] M. Yamaguchi, H. Inui, K. Ito, High – temperature structural intermetallics, Acta Mater. 48 (2000) 307322. [4] S. Djanarthany J.-C. Viala, J. Bouix, An overview of monolithic titanium aluminides based on Ti3Al and TiAl, Mater. Chem. Phys. 72 (2001) 301–319. [5] Y. Shida, H. Anada, The effect of various ternary additives on the oxidation behavior of TiAl in high temperature air, Oxid. Met. 45 (1996) 197–218. [6] D. Vojtěch, T. Popela, J. Kubásek, J. Maixner, P. Novák, Comparison of Nb- and Ta-effectiveness for improvement of the cyclic oxidation resistance of TiAl-based intermetallics, Intermetallics 19 (2011) 493-501. [7] M. Mitoraj, E. Godlewska, O. Heintz, N. Geoffroy, S. Fontana, S. Chevalier, Scale composition and oxidation mechanism of the Ti-46Al-8Nb alloy in air at 700 and 800 oC, Intermetallics 19 (2011) 39-47. [8] M. Mitoraj, E. Godlewska, Oxidation of Ti–46Al–8Ta in air at 700 °C and 800 °C under thermal cycling conditions, Intermetallics 34 (2013) 112–121.
[9] E. Godlewska, M. Mitoraj, F. Devred, B. E. Nieuwenhuys, Reactivity of a Ti–45.9Al–8Nb alloy in air at 700–900°C, J. Therm. Anal. Calorim. 88 (2007) 1, 225–230. [10] E. Godlewska, M. Mitoraj, R. Mania, R. Gajerski., Oxidation of a Ti-46Al-8Nb alloy modified with some Ni-base coatings, Ann. Chim. - Sci. Mat. 33 (2008) 9–16. [11] D.B Wei, P.Z. Zhang, Z.J. Yao, J.T. Zhou, X.F. Wei, P. Zhou, Cyclic oxidation behavior of plasma surface chromising coating on titanium alloy Ti–6Al–4V, Appl. Surf. Sci. 261 (2012) 800 – 806. [12] J. Kong, C. Zhou, S. Gong, H. Xul., Low-pressure plasma-sprayed Al–Cu–Fe–Cr quasicrystalline coating for Ti-based alloy oxidation protection, Surf. Coat. Tech. 165 (2003) 281–285. [13] M. Góral, L. Swadzba, G. Moskal, M. Hetmanczyk, T. Tetsui, Si-modified aluminide coatings deposited on Ti46Al7Nb alloy by slurry method, Intermetallics 17 (2009) 965–967. [14] S. Teng, W. Liang, Z. Li, X. Ma, Improvement of high-temperature oxidation resistance of TiAl-based alloy by sol–gel method, J. Alloy. Compd. 464 (2008) 452-456. [15] E. Godlewska, M. Mitoraj, K. Leszczynska, Hot corrosion of Ti-46Al-8Ta (at. %) intermetallic alloy, Corros.Sci. 78 (2014) 63-70. [16] K. Zhang, Z. Li, W. Gao, Hot corrosion behaviour of Ti–Al based intermetallics, Mater. Lett. 57 (2002) 834–843. [17] Z. Tang, F. Wang, W. Wu, Effect of a sputtered TiAlCr coating on hot corrosion resistance of gammaTiAl, Intermetallics 7 (1999) 1271-1274. [18] Z. Tang, F. Wang, W. Wu, Effect of Al2O3 and enamel coatings on 900°C oxidation and hot corrosion behaviors of gamma-TiAl, Mat. Sci. Eng. A-Struct. 276 (2000) 70 - 75. [19] M. Bacos, M. Thomas, J.-L. Raviart, A. Morel, S. Mercier, P. Josso, Influence of an oxidation protective coating upon hot corrosion and mechanical behavior of Ti-48Al-2Cr-2Nb alloy, Intermetallics 19 (2011) 1120-1129. [20] X. Ma, Y. He, J. Lin, D. Wang, J. Zhang, Effect of a magnetron sputtered (Al 2O3– Y2O3)/(Pt– Au) laminated coating on hot corrosion resistance of 8Nb – TiAl alloy, Surf. Coat. Tech. 206 (2012) 2690– 2697. [21] H. Zheng, L. Xiong, Q. Luo, S. Lu, Microstructure and cyclic oxidation behavior of silicide coating and Si-Al coating on Cr-50Nb alloy, Corros.Sci. 89 (2014) 326-330. [22] N. Chaia, Y. Bouizi, S. Mathieu, M. Vilasi, Isothermal and cyclic oxidation behaviour of hot-pressed MSi2 compounds (with M = V, Ti, Cr), Intermetallics 65 (2015) 35 – 41. [23] S. Majumdar, J. Kishor, B. Paul, R.C. Hubli, J.K. Chakravartty, Isothermal oxidation behavior and growth kinetics of silicide coatings formed on Nb–1Zr–0.1C alloy, Corros. Sci. 95 (2015) 100–109. [24] T. Ishitsuka, K. Nose, Stability of protective oxide films in waste incineration environment – solubility measurement of oxides in molten chlorides, Corros. Sci. 44 (2002) 247-263. [25] E. Godlewska, M. Mitoraj, S. Kozinski, Influence of diffusion silicide coatings on the oxidation behavior of a Ti-6Al alloy in a cyclic oxidation test, Polish Ceramic Bulletin, Ceramics, 91/1 (2005) 551-558.
[26] E. Godlewska, M. Mitoraj, S. Kozinski, Deposition of silicide coatings on Ti-6Al alloy using pack cementation method, Polish Ceramic Bulletin, Ceramics, 91/1 (2005) 543-550. [27] B. Cockeram, R. Rapp, The kinetics of multilayered titanium-silicide coatings grown by the pack cementation method, Metall. Mater. Trans. A 26 (1995) 777- 791. [28] T. Munro, B. Gleeson, The deposition of aluminide and silicide coatings on γ-TiAl using the halideactivated pack cementation method, Metall. Mater. Trans. A 27 (1996) 3761- 3772. [29] Z. Zhang, H.M. Flower, Composition and lattice parameters of silicide and matrix in cast Ti–Si–Al–Zr alloys, Mater. Sci. Technol. 7 (1991) 812-817. [30] A.M. Chaze, C. Coddet, Influence of silicon on the oxidation of titanium between 550 and 700oC, Oxid Met. 27 (1987) 1-20. [31] R.T. Downs, D.C Palmer, The pressure behavior of α cristobalite, Am. Mineral. 79, (1994) 9-14. [32] A. Abba, A. Galerie, M. Caillet, High-temperature oxidation of titanium silicide coatings on titanium, Oxid. Met. 17 (1982) 43-54. [33] F.N. Schwettmann, R.A. Graft, M. Kolodney, Mechanism of the oxidation of titanium disilicide, J. Electrochem. Soc. 118 (1971) 1973-1977. [34] P. Schmidt, M. Binnewies, R. Glaum, M. Schmidt, Chemical Vapor transport reactions–methods, Materials, Modeling, in: S.O. Ferreira (Edt), Advanced Topics on Crystal Growth, InTech, Rijeka, Croatia, 2013, pp. 227- 305. [35] M. Mitoraj-Krolikowska, E. Godlewska, Hot corrosion behaviour of (γ + α2)-Ti-46Al-8Nb (at.%) and α-Ti-6Al-1Mn (at.%) alloys, Corros.Sci. 115 (2017) 18–29. [36] Y. Qiao, X. Guo, X. Li, Hot corrosion behavior of silicide coating on an Nb–Ti–Si based ultrahigh temperature alloy, Corros.Sci. 91 (2015) 75–85. [37] I. Barin, O. Knacke, Thermochemical properties of inorganic substances, Springer-Verlag, Berlin, 1973. [38] I. Barin, O. Knacke, Thermochemical properties of inorganic substances - Supplement, SpringerVerlag, Berlin, 1977. [39] M.W. Chase, Nist-Janaf thermochemical tables, Fourth Edition, Monograph 9 (Part I and Part II) (1998)
Figures
Fig. 1 Backscattered electron image of Ti-46Al-8Ta (at. %) with a Ti-10Si layer deposited by magnetron sputtering.
Fig. 2 Top view of as-received silicide coating a) digital picture, b) SEM image of the surface.
Fig. 3 Cross-section of a Ti–46Al–8Ta (at. %) with a silicide coating with EDS analysis in marked points (Table 1).
Fig. 4 XRD spectra from the surface of the as-coated sample and after removal of consecutive layers of the silicide coating.
Fig. 5 Gross mass changes (sample with a crucible) of the Ti–46Al–8Ta (at. %) samples with and without silicide coating after hot corrosion test in air at 800 oC with different salt deposits.
Fig. 6 Backscattered electron images of samples with a silicon-rich coating contaminated with NaCl (a, b); Na2SO4 (c, d) and Na2SO4 + NaCl (e, f) – at different magnifications - after 80 h of oxidation in air at 800 oC. EDS analysis in marked points (see Table 1)
Fig. 7 XRD spectra from the surface of coated sample after 80 h of exposure; experimental conditions: 800 oC, 20-h cycles, laboratory air, NaCl or Na2SO4 deposits.
Fig. 8 XRD spectra from the surface of coated sample after 80 h, 160 h and 300 h of oxidation tests at 800 oC with a Na2SO4 + NaCl deposit.
Fig. 9 Cross-sections of coated samples with a) NaCl, (b, c) Na2SO4 and d) Na2SO4 + NaCl deposits after 80 h of oxidation in air at 800 oC with EDS analysis in marked points (Table 1).
Fig. 10 Net mass changes of the Ti–46Al–8Ta (at. %) samples with silicide coating after hot corrosion test in air at 800 oC with different salt deposits.
NaCl
Na2SO4
Na2SO4+NaCl
a)
b)
Fig. 11 Digital pictures of samples contaminated with NaCl or Na2SO4 or Na2SO4+ NaCl after 300 h of oxidation in air at 800 oC without (a) and with (b) silicide coating.
Without coating
With coating
Fig. 12 Cross-sections of samples with and without silicide coating contaminated with a NaCl deposit after 300 h of oxidation in air at 800 oC (20 h cycles).
Fig. 13 XRD spectra from the surface of uncoated sample and spalled scale collected in alumina crucible (sample with NaCl) after 300 h of exposure at 800 oC, 20-h cycles, laboratory air, NaCl, Na2SO4 or Na2SO4+NaCl deposits.
Tables Table 1 Local or average chemical composition of samples shown in Figs. 3, 6, 9 (EDS). Fig. No
Type of analysis Point/ average
Ti
Fig. 3
1 2 3 4 5 6
37.6 35.2 59.9 59.2 28.5 41.8
14.0 0.4 48.0 20.7 7.2 36.9 3.1 13.1 23.9 5.4 12.1 23.3 59.2 6.8 5.5 49.6 8.6 0.0
Fig. 6a
average
8.3
3.8
Fig. 6b
1 2
Fig. 6c
average
5.1
1.4
-
10.7 18.5 55.6
Fig. 6d
3 4
6.1 12.7
0.8 0.8
-
5.2 22.5 47.2 18.2 10.8 8.8 62.8 4.1
Fig. 6e
average
23.7
2.7
1.0
5.8
3.5
63.4
-
Fig. 6f
5
2.5
31.8
0.0
1.2
1.2
63.3
-
Fig. 9c
1 2 3
57.6 1.9 12.7 27.8 62.9 3.9 12.7 20.5 30.0 65.5 4.5 0.0
-
-
-
Fig. 9d
1 2 3 4 5 6 7 8 9 10
7.7 17.0 21.6 55.2 52.7 64.7 26.8 28.1 35.3 42.1
2.0 1.3 0.2 1.1 1.2 2.3 0.0 0.0 0.0 0.0
64.3 58.6 42.4 19.1 16.5 -
-
Al
12.4 3.4 3.8 10.8
Ta
at. % Si
Na
O
S
-
-
-
-
20.2
5.5
62.2
-
0.0 -
11.1 19.6
7.5 8.2
65.6 57.6
-
2.0 2.4 23.1 2.5 1.8 9.5 2.7 4.8 13.9 12.0 67.5 5.7 67.4 4.5 55.6 9.1 49.6 8.3
24.0 20.7 10.2 13.3 22.1 7.1 0.0 0.0 0.0 0.0
8.7
Table 2 Standard Gibbs free energy of formation at 1100 K in kJ/mol Cl2 [39]. ΔGo [kJ/mol Cl2] AlCl3 TiCl4 TaCl5
-351.58 -315.32 -226.80
SiCl4
-259.04
Table 3 Standard Gibbs free energy of reactions at 1100 K [37, 38, 39]
Reaction number 1 2 3 4 5 6 7 8 9 10 11
Reaction ΔGor at 1100K [kJ] -495.9 TiSi2 + 3O2 → TiO2 + 2SiO2 -528.8 TiSi + 2O2 → TiO2 + SiO2 -463.0 2TiSi2 + 4O2 → TiO2 + 3SiO2 + TiSi -202.8 6TiSi2 + 10 O2 → TiO2 + 9SiO2 + Ti5 Si3 396.5 Na2 SO4 → Na2 O + SO3 N/A Ta2 O5 + Na2 O → 2NaTaO3 1 -439.2 TiO2 + 2NaCl + O2 → Na2 TiO3 + Cl2 2 -396.9 TiO2 + Na2 SO4 → Na2 TiO3 + SO3 1 -698.4 2SiO2 + 2NaCl + O2 → Na2 Si2 O5 + Cl2 2 -656.0 2SiO2 + Na2 SO4 → Na2 Si2 O5 + SO3 N/A SiO2 (s) + 2TaCl5 (g) → SiCl4 (g) + 2TaOCl3 (g)