Ceramics in turbine applications

TWI Bulletin, September/October 1994

Wendy Hanson joined TWI in 1992 and is a Senior Research Engineer in the Ceramics Group in the Advanced Materials and Processes Department. Before this, she studied Metallurgy and Engineering Materials at the University of Strathclyde. She gained a PhD at Strathclyde for a study of the fabrication and properties of silicon carbide ceramic matrix composites. Her TWI interests involve joining ceramics to metals, polymer derived ceramics, high temperature inorganic adhesives and glass-ceramics.

John Fernie joined the Ceramics Group as a Senior Research Engineer in 1990. Now Technology Manager - Ceramics in the Advanced Materials and Processes Department, he has responsibility for a number of projects concerned with ceramic joining and coating technology. His current research interests include microwave heating, glass-ceramics and polymer derived ceramics. Before joining TWI, John studied Metallurgy and Microstructural Engineering at Sheffield City Polytechnic (now Sheffield Hallam University), and went on to gain a PhD at The Centre for Advanced Materials Technology at the University of Warwick. John is also a Chartered Engineer and a Chartered Physicist.

Ceramic-metal joining is essential if the excellent high temperature strength, low weight, good corrosion and wear resistance of ceramics are to be exploited in manufacturing industry.

As Wendy Hanson and John Fernie report, TWI has developed a novel flexible interlayer, which may be brazed into position, which will give ceramic-metal bonds with good strength and corrosion resistance at working temperatures up to 600°C.

These are of great benefit to turbine technologists who seek reduced weight and increased temperature capability to increase engine efficiency. Increased fuel efficiency, a consequence of higher temperature combustion, leads directly to the reduction of NO _x emissions and is welcome as a contribution to the reduction in environmental pollution. When considering the implementation of ceramics into turbine engine environments, design requirements must also be taken into consideration; these include factors such as load distribution, environment, reliability and life predictions.

It is for these reasons that the inclusion of ceramic components into critical components has been disappointingly slow.

However, the increased toughness of ceramic matrix composites and greater reliability of current ceramics makes them viable in a number of components in the turbine engine, including the combustor can, burner nozzles, flame holders and turbine blades.

One particular target area is the replacement of nickel alloy turbine blades with ceramics, such as silicon nitride (Si ₃N ₄). Ceramics confer a number of advantages and with the introduction of the interlayer proposed here, there is also the ability to re-design the blade to be less complicated in shape, since the intricate cooling channels required in metallic blades are no longer necessary.

Ceramic properties

The various advantages of ceramics, stemming directly from their covalent bonding, also generate significant engineering disadvantages, foremost amongst which are mechanical and thermal brittleness. The inherent stability of ceramics also creates difficulties in the manufacture of joints between both similar and dissimilar materials. ^[1] These arise because of the chemical stability of the ceramic surface ( i.e. low reactivity) and the inherently low coefficient of thermal expansion (CTE). It is these disadvantages coupled with low fracture toughness and Weibull (reliability) modulus, which have restricted the use of engineering ceramics. A major problem is the CTE mismatch between the metal and ceramic substrates. Typical values of CTE for a nickel alloy substrate are -11 x 10 ^-6/°C and for a ceramic, such as silicon nitride ~3 x 10 ^-6/°C. To overcome this, a combination of good design and use of interlayers is required. The design problem is addressed via finite element analysis to model the stresses at the joint region and to optimise the design of the interlayer. These interlayers are usually thin shims of metal whose properties lie between those of the ceramic and metal. Typical designs of such interlayers are given in Fig.l.

Designs i) to iii) are all solid interlayers, which have disadvantages related to their weight and configuration. Design iv), a flexible interlayer, offers a number of potential advantages:

• lower joint stress, giving greater reliability;
• inclusion of cooling passages at the joint area;
• simplified joint designs, allowing lower production costs;
• ability to use the ceramic at a higher operating temperature than the metal;
• accommodation of residual stress by interlayer deformation;
• less volume of interlayer, therefore weight saving.

One of the notable features of this design is the ability for the flexible interlayer to allow a cooling gas to pass through. For application in turbine engines, this allows the ceramic component to operate at a significantly higher temperature than either the braze (used as the joining media) or metallic substrate. This is because the air which would normally have travelled through the blade will now stream past the flexible interlayer region, thus cooling it.

Materials

Ceramic and metallic substrates

The metallic substrate used was a standard nickel alloy commonly found in turbine applications, Inconel 939.

This programme investigated two temperature ranges. A low temperature regime where the joint would only experience an operational temperature of less than 500°C, and a high temperature regime, where the joint would endure temperatures up to 800°C. For this a number of different interlayers and braze alloys were investigated, see Table 1.

Table 1 Combination of materials

Braze alloys

Since ceramics are not readily wettable by many of the elements which constitute most brazing alloys, the ceramic must have its surface activated in order that the braze wets the substrate. An Ag-Cu-Ti alloy already contains a surface activating agent, i.e. Ti. However, for the high temperature braze alloy, Ni-Cr-Si, the Ti must be deposited on the surface of the Si ₃N ₄ before brazing.

Configuration of interlayers

Two designs were examined, see Fig.2, a corrugated and a dimpled. ^[3] The dimpled interlayers were produced in-house, the corrugated design was sourced commercially. The dimensions of these interlayers are given in Table 2. The interlayers were produced from commercially pure iron, stainless steel (EN 58B) and nickel alloys (Nimonic 75 and Haynes 230).

Table 2 Interlayer dimensions

Production of bonded joints

All samples were produced by a vacuum brazing procedure. For low temperature samples, and Ag-Cu-Ti braze, a temperature of 850°C was used. The higher temperature brazing cycle, for Ni-Cr-Si, used a temperature of 1180°C.

Results

Ag-Cu-Ti

The dimpled interlayer was produced in iron which had been dimpled with either four large, or five small indents, positioned to mimic the sides of a conventional dice. The samples were examined using microscopy and also by shear testing. An SEM micrograph of a dimpled iron interlayer showing wetting of Ag-Cu-Ti braze on to both metallic and ceramic substrates is given in Fig.3. The shear testing results for these interlayers are given in Table 3. Results were also obtained using Nimonic 75 corrugation brazed using Ag-Cu-Ti. Examples of these types of sample are given in Fig.4. Unlike the dimpled design, the corrugated interlayer has an orientation effect and the strength of the interlayer varies dependent on its position with respect to the testing device. The results of this are shown in Table 4. The orientations described are shown in Fig.5. With the interlayer parallel to the testing machine, these results are comparable with the highest results achieved for large, dimpled, iron interlayers.

Table 3 Results of shear testing dimpled Fe interlayer joints

Table 4 Results of shear testing corrugated Nimonic 75 interlayer joints.

Ni-Cr-Si

Two types of interlayer were used for these experiments - corrugated Nimonic 75 and dimpled Haynes 230 alloy. These were brazed into position using Ni-Cr-Si braze alloy after the ceramic had been sputter coated with ~1µm of Ti.

It was found that whilst the design of the corrugated interlayer was promising, the Nimonic 75 alloy itself could not withstand the high temperatures required for brazing. However, the Haynes 230, whilst showing slightly lower shear strength values on testing than the low temperature bonds produced in iron ( Table 3), is better suited to the high temperatures. Table 5 gives the shear strength results for these high temperature joints.

Table 5 Shear strength results for corrugated Nimonic 75 and dimpled Haynes 230 interlayered joints.

The Nimonic 75 interlayer has failed with the high temperature braze because the side walls of the interlayer have become thinner, probably because of dissolution of the nickel into the braze alloy. An example of this failure mode is given in Fig.6. This SEM micrograph shows the deformation and failure of a corrugated interlayer after shear testing. The interlayer has broken, but the ceramic-braze and metallic substrate-braze bonds have not been compromised.

The Haynes 230 interlayer has withstood the higher brazing temperature and failed in areas where the braze alloy has run on to the nickel alloy substrate, thereby forming a solid bond. Therefore, the results obtained are actually from failure in the braze since the load applied to fail the interlayer material has not been reached during the test.

Discussion

The application of corrugated Nimonic 75 interlayers has also allowed comparison between both use of low and high temperature brazes and between dimpled and corrugated designs. Nimonic 75 has potential as a low temperature interlayer; however, at temperatures above 800°C, this material is unsuitable.

Identification of a further high temperature interlayer, Haynes 230, makes this likely as a possible material for use in future. It shows comparable shear strength values with those achieved through the low temperature iron interlayers. Full implementation into commercial applications could require production of the alloy in the more promising corrugated design (since Haynes 230 is difficult to produce in large dimple configuration, because of tearing of the interlayer on formation). Other alloys which may also be considered for such applications are stainless steel and titanium alloys.

Conclusions

A brazing procedure has been established which forms metal-ceramic bonds for both low and high temperature regimes.

For low temperatures, use of Ag-Cu-Ti braze alloy and either dimpled iron or corrugated Nimonic 75 interlayers, have been demonstrated to be feasible.

For high temperatures, only the Haynes 230 dimpled interlayer could be successfully brazed using Ni-Cr-Si alloy. The shear strengths of these bonds were comparable with those obtained for the low temperature regime.

Acknowledgements

This programme was carried out with funding from European Gas Turbines Limited.

References