SOME SPECIALLY BUILT semiconductors have been found to display important magnetic properties which could lead to the development of smaller, faster gadgets.
Boffins from the National Institute of Standards and Technology (NIST), Korea University and the University Of Notre Dame carried out tests which demonstrated the rare magnetic property, prompting hopes for future magnetic data storage in semiconductor material. This potentially means faster data processing through built-in logic circuits controlled by electric fields because Semiconductors with magnetic properties would not only be able to process data, but also store it.
Magnetic data storage isn’t something new. It is currently used in computer hard drives and most MP3 players, although the actual storage device is based on metallic rather than semiconductor materials, which slows performance.
What researchers have now discovered is that, in certain circumstances, thin magnetic layers of the semiconductor material gallium arsenide (GaAs) show antiferromagnetic coupling – where a layer aligns its magnetic pole in the opposite direction to the next magnetic layer.
The research team beamed neutrons onto multilayer stacks of GaAs using a technique called ‘polarised neutron reflectometry’. Because neutrons are magnetic and can easily beam through the whole stack, any reflected neutrons demonstrated magnetic properties of individual layers clearly.
Scientists found that at very cold temperatures and small magnetic fields, the polarised neutron data clearly showed signs of antiparallel magnetic alignment in neighbouring layers occurring. Increasing the magnetic field demonstrated a parallel alignment of all layers.
The research is still some way off from finding ways to re-create the results at room-temperature, but the NIST team hopes its breakthrough will be a first definitive step in the right direction. µ
L'I
nq
NanoTechWire
I'll just reverse the polarity of the neutron flow...
These properties have been under study for years since the early 80's with IBM developing MRAM which is very fast compared to DDR3 DRAM for reads and low power its just densitys which they need to crack now
This is along ways off though. I don't think people understand that to get 1tb into a 3.5inch hard disk is quite a task. 

For these circuits to have any real value, the density of bit storage would have to be quite a few orders of magnitude of what their sample shows.

It may pan out or it may become something impractical.