Spin polarisation of ultrashort spin current pulses injected in semiconductors

TL;DR

This paper investigates how different metallic and semiconductor interfaces affect the amplitude, spin polarization, and duration of ultrashort spin current pulses, aiming to improve their propagation in spintronic devices.

Contribution

It expands the study of spin pulse injection to various interfaces, providing guidelines for selecting efficient combinations to enhance ultrashort spin current propagation.

Findings

Different interfaces significantly influence pulse amplitude and polarization.

Certain metallic and semiconductor combinations optimize pulse duration.

Guidelines for experimental interface design are proposed.

Abstract

Ultrashort spin current pulses have a great potential of becoming the carriers of information in future ultrafast spintronics. They present the outstanding property of an extremely compressed time profile, which can allow for the building up of spintronics operating at the unprecedented THz frequencies. The ultrashort spin pulses, however still lack other desirable features. For instance the spatial profile resembles more that of a spill rather than that of a spatially compressed pulse. Moreover the ultrashort spin current pulses can travel only across small distances in metals. The injection of the ultrashort spin pulses from the metallic ferromagnet, where they have to be generated, into a semiconductor is proposed as the first step to overcome both issues by allowing the exited electrons to propagate in a medium with few scatterings. However designing efficient interfaces for the injection is challenging due to practical constraints like chemical and structural stability. This work therefore expands the study of injection to a broader range of interfaces, and analyses how different metallic layers and semiconductors influence the amplitude, the spin polarisation and duration of the ultrashort pulses. This provides guidelines for the selection of efficient interfaces and, equally importantly, experimentally testable trends.