Nicolas Aimon develops pulsed laser deposition techniques for mixed multiferroic oxide films.
MIT Materials Science and Engineering graduate student Nicolas M. Aimon, above, inserts a holder into a pulsed laser deposition chamber. The system has an array of mirrors that focus the laser light on a chamber in which a target metal is vaporized and lands on a substrate forming a crystal layer. The ablated, or vaporized, material, crystallizes with the same crystal orientation as the substrate at high temperature, which is called epitaxial growth. Aimon uses the pulsed laser deposition system to create thin films of complex metal oxides, such as cobalt ferrite pillars in a bismuth ferrite matrix. His work has demonstrated CFO's potential as a magnetic material for data storage and computation.
Photos: Denis Paiste, Materials Processing Center
MIT materials science and engineering graduate student Nicolas M. Aimon has developed new methods of making mutiferroic complex metal oxide thin films by pulsed laser deposition and controlling their magnetic properties. The work could lead to a new generation of smaller, more energy efficient devices for computing and data storage.
In a 2012 paper in Applied Physics Letters, Aimon reported how he grew nanopillars of cobalt iron oxide (CFO) in a matrix of bismuth iron oxide (BFO) on a strontium titanate (SrTiO3) substrate. Of key interest is the coupling between magnetic and electrical properties in these complex metal oxides, with the CFO/BFO nanocomposite film showing promise as part of an electrically-switchable magnetic data storage device. Co-authors were MIT Professor Caroline Ross, and post-doctoral associates Dong Hun Kim and Hong Kyoon Choi. Because the pillars are magnetic, they could be magnetized up or down, representing zero or one in a storage device. "It's not enough to be able to store information though, you want to also be able to write and read this information," Aimon said. "What we're trying to do is define processes and fabrication techniques that allow us to grow these materials so that the pillars grow where we want them to grow," Aimon said.
Aimon, Ross and other collaborators analyzed the structure of the complex metal oxide films, studied their magnetic characteristics and modeled magnetic switching through application of an electric field, which they have reported in multiple papers. Aimon works in Ross’s Magnetic Materials and Devices Group lab. (See related article.)
Switchable magnetic pillars
Aimon explained that the magnetization in pillars of the ferrimagnetic cobalt iron oxide material, which were grown in a matrix of the ferroelectric material bismuth iron oxide, could be controlled through application of an electric field. An electric field applied to the BFO produces a strain, or mechanical compression of its atomic crystalline structure, which is transferred to the CFO. In the CFO, that strain causes a realignment, or change in directionality, of the CFO's magnetization. Unlike transistor-based computers in which an electric current drives information flow, "In this case, it's magnetostatic interactions that drive the propagation of information in the material," Aimon said. Ma2gnetostatic interactions are the attraction or repulsion between the north and south poles of magnets. "For all these applications, both the storage and the computation, you need to be able to control the position of the pillars. If they are randomly placed, there is no way you can say, 'read the information, write the information, on this specific location.' The location, that's the basic thing you need to control so you can read and write and define the shapes to be able to make the building blocks, logical blocks."
"But then you want to choose the material right so that it has the right magnetic properties, for example, if pillars are too strongly set in a particular magnetic orientation, which we call if they have too strong magnetic anisotropy, this information propagation – the fact that the neighbors interact – is not going to be happening because they just are too strongly established in their magnetic state. We want to tune the material of the pillars so the magnetic properties are right so this happens," Aimon said. The CFO/BFO thin film appears to have those properties. "What makes the strength of this material, is that the writing process is done thanks to the matrix that is in between the pillars," Aimon said. The bismuth ferrite matrix is piezo-electric or ferro-electric, which means that it is strained, that is it changes shape, when a voltage or electric field is applied to it. The strain, or change of shape, in the BFO is transferred to the cobalt pillars – which change their magnetic properties allowing them to be reoriented."
Like oil and water
The cobalt ferrite and the bismuth ferrite oxides, have different crystal structures, a spinel configuration for the cobalt and a perovskite configuration for the bismuth. These two materials naturally separate – like oil and water – which gives them the desirable property of self-assembling. But left to their own devices the materials will form separate clusters randomly.
Articles of Interest:
Aimon holds a ceramic pellet, above, that will be ablated, or vaporized, by a laser in a pulsed laser deposition chamber to form single crystal complex metal oxides, such as cobalt iron oxide or bismuth iron oxide. Below, Aimon peers into the chamber of a pulsed laser deposition chamber, which is used to create thin films of complex metal oxide.
Controlling growth of the pillars through the topography of the substrate is key to making complex metal oxide thin films with magnetic properties suitable for data storage and computation. That's where Aimon's work has made progress, in controlling the position of the cobalt pillars. Aimon has templated materials by making pits in the single crystal strontium titanate substrate that attract the CFO. One technique is to use an ion beam to pit the substrate, another is to use block co-polymer templating, which is a specialty of the Ross lab. Without intervention, "they just grow randomly, in random locations and sizes on the substrate," Aimon said. "Since we control the position of the pits, we can control the position of these small islands." Aimon will report more detail on the templating in a forthcoming paper, which is under review. Aimon also will argue that the techniques he has developed will have wide applicability to thin film materials not only for computing but also for electrocatalysis, sensors and other uses.
The group is making significant progress in understanding the growth of complex metal oxides and their magnetic properties, Ross said. "We are making a lot of progress in terms of characterizing the materials. I think you can say that templating the locations of these self-assembled structures is a significant step in the sense that if you are going to make any kind of device out of them, you really need to control exactly how it grows and where it grows, so that's been the problem that we have focused on for the last few years, and Nicolas now has a very nice solution to that particular problem."
"Self-assembly reduces production times and then costs and also potentially gives you access to structures that you would not be able to do otherwise," Aimon said. The self-assembly also allows tiny length scales that cannot be done without very lengthy and costly processes.
In Applied Physics Letters in February 2012, Aimon, Ross, Kim and Hong Kyoon Choi reported three different deposition processes for making the nanocomposites. In the process, the vaporized CFO or BFO takes on a single crystal orientation matching the crystal underneath, which is known as epitaxial growth, and the fractions of each phase could be varied. In a Journal of Applied Physics paper published in March 2013, Aimon, Ross and Dong Hum Kim described the fabrication by pulsed laser deposition of magnesium ferrite (MgFe2O4, or MFO), a spinel, grown in a bismuth ferrite matrix (BFO). Analysis of the magnetic properties of MFO found it has a much weaker response to strain than cobalt iron oxide (CFO), which was shown to have promise in the 2012 Applied Physics Letters paper. "Because the Co2+ ion in CFO is highly magnetoelastic, BFO-CFO nanocomposites show a strong out-of-plane easy axis due to the out-of-plane compressive strain that the BFO matrix imposes on the CFO, and the shape anisotropy of the pillars," the report said. The easy axis is the direction that the magnetization lies when there is no applied field, or its preferred direction of magnetization. For the CFO pillars, their preferred direction of magnetization is out of plane, which is along the long axis of the pillars. If two pillars are close together, each one acts like a little magnet with North and South poles at its ends, and they interact with each other. Since they are embedded in the solid matrix, they cannot pull towards each other, but the mutual interaction affects the way they switch and allows them to "communicate" or pass information from one to the next. "It's a bit like static electricity allowing charged objects to attract or repel each other, which is called electrostatic interaction. The analogous magnetostatic interaction between magnetic objects allows fridge magnets to stick to fridges, for example," Ross said.
Seeding material growth
Building on the idea of forming small pits in the substrate to attract the CFO during growth, Aimon said, those positions become a seed layer for later deposits in composite growth. "The pillars that grow off of these seeds are going to be in the positions of the initial pits. In between the pillars you still are going to have the matrix. Instead of having to pattern the pillars individually, and then grow the matrix in between, the fact that you just have to pattern the substrate itself, we argue, is a simpler way and more general way to control the position of the pillars and to control the structure of these composites than if you had to pattern each material individually.
Close up shows triangular shape of a complex metal oxide crystal in a spinel structure. Summer Scholar Joanna Denton worked under Aimon in the Ross lab to characterize the magnetic properties of triangular nanopillars of metal ferrites (cobalt, nickel, magnesium or iron). Photo: Denis Paiste, Materials Processing Center
This idea of seeding the growth of the material can be applied to other systems, other composites, that have a similar growth process," Aimon said. "There are other composites that also self-assemble in a relatively similar fashion which have properties that make them interesting for other applications. But these applications are also probably going to need some control over the structure – over the location of the pillars – and that's where we say we think it's a generally applicable method, this controlling the topography of the substrate. It's future work to show that this can be used to pattern other composite systems, other thin films that self-assemble."
Aimon, 27, is a native of Niort in France and received his undergraduate degree in Paris at the Ecole Polytechnique. He anticipates completing his Ph.D. at MIT in spring 2014. His next work will focus on demonstrating that the pulsed laser deposition methods he developed for complex metal oxides are generally applicable to other thin film systems. There also is much work to be done on the oxides' theoretical physics, he said. Being able to control these composites and their assembly also opens a path to create novel structures that researchers want to study at a fundamental level. Promising areas for study include phenomena that happen at the interfaces between two different complex metal oxides such as high conduction of current or magnetic properties. "Even if you have a very boring material on the left and a very boring material on the right, there might be something interesting happening at the interface itself. We also argue that these composites are a really nice playground for studying these interfaces at a fundamental level," Aimon said.
|Left, a nanocomposite consisting of pillars of CoFe2O4 (CFO) spinel in a matrix of BiFeO3 (BFO) perovskite. Center, a similar sample in which the positions of the CFO were templated by focused ion beam patterning. Right: Schematic cross-section of the nanocomposite.
Image courtesy of the researchers, Copyright American Institute of Physics
Summer Scholar contribution
In summer 2013, Aimon supervised the work of Summer Scholar Joanna Denton on characterizing the magnetic properties of equilateral triangular, magnetically anisotropic nanopillars of metal ferrites (cobalt, nickel, magnesium or iron)
produced in the Ross lab. Cobalt ferrite (CFO) has much more magnetocrystalline anisotropy and much higher coercivity than do magnetite (FFO), magnesium ferrite (MFO) or nickel ferrite (NFO), Denton's studies found. "Because their anisotropies are so weak, FFO, MFO, and NFO have little hysteresis. Thus, independent of
magnetoelastic anisotropy (anisotropy produced by strain), in terms
of coercivity, CFO would be the best material for memory devices," Denton reported in a poster presentation. Denton was one of 18 Summer Scholars hosted at MIT by the Materials Processing Center and the Center for Materials Science and Engineering from late June through mid August 2013.
"If you grow under different conditions, instead of getting pillars of one material in the matrix of the other, you get these little triangular facets, triangular particles, and we've noticed this by accident," Ross said. "We've actually had a couple of summer students work on topics related to that over the last couple of years and Joanna did some nice work on analysis of these triangular particles." Ross hopes to publish a paper on the results of their work.