Cyclic Oxidation of Hastelloy X

Abstract

This investigation has been conducted as a part of a program to evaluate and select high-temperature alloys which are particularly suitable for fabricating transpiration-cooled (porous) materials used in turbine blades and similar engine components. The kinetics of cyclic oxidation of Hastelloy X in the form of sheet and wire were studied over the temperature range 760~176 The chemical composition of the alloy is listed in Table I, as provided by the manufacturer. Sheet specimens were cut to nominal dimensions of 6 • 0.5 • 0.060 in. Wire specimens were 0.005 _ 0.0005 in. in diameter. Surface contamination of specimens was removed by sonic-cleaning in hot trichloroethylene followed by acetone rinsing. After cleaning the specimens were annealed for 8 hr at 1150~ in dry hydrogen (dew point below-62~ Wires were loosely wound around prefired mullite tubes for support during the annealing cycle. The specimens appeared clean and bright after the anneal. The sheet specimens were then polished with successively finer grades of abrasives, and were finished with 320 grit silicon carbide. A bright, smooth surface of approximately 10 ~in. RMS roughness was produced. Final sonic cleaning was accomplished in hot trichloroethylene with an acetone rinse. Sheet specimens were contained in zircon ceramic thimbles with four-point minimum contact at the corners of the specimen. Wire specimens consisted of loose bundles about 6 in. long and 11 89 in. in diameter containing approximately 100 ft of 0.005 in. diameter wire which weighed about 3g. Each wire bundle was placed in a separate zircon thimble. All thimbles were fired at 1600~ and baked out at 760~ to constant weight before using. Separate specimens (a total of 9 for each temperature) were subjected to cyclic oxidation in air for times of 4, 16, 64, 100, 200, 300, 400, 500, and 600 hr at each given temperature. After each exposure cycle, a tray containing all specimens was removed from the furnace and air-cooled. One specimen was removed for evaluation and the rest returned to the furnace. Temperatures are generally __5 ~. Calibrated Chromel-Alumel thermocouples of heavy wire were placed at several locations on each tray and temperatures recorded continuously. At higher temperatures , thermocouples were replaced frequently. Specimen weight and oxidation weight gain were determined to _0.1 mg with an analytical balance after each oxidation cycle. At high temperatures where the amount of oxide spall was large, the ceramic thimble and the specimen were weighed together before and after oxidation, and this method was found to be more reproducible than direct weighing. Further details of the experimental procedure together with the resulting data, which is analyzed herein, have been reported elsewhere (1, 2). The weight-gain vs. oxidations time curves for the sheet specimens and for the wire specimens are shown in Fig. 1 and 2, respectively; in the case of wire specimens , the mean surface area was used for calculating the specific weight gain, based on the initial and final diameters of the metal. In all cases, initially the weight increases parabolically in a manner similar to that often found for continuous oxidation (3, 4). At higher temperatures and longer times sheet specimen behavior was observed to be erratic. In these cases, the differing spall behavior of the individual specimens explains the observed results because fresh surfaces oxidize rapidly. Also, a few points seem to show a large experimental error at lower temperatures, partly because weight gain was small. The oxidation rate constants calculated for the sheet specimens are given in Table II, and for the wire specimens in Table III. The rate constants are plotted as an inverse function of temperature in Fig. 3. The rate constants obtained from continuous oxidation of Hastelloy X wire (5) and sheet (6) are also included in Fig. 3 for comparison. A least-squares fit to the Ar-rhenius plots yields an activation energy of 49.9 _ 2.6 kcal/mole for the sheet specimens and 51.6 _ 1.5 kcal/ mole for the wire specimens, respectively. Taking into account the estimated experimental errors (• the activation energies become 49.9 __. 5.1 kcal/mole, and 51.6 _-4-4.1 kcal/mole, respectively. The activation energies are practically equal for both types of speci-Table I. Chemical composition of Hastelloy X, weight per cent