Ultra-thin ferromagnets, when coupled with magnetoelectric or multiferroic materials, could potentially enable highly energy-efficient electric field control of the magnet for use in nanoelectronic memories. Substitutional doping of magnetic impurities in monolayer transition metal dichalcogenides (TMDs) may be a promising way to create 2D ferromagnets but, according to theoretical calculations, require high doping levels (10-20 %) to achieve above room temperature (RT) Curie temperature. Room-temperature ferromagnetism has been reported for very low doping levels (0.1-1%), in conflict with the theoretical calculations, and always in the presence of high concentrations of defects, making it unclear if the dopants alone are responsible for this ferromagnetism. In this work, we use a combination of molecular beam epitaxy (MBE), plan-view TEM, XRD, XPS, Raman, and magnetometry to study iron and vanadium doping in monolayer WSe2. We show that optimally grown monolayers with up to 35% doping and low defect density show no ferromagnetism. Interestingly, ferromagnetism is observed when these monolayers contain a significant amount of selenium vacancies (Sevac), intentionally created via a post-growth heat treatment process, and magnetism is seen to scale with heating time/vacancy concentration. Moreover, even undoped WSe2 shows similar ferromagnetism for Sevac > 1014 cm-2. This ferromagnetism can later be quenched by filling these vacancies via Se annealing. For films with low Se vacancy concentration, TEM analysis reveals significant clustering of the V and Fe dopants which likely suppresses the ferromagnetism predicted by theory - a problem that has previously plagued III-V based dilute magnetic semiconductors as well. We go so far as to claim that all of the above room-temperature ferromagnetism reports in the literature thus far are due to Se vacancies and not magnetic doping.
Read full abstract