J. Stringer

The high-temperature oxidation of nickel-20 wt. % chromium alloys containing dispersed oxide phases

Alloys of Ni-20 wt. % Cr containing 3 vol. % of a dispersed oxide phase have been prepared by a mechanical alloying method and oxidized in oxygen at 100 Torr in the temperature range of 900 to 1200°C. It appears that the dispersed oxide has four distinct effects on the oxidation: (1) the selective oxidation of chromium to form a continuous protective Cr2O3scale is promoted; (2) the rate of growth of Cr2O3is reduced compared with particle-free alloys; (3) the adhesion of the Cr2O3is greatly improved; and (4)the scale-forming reaction appears to be at the scale-metal interface in alloys containing a dispersion, but at the scale-oxygen interface in alloys without a dispersion. It appears that the nature of the dispersed oxide is not important, since very similar effects can be obtained with ThO2,Y2O3,and CeO2dispersions. It is demonstrated that a logical deduction from this evidence is that the growth of Cr2O3scales on dispersion-free systems must involve short-circuit diffusion of chromium through the scale, and that it seems probable that an effect of the dispersion must be to retard or eliminate this short-circuit process. It is suggested that the oxide particles act as nucleation centers for the oxide, thus reducing the oxide grain size; and it is shown that this simple hypothesis is sufficient to explain a number of the experimental observations.

Stress generation and relief in growingoxide films

Abstract The experimental techniques for determining the growth stresses in oxide scales are described, and the results briefly summarized. The theories proposed by various workers to account for the existence of growth stresses are critically reviewed. The reasons for suggesting that plastic flow takes place during the oxidation process are discussed, and the various mechanisms proposed for creep in oxides are briefly described; the apparent importance of extrinsic processes is emphasized. A detailed survey of the effect of stoichiometry on oxide plasticity is given, and the implications of this in terms of mobile lattice defects is discussed. The problems of vacancy injection and dislocation generation in the metal substrate are also examined. Finally, a number of suggestions are made for further experimental investigations.

The oxidation behavior of CoCrAI systems containing active element additions

The effect of small amounts of yttrium (up to 1 wt. %) and hafnium (up to 1.5 wt.%) on the oxidation behavior of Co-Cr-Al alloys in the temperature range 1000–1200°C for times up to 1000 hr in air has been studied. The major portion of the study has been concerned with Co-10Cr-11Al base alloys. Both isothermal and cyclic tests have been carried out; the cycle used consisted of 20 hr at temperature, followed by cooling to room temperature. Both additions reduce the overall oxidation, Hf somewhat more so than Y. In part, this is due to the improved adhesion between scale and alloy reducing scale spallation at temperature, and in part due to possible modification of the Al2O3 grain size. The former factor is far more critical under thermal cycling conditions. Under isothermal conditions the oxidation rate increases with increasing Hf content with all but the 1.5 wt.% alloy oxidizing more slowly than the Hf-free alloy; increase in Y content has the reverse effect. Under thermal cycling conditions the 0.3 and 1.0 wt.% Hf alloys show the lowest overall weight gain. Metallographic evidence suggests that the improved scale adhesion is due principally to a pegging mechanism; the active elements promote the growth of intrusions of Al2O3 into the alloy. However, if the intrusions are too large, they can act as initiators of scale failure.

The high-temperature oxidation of cobalt-21 wt.% chromium-3 vol.% Y2O3 alloys

Alloys of Co-21 wt. % Cr-3 vol. % Y2O3have been prepared by a mechanical alloying method, and oxidized in oxygen at 100 Torr in the temperature range 900–1200°C. The general effects of the dispersed oxide phase are similar to those reported for nickel-base alloys: the selective oxidation of chromium to form a continuous protective Cr2O3scale is promoted; the rate of growth of Cr2O3is reduced compared to dispersoid-free alloys; the adhesion of the Cr2O3is greatly improved; and the scale-forming reaction is probably at the scale-metal interface in the alloys containing the dispersoid, whereas it is at the scale-oxygen interface in dispersoid-free alloys. This last point has not been positively demonstrated. The improvement in adhesion is of particular significance, since the scales on cobalt-base alloys are prone to spallation, and it has been possible to study the mechanism of adhesion in more detail. It appears that in dispersoid-free material the metal recedes from the scale-metal interface, leaving the scale supported on the tops of metal “peaks” but this does not happen in the alloy containing the dispersoid, either because the growth direction of the scale has been changed, or because of changes in the substrate grain size. In general, the observations support the model proposed in an early study for the oxidation of Ni-20 wt.% Cr alloys containing oxide dispersions.

Mechanisms of Na2SO4-induced accelerated oxidation

Abstract The hot corrosion behaviour of various nickel and cobalt-base alloys has been studied using the coating test, in which alloy samples were coated with approx. 1 mg/cm 2 Na 2 SO 4 and oxidized in slowly flowing oxygen at 1 atm. pressure. The reaction kinetics were followed thermogravimetrically, and the corrosion products were examined in detail using conventional metallographic techniques, X-ray diffraction and electron probe microanalysis. In both nickel and cobalt-base alloys, the presence of solid solution strengthening elements, such as molybdenum, tungsten or niobium, are important in promoting accelerated oxidation in the presence of a condensed layer of sodium sulphate on the surface of the alloy. Surprisingly, tantalum does not show a similar effect. The oxides of these elements react with the salt forming molybdates, tungstates or niobates, removing oxide ions and thereby increasing the acidity of the sulphate layer. Both NiO and CoO are then soluble in this acidified salt layer, CoO apparently more so than NiO, and accelerated oxidation can proceed. Alloys containing sufficient chromium to form a continuous. protective Cr 2 O 3 layer are not subject to accelerated oxidation: acid fluxing of Cr 2 O 3 does not occur. However, aluminium containing alloys which under normal circumstances are protected by an Al 2 O 3 scale, do suffer accelerated oxidation because Al 2 O 3 is fluxed from the surface by the molten sulphate layer. Other factors, apart from alloy composition, affect the hot corrosion behaviour. The presence of NaCl can initiate attack in normally resistant high chromium alloys by causing cracking of the protective Cr 2 O 3 scale. Temperature is also important: at low temperatures the influence of the Na 2 SO 4 layer is small because of the slow rates of reaction; at high temperature it is also small because the salt layer is evaporated off the surface. Locally reducing, or low oxygen potential atmospheres can also be important in promoting accelerated oxidation by Na 2 SO 4 , in that they allow sulphur penetration into the alloy and the resultant precipitation of chromium-rich sulphides causes a depletion of chromium such that a protective Cr 2 O 3 scale is not developed.

Oxidation and hot corrosion of nickel-based alloys containing molybdenum

Abstract The oxidation of Ni-15% CrMo alloys has been studied at 900°C in flowing and static oxygen atmospheres. In flowing atmospheres, molybdenum has no effect: all the alloys oxidize in a protective manner. However, in static atmospheres the oxidation rate of alloys with > 3% Mo eventually accelerates, and catastrophic destruction of the alloy takes place. Under these circumstances a molybdenum-rich oxide layer is detected adjacent to the alloy. When specimens are coated with Na 2 SO 4 prior to oxidation, alloys containing > 3% Mo again suffer catastrophic degradation, in either flowing or static atmospheres, and again a molybdenum-rich oxide layer is observed. This suggests that the principal role of the salt coating is to prevent the escape of MoO 3 to the atmosphere. The morphology of the attack in the rapid propagation region is very similar to that obtained in pre-sulphidation/oxidation experiments in the absence of salt and that particular aspect of the reaction is not greatly affected by molybdenum; the aluminium content is more important in determining the nature of the propagation. Attack similar to that exhibited by molybdenum-containing alloys can be obtained with Ni-15%Cr binary alloys in the presence of MoO 3 vapour in the atmosphere, and this might suggest that the MoO 3 reacted with the Na 2 SO 4 to produce an acid (SO 3 -rich) salt, leading to acidic fluxing. However, very similar types of attack were obtained when Na 2 MoO 4 was added to the Na 2 SO 4 , and this should not have affected the acidity of the salt at all. These experiments suggest that acidic fluxing may not be important in the hot corrosion of alloys of this type (molybdenum-containing) and that when catastrophic corrosion is observed, its initiation is probably due to the formation of a molybdenum-rich oxide layer, molten during the reaction. There appears to be a threshold molybdenum content below which attack does not occur, and this seems insensitive to an increase in the chromium content from 15 to 25%.

The effect of a thoria dispersion on the hightemperature oxidation of chromium

Abstract The oxidation of chromium and a Cr-3·12wt% ThO2 alloy has been examined at 1100 and 1200°C in oxygen at 100 torr for 25 h. The rate of oxidation is significantly reduced by the thoria dispersion. The morphology of the oxide has been studied using stereo scanning electron microscopy, coupled with back-scattered electron photographs to locate the thoria particles. At least during the early stages of the oxidation no thoria particles are incorporated in the scale; instead, they accumulate at the metal/oxide interface. The continuous oxide formed on the thoriated alloy has a rather smaller grain size than that formed on pure chromium. The morphology shows that oxygen can diffuse slowly through the oxide scale on the thoria-containing alloy, at least after it detaches from the metal, whereas that formed on the pure chromium appears impervious to oxygen.

The effect of small amounts of silicon on the oxidation of high-purity Co-25 wt. % Cr at elevated temperatures

Some investigators have reported that Co-25 wt.% Cr oxidizes slowly at temperatures in the range 1000–1200°C forming a protective Cr2O3 scale; and this is the normal behavior of cobalt-base superalloys. Others have reported very rapid oxidation, forming a two-layer scale: an outer CoO layer and an inner mixture of Cr2O3 and CoCr2O4 particles in a CoO matrix. This investigation shows that the principal reason for this behavior is the purity of the material; it appears that the rapid mode of oxidation is the intrinsic behavior for high purity material. The most probable impurity to produce the slower mode is silicon, and it is shown that as little as 0.05 wt. % Si is sufficient to change the mode of oxidation provided sufficient oxygen is also present in the alloy: it seems probable therefore that a fine dispersion of SiO2 is responsible.

The oxidation of cobalt—tungsten alloys

Abstract The oxidation behaviour of CoCrW alloys containing from 0–25%Cr and up to 30%W in oxygen at 900–1100°C has been studied. In CoW alloys there is a slight reduction in the oxidation rate as the tungsten content is increased, hwoever this is much mor emarked in Co15CrW alloys. Tungsten has little effect in Co25CrW alloys. On the binary alloys and CoCrW alloys which do not form Cr 2 O 3 , the scale has two layers: an outer, tungsten-free layer of columnar-grained CoO, and an inner layer of CoO containing CoWO 4 precipitates together with CoCr 2 O 4 particles in the ternary alloys. The relative thicknesses of the two layers and the distribution of the constituents in the inner layer depends in temperature and alloy composition. The CoWO 4 and CoCr 2 O 4 particles appear to be responsible for the reduction in oxidation rate by a blocking mechanism in the inner layer. There is some evidence to suggest that tungsten additions to Co−25%Cr alloys assist the exclusive formation of Cr 2 O 3 .

The functional form of rate curves for the high-temperature oxidation of dispersion-containing alloys forming Cr2O3 Scales

It has been shown recently in a number of investigations that the presence of dispersed stable oxides in chromium-containing alloys can result in the formation of protective Cr2O3scales, which appear to grow considerably slower than similar scales on alloys not containing dispersoids. In addition, Cr2O3is removed by further oxidation to the volatile species CrO3,and the rate of this process is unaffected by the dispersoid. Simple kinetic models have been used to describe the results, but it is suggested, on the basis of a curve-fitting analysis, that these approaches are incorrect.