Introduction

Magnetic ferrite core memory was invented and produced in the 1960s, prior to semiconductor memory. Ferrite cores are made from a paste of ferrite powers, which are sintered at high temperature. The process of forming a discrete core is not as scalable as the integrated circuit process on a silicon wafer. The product life of a magnetic core was short, and in the 1970s this technique was replaced by semiconductor memory. A similar fate happened to magnetic bubble memory, another type of magnetic memory, which was built on a magnetic garnet material substrate (gadolinium gallium garnet, Gd3Ga2(GaO4)3). The bit density of bubble memory technology is scalable since it is made with a planar process, similar to the silicon integration circuits. However, because it is on garnet, it is passive and cannot perform logic functions (such as address decoding), and it requires a companion silicon chip to provide the logic function to complete the memory access function. Even with better memory performance, magnetic bubble memory could not compete against magnetic hard disk and semiconductor memory, which continue to show a clear path of scaling for a lower cost. By the mid 1980s, commercial magnetic bubble production had ended.

Subsequent efforts in the development of magnetic memory have been focused on the integration of magnetic thin-film memory devices into silicon wafer processes. Magnetic memory devices exist in the form of thin-film stacks, which can easily be integrated into the back-end metal wiring metallurgy process.

The advent of semiconductor technology has impacted the lives of many of us since the 1970s. Silicon CMOS (complementary metal-oxide-semiconductor) devices are practically ubiquitous, and by the year 2000, the value of the semiconductor industry exceeded that of the automobile industry. The magnetic industry, on the other hand, is much smaller than the semiconductor industry. Engineering schools of universities rarely cover any courses in this discipline. Nonetheless, a tiny magnetic recording device is in the hard disk of every computer. Like CMOS devices, magnetic recording technology is being scaled down from generation to generation. At the time of writing, the physical size of the magnetic bit remains smaller than a DRAM bit on silicon chips.

Researchers working in these two communities had little in common until the development of the modern magnetic random access memory, or MRAM. A MRAM chip is built by integrating magnetic tunneling junction (MTJ) devices onto the silicon CMOS circuits. The research activity of MTJs in academia and industry, both hard disk and semiconductor, has been very active since it first showed signs of technology implication in the mid 1990s. That effort led to the mass production of the MTJ recording head in hard disk in 2006. In the same year, the semiconductor industry announced the first successful introduction of an MTJ memory product. The viability of MTJ technology is proven. It is expected that research activities will develop further, which will increase cooperation between these two research communities.

Introduction

Two thousand years ago, the Chinese invented the compass, a special metallic needle with one end always pointing to the North Pole. That was the first recorded human application of magnetism. Important understandings and developments were achieved in the mid 19th century and continue into the present day. Indeed, today, magnetic devices are ubiquitous. For example, to name just two: energy conversion devices provide electricity to our homes and magnetic recording devices store data in our computers. This chapter provides an introduction to basic magnetism. Starting from the simple attractive (or repelling) force between magnets, we define magnetic field, dipole moment, torque, energy and its equivalence to current. Then we will state the Maxwell equations, which describe electromagnetism, or the relationship between electricity and magnetism.

A great tutorial is provided by Kittel, which may be used to support students studyingChapters 1–4.

Magnetic forces, poles and fields

In the early days, magnetic phenomena were described as analogous to electrical phenomena: like an electric charge, a magnetic pole was considered to be the source of magnetic field and force. The magnetic field was defined through the concept of force exerted on one pole by another.

Introduction

The best description of spin-torque transfer can be found in the patent issued to John Slonczewski of IBM:

The subject of spin-torque transfer was not widely known until 1996. Due to its enormous technology potential, both academic and industrial research activities had been very active, and very rapid progresses have been made in recent years: from the first experimental verification of spin-torque transfer in giant magnetoresistance (GMR) film, to the implementation of this mechanism to magnetic tunneling junction devices. A large portion of this effort was directed towards the development of practical magnetic RAM chips based on the spin-torque-transfer mechanism.

Nanoscale magnets have at least one dimension in the nanometre range. They exhibit size-specific magnetic properties such as superparamagnetism, remanence enhancement, exchange averaging of anisotropy and giant magnetoresistance when the small dimensions become comparable to a characteristic magnetic or electrical length scale. Thin films are the most versatile magnetic nanostructures, and interface effects such as spin-dependent scattering and exchange bias influence their magnetic properties. Thin-film stacks form the basis of modern magnetic sensors and memory elements.

Matter behaves differently down in the nanoworld, where the length scales of interest range from about 1 nm up to about 100 nm. The atomic-scale structure of matter can usually be ignored, but the mesoscopic dimensions of the magnetic nano-objects are comparable to some characteristic length scale, below which the physical properties change. We have already encountered one important nanoscale object in bulk magnetic material – the domain wall. It is extended in two directions, but not in the third; the domain wall width δw is one of the characteristic lengths that concern us here.

The number of small dimensions in a nanoscale magnet may be one, two or three. Some examples of each are illustrated in Fig. 8.1. The one-small-dimension class includes magnetic thin films, which are at the heart of many modern magnetic devices. Magnetic and nonmagnetic layers can be stacked to make thin-film heterostructures, such as spin valves and tunnel junctions. The films are usually grown on a macroscopic substrate.

Resonance arises when the energy levels of a quantized system of electronic or nuclear moments are Zeeman split by a uniform magnetic field and the system absorbs energy from an oscillating magnetic field at sharply defined frequencies, which correspond to transitions between the levels. Classically, resonance occurs when a transverse AC field is applied at the Larmor frequency. Resonance methods are valuable for investigating the structure and magnetic properties of solids, and they are used for imaging and other applications. The resonant moment may be an isolated ionic spin or free radical, as in electron paramagnetic resonance (EPR), or a nuclear spin as in nuclear magnetic resonance (NMR). Otherwise it can be the ordered magnetization as in ferromagnetic resonance (FMR). Resonant effects are also associated with spin waves, and domain walls. The related techniques of Mössbauer spectroscopy and muon spin resonance provide further information on hyperfine interactions in solids.

A magnetic system placed in a uniform magnetic field B0 may absorb electromagnetic radiation at a precisely defined frequency ν0 = ω0/2π which falls in the radio-frequency or microwave range. The phenomenon is related to the Larmor precession of the magnetic moment, introduced in §3.2.2. In order to observe the resonance, an experimental geometry with crossed magnetic fields is needed. The steady uniform field defines the z-direction, while a high-frequency AC field bx = 2b1 cos ωt is applied in the perpendicular plane.