"transistor-like" spin nano-switches: Physics and applications
Progress in the last two decades has effectively integrated spintronics and nanomagnetics into a single field, creating a new class of spin-based devices that are now being widely used in magnetic memory devices. However, it is not clear if these advances could also be used to build logic devices. The objective of this thesis is three-fold: The first is to describe a general paradigm for combining Read and Write units used in memory devices into transistor like nano-switches having input-output isolation and gain. Such switches could be used to build logic circuits without the need of any external circuits or amplification. The second is to describe an experimentally benchmarked simulation model for designing a concrete implementation of a transistor-like switch based on: Giant Spin Hall Effect (Write), Magnetic Tunnel Junction (Read) and magnetic coupling for isolation. It turns out that the model can also be used to analyze/design stray fields in perpendicular magnetic tunnel junction (pMTJ), an important problem in scaled pMTJ devices. The third is to discuss the novel features and possible new class of circuits of spin nano-switches. We will first describe a spin switch nano-oscillator based on the standard principle of incorporating feedback into a device with gain. We then discuss how spin nano-switches can be used to implement different types of neural networks such as second generation, third generation and deep belief neural networks.
Datta, Purdue University.
Off-Campus Purdue Users:
To access this dissertation, please log in to our