A new sensing device is emerging that converts changes in magnetic field to voltage, and vice versa. Potential applications for what are termed multiferroic or magnetoelectric (ME) materials range from very low cost magnetic sensors for the automatic sector, to better read heads for hard disk drives. Since the devices are built up from two or more different materials, they can either be made very large, or as in the case of a hard disk read head, just 40nm thick. The base material is magnetostrictive. This means it changes shape in a magnetic field, and conversely changes its magnetic property when deformed. Piezoelectric materials, however, change shape in an electric field and produce a voltage when deformed. The clever bit is forcing the two to interact across a rigid interface. The basic effects have been known for many years, and the idea of combining them dates back to 1884. But there is now a plethora of research and development, patents, and papers being published in academic journals on the subject. Efforts by industry have stepped up strongly in the last six months and it was one of the subjects of discussion during the UK Magnetics Society's, Advanced Functional Materials, meeting at the National Physical Laboratories (NPL) in Teddington. NPL materials scientist Dr Markys Cain is leading a project that aims to measure the magneto electric coupling coefficients produced from these new materials in order to generate reliable design data. The team has already developed an instrument usually used to measure the piezoelectric coefficient of ferroelectrics. They have added the capability so it can now apply a combination of steady and alternating magnetic fields. But other measurement technologies are being looked at. The purpose of the project is to generate design data to accelerate the development process for the devices, whether they are made up of layers bonded together, or composite materials where one phase is dispersed within another. Since coupling forces are transmitted through stress at the interfaces, it can be made much larger in composites than in devices made from layers bonded together. A simple structure can be made up of from an 8mm diameter disk 320µm thick lead piezoelectric zirconate titanate sandwiched between two 160µm thick disks of cobalt iron vanadium. This is both ferromagnetic and strongly magnetostrictive. The whole arrangement is bonded together with silver loaded epoxy, which is simple enough to be mathematically analysed in order to validate modelling. Devices being proposed and developed for real world, commercial applications tend to be more complicated. A sensor stack proposed for a read write head by the NPL teams, and probably being developed somewhere, consists of a magnetic shield, a gap, a seed layer, an anti ferromagnetic layer, a ferromagnetic layer, a ferroelectric layer, another ferromagnetic later, another anti ferromagnetic layer, a capping layer, another gap, and a second shield. The head then has seven layers, which is simpler than typical magnetoresistive (MR) read heads currently in service that have 15 layers and have to be biased in order to measure the changing resistance as opposed to producing a voltage. Furthermore, in an ME device, there is no need for permanent magnets to produce a bias field as there is an MR device. However, it is not necessary to have anything so sophisticated or complicated to produce useful results. Dr Neil Mathur, from the Department of Materials Science at the University of Cambridge, described a one cent room-temperature magnetoelectric magnetic field sensor. The device uses a multi layer ceramic chip capacitor made by AVX, which in quantity sell for about $0.01 each. This is rated at 0.6 micro Farads and 16V. Inside are 81 layers, made of alternating 1.5µm thick nickel based electrodes and 9.8µm thick layers of dielectric barium titanate. The nickel alloy is ferromagnetic and magnetostrictive and the barium titanate is piezoelectric. If about 300V is applied to the electrodes, and the capacitor is then discharged, the high voltage is sufficient to pole the barium titanate without blowing the device. The result then is that along the "Easy" axis - the energetically favourable direction of spontaneous magnetisation in a ferromagnetic material - a magnetic field of 100mT develops 7mV at the electrodes, rising to 10mV at 200mT. Output falls as the temperature rises, until reaching the Curie temperature, which in this case is about 400K. The devices cost much less than Hall sensors in volume, which cost about $0.03 each, or reed switches, that cost around $0.10. These are very small and rugged and can work at automotive engine bay temperatures. The down side is that the maximum 10mV output is rather small, and show a strong temperature dependence. Nonetheless, they show just how cheap and rugged commercial multiferroic-magnetoelectric devices are going to be. Dr Cain says: "The number of potential applications are multiple and numerous. But, so far these are just ideas. However, piezoelectric materials have led to so many applications, so who knows what these new devices will be used for?" In his list, there are: ultra high sensitivity magnetic field detectors that can accurately measure down to 1nT, new memory elements, new types of spin valves for spintronic applications such as quantum computing, novel actuators and transducers, and electrically tuneable microwave filters, oscillators and phase shifters. And with the high rate of development in the technology and the intense competition between the participating companies, we do not imagine users will have too long to wait.
Pointers * Devices combine magnetostrictive and piezoelectric materials * They can be made as separate layers, bonded together or as composite materials * The main potential application is to turn magnetic field variations into voltages * Research and development is intense. It is possible that some are already in use, embedded inside products such as hard disk drives