Skip to main content

Development of Bonding for Diffusion Barriers for the High Temperature Blades of Gas Turbines

Exploitable results

The goal of the present basic research project was to develop suitable and reliable bondings for the diffusion barrier layers developed in a previous project of the European concerted action COST 501/II, WP 7A for delaying the detrimental diffusion of nickel and aluminium between the nickel base super-alloy blades and their MCrAlY type overlay at temperatures at and below 1100 degrees Celsius. This detrimental diffusion between the mentioned materials starts actually at about 950 degrees Celsius depending on the composition of the super-alloy used in the blades. In the present project, the diffusion barriers inherited from the earlier project were an amorphous layer of Al-O-N and a triple layer of TiN+AlN+TiN, the former developed by RWTH.LFWW, and the latter by TUT.DMS. The substrate super-alloys were IN738 LC in the polycrystalline form and SRR 99 in the single crystalline form, the former one used in land-base as well as in aero gas turbines and the latter one mostly in aero gas turbines. For both of the super-alloys, the dissolving temperature of their Ni(3)Al precipitates is about 1050-1150 degrees Celsius. This means that the super-alloys lose their strength at and above these temperatures. This is why the temperature of 1100 degrees Celsius was taken as the maximum checking temperature for the developed bonding. As IN738 at this temperature is already in its softened state, i.e., it strains more easily than in its hardened state, this extra strain causes extra requirements for the bonding and in that way insures the reliability of the bonding. The overlay materials were LCO22 and Amdry 995, which have almost the same composition. A continuously changing gradient bonding appeared not to be suitable for either of the cases. Therefore, a solution was searched from special alloying of the barrier layer and from using a separate bonding layer or layers. On the basis of annealing tests, a (semi)crystalline barrier layer of Al(2)O(3) appeared to be as good as or better barrier layer than the amorphous Al-O-N layer. Thus, this new layer replaced the original one. For this new layer, an alloying with 1% Y(2)O(3) or a separate bond layer of Y(2)O(3) appeared to be promising solutions for the bonding. These solutions were tested by mechanical bending at different temperatures, by scratch tests at room temperature, by annealing at 1100 degrees Celsius and by thermal cycling between room temperature and 1100 degrees Celsius. In perpendicular blasting at room temperature by glass bolls under a pressure of 5 bars, the bonding failed and the diffusion barrier and its sputtered overlay spalled off entirely. Thus, the roughening of the surface of the sputtered overlay required for the further thickening of the overlay by means of plasma spraying is not an applicable method for the developed bonding. In the triple type barrier layer, the outermost layers were of TiN. Thus, the search of the bonding was directed to the bonding of TiN to the substrate and to the overlay materials. After thermodynamic evaluations and extensive laboratory scale tests on small specimens, seven promising candidates were found for the bond layer. The bonding quality of these candidates were tested first by tensile tests at room temperature, then by long term annealing, and by long term thermal cycling between room temperature and 1100 degrees Celsius. Additional reliability was searched through cyclic mechanical stressing under different constant amplitudes at 1100 degrees Celsius, and by cycling of temperature between 850 degrees Celsius and 1100 degrees Celsius under a constant loading of 100 MPa. In the simulating burner rig tests, maximum temperatures were 1100 degrees Celsius and the minimum 850 degrees Celsius. The simulating high temperature corrosion tests (the salt spray tests) were done at 900 degrees Celsius, at the temperature of maximum corrosion rate, and the spraying of salt liquid took place daily at room temperature. From the candidate bonding, the layers of 6bv2 and 6bv3 (as such or as nitrides towards the surface against TiN) passed these tests and the above mentioned blasting test successfully. Therefore, these bond layers as suggested for the bonding of the diffusion barriers in the subsequent industrial development project. In the long term annealing as well as in the long term thermal cycling tests, the single layer of TiN appeared to be a sufficient diffusion barrier comparable to that of the earlier developed triple layer. Thus, no need exists for using the more complicated diffusion barrier at or below the temperature of 1100 degrees Celsius. An additional goal in the original project plan was the development of a non-destructive testing method, which could be used, with certain reliability, as a quantitative test of the adhesion strength (bonding strength) between the super-alloy substrates and their different coatings. As neither the ultrasonic nor the thermal vibrations used were capable of creating detectable yielding at the interfaces, the developed ultrasonic method remained on qualitative level. An inverse relationship was, however, detected between the non-linear parameters of the ultrasonic measurements and the tensile strengths of interfaces measured at room temperature.