Engineering topics
Publications and source records attributed to Smialek, James L..
A barrier layer for a silicon containing substrate which inhibits the formation of gaseous species of silicon when exposed to a high temperature aqueous environment comprises a barium-strontium alumino silicate.
Turbine superalloy components are subject to high-temperature oxidation during operation. Protection is often conferred by coatings designed to form slow-growing, adherent oxide scales. Degradation by oxidation is exacerbated by the thermal cycling encountered during normal aircraft operations. Cooling has been identified as the major contributor to stresses in the oxidation scales, and it may often cause some oxide scale spallation with a proportional loss of protective behavior. Overall oxidation resistance is, thus, studied by the weight change behavior of alloy coupons during high-temperature cyclic oxidation in furnace or burner rig tests. The various characteristics of this behavior are crucial in understanding the performance of alloys at high temperatures. This new modeling effort helps in the understanding of the major factors involved in the cyclic oxidation process. Weight change behavior in cyclic oxidation is typified by an initial parabolic weight gain response curve that eventually exhibits a maximum, then transitions into a linear rate of weight loss due to spalling. The overall shape and magnitude of the curve are determined by the parabolic growth rate, kp, the cycle duration, the type of oxide scale, and the regular, repetitive spalling process. This entire process was modeled by a computer program called the Cyclic Oxidation Spalling Program (COSP) previously developed at the NASA Glenn Research Center. Thus, by supplying appropriate oxidation input parameters, one can determine the best fit to the actual data. These parameters describe real behavior and can be used to compare alloys and project cyclic oxidation behavior for longer times or under different cycle frequencies.
COSP is a publicly available computer program that models the cyclic oxidation weight gain and spallation process. Inputs to the model include the selection of an oxidation growth law and a spalling geometry, plus oxide phase, growth rate, spall constant, and cycle duration parameters. Output includes weight change, the amounts of retained and spalled oxide, the total oxygen and metal consumed, and the terminal rates of weight loss and metal consumption. The present version is Windows based and can accordingly be operated conveniently while other applications remain open for importing experimental weight change data, storing model output data, or plotting model curves. Point-and-click operating features include multiple drop-down menus for input parameters, data importing, and quick, on-screen plots showing one selection of the six output parameters for up to 10 models. A run summary text lists various characteristic parameters that are helpful in describing cyclic behavior, such as the maximum weight change, the number of cycles to reach the maximum weight gain or zero weight change, the ratio of these, and the final rate of weight loss. The program includes save and print options as well as a help file. Families of model curves readily show the sensitivity to various input parameters. The cyclic behaviors of nickel aluminide (NiAl) and a complex superalloy are shown to be properly fitted by model curves. However, caution is always advised regarding the uniqueness claimed for any specific set of input parameters,
This paper summarizes the main effects of sulfur impurity content on the cyclic oxidation resistance of single crystal superalloys, with emphasis on scale and TBC adhesion. Eleven hundred degrees C cyclic oxidation of PWA 1480 produces scale spallation leading to a weight loss of more than 30 Mg/sq cm after 500 one-hr cycles for a sulfur content of 6 ppmw. The sulfur content was reduced to levels below 0.1 ppmw by hydrogen annealing, resulting in weight gains of only 0.5 to 1.0 Mg/sq cm after 1000 one-hr cycles. Samples were produced with various sulfur contents by adjusting the annealing temperature, time, and sample thickness (i.e., diffusion product Dt/L(exp 2)). The subsequent cyclic oxidation behavior, mapped over a sulfur content/thickness diagram, shows a transition to adherent behavior at sulfur levels equivalent to 1 monolayer of total segregation. Additional information is contained in the original extended abstract.
A barrier layer for a silicon containing substrate which inhibits the formation of gaseous species of silicon when exposed to a high temperature aqueous environment comprises a yttrium silicate.
A barrier layer for a silicon containing substrate which inhibits the formation of gaseous species of silicon when exposed to a high temperature aqueous environment comprises a calcium alumino silicate.
A barrier layer for a silicon containing substrate which inhibits the formation of gaseous species of silicon when exposed to a high temperature aqueous environment comprises a yttrium silicate.
A barrier layer for a silicon containing substrate which inhibits the formation of gaseous species of silicon when exposed to a high temperature aqueous environment comprises a calcium alumino silicate.
Four single crystal EPM (enabling propulsion materials) developmental airfoil superalloys were hydrogen annealed at 1300 C for up to 100 hours to remove sulfur and improve oxidation resistance. Although the 1100 and 1150 C cyclic oxidation resistance was remarkably improved by annealing for 24 or 100 hours, the behavior was still considerably inferior to that of commercially available single crystal superalloys, especially those that are either Y-doped or hydrogen annealed. Excessive degradation in the developmental alloys appeared to be correlated with low Cr contents and, to a lesser extent, high Co and Re contents.
The single crystal superalloy Rene'N5 (with or without Y-doping and hydrogen annealing) was cyclically oxidized at 1150 C for 1000 hours. After considerable scale growth (>= 500 hours), even the adherent alumina scales formed on Y-doped samples exhibited delayed interfacial spallation during subsequent water immersion tests, performed up to one year after oxidation. Spallation was characterized by weight loss, the amount of spalled area, and acoustic emission response. Hydrogen annealing (prior to oxidation) reduced spallation both before and after immersion, but without measurably reducing the bulk sulfur content of the Y-doped alloys. The duration and frequency of sequential, co-located acoustic emission events implied an interfacial crack growth rate at least 10(exp -3) m/s, but possibly higher than 10(exp 2) m/s. This is much greater than classic moisture-assisted slow crack growth rates in bulk alumina (10(exp -6) to 10(exp -3) m/s), which may still have occurred undetected by acoustic emission. An alternative failure sequence is proposed: an incubation process for preferential moisture ingress leads to a local decrease in interfacial toughness, thus allowing fast fracture driven by stored strain energy.
Second-order spallation phenomena have been noted for Y-doped Rene'N5 after long term oxidation at 1150 degrees C. The reason for this behavior has not been conclusively identified. A mass equivalence analysis has shown that the surface recession resulting from oxidation has the potential of releasing about 0.15 monolayer of sulfur for every 1 mg/sq cm of oxygen reacted for an alloy containing 5 ppmw of sulfur. This amount is significant in comparison to levels that have been shown to result in first-order spallation behavior for undoped alloys. Oxidative recession is therefore speculated to be a contributing source of sulfur and second-order spallation for Y-doped alloys.
To improve scale adhesion, single crystal superalloys have been desulfurized to levels below 1 ppmw by hydrogen annealing. A transition to fully adherent behavior has been shown to occur at a sulfur level of about 0.2 ppmw, as demonstrated for PWA 1480, PWA 1484, and Rene N5 single crystal superalloys in 1100-1150 C cyclic oxidation tests up to 2000 h. Small additions of yttrium (15 ppmw) also have been effective in producing adhesion for sulfur contents of about 5 ppmw. Thus the critical Y/S ratio required for adhesion was on the order of 3-to-1 by weight (1-to-1 atomic), in agreement with values estimated from solubility products for yttrium sulfides. While hydrogen annealing greatly improved an undoped alloy, yielding <= 0.01 ppmw S, it also produced benefits for Y-doped alloys without measurably reducing the sulfur content.
The 1200 C oxidation behavior of NiAl was examined at various levels of sulfur and zirconium dopants to test the possibility of a critical S/Zr ratio required for adhesion. Cyclic furnace testing for 200 1 -hr cycles and interrupted testing for 500 hr were used as screening tests. Pure NiAl and NiAl(Zr) with 0. 14 at.% Zr were chosen as model base compositions; they exhibited normal, slow-growing scales (3 Mg/sq cm) with excellent adhesion for the Zr-doped alloys. NiAl with about 120 ppma S exhibited a substantial weight loss (-20 Mg/sq cm) in cyclic tests and a very large weight gain (+60 Mg/sq cm) in interrupted tests. The major surface phase remained as alpha -Al2O3. Sulfur doping the NiAl(Zr) alloy caused massive weight gains of 80 - 100 Mg/sq cm, swelling, cracking, and nearly complete conversion into NiAl2O4, and alpha- Al2O3. The initial objective of determining critical S/Zr ratios for adhesion was therefore unattainable. Initiation of the catastrophic attack was examined after a 10 hr exposure, revealing a few sites of broad, raised, and cracked ridges. In cross-section, the ridges appeared as modular intrusions, with a complex, fractal, oxide-metal interface. They were primarily alumina (with occasional entrapped islands of NiAl2O4 or pure Ni metal). They possessed a unique microstructure consisting of 0.3 microns lamellae, separated by 0.1 microns open channels. This allowed for rapid growth controlled by gaseous diffusion. The microstructure is discussed in terms of SO2 evolution and a sulfur-driven de-passivation process.
Titanium aluminides (TiAl) are of great interest for intermediate-temperature (600 to 850 C) aerospace and power-generation applications because they offer significant weight savings over today's nickel alloys. TiAl alloys are being investigated for low-pressure turbine blade applications, exhaust nozzle components, and compressor cases in advanced subsonic and supersonic engines. Significant progress has been made in understanding the fundamental aspects of the oxidation behavior of binary TiAl alloys. However, most of this work has concentrated on short term (<1000 hr), high-temperature (900 to 1000 C) exposures. Also, there is not much data available in the literature regarding the oxidation behavior of the quaternary and higher order engineering alloys. This is especially true for the very long term, low-temperature conditions likely to be experienced during aerospace applications. An investigation at the NASA Glenn Research Center at Lewis Field was undertaken to characterize the long-term oxidation behavior of various model and advanced titanium aluminides for periods up to 7000 hr at 704 C in air using a high-resolution field emission scanning electron microscope. Also, a unique surface treatment technique developed to improve the oxidation resistance of TiAl was evaluated. The alloys included in this investigation are listed in the table. The table also shows typical alloy compositions and the specific weight changes and scale thickness measured for each alloy after exposure to 700 C for 7000 hr in air.
Previously, the effect of hydrogen annealing on increasing the adhesion of Al2O3 scales had been related to the effective desulfurization that occurred during this process. The simultaneous reduction of other impurities has now been re-examined for up to 20 impurity elements in the case of five different alloys (NiCrAl, FeCrAl, PWA 1480, Rene'142, and Rene'N5). Hydrogen annealing produced measurable reductions in elemental concentration for B, C, Na, Mg, P, K, Sr, or Sn in varying degrees for at least one and up to three of these alloys. No single element was reduced by hydrogen annealing for all the alloys except sulfur. In many cases spalling occurred at low levels of these other impurities, while in other cases the scales were adherent at high levels of the impurities. No impurity besides sulfur was strongly correlated with adhesion.
The requirements for hot-gas-path materials in gas turbine engines are demanding. These materials must maintain high strength and creep resistance in a particularly aggressive environment. A typical gas turbine environment involves high temperatures, rapid gas flow rates, high pressures, and a complex mixture of aggressive gases. Figure 26.1 illustrates the requirements for components of an aircraft engine and critical issues [1]. Currently, heat engines are constructed of metal alloys, which meet these requirements within strict temperature limits. In order to extend these temperature limits, ceramic materials have been considered as potential engine materials, due to their high melting points and stability at high temperatures. These materials include oxides, carbides, borides, and nitrides. Interest in using these materials in engines appears to have begun in the 1940s with BeO-based porcelains [2]. During the 1950s, the efforts shifted to cermets. These were carbide-based materials intended to exploit the best properties of metals and ceramics. During the 1960s and 1970s, the silicon-based ceramics silicon carbide (SiC) and silicon nitride (Si3N4) were extensively developed. Although the desirable high-temperature properties of SiC and Si3N4 had long been known, consolidation of powders into component-sized bodies required the development of a series of specialized processing routes [3]. For SiC, the major consolidation routes are reaction bonding, hot-pressing, and sintering. The use of boron and carbon as additives which enable sintering was a particularly noteworthy advance [4]. For Si3N4 the major consolidation routes are reaction bonding and hot pressing [5]. Reaction-bonding involves nitridation of silicon powder. Hot pressing involves addition of various refractory oxides, such as magnesia (MgO), alumina (Al2O3), and yttria (y2O3). Variations on these processes include a number of routes including Hot Isostatic Pressing (HIP), gas-pressure sintering, sinter-HIPing, and Encapsulation-HIPing. It is important to note that each process involves the addition of secondary elements, which later were shown to dramatically influence oxidation and corrosion behavior. As dense bodies of silicon-based ceramics became more readily available, their desirable high temperature properties were confirmed. These materials retained strength to very high temperatures (i.e. 1300-1400 C). Further, they were lightweight and made from abundant materials. SiC and Si3N4 therefore emerged as leading ceramic candidates for components in heat engines, designed to operate at higher temperatures for better performance and fuel efficiency. The first US programs for ceramics in heat engines have been reviewed [6]. Selected programs on ceramic engine parts are summarized here in regard to their contributions to understanding the corrosion behavior of a heat engine environment.
Silicon carbide (SiC) and Si3N4 materials were tested in various turbine engine combustion environments chosen to represent either conventional fuel-lean or fuel-rich mixtures proposed for high-speed aircraft. Representative chemical vapor-deposited (CVD), sintered, and composite materials were evaluated by furnace and high-pressure burner rig exposures. Although protective SiO2 scales formed in all cases, the evidence presented supports a model based on paralinear growth kinetics (i.e., parabolic growth moderated simultaneously by linear volatilization). The volatility rate is dependent on temperature, moisture content, system pressure, and gas velocity. The burner tests were thus used to map SiO2 volatility (and SiC recession) over a range of temperatures, pressures, and velocities. The functional dependency of material recession (volatility) that emerged followed the form A[exp(-Q / RT)](P(sup x)v(sup y). These empirical relations were compared with rates predicted from the thermodynamics of volatile SiO and SiOxHy reaction products and a kinetic model of diffusion through a moving boundary layer. For typical combustion conditions, recession of 0.2 to 2 micrometers/hr is predicted at 1200 to 1400 C, far in excess of acceptable long-term limits.
SiC and Si3N4 materials were tested under various turbine engine combustion environments, chosen to represent either conventional fuel-lean or fuel-rich mixtures proposed for high speed aircraft. Representative CVD, sintered, and composite materials were evaluated in both furnace and high pressure burner rig exposure. While protective SiO2 scales form in all cases, evidence is presented to support paralinear growth kinetics, i.e. parabolic growth moderated simultaneously by linear volatilization. The volatility rate is dependent on temperature, moisture content, system pressure, and gas velocity. The burner tests were used to map SiO2 volatility (and SiC recession) over a range of temperature, pressure, and velocity. The functional dependency of material recession (volatility) that emerged followed the form: exp(-QIRT) * P(exp x) * v(exp y). These empirical relations were compared to rates predicted from the thermodynamics of volatile SiO and SiO(sub x)H(sub Y) reaction products and a kinetic model of diffusion through a moving, boundary layer. For typical combustion conditions, recession of 0.2 to 2 micron/h is predicted at 1200- 1400C, far in excess of acceptable long term limits.