TMS QUICK LINKS:
  • TMS ONLINE
  • MEMBERS ONLY
  • MEMBERSHIP INFO
  • MEETINGS CALENDAR
  • PUBLICATIONS

JOM QUICK LINKS:
  • JOM HOME PAGE
  • MORE HTML ARTICLES
  • COMPLETE ISSUES
  • BOOKS REVIEWS
  • SUBSCRIBE

Engine designers show continued interest in titanium aluminides based on the intermetallic γ-TiAl phase as lightweight structural materials to be used at moderately elevated temperatures. Although alloy development has made significant progress in terms of mechanical properties and environmental resistance, protective coatings have been developed that help to extend the lifetime of these alloys significantly. The major challenge of coating development is long-term stability of a protective oxide scale that forms during service for which purpose alumina formation is essential. Furthermore, changes of coating chemistries at high temperatures must be controlled to avoid rapid degradation of the coatings due to diffusional losses into the substrate material and vice versa.

INTRODUCTION

Titanium aluminide alloys based on γ-TiAl have convincingly demonstrated that they can meet the requirements for automotive combustion engines as well as for aero engines.^1,2 Titanium aluminides are already in use in small-series automotive applications. However, in some cases the risk and cost issues of γ-TiAl implementation into aero engines, for example as high-pressure compressors or low-pressure turbine airfoils (see Figure 1), are seen as higher than tolerable in current engine programs,¹ thus preventing these alloys from aero-engine application. Nevertheless, the potential of titanium aluminides is indisputable³ and has spurred R&D activities worldwide. Recent developments indicate that the past problems related to the supply chain of γ-TiAl aero-engine parts might have been overcome, considerably enhancing the trust of designers in intermetallics. Therefore, today it appears more likely than ever that aero engines will be equipped with γ-TiAl parts in the near future.

At the same time, materials and technology development continues. Alloy development aims at increased temperature capabilities up to temperatures in excess of 900°C. Although γ-TiAl alloys possess better intrinsic resistance against environmental attack than, for example, conventional titanium alloys, protective coatings are required for both environmental protection and thermal insulation of the structural material when targeting higher temperatures. Even third-generation γ-TiAl alloys such as Ti-45Al-8Nb suffer from mixed oxide scale formation which results in unacceptably high oxidation rates and short lifetimes. Despite the high aluminum content of γ-TiAl alloys, they do not exclusively form protective alumina (Al₂O₃) scales but always build titania (TiO₂) as well (Figure 2), which is a fast-growing oxide that does not provide long-term oxidation protection. Therefore, coatings have to be developed that form and maintain protective alumina scales over their entire anticipated lifetime.

In this study, metallic and nitride overlay coatings produced by different magnetron sputtering techniques were investigated. The coatings were deposited onto third-generation titanium aluminide alloy Ti-45Al-8Nb to improve its environmental resistance. Further, thermal barrier coatings (TBCs) deposited by electron-beam physical vapor deposition (EBPVD), which are typically deposited onto nickel-based superalloys, were tested on γ-TiAl. In addition to the sputtered metallic and nitride overlay coatings, diffusion aluminide coatings were used as bond coats for TBC application.

Metallic Coatings
Alloys from the Ti-Al-Cr system were reported to have good oxidation resistance combined with reasonable mechanical properties when compositions are chosen from the two-phase region γ-TiAl and Ti(Cr, Al)₂.¹¹ When applied as coatings to Ti-45Al-8Nb by magnetron sputtering,¹² Ti-Al-Cr coatings provide excellent oxidation resistance at 750°C^13,14 for up to several thousand hours. Compared with the uncoated reference material, at 900°C, the Ti-Al-Cr coatings improve oxidation resistance up to about 1,000 1-h cycles before rapid oxide scale formation occurs (Figure 3). At 950°C, where the oxide scale on the uncoated reference sample readily spalls after a few cycles, Ti-Al-Cr coatings provide reasonable protection for up to 350 1-h cycles before breakaway oxidation occurs. Good oxidation resistance of the Ti-Al-Cr coatings at temperatures between 750°C and 900°C is obviously caused by the two-phase microstructure of the coating with the Laves phase being most oxidation resistant. Scanning-electron microscopy (SEM) cross-section analysis after 2,000 h oxidation at 750°C reveals that basically the initial microstructure consisting of γ-TiAl and Laves phase is maintained (Figure 4). Compositions of the different phases are given in Table I.

The oxide scale formed under these conditions has a thickness of about 2 μm and consists of a continuous alumina layer providing excellent protection. An Oxford energy-dispersive x-ray spectrometry (EDS) cross-section analysis revealed¹⁴ that in addition to alumina the outer oxide scale contained some TiO₂ and Cr₂O₃. However, alumina was clearly the dominating oxide phase (see Table I). The interdiffusion zone is characterized by a zone depleted in Laves phase with the remainder being γ-TiAl and a more or less continuous zone of Laves phase interface precipitates that formed due to chromium inward diffusion from the coating into the substrate alloy. Neither cracks nor pores were observed in the interdiffusion zone.

Experimental results clearly indicate that the sputtered Ti-Al-Cr coatings can improve oxidation resistance of γ-TiAl alloy Ti-45Al-8Nb up to a maximum temperature somewhat below 900°C. At higher temperatures, the coatings tend toward strong interdiffusion with the substrate alloy, which obviously promotes the formation of less protective oxide scales. Once the continuous alumina scale is interrupted, titanium oxide rapidly forms and controls the rate of oxide scale growth (breakaway oxidation). Future Ti-Al-Cr coating development will consider this issue more stringently.

Nitride Coatings
Ti-Al-Cr-Y-N coatings have excellent oxidation resistance at 750°C¹⁵ and effectively reduce the mass gain caused by oxide scale formation by a factor of four compared to the uncoated reference material (Figure 5, reference time: 1,000 h). Scanning-electron microscopy/EDS investigations of the oxide scale revealed a duplex oxide scale consisting of an outer layered zone with alternating enrichment of Al₂O₃ and TiO₂ and a mixed titania-alumina inner layer (Figure 6). At 750°C, the long-term stability of the Ti-Al-Cr-Y-N coating was proven up to 3,000 h in isothermal tests.¹² The oxidized surface of the nitride coating was fairly porous. Few agglomerations of oxides were observed. Chromium and yttrium were also detected in the oxide scale, however, at relatively small levels (Cr ≤ 1.5 at.%, Y ≤ 0.8 at.%). Yttrium was not found in the outer part of the oxide scale. Oxygen was not detected in and beyond the remaining nitride coating, which exhibited columnar grain morphology.¹² Scanning-electron microscopy studies¹² revealed a base layer at the lower end of the nitride coating. The initial composition of this layer was different from that of the protective coating. In the transition zone, the titanium concentration increased whereas aluminum was depleted. The concentration of nitrogen decreased toward the substrate, revealing an intermediate plateau near the coating layer. The enrichment of titanium was related to the formation of titanium nitride. With decreasing nitrogen and increasing aluminum content, Ti₂AlN might be formed adjacent to the substrate.

At 850°C, the nitride coating still provided good oxidation resistance despite the fact that it experienced relatively high initial mass gain during the first 100 1-h cycles (Figure 5). Once a protective oxide scale formed, the oxidation rate dropped significantly. Clearly, the oxidation curve of the Ti-Al-Cr-Y-N-coated Ti-45Al-8Nb sample falls below that of the uncoated reference after 1,200 1-h cycles. At 900°C, the nitride coatings failed after about 300 1-h cycles by oxide scale spallation (Figure 5). Post-oxidation SEM investigations revealed that at this point the entire nitride layer was oxidized. From these results the maximum useful temperature is concluded to be 850°C for the Ti-Al-Cr-Y-N coating.

It can be concluded from these and earlier results that protective Ti-Al-Cr and Ti-Al-Cr-Y-N coatings have been developed which might be useful to protect γ-TiAl components up to around 900°C. Future coating development aims at further improvement of the high-temperature capability and lifetime extension of the coated component.

As a note, quite obviously coatings add costs to a fully finished component which must be considered. With this in mind, the authors have chosen industrially available processing methods that are already widely used in industry (partly even in the aero-engine industry), thus keeping the additional costs for coating deposition at a reasonable level. However, once coatings have matured, the lifetime benefit and the allowable increase in service temperature provided by the use of environmentally resistant coatings are expected to pay off fairly quickly. Even if the service temperature can be increased only by a few 10K, this might be enough to replace heavier nickel-based superalloys, which results in significant weight reduction.

Thermal barrier coatings have been widely used on gas turbine hardware to insulate thermally loaded nickel-based superalloy airfoils.^16–21 This technology was recently successfully applied to γ-TiAl-based alloys by the present authors.^12,13 The concept of TBCs is illustrated in Figure 7. Due to the presence of a ceramic coating with low thermal conductivity (typically yttria-stabilized zirconia), a temperature gradient is established across the wall thickness of an internally cooled component such as a turbine blade or an actively cooled flat structure. As the schematic temperature profile suggests, in the presence of such a TBC, the useful surface temperature of the component can be increased by up to 150K, depending on the thickness and the thermal conductivity of the ceramic coating, while the metal temperature remains low. Since the lifetime or the temperature limit of a γ-TiAl component is largely determined by its oxidation resistance and/or its mechanical properties at elevated temperatures, the use of TBCs on γ-TiAl might further push the useful service temperature limit of γ-TiAl components into a range which is yet dominated by superalloys.

Figure 8 shows mass change curves of four different TBC systems on γ-TiAl alloy Ti-45Al-8Nb exposed to cyclic testing at 900°C. The reference system (pre-oxidized Ti-45Al-8Nb substrate plus TBC top coating) shows reasonable performance up to a total number of 950 1-h cycles. Except for some edge-chipping phenomena leading to occasional, small irregularities in the mass change curve, the TBC was well adherent to the substrate but then failed by sudden spallation. As known from earlier work,^12–14 TBCs on γ-TiAl alloys can accept the formation of fairly thick oxide scale (20–30 µm) without failure, which is much more than allowable for TBC-coated nickel based superalloys; the critical oxide scale thickness for the onset of TBC spallation is typically reported to be about 7 μm for nickel-based superalloys.²⁰ The current understanding for TBC on nickel-based alloys is that once the thermally grown oxide scale has reached a critical thickness, the stresses generated in this thin oxide layer during formation and growth (typically 3–4 GPa²⁰) promote spallation of the oxide scale, and along with it goes the TBC. Stress generation in the oxide scale is mainly due to growth stresses (approximately 1 GPa in typical TBC systems) within the oxide scale and due to the differences in the coefficients of thermal expansion between the oxide scale grown at high temperatures and the metal substrate.

Growth stresses have not yet been measured for thermally grown oxide scales on TBC-coated γ-TiAl alloys, but the main reason for good TBC adherence to these intermetallics is thought to be the lower mismatch in thermal expansion coefficient compared with nickel-based superalloys, which leads to lower stresses during thermal cycling. Therefore, thicker oxide scales can be developed on γ-TiAl alloys before oxide scale (and TBC) spallation occurs.

The lowest mass change was measured for an aluminide-coated TBC system on Ti-45Al-8Nb (Figure 8, TiAl₃+TBC) which was tested up to 1,000 1-h cycles. Processing of diffusion aluminide coatings is crucial with regards to performance. Figure 9a shows a cross section SEM micrograph indicating rapid oxidation of a poorly processed aluminide coating acting as a bond coat for an EB-PVD thermal barrier top coating. After 1,000 1-h cycles, a massive oxide scale was formed with a complex oxide composition and a porous microstructure. Failure of the TBC occurred within the oxide scale along an area with highest porosity (see the gap in Figure 9a). In contrast to that, the oxide scale formed on a properly processed TiAl₃ coating is very thin (1 μm after 1,000 1-h cycles, see Figure 9b). The oxide scale formed is alumina, which contains some porosity. The TBC is well attached to the alumina scale, which has good adherence to the TiAl substrate. Aluminide coatings provide excellent oxidation protection, but their inherent brittleness is considered crucial in practical applications. Even without external mechanical loading these coatings tend toward crack formation just through stresses induced by thermal mismatch between the coating and the substrate.

The mass change curves in Figure 8 indicate that, at 900°C, the metallic Ti-Al-Cr coating provides reasonable protection while the Ti-Al-Cr-Y-N tends toward rapid oxidation, leading to early TBC spallation. This is in general agreement with the findings discussed previously for these coating systems. For the Ti-Al-Cr-coated TBC, mass change exhibits a plateau of slow oxide growth between 200 and 1,000 1-h cycles and, thereafter, increases significantly, finally leading to breakaway oxidation and spallation. While after 120 1-h cycles at 900°C the oxide scale formed underneath the ceramic top coating is continuous, the Ti-Al-Cr coating is already depleted in Laves phase and substantial interdiffusion with the substrate has occurred (see markers 1 and 2 in Figure 10a). At 950°C, which is also representative for longer exposure times at 900°C, the coating is locally strongly oxidized which leads to considerable mass gain (Figures 8 and 10b). Longer exposure or higher temperatures lead to aluminum depletion below the oxide scale, resulting in the formation of fast-growing non-protective titania (Figure 10b, markers 3 and 4).

Current work on TBCs for titanium aluminides indicates significant potential for practical application on aero engine hardware. However, the oxidation issue of the TiAl structural material must be overcome, possibly by the use of protective coatings that provide long-term stability of a protective alumina scale which serves as a bond coat between the substrate and the TBC. Finally, the internal cooling of components such as airfoils marks a potentially insurmountable challenge due to the inherent brittleness of titanium aluminides. However, once titanium aluminides have made their way into aero engines as airfoils, flat structures such as casings and cones might follow. Some of these structures could be actively cooled from the backside, making use of TBCs desirable.

Metallic Ti-Al-Cr and ceramic Ti-Al-Cr-Y-N coatings provide good oxidation protection to intermetallic γ-TiAl alloys. Yet their long-term stability is insufficient at temperatures around and beyond 900°C, both with regards to maintenance of a protective oxide scale as well as their tendency for interdiffusion with the substrate material. Recent work on TBCs on intermetallic titanium aluminides has proven significant potential for actively cooled turbine hardware. The major challenge is improved oxidation resistance of the substrate material. Oxidation resistant coatings with long-term ability to form protective alumina might serve as bond coats that provide good adherence of the ceramic top coatings to the substrate, which is particularly needed under thermal cyclic conditions.

ACKNOWLEDGEMENTS

Part of this work is continued and extended in the European project INNOVATIAL–Innovative processes and materials to synthesize knowledge based ultra-performance nanostructured PVD thin films on gamma titanium aluminides, which is supported by the European Commission within the Sixth Framework Programme for Research and Technological Development.

REFERENCES

C. Leyens holds the Chair of Physical Metallurgy and Materials Technology at the Technical University of Brandenburg at Cottbus, Germany, and is associate director of the Institute of Materials Research at the German Aerospace Center (DLR), Cologne, Germany. R. Braun and M. Fröhlich are with the Institute of Materials Research at DLR. P. Eh. Hovsepian is head of the Nanotechnology Center for PVD Research at Sheffield Hallam University, United Kingdom.

For more information, contact Christoph Leyens, Technical University of Brandenburg at Cottbus, Konrad-Wachsmann-Allee 17, 03046 Cottbus, Germany; +49-355-69-2815; e-mail leyens@tu-cottbus.de.