Materials Science and Technology

The performance of gas turbines has been improved by the development of alloys with progressively increasing high-temperature capabilities. While both strength and corrosion resistance are important, the strength requirements have a higher priority, and alloy developments which led to higher strengths also had the effect of reducing the corrosion resistance, particularly with nickel-base alloys. The most important form of corrosion is the accelerated oxidation which takes place when the air or fuel is contaminated with certain impurities, of which alkali metal salts are the most important. This type of attack is generally known as ‘hot corrosion’. Two different forms of hot corrosion have been distinguished. Type I, which is present over a temperature range of about 800–950°C, and type II, which is present over the range 700–800°C. Both processes involve an incubation period, an initiation step, and a propagation stage. Most attention has been given to the propagation stage but, from a technical point of view, the initiation step is the most important process. Mechanisms suggested include the salt fluxing model, the electrochemical model, and the sulphidation–oxidation model. Both the practical and theoretical aspects of the problem will be reviewed.

The importance and complexity of environmental influences on the hot corrosion process in gas turbines is now recognized. The present paper seeks to emphasize that the hot corrosion phenomenon is the end product of a chain of events which starts at the air / sea interface (encompassing wind speed, relative humidity, intake siting, etc) and which extends through filter performance, compressor operating characteristics, and flame tube design before terminating at the hot corrosion site, i.e. the high-pressure rotors. Attention is particularly drawn to the influence exerted by both salt and carbon particulate material and their interrelationship with each engine stage upstream of the turbine.

The hot corrosion of alloys comprises a complex series of interacting processes. Despite the difficulties involved, the thermodynamic analysis of corrosion ought to take into account the multicomponent nature of the alloy, the constitution of the gas, and the microenvironments present in the corrosion layer. Most importantly the metallic, oxide, sulphide, and salt phases ought to be treated as solutions (mixtures). The methods for assessment and calculations involving solution phases are demonstrated by reference to the salt system NaCl-NaOH-Na₂CrO₄-Na₂S0₄ and its equilibrium with Cr₂O₃ as the protect!ve oxide layer and the gaseous combustion products. The data requirements assa.ciated with these developments and their possible application to kinetic analysis are discussed.

The redox potential in sulphate melts is controlled by the partial pressures of oxygen and SO₃ in equilibrium with the melt. The rate for the reduction of oxygen is limited by the solubility. Higher rates are observed for the reduction of SO₃ which is identified as the principal oxidizing species. Equilibrium potentials for reversible metal-metal ion couples for the principal components of superalloys are negative relative to the redox potential for the melt. The anodic oxidation of the metal therefore proceeds irreversibly in conjunction with the cathodic reduction of SO₃ An oxide layer is formed on the surface of the metal when the metal ion concentration at the surface, combined with the oxide ion concentration in the melt, which is related to the partial pressure of SO₃, exceeds the solubility limit for the oxide. The corrosion behaviour will depend on the mass transport processes through this oxide layer. Temperature gradients through the molten sulphate, and oxide ion concentration gradients established as a consequence of the corrosion reaction, may influence the morphology of the corrosion product. The gradient between the oxygen chemical potential at the outer surface of the molten sulphate, defined by the oxygen partial pressure in the gas and the oxygen chemical potential at the metal/metal oxide interface, defined by the dissociation equilibrium for the metal oxide, combined with the transport of SO₃ through the molten sulphate, increases the sulphur chemical potential at the metal surface and leads to the formation of the metal sulphide. Stress in the protective oxide layer caused by the growth of the sulphide phase at the interface between the metal and the oxide will eventually fracture the oxide and cause accelerated corrosion.

Key principles of alloy oxidation are discussed, with emphasis on Cr₂O₃ and Al₂O₃ formation on nickel-, cobalt, and iron-base alloys. The various special cases of alloy oxidation, which are quantifiable to varying degrees, are presented schematically. The important competition between surface scale development and internal oxidation is emphasized and extended to explain transient oxidation. The ability to measure and model the distribution of alloying elements in steady-state scale and substrate is described. The priority now is to understand further alloying element and defect segregation and transport in grain boundaries, and also other short-circuit paths including pores, in Cr₂O₃ and Al₂O₃ scales.

Interaction between alloy depletion, void formation, and phase-boundary oxidant transport in single- and multi-phase alloy substrates requires further elucidation. Brief consideration of ternary and quaternary alloy oxidation illustrates the, ability partially to explain complex alloy behaviour. The role of reactive element additives and dispersoids is reviewed concisely in light of recent work. MST/442

Many applications of high-temperature alloys involve high-temperature oxidation under thermal cycling conditions. The oxide scales formed have lower coefficients of thermal expansion than the metallic alloys and significant thermal stresses can arise during temperature variations. These thermal stresses add to the residual growth stresses which accompany oxide formation. Under certain conditions, stresses in oxide scales may be partly relieved by plastic deformation of scale or alloy. However, when these mechanisms cannot be operative, scale buckling or cracking occur depending on interfacial and oxide fracture strengths. Eventual oxide spallation causes rapid degradation since depleted regions of alloy are in contact with the oxidizing atmosphere. Incorporation of ‘active elements’ such as yttrium, in reducing the residual growth stresses significantly improves the cyclic oxidation resistance of high-temperature nickel-, cobalt-, and iron-base alloys. The present paper attempts to review briefly the mechanisms involved in these phenomena and the tentative cyclic oxidation models. MST/443

The high-temperature corrosion of superalloys is associated with contaminants. When comparing contaminant conditions the contaminant flux rate (CFR) should be considered rather than the contaminant level in the fuel or environment. At temperatures above 700° C, vanadates cause fluxing of the protective oxide scales and it is shown that corrosion is determined by the CFR and temperature rather than by material selection. The effects of sulphur level in the fuel on the efficiency of magnesium additives are also considered. Chloride contamination is shown to produce scale rupture and the influence of chloride contamination under gaseous and deposit conditions is examined. In particular the differences between marine gas turbine conditions and laboratory tests to simulate hot corrosion are evaluated. It is suggested that in marine turbines fluxing mechanisms are more appropriate to the alloy than to the protective scale. Finally, the influence of chlorides on low-temperature type II hot corrosion is considered. MST/445

Some of the more recent experiments concerning the interactions between corrosion and creep in nickel-base superalloys have been reviewed and placed into three categories. The most systematic work, from the viewpoint of model development for the prediction of the service life, has concerned the exposure to a damaging environment before creep testing; but there is still insufficient work of this type. The ‘sharp interface’ composite model of corrosion-creep interactions has been developed to unify and extend the few attempt that have been made in the literature to provide predictive equations in this area. The virtue of such an approach is that it keeps the physics of the various processes well to the fore, but it is approximate and its success can only be gauged when well-defined systematic experiments have been performed in this important area. MST/446

The reliability of industrial gas turbines can be limited by the premature degradation of critical hot gas path components. Cost effective operation of these turbines therefore requires a knowledge of component life and its limitations according to service requirements. Through this knowledge, ways to restrict the degradation of components can also be determined. Creep, corrosion, and fatigue are considered to be the principal life limiting factors with the hot gas components. The present paper discusses the influence of these three processes on component life, and examples of associated service problems are presented. Specific reference is made to the corrosion resulting from the presence of sodium sulphate deposits in offshore turbines. Control of this corrosion through the use of fuel additives and protective coatings is considered. MST/448

The purpose of this paper is to review corrosion processes and experience related to coal-fired gas turbines. For over 40 years there have been major programmes of research directed towards burning coal in a gas turbine either directly or as a coal-derived fuel. This history is briefly reviewed, demonstrating the importance which corrosion of the hot section has had in limiting the achievements and defining the systems. The probable corrosion mechanisms are identified, and because of their synergistic interaction with corrosion, erosion and deposition are also considered. The discussion identifies pressurized fluidized bed combustion (PFBC) as the direct combustion technique most likely to be able to avoid serious corrosion problems in the immediate future. Recent results from PFBC pilot plant investigations related to turbine materials are compared, and thereby future directions for overcoming corrosion problems are proposed. MST/449

Hot corrosion studies were conducted at 900°C in a simulated combustion gas atmosphere containing sulphur trioxide at a partial pressure of 750 Pa on a nickel-base super alloy (Nimocast 713) partially immersed in various ZnSO₄–10wt-%NaCl–Na₂SO₄ molten salt mixtures. Weight changes were measured thermogravimetrically with time and it was found that the overall effect of the zinc sulphate addition was to reduce the degree of hot corrosion. Quantitative scanning electron microscopy investigations revealed no evidence of zinc in the attacked zone. It is postulated that the zinc reduces the chlorine activity in the system by forming a volatile chloride, thereby reducing the corrosive effect of the gas phase/molten salt mixtures. MST/507

The vanadate-induced corrosion of M Cr AIY coatings (M = Ni/Fe) on Inconel alloy 600 has been studied by using samples with preformed vanadate deposits. The samples were exposed to oxygen at 650 to 800°C for a maximum of 25 h in a thermobalance. The presence of molten vanadate ((Na₂O)_x·V₂O₅) on the surface causes accelerated corrosion. The corrosion rates increase with increasing V₂O₅ content of the vanadates. In the presence of molten vanadate, protective scales of Cr₂O₃-Al₂O₃ fail to develop; rather the oxide layer next to the coating consists of the oxides of all components in the coating and of vanadium both as vanadate and as vanadium sesquioxide. In most cases the molten vanadate phase fluxes the oxide scale and particularly the NiO phase in the oxide layer. Under these conditions the corrosion is probably governed by inward transport of oxygen (or oxygen species) in the vanadate melt. When the NiO concentration in the oxide layer next to the metal is high (as on nickel-base coatings with more than 75 wt-%Ni) a Ni₃(VO₄)₂ layer gradually develops between the oxide and the melt. This is accompanied by compositional changes in the oxide layer beneath the vanadate and results in a slow oxidation rate. MST/510