An internally circulating fluidized-bed boiler is composed of a combustion chamber and a heat recovery cell. The heat exchange tubes are exposed to complex erosive and corrosive conditions due to impact of sand particles under high-temperature corrosion in chlorine containing atmospheres, which results in the severe degradation of the materials. In general, protective coatings are applied on the heat exchange tubes to improve an erosion-corrosion (E-C) resistance. In our previous study, high-temperature E-C resistance of Ni-xFe alloys (x = 0, 10, and 30 wt.%) was found to increase with increasing Fe content because of a formation of the Fe-rich oxide scale, which is harder than a Ni oxide scale. However, factors affecting an E-C behavior such as the impact energy of sand particles and the salt concentration have yet to be sufficiently evaluated. In this study, the effect of the impact energy of silica sand on the high-temperature erosion-corrosion of Ni-xFe alloys was investigated.Pure Ni, Ni-10Fe and Ni-30Fe (wt%) alloys were prepared by Ar arc melting. Samples with a thickness of approximately 1mm were cut from the homogenized ingots and polished down to a 1200 grit using SiC papers. Figure 1 shows the schematic of the E-C test rig. The sample holder with affixed samples was inserted into the quartz tube containing silica sand (average particle size 500μm). The silica sand was heated to 700°C, and fluidized by compressed air preheated to 400°C at a flow rate of 25 L/min. The impact angle of the sand to the surface was set to be 45°, and the sample temperature was maintained at 360-380°C by water cooling of the sample holder. The salt vapor of KCl-40 mol.%NaCl-20 mol.%CaCl2 was continuously supplied from the bottom of quartz tube into sand by a compressed air. The impact energy of sand particles was set at 540 J/m2 and 1150 J/m2, which were calculated by a fluidized bed analysis. The high-temperature E-C test was stopped every 25 h or 50 h to change the silica sand and salt crucible , and continued for up to 100 h. After the E-C tests, the samples were analyzed by XRD, FE-SEM, EDS, and STEM. The STEM samples were prepared by using FIB.Figure 2 shows the relationship between the impact energy of silica sand and the mass change of each sample after 50 h of E-C test. The mass change at 0 J/m2 was defined as the mass change after the embedded corrosion test for 50 h in silica sand with 0.01 wt.% of KCl-40 mol.%NaCl-20 mol.%CaCl2 salts. The slope of the graph decreased with increasing Fe content, suggesting that the effect of impact energy on the E-C resistance became less pronounced for the alloy with higher Fe content. Regardless of the impact energy, no oxide scale formation was observed on the pure Ni after the E-C test, which indicates that mass loss might be caused by the erosion of the metal substrate. In Ni-10Fe, a thin (Fe, Ni) spinel was remained on outermost surface of the oxide scale and NiO containing several percent of Fe was remained beneath the spinel layer after the test at 540 J/m2, but no oxide scale was remained after the test at 1150 J/m2, which indicates that the mode of high temperature E-C might have changed from erosion of the oxide scale to that of the metal substrate as impact energy increased. In Ni-30Fe, Fe3O4 and (Fe, Ni) spinel was remained on most of the alloy surface after the test at 540 J/m2, suggesting that the E-C is predominated by erosion of oxide scale. However, some scratches due to the impact of sand particles were observed after the test at 1150 J/m2, and EDS analysis confirmed that the signals from oxygen was very weak on the scratched area, which indicates that not only the erosion of oxide scale but also the erosion of alloy substrate could occur in Ni-30Fe. These results suggest that the influence of impact energy depends on the type of oxide scale formed, and the dependence of the erosion rate constant on impact energy could be smaller for the Fe-rich oxide scale. Figure 1
Read full abstract