Magnetism in Ge Produced by Implantation with Fe and Mn ions

Since several years there have been strong efforts to produce a new class of semiconductors: diluted magnetic semiconductors (DMS). They should expand the functionality of the common semiconductors by using not only the electrical charge but also the spin of the electrons for the information transfer. Ferromagnetism in semiconductors can be provided by doping with transition metal atoms. One possible doping method is ion implantation.

Besides silicon, germanium is one of the most important elements for semiconductor industry. Recent studies have shown that Ge is generally suitable as base material for DMS. But, doping with only one dopant (Mn or Fe) often results in the formation of clusters which are assumed to be inappropriate for the electronic and magnetic properties. Doping with two dopants, however, may prevent phase separation and the dopant atoms are placed either interstitial or substitutional in the Ge lattice.

We present a method of Fe and Mn double implantation and subsequent flash lamp annealing for Ge(Mn,Fe) preparation. For characterization of their properties different analysing methods were applied: conversion electron Mössbauer spectroscopy (CEMS), Auger electron spectroscopy, super conducting quantum interference device magnetometry and transmission electron microscopy (TEM).

The implantation conditions were chosen in a way that the maximum Fe or Mn concentrations reached about 6 atomic percent and that the peak of the concentration profile lay at about 50 nm. Implantations were performed both at room temperature (RT) and at 260°C.

Room temperature CEM spectra of all as implanted samples consisted of a broad doublet which can be decomposed in two single doublets. This indicates that iron atoms are fixed at two different lattice positions. Magnetic splittings did not occur. Magnetization measurements of the RT-implanted samples showed no indication of ferromagnetism even at 4 K while the magnetization curves of the samples implanted at elevated temperature reveal a clear ferromagnetic hysteresis. The shape of the curves has a typical signature for a growth of magnetic clusters which are undesirable, i.e. implantation at elevated temperature seems to be an improper preparation method for DMS.

Further treatment of the samples implanted at RT is more promising. CEM spectra of the annealed samples were doublets too. However, the CEM spectrum of one sample contained also a singlet. A CEM spectrum of this sample recorded at 80 K revealed a magnetic hyperfine field distribution and a discrete sextet.

Magnetisation measurements showed clear hysteresis loops indicating magnetism but no clustering which was confirmed by TEM. Hence it can be concluded that double ion implantation at RT and subsequent flash lamp annealing could be a way for producing DMS.