Spinning into a new era of computing

Flipping a spin is easier and faster than moving around electrons. So are we going to switch to spintronics? Michael Gross investigates.

Weird things happen when you make semiconductor particles smaller and smaller until you reach the nanometre scale. In these so-called Q particles or quantum dots, physical properties like the band gap and even the colour of the particle will depend on its size. The same material may form 6nm grains that are red, and 4nm dots that turn out green ( Chem. Br.  , September 2003, p21). Apart from applications in the molecular scale - dyes and tags - quantum dots are also promising candidates for quantum informatio n processing and particularly for spintronics devices, where the information is carried by the spin rather than the charge of an electron. 

Spintronics (short for spin-based electronics) or magnetoelectronics comes in three main flavours: based on manipula ting the spins in metals, magnetic semiconductors, or single electrons. The metal variety, based on an effect known as giant magnetoresistance, is already used in the high-capacity hard drives of today’s computers. The other two varieties are still in thei r infancy, but they are catching up, thanks in part to the intriguing potential of quantum dots. 

Min Ouyang and David Awschalom at the University of California at Santa Barbara, US, have recently succeeded in using a self-assembly approach to connect different kinds of quantum dots with conjugated molecular chains. The researchers formed a multi-layered network that allowed them to probe in detail the behaviour of electron spins in this system. 

The quantum dots used in this study were CdSe nanoparticles of 3.4 and 7.0 nm diameter, referred to as A and B, respectively. Starting from a fused silica surface functionalised with amine or thiol groups, the researchers applied the first monolayer of quantum dots by immersing the surface in a suspension of the required particles. To link the dots to each other and allow them to add on the next layer, they added a dithiol compound, in this case 1,4-benzenedimethanethiol. By alternating addition of quantum dots and dithiols, they created layer cakes of different quantum dots, with layerings such as ABAABA. 

Because of the size-dependent properties of quantum dots, A and B can easily be addressed and distinguished spectroscopically. Ouyang and Awschalom used circularly polarised laser pulses to excite electron spins in one specific type of particle. 

After a delay time, they measured the absorption of a second kind of laser pulses to check up on the fate of the spin changes. Using this method, known as time-resolved Faraday rotation measurements, the researchers detected significant transfer of spins from the B-layer to the A-layer, that is from the larger to the smaller quantum dots, over a wide range of temperatures, from 4.5K up to room temperature. Spin transfer efficiency is 12 per cent at very low temperatures, but reaches up to 20 per cent at 20?C. 

While the mechanism of this spin transfer remains speculative, the authors propose that it occurs via the delocalised p -orbitals of the aromatic dithiols, rather than through physical mechanisms such as tunnelling or dip ole-dipole interaction. It appears that molecules like the benzene derivative used here not only tie the quantum dots together, but also couple them electronically in a very useful way. 

Shortly after this bottom-up approach, K. V. Rao’s group at the Royal Institute of Technology, Stockholm, Sweden, reported in Nature  the successful conclusion of a top-down approach to room temperature spintronics, the feasability of which had been predicted three years earlier by Tomasz Dietl (Polish Academy of Sciences, Warsaw). Rao and coworkers succeeded in creating a magnetic semiconductor by doping ZnO with Mn, sintering at relatively low temperatures (below 700?C). They showed that the resulting material is ferromagnetic both in bulk and in transparent thin films. 

Semiconductors doped with individual atoms for spintronic use have a head start on the other approaches, because this technology can be directly hitched to all the manufacturing marvels that the computer industry has developed in its race against Moore’s law. Thus, in the mid-term, the next spintronic devices after the already existing metallic ones might well be based on the achievements of Dietl’s and Rao’s groups. 

In the long-term, however, the bottom-up approach and use of individual electrons trapped in quantum dots opens up the weird but also extremely promising perspective of large-scale quantum computers.