DMS and nano-particles 

In order to utilize the semiconductors for logic functions using spin of an electron, the semiconductor need to be made magnetic. The workhorse of the industries – Si, GaAs, Ge are non-magnetic semiconductors. There are two possible ways to make them magnetic-doping them with magnetic ions and injecting spins from magnetic metals.  There are other ways to achieve spin imbalance in non-magnetic semiconductors-utilizing the intrinsic spin-orbit coupling as a source of the spin-deflection to the external electric field or spin Hall effect (intrinsic or extrinsic).

Our research effort is directed towards making bulk Si magnetic. We dope it with a dilute concentration of Mn (less than a percent), and holes (p-type host). These are experimental conditions used in probing the magnetism in single-crystalline Si.  Various experiments on magnetism in Mn doped bulk Si claim the observation of room temperature ferromagnetism. The source of uncertainty in the conclusion was attributed to the fact that the magnetic signal they measure may be due to the implantation damage (ion implantation is one of the way to insert Mn) or clustering of Mn ions, vacancies etc Our work [1] on boron doped Si and co-doped with two Mn ions along [111] direction (to avoid clustering, we placed them far apart) shows that they couple in an anti-ferromagnetic way. We considered the two Mn replacing two Si (substituitional), no Si (interstitial) and one Si (one substituitional and one interstitial). We also considered possibility of point vacancies created by ion-implantation. We find that the two interstitial Mn atoms with anti-ferromagnetic coupling is the lowest energy structure. We believe that if Mn ions are placed along any other crystallographic direction or placed close to each other (i.e. clustering), there may be a ferromagnetic order. It may be possible that moments on Mn are oriented in non-collinear fashion and the Mn ions are distributed at random (disorder). In this case, the saturation magnetization must decrease. We have not taken into account these possibilities in our calculation.

We identified high-spin and low-spin magnetic states (high spin means more magnetic moment) in Mn doped Si nano-crystals [2]. The spherical Si nano-crystals (diameter ~ 2.5 nm), whose surfaces were passivated by hydrogen atoms, were built from the bulk Si.   Depending upon whether the Mn ion is located at the central or near sub-surface region, we get high spin or low-spin states. This cross-over from high-spin to low-spin states or vice-versa can be explained by invoking the energetic arguments- competition between magnetic exchange energy and coulomb energy due to local ligand-field from surrounding Si atoms. If magnetic exchange energy wins, then we get high-spin state otherwise, the low-spin state. The energy barrier between the two states is couple of hundreds of meV, far greater than the room temperature thermal energy. This precludes the thermal transition between these two states thereby opening up a possibility of tuning these states with an external electric field. This is a daunting task as the coupling between the local moments and the external field needs to be established either directly or indirectly.