Search for: All records

Computations have predicted good thermoelectric performance for a number of Zintl phases when doped n-type. Combined with the successful experimental realization of n-type KGaSb 4 , KAlSb 4 , and Mg 3 Sb 2 with zT ≳ 1, this has fueled efforts to discover novel n-type dopable Zintl phases. However, a majority of Zintl phases exhibit strong proclivity toward p-type doping and prior successes in finding n-type dopable Zintls were largely serendipitous. Herein we use modern first-principles defect calculations to study trends in the dopability of Zintl pnictides and find that the average oxidation state of the anion is a useful chemical guide to identify novel n-type dopable phases. Specifically, we observe that Zintl pnictides with average oxidation of the anion near −1 are n-type dopable. The trend is mainly a consequence of the high formation energy of native acceptor defects ( e.g. cation vacancies) and the resulting absence of charge (electron) compensation. Using the oxidation state guide in conjunction with a descriptor of thermoelectric performance, we conduct a large-scale materials search and identify promising candidates that are n-type dopable. 

Semiconductor engineering relies heavily on doping efficiency and dopability. Low doping efficiency may cause low mobility and failure to reach target carrier concentrations or even the desired carrier type. Semiconducting thermoelectric materials perform best with degenerate carrier concentrations, meaning high performance in new materials might not be realized experimentally without a route to optimal doping. Doping in the classic PbTe thermoelectric system has been largely successful but reported doping efficiencies can vary, raising concerns about reproducibility. Here, we stress the importance of phase equilibria considerations during synthesis to avoid undesired intrinsic defects leading to sub-optimal doping. By saturation annealing at 973 K, we decidedly fix the composition in single crystal iodine-doped PbTe samples to be Pb-rich or Te-rich without introducing impurity phases. We show that, regardless of iodine concentration, degenerate n-type carrier concentrations with ideal doping efficiency require Pb-rich compositions. Electrons in Te-rich samples are heavily compensated by charged intrinsic Pb vacancy defects. From Hall effect measurements and a simple defect model supported by modern defect calculations, we map out the 973 K ternary Pb–Te–I phase diagram to explicitly link carrier concentration and composition. Furthermore, we discuss unintentional composition changes due to loss of volatile Te during synthesis and measurements. The methods and concepts applied here may guide doping studies on other lead chalcogenide systems as well as any doped, complex semiconductor.