James R. Keiser

Exposure of Ceramics and Ceramic Matrix Composites in Simulated and Actual Combustor Environments

A high-temperature, high-pressure, tube furnace has been used to evaluate the long term stability of different monolithic ceramic and ceramic matrix composite materials in a simulated combustor environment. All of the tests have been run at 150 psia, 1204 degrees C, and 15% steam in incremental 500 h runs. The major advantage of this system is the high sample throughput; >20 samples can be exposed in each tube at the same time under similar exposure conditions. Microstructural evaluations of the samples were conducted after each 500 h exposure to characterize the extent of surface damage, to calculate surface recession rates, and to determine degradation mechanisms for the different materials. The validity of this exposure rig for simulating real combustor environments was established by comparing materials exposed in the test rig and combustor liner materials exposed for similar times in an actual gas turbine combustor under commercial operating conditions.

Experimental determination of the residual stresses in a spiral weld overlay tube

Neutron diffraction was used to determine the residual stresses in a spiral weld overlay tube. The specimen was a 2.5 in. OD carbon steel tube covered with a layer of alloy 625 weld overlay. Residual strains in the carbon steel and weld overlay layers were determined using the ferritic (211) and austenitic (311) reflections, respectively. The residual stresses in each material were derived from the measured strains using Hookes law and appropriate elastic constants. Tensile stress regions were found not only in the weld metal but also in the heat-affected zone in the carbon steel. The maximum tensile stress was located in the weld overlay layer and amounted to 360 MPa, about 75% of the 0.2% yield strength of the weld metal. The experimental data were compared with a finite element analysis based on an uncoupled thermal-mechanical formulation. Overall, the modeling results are in satisfactory agreement with the experimental data, although the hoop strain (stress) seems to have been overestimated by the finite element model. Additional neutron diffraction measurements reveal that these welding residual stresses vanished after annealing at 900°C for 20 min.

Nitrogen impurity gettering in oxide dispersion ductilized chromium

Work by Scruggs in the 1960s demonstrated that tensile ductility could be achieved at room temperature in powder metallurgically-produced Cr alloyed with MgO. During consolidation, much of the MgO converted to the MgCr{sub 2}O{sub 4} spinel phase, which was hypothesized to getter nitrogen from the Cr, rendering it ductile. We have duplicated this effect, achieving room temperature tensile elongations of 4% for hot-pressed Cr-6MgO-(0-1)Ti (wt.%) and 10% for hot-pressed and extruded Cr-6MgO-0.75Ti. Direct incorporation of nitrogen into the MgCr{sub 2}O{sub 4} phase was not detected; however, impurities, particularly nitrogen and sulfur, were observed to segregate to and/or precipitate at interfaces between the MgO/MgCr{sub 2}O{sub 4} phases and the Cr matrix. Exploratory studies of other non-spinel forming oxide dispersions (La{sub 2}O{sub 3}, TiO{sub 2} and Y{sub 2}O{sub 3}) showed a similar pattern of impurity segregation/precipitation, suggesting that there is nothing unique about spinel dispersions in Cr with regards to impurities. However, none of these other dispersions resulted in similar levels of tensile elongation.

Influence of residual stress on thermal expansion behavior

We demonstrate that the thermal expansion behavior of a material can be substantially modified by the presence of residual stresses. In the case of a composite tube made of two layers of dissimilar steels, in situ neutron diffraction measurements revealed a significant difference in the coefficients of thermal expansion along the radial and tangential directions. It is shown that the observed difference in thermal expansion is due to the change of residual stresses with temperature.

Characterization of chromia scales formed in supercritical carbon dioxide

Abstract Initial experimental work at 700°–800 °C is in progress to develop a lifetime model for supercritical CO2 (sCO2) compatibility for a 30-year lifetime of a >700 °C concentrated solar power system. Nickel-based alloys 282, 740H and 625 and Fe-based alloy 25 are being evaluated in 500-h cycles at 1 and 300 bar, and 10-h cycles in 1 bar industrial grade CO2. The alloys showed similar low rates of oxidation in 1 and 300 bar CO2 in 500-h cycles at 750 °C. However, in 10-h cycles, alloy 25 showed accelerated attack at 700° and 750 °C. Transmission electron microscopy scale cross-sections on alloy 25 after 1000 h at 700 °C in sCO2 and in air only showed a small row of carbides beneath the scale in the former environment. Similar characterisation was performed on alloys 625 and 282 after sCO2 exposure at 750 °C.

Mechanical behavior of erosion-corrosion scales on steels as characterized by single-particle impacts

Samples of 2.25Cr-1Mo (less than 0.5 Si) and 2.5Cr-0.55Mo-1.4Si steels were eroded-corroded at 450 and 650 °C using fluidized bed combustor bed particles at velocities of 10 and 20 m s−1. The steel with higher silicon content showed significantly lower metal loss rates under all conditions. The samples were subsequently subjected to single-particle impacts using spherical WC particles at velocities around 50 m s−. The impact response of the scales could be explained in terms of a combination of substrate hardness and scale morphology effects but could not be consistently related to the superior erosion-corrosion resistance of the steel with higher silicon content. All scales were composed of oxidation product and deposited bed material erodent. Samples eroded-corroded at 450 °C had denser, more mechanically stable scales which could be associated with the generally lower erosion-corrosion rates at this temperature. At 650 °C the scales were more loosely packed, especially at the lower erosion-corrosion velocity, which resulted in apparent ductility by permitting them to densify under impact. Scales were either segmented or continuous in appearance. Thick continuous scales maintained their integrity under the lower velocity conditions of the erosion-corrosion tests, thus leading to low metal losses, but spalled catastrophicaliy under the single impacts. Segmented scales spalled in smaller pieces under single impacts. It is proposed that the segmented scales would exhibit significant failure under low velocity conditions, thus providing less protection to the steels than continuous scales under similar conditions.

Residual stresses due to processing of composite tubes

X-ray and neutron diffraction were used to characterize residual stresses in composite tubing of a corrosion-resistant clad alloy on carbon steel. A useful X-ray method, based on the measurement of the fcc (3 1 1) reflection using Cr K{sub {beta}} radiation, was developed which allowed precise determination of surface residual stresses in the textured clad layer. Neutron diffraction measurements were carried out in both the carbon steel core and the clad layer, using the bcc (2 1 1) and fcc (3 1 1) reflections, respectively. The neutron diffraction results are consistent with surface residual stresses determined with X-ray diffraction. However, the through-thickness stress profiles established by X-ray and neutron diffraction do not agree with elastic calculations based on the thermal expansion mismatch between the carbon steel and clad alloy. The differences between the calculation and experimental results are discussed.

A comparison of three microindentation hardness scales at low and ultralow loads

Abstract Indentation hardness tests were performed on thick, fine-grained, electro-formed deposits of copper and nickel using Knoop, Vickers, and Berkovich indenters. The latter type of indenter was used for shallow penetrations (85–1750nm), and results are reported in terms of nanoscale hardness (NH) numbers. Knoop and Vickers indenters were used with applied loads of between 0.15 and 0.98 N, and at the lowest load, produced indentation depths comparable to the larger ones obtained with the Berkovich indenter. The NH numbers became very sensitive to penetration depth when the penetration depth was less than certain critical values. NH numbers for Cu and Ni were higher than those for Knoop and Vickers testing at comparable penetration depths. Applying indenter area function corrections to calculate hardness numbers (i.e., considering projected area versus facet contact area) resulted in a closer correlation between microhardness and nanohardness scales; however, changes in the tip shape because of wear or other imperfections can lead to inaccurate calculation of NH numbers at the lowest loads. Results also suggest that the interconversion of lowload hardness numbers from one scale to another can be material-dependent.

Characteristics of individual impact craters on selected aluminum alloys

Abstract When an erosive particle impacts the surface of a target, an appreciable amount of energy is deposited in the deformed region. The deformation generally causes strain hardening of the surface but local heating, that is manifested as melting or softening of the surface, has been reported. To assess the effect of individual impacts on the surface of a target, a series of aluminum alloys were impacted by tungsten carbide spheres, 343 /gmm in diameter, at 28.5 m s−1 and 30° incidence angle. A mechanical properties microprobe was used to measure the hardness of the crater bottoms and the material just below the surface of the craters. Strain hardening was observed in these materials but there was no evidence of softening.

Corrosion of type 316 stainless steel in molten LiF-LiCl-LiBr

Abstract The properties of LiF-LiCl-LiBr salt make it attractive as a solvent for extracting tritium from a fusion reactor lithium blanket. Consequently, the corrosion of type 316 stainless steel by flowing (about 15 mm/s) LiF-LiCl-LiBr at a maximum temperature of 535°C was studied to determine whether compatibility with the structural material would be limiting in such a system. The corrosion rate was found to be low (