热电效应
结构精修
晶体结构
材料科学
结晶学
粉末衍射
热电材料
塞贝克系数
化学
热力学
物理
作者
Jianli Mi,Pingjun Ying,Mattia Sist,Hazel Reardon,Peng Zhang,Tiejun Zhu,Xinbing Zhao,Bo B. Iversen
标识
DOI:10.1021/acs.chemmater.7b01768
摘要
Gaining insight into crystal structure is essential for understanding thermoelectric transport mechanisms and predicting thermoelectric properties. The main challenge in studying thermoelectric mechanisms is often imprecise or wrong models of the crystal structure. This work examines the structure modifications observed in MgAgSb thermoelectric materials by multitemperature high-resolution synchrotron radiation powder X-ray diffraction (SR-PXRD). Rietveld refinement reveals large atomic displacement parameters (ADPs) of the Ag1 atoms at the 4a position indicating possible atomic disorder, which may contribute to the low thermal conductivity observed in α-MgAgSb. The temperature dependence of anisotropic structural parameters indicates a tendency of increasing structural symmetry in α-MgAgSb with increasing temperature, largely contributing to the temperature evolution of the thermoelectric properties. Two MgAgSb polymorphs (β-MgAgSb and γ-MgAgSb) coexist at 700 K, and only the γ-MgAgSb crystalline phase is found at high temperatures (800–1000 K). The content of γ-MgAgSb phase decreases with temperature due to the increase of liquid impurities, and the sample is only 43.8% crystalline at 1000 K. At 800 K, the high resolution powder data are fitted equally well using type I (with Mg, Ag, and Sb on the 4b, 4c, and 4a sites, respectively) and type II (with Mg, Ag, and Sb on the 4a, 4b, and 4c sites, respectively) half-Heusler crystal structure models. Nonetheless, maximum entropy method (MEM) analysis carried out on the extracted factors shows that the type II structure gives a more physically sound MEM electron density. The disorder in γ-MgAgSb consists of mixed sites of Mg and Ag as well as vacancies, and the strong disorder of the cation sublattice contributes to the low thermal conductivity.
科研通智能强力驱动
Strongly Powered by AbleSci AI