Nanomagnetic Logic

The exponential improvements in speed, energy efficiency, and cost that the computer industry has relied on for growth during the last 50 years are in danger of ending within the decade. These improvements all have relied on scaling the size of the silicon-based transistor that is at the heart of every modern CPU down to smaller and smaller length scales. However, as the size of the transistor reaches scales that are measured in the number of atoms that make it up, it is clear that this scaling cannot continue forever.As a result of this, there has been a great deal of research effort directed at the search for the next device that will continue to power the growth of the computer industry. However, due to the billions of dollars of investment that conventional silicon transistors have received over the years, it is unlikely that a technology will emerge that will be able to beat it outright in every performance category. More likely, different devices will possess advantages over conventional transistors for certain applications and uses.One of these emerging computing platforms is nanomagnetic logic (NML). NML-based circuits process information by manipulating the magnetization states of single-domain nanomagnets coupled to their nearest neighbors through magnetic dipole interactions. The state variable is magnetization direction and computations can take place without passing an electric current. This makes them extremely attractive as a replacement for conventional transistor-based computing architectures for certain ultra-low power applications.In most work to date, nanomagnetic logic circuits have used an external magnetic clocking field to reset the system between computations. The clocking field is then subsequently removed very slowly relative to the magnetization dynamics, guiding the nanomagnetic logic circuit adiabatically into its magnetic ground state. In this dissertation, I will discuss the dynamics behind this process and show that it is greatly influenced by thermal fluctuations. The magnetic ground state containing the answer to the computation is reached by a stochastic process very similar to the thermal annealing of crystalline materials. We will discuss how these dynamics affect the expected reliability, speed, and energy dissipation of NML systems operating under these conditions.Next I will show how a slight change in the properties of the nanomagnets that make up a NML circuit can completely alter the dynamics by which computations take place. The addition of biaxial anisotropy to the magnetic energy landscape creates a metastable state along the hard axis of the nanomagnet. This metastability can be used to remove the stochastic nature of the computation and has large implications for reliability, speed, and energy dissipation which will all be discussed.The changes to NML operation by the addition of biaxial anisotropy introduce new challenges to realizing a commercially viable logic architecture. In the final chapter, I will discuss these challenges and talk about the architectural changes that are necessary to make a working NML circuit based on nanomagnets with biaxial anisotropy