Final Report Summary - REFREEPERMAG (RARE EARTH FREE PERMANENT MAGNETS)
Executive Summary:
4.1 Final publishable summary report
There exist a number of families of PMs, each one possessing specific characteristics. The most powerful PMs are the ones based on rare-earths and cannot be superseded by any transition- metal alloys. With respect to the energy product the NdFeB magnets are by far the best, however with a relative high cost and limited maximum operational temperature of 200 0C. For higher operational temperatures and at a much higher cost the Sm-Co 1:5 and 2:17 families are available. Ferrites are the most widely used by a very wide margin due to their very low price. Fe based alloys containing (Al,Ni,Co) that is AlNiCo magnets find their uses in electronic instruments, sensors, loudspeakers, TWT amplifiers, etc.
Export quotas posed recently by the rare-earth producers, mainly China which is currently the source of more than 95 % of the world’s supply on rare-earths, has resulted in shortages and a dramatic prize increase of these strategic materials. The rare- earth shortage has stimulated intense effort towards the search for new non-rare-earth containing magnets either by modifying the Alnico and Ferrite type magnets or discovering new materials containing no rare-earths. The search is driven not only by scientific interest but mainly from the applications sector’s needs, especially for the electric car and the wind energy industries.
This project aimed and succeeded at developing a new generation of high-performance permanent magnets (PM) without rare-earths. Our approach was based on:
a) novel production of high-aspect-ratio (>5) nanostructures (nanowires, nanoparticles, nanorods, nanoflakes) by exploiting the magnetic shape anisotropy of the constituents that can be produced via chemical nanosynthesis -polyol process or electrodeposition- in Co-based novel magnetic phases, which were consolidated with novel compaction process for a new generation of rare-earth free permanent magnets with energy product in the range of (BH)max > 160 KJ/m3 at room temperature.
b) using a high-throughput thin film synthesis and high-through-put characterization approach we have identified novel magnetic phases of the type Fe-Co-X (X= C, B) with very high anisotropy in excess of 0.5 MJ/m3 and coercive field of ~ 1 KOe stabilized in a tetragonal or hexagonal structure by epitaxial growth on Au1-xCux combinatorial searched suitable substrate, under various conditions of pressure, stoichiometry and temperature, and
c) Mn-X based alloys (X=Ga, Bi) as bulk processed permanent magnets
To obtain these remarkable results we were based our approach on combinatorial modelling using state of the art approaches and software such as WIEN2k, Elk, muffin-tin orbitals (EMTO) pseudopotential approaches with plane waves (VASP) or numerical local orbitals (SIESTA, OpenMX) for DFT calculations.Most of these codes provide MPI-parallel versions suitable for large-scale computations. MAGPAR, FEMME will be used for micromagnetics together with finite element discretisation codes. We have combined modeling with unique facilities for preparation and characterization that existed in the partner’s labs, such chemical or electrochemical synthesis of nanoparticles and nanowires, combinatorial sputtering for thin films and multilayers, ultrasensitive FMR for anisotropy evaluation and consolidation techniques for bulk PMs. We have developed new tools for high-throughput characterization such as M-O Kerr effect, combi- XRD , combi 3D-scanner , polarization set-up for thermal neutrons and a procedure to correlate texture maps and microstructure.
In addition, due to the largely unknown environmental and toxicological impact of nanomaterials, a comprehensive Life Cycle Assessment was done for chemically synthesized nanorods and sputtered thin films and the magnets produced. The potential environmental impacts of these modules over their entire life were quantified.
This activity has resulted in 43 journal publication, 61 conference publications , 2 book chapters and 3-PhD and more than fifty invited-oral presentation in conferences, summer schools or special workshops. The developed concepts were well accepted by the scientific community and frequently cited.
In addition, we are among the pioneers in developing permanent magnets based on nanorods with very high energy product and among the first groups that tested Mn-Bi based magnets in a motor with very promising performance. A techno economical study revealed that the nanorods-based magnets are for niche markets due to the high cost although the Mn-Bi magnets with some improvement can compete with ferrites. No patents have been filled until the end of project
Project Context and Objectives:
A summary description of project context and objectives (not exceeding 4 pages).
Main Ideas: The scientific base or our proposal
Magnetic anisotropy is the directional dependence of a material's magnetic properties. In the absence of an applied magnetic field, a magnetically isotropic material has no preferential direction for its magnetic moment while a magnetically anisotropic material will align its moment along one of the easy axes. Magnetic anisotropy is a prerequisite for hysteresis in ferromagnets: without it a ferromagnet is magnetically soft. Magnetocrystalline anisotropy arises from quantum-mechanical interactions between electrons and the lattice. The spin-orbit interaction couples the spin degrees of freedom to the lattice and leads to the alignment of the magnetization along certain preferred crystallographic orientations. There are different sources of magnetic anisotropy:
• Magnetocrystalline anisotropy: the atomic structure of a crystal introduces preferential directions for the magnetization (due to spin-orbit interaction)
• Shape anisotropy: when a particle is not perfectly spherical, the demagnetizing field will not be equal for all directions, creating one or more easy axes (due to intrinsic dipolar interaction).
• Magnetoelastic anisotropy: stress may alter magnetic behavior, leading to magnetic anisotropy.
• Surface/interface anisotropy: the lower coordination at interfaces and its associated reduced symmetry lead to strong modification of the magnetocrystalline anisotropy
• Exchange anisotropy: appears at the interface of magnetic phases with different character (i.e. antiferromagnetic and ferromagnetic materials)
• Externally imposed (induced) anisotropy
The total magnetic anisotropy can be expressed as the sum of the various contributions
HTotal = Hcr + Hshape + H stress + Hexc +Hother (Eq. 1)
where, Hcr = h ( 2K1/ Ms), Hshape = h (Na -Nc) Ms , Hstress = h (3λsi σ/Ms), Hexc = h(-J12 s1s2)
and Na , Nc are the demagnetizing coefficients parallel to a and c axes of an “ellipsoid”, λsi the saturation magnetostriction, σ is the stress and K1 the crystal anisotropy constant. The Na and Nc are expressed as a function of the c/a ratio. Assuming a coherent rotation magnetization reversal mechanism, the coercivity for an assembly of elongated particles can be found from the relation
Ηc = (1-p) (Νb – Νa) Μs (Eq.2)
where p is the packing fraction and Na and Nb the demagnetization factors along and perpendicular to the length of the particles. Assuming an Ms somewhere in the range of 1700 - 1900 G (for Fe or Fe-Co-X alloys), Nb - Na = 2 and p ~ 0.4-0.6 an Hc of 4-6 kOe is expected with (BH)max ~ 240 KJ/m3. Having this in mind we propose to exploit this possibility to prepare a novel series of magnets covering the range 60
KJ/m3 < (BH)max < 160 KJ/m3, depending on optimization of packing, magnetization and intrinsic properties of the high-aspect-ratio nanostructures.
The best example of exploitation of the shape anisotropy is the AlNico series. AlNico magnets are alloys of Fe, Co, Ni, and Al with small amounts of other elements (Cu, Nb, Ta, and Ti) which are used to optimize coercivity at the expense of magnetization. These magnets develop their coercivity after a carefully controlled heat treatment which leads to precipitation hardening. However, in this case the coercivity is due to the shape anisotropy of {Fe-Co}-rich precipitates ( ~30 nm lateral dimension) instead of magnetocrystalline anisotropy as in the case of 2:17 magnets. We propose to develop a new type of nano-composite material for the fabrication of permanent magnets by a bottom-up approach. The new material will be based on assembled metallic nanostructures with high aspect ratio (nanowires, nanoparticles, nanorods, nanoflakes). We will use state-of-the-art techniques to fabricate high-aspect-ratio nanostructures and consolidate them after alignment into bulk permanent magnets. This is a disruptive approach for new permanent magnets relying on shape anisotropy and not an incremental improvement of AlNiCo magnets. We propose to develop the synthesis of metallic Fe-Co based nanoparticles with high aspect ratio by a chemical organometallic synthesis, gas phase methods and ball milling. The fine control of the chemical processes will allow to control the morphology of the nano-objects (aspect ratio – diameter – composition) with high precision and reproducibility. These approaches can potentially be scaled up to large volume production. Nanostructures combine several advantages compared to existing permanent magnet materials. The anisotropy will be an extrinsic property based on the shape anisotropy. This property is rather insensitive to temperature effects contrary to magneto-crystalline anisotropy. We think that it will be possible to reach high energy products. Our new composite material could be used as a high performance replacement for permanent magnets, which are operating at high temperature (T>200°C).
The scientific base- the Bain Path- for the search of new non rare-earth containing phases with large magnetocrystalline anisotropy is based on the previous success of growing phases –with phase instability- by epitaxy on selected substrates. Strain effects in functional materials are of great interest currently for modifying material properties. By strained epitaxial film growth, physical properties can be controlled and improved, e.g. in semiconductors, multiferroic materials, and ferromagnets . Moreover, there are even materials which exhibit functional properties such as ferroelectricity only in strained films. During epitaxial growth, the film orientation is determined by the substrate onto which the material is deposited and the temperature used. The thin film adapts its in-plane lattice parameters to the substrate, even if its equilibrium lattice parameters differ considerably. The particular case in which the lattice parameters of the film material are strained such that they are identical to those of the substrate is called strained coherent film growth. In common rigid metals, straining of the crystal lattice by coherent film growth requires substantial elastic energy. As a consequence, already at strains of a few percent, coherent growth is limited to ultrathin films with thicknesses of up to several atomic layers. For the growth of coherent epitaxial films with high strains and large thicknesses, soft materials should be used, as suggested by van der Merwe. Such softening is observed in crystals with lattice instabilities, e.g. in materials with a martensitic transformation. As an example of how to exploit such a martensitic instability, Godlevsky and Rabe predicted the possibility of inducing a cubic to tetragonal distortion with c/a ratios from 0.95 to 1.25 in the magnetic shape memory material Ni2MnGa. In fact, in experiments Dong et al. demonstrated a considerable epitaxial strain of 3% in a Ni-Mn-Ga film. This shows that for improved tunability of crystal lattice and functional properties in thicker films, it is advantageous to exploit alloy lattice instabilities
In particular, the possibility of stabilizing intermediate lattice geometries-fct- and varying smoothly between stable phases along Bain transformation paths offers new routes to modify and investigate structure-property relations in functional materials. To achieve this cumbersome and difficult task we will employ a novel approach, which offers flexibility, multiplicity and high speed towards a first identification of materials with the desired properties.
While discoveries in materials science have made a major impact they are mostly based on simple materials - either elemental as Si or binaries, which are synthesized in a traditional way, where experience and intuition are the drivers. Serial processing is the norm due to time/manpower constraints. However, the high-tech industries demand more and more tailor-made materials with a complex list of requirements. Most often only optimized materials having four to five components are able to fulfil the requirements, which increases the number of required experiments by orders of magnitude. The solution to these problems and limitations are the corner stones of this proposal.
The Combinatorial Technique (CT) or High-Throughput (HT) approach is a revolutionary step forward in the development of new materials. It involves the development and
application of new tools for systematic and parallel synthesis focussed on the characterization to the most industry relevant and innovative of multi-functional classes of magnetic materials. Automatic well- controlled production and characterization of binary, ternary and quaternary systems yielding thousands of new materials becomes possible and it is the most effective way to search for new phases. Further development in bulk form is beyond this proposal. The high-throughput synthesis (HTS) approach involves the parallel production of huge number of samples in form of (arrays) of samples, also
called material libraries. For example in an array of 10x10 samples it is possible to vary composition parameter A in the X direction in 10 steps, and composition or fabrication parameter B (e.g. composition, substrate temperature) by 10 steps in Y direction. This produces a ‘library’ of 100 different but very well- comparable samples with two independently varied production parameters within one single experiment. The next step is to have rapid high-quality characterization techniques, which can, in one operation, measure the whole material library for the specific properties of interest. These are termed ‘High Throughput Characterization’ (HTC) measurements. The final stage is the appropriate informatics systems - data handling (visualization) and mining to aid in the analysis of the large quantities of data produced.
Coupling this (HT)approach with the possibility of epitaxial multicomponent thin film growth at various temperatures and on various substrates, to stabilize new types of structures in an analogous way for bulk materials- e.g. spinodal or similar type- using (HTS) techniques will permit us to explore a wide range of promising high-magnetization/ high anisotropy/Curie temperature alloys a)of the type {Fe-Co}- X-Y (X=other 3d or 4d metals and Y=B,C,P or N) and b) of Heusler alloys with the general formula X2YZ (where X is usually Fe, Co, Ni, Cu; Y other transition metals, most often Mn; and Z a group-B element (Al, Ga, Ge, Sn...).
1.1.3 Scientific and technical objectives
Through an integration of professional skills (chemists, physicists, engineers, materials scientists) from academic and industrial groups, the activity will approach different aspects of modelling of permanent magnet materials and devices, materials preparation and characterization of their electric, magnetic and other relevant properties e.g. porosity, strength, thermal conductivity, core fabrication and device assembly, performance evaluation and optimization of materials and processing steps to exploitation plans. The know-how on the different aspects will be transferred from the lab size to commercial size products aiming at filling the gap between research laboratories and industrial production. A prerequisite for achieving the above technological goal is an improved understanding of magnetic phenomena at the nanoscale (in thin films, nanoparticles, nanowires, nanoflakes), which thus represents the intrinsic strategic baseline of the project.
The strategic objectives of the project are:
a) use of the shape anisotropy as the dominant mechanism to develop novel high-aspect-ratio nanostructures as building blocks to fabricate novel non-rare-earth containing permanent magnets using a bottom–up approach, and b) use HT techniques to exploit strain-induced epitaxial growth of distorted cubic to tetragonal high magnetization alloys that will result in the high anisotropy that is required for the production of novel permanent magnets e.g. Fe-Co based alloys, leading to the identification and characterization of new magnetic multifunctional materials of scientific and technological interest, c) to use nanoscale building blocks to exploit their large surface/interface anisotropies at artificial grain boundaries, and d) to mix particles with different high-aspect shapes to create well defined pinning centers for domain wall motion and thereby enhance the coercive field.
Thus the project will focus on the following specific scientific objectives:
SO1 To develop novel modelling techniques for optimization of hard magnetic properties of high- aspect-ratio nanostructures (inducing strong non-linear edge effects)
SO2 To develop cutting-edge basic research into the intrinsic -atomically resolved- properties of hard magnetic high-aspect-ratio nanostructures.
SO3 To develop online monitoring techniques during and after synthesis of the starting materials before consolidation
SO4 To understand the correlation of shape, composition, structure and morphology with magnetic anisotropy, magnetic moment
SO5 To provide from atomistic approach calculation the magnetocrystalline anisotropy of novel distorted cubic materials of the type Fe-Co and Heusler alloys in a single and a high-
throughput way.
SO6 To understand the conditions of stabilization of the tetragonal –phase along the “Bain-path” in
the Fe-Co and in Heusler alloys
SO7 To identify best practice examples for the full life cycle of the materials from synthesis to re-
cycling taking into account sustainability and cost issues
Technical objectives
The technical objectives of REFREEPERMAG are derived from the scientific objectives and are expected to lead to feasible materials for the fabrication of novel permanent magnets suitable for replacement of some applications of -earth based products or the development of a novel class of permanent magnets, first through thin film combinatorial approach with a possible extension or suggestion of bulk tetragonally-distorted Fe-Co or Heusler Alloys. More specifically:
TO1 To develop new technologies for high-aspect-ratio nanostructures production of new doped
{Fe-Co}-X-Y intermetallics with high magnetization (Br >1 T) and anisotropy (Hc>1T) using novel environmentally friendly fabrication techniques.
TO2 Use novel processing techniques for the consolidation of nanowires into Permanent Magnet compacts, with a view to realising energy products 60 KJ/m3 <(BH)max < 160 KJ/m3 at room temperature , better than Alnico and Ferrites and approaching those of rare-earth bonded permanent magnets.
T03 To discover new hard magnetic phases using a HT approach of thin films of {Fe-Co}-X-Y or
Heusler alloys, deposited epitaxially on various substrates with K1 > 107 ergs/cm3
TO4 Develop and establish advanced high-throughput-characterization techniques for structural
and magnetic properties resulting in a database of complex alloys and composites
TO5 Produce thick films/foils of {Fe-Co}-X-Y or Heusler alloys using the interlayer libraries approach
TO6 Demonstrate as a proof-of-principle one demonstration of the novel PMs into a micromotor
TO7 Suggest DIN standards and quality as well as sustainability control procedures for these novel
PMs
Project Results:
A description of the main S&T results/foregrounds (not exceeding 25 pages)
The major S&T task of the REFREEPERMAG project were divided into seven workpackages that are listed below with extra three workpackages for management, scientific coordination and exploitation -dissemination:
WP
Number WP Title Type of Activity
WP1 Ab-initio and Micromagnetic Modelling RTD
WP2 Ab-initio and Micromagnetic Modelling RTD
WP3 High-Throughput synthesis of Rare-Earth Free permanent magnet materials libraries RTD
WP4 Characterization –structural and magnetic of high-aspect-ratio nanostructures and Rare-Earth-Free thin films/multilayers and bulk PMs RTD
WP5 Up-scaling and Consolidation Techniques of nanocomposites-sintered magnets RTD
WP6 Towards Applications RTD
WP7 Life Cycle Impact Assessment (LCIA) RTD
WP8 Management MGT
WP9 Scientific Coordination RTD
WP10 Exploitation and dissemination OTHER
Advances that this project brought about beyond the state-of-the-art are:
1. With respect to Modelling
Employed for the first time combinatorial atomistic modelling – an extension of the work of Partner # 3- to search for unexplored compositional spreads of candidate magnetic phases of the type {Fe- Co}-X-Y and Heusler alloys of the type X2YZ that were stabilized in lower than cubic symmetry in conjunction with our combinatorial and high-throughput experimental deposition and characterization techniques, that were employed for the first time into the search of novel candidate materials for permanent magnets. Finite element micromagnetic modelling and micromagnetic- Monte Carlo simulations were used to simulate the best geometrical configurations for high coercivity and thermal stability of the switching processes.
2. With respect to fabrication approaches
2.1 The polyol synthesis, an environmentally friendly process, was extended to systems beyond the Co and Co-Ni system to more complicated ones with larger magnetization such as Fe and Fe-Co. These systems are more interesting because they will have higher magnetization (fcc Fe has a 30% larger atomic magnetic moment than bcc Fe), and consequently higher anisotropy when employing the shape anisotropy. Partner # 1 has successfully fabricated air stable FeCo nanoparticles with high magnetization using a modified polyol process.
Organometallic chemistry was developed by partner # 4 for the synthesis of air-stable Fe and FeCo nanoparticles with spherical and cubic shapes. This method was extended to the synthesis of iron-based anisotropic particles. The growth of iron-based nanorods and nanowires by a chemical process was a real breakthrough
2.2 The fabrication of nanowires using electrodeposition (Partner #7) were optimized firstly with respect to the materials composition as above. Secondly, the different parameters of fabrication (e.g. pH of the bath, current density, or temperature) were systematically checked for a controlled tuning of crystalline phases of nanowires as well as the plating time that determine the length of nanowires. In addition, crystal size and new crystal phases were induced by appropriate annealing. Finally, the distance between self-assembled pores in the alumina template (thickness of walls between the holes) were tuned serving the purposes of achieving the maximum production of nanowires, and if the structure is right wereused as prepared (e.g. as a monolithic sensor) to minimize or control the effect of dipolar interactions.
2.3 For the first time we explored the possibilities of shape anisotropy of nanoflakes, were produced either by modifying the ratio and kind of solvents in the polyol process (Partner #4) or by intensive ball milling in the presence of surfactants as in the rare-earth containing magnets (Partner #13).
3. With respect to the non-equilibrium techniques for stabilizing the fct-phase along the Bain path of high-Ms materials
3.1 We advanced the search of novel magnetic materials to a state beyond the state-of-the-art by using a high-throughput synthesis (HTS) technique in a very systematic way never used before as far we know. (HTS) is required due to the large number of parameters which need to be varied in order to achieve the conditions for stabilization of the tetragonal –fct- phase along the Bain path. These parameters were the substrate lattice constant, temperature, third or fourth element addition to the {Fe-Co}–base alloy. The (HTS) approach was the best approach to achieve the goal to identify efficiently the range of parameters for further detailed studies. Combined with the (HT) computational modelling, we introduced the search for novel permanent magnets a fast and reliable process never tried before.
3.2 The high-throughput approach was ideal as a fast and reliable method to stabilize novel Heusler alloys, which in the bulk possess already high moment and high Curie temperature, but yet lack large anisotropy. We carried out an intensive composition spread study of selected phases extending our studies from thin to thick films as well and succeeded to stabilize a tetragonal-like Fe2GuGa –phase on Si(111) at 300 0C.
4. With respect to intrinsic and extrinsic properties characterization
4.1 The vast number of materials libraries produced with the high-throughput synthesis need adequate high-throughput characterization techniques. We used existing and develop new screening techniques as non-destructive techniques assessing chemical (automated EDX in SEM), structural and microstructural (automated XRD) and magnetization properties. We optimized especially the two techniques for magnetization and anisotropy measurements, using a high- throughput high-temperature SQUID magnetometer for magnetization measurements and high- throughput Kerr-effect set-up, one of them with up to 2T, for high-throughput anisotropy-coercivity measurements. A correlation with structural data w lead us to conventional techniques for structure, microstructure and magnetization measurements for selected samples with promising properties within the preselected ranges. It has to be noted that the combinatorial approach is also here beneficial as we have complete data sets of very comparable samples from the materials libraries. From this we can select efficiently the samples which were studied in more detail e.g. for investigating the nanostructure of a discovered interesting material we made targeted preparations of cross-sectional samples for TEM (transmission electron microscopy) from a materials library by using a FIB (focused ion beam).
4.2 Newly developed anisotropic nanomagnets with high aspect ratios need to be magnetically characterized on an individual basis to understand the correlation of surface chemistry [23] (ligands for example), shape, morphology crystalline structure and magnetic properties. Single particle magnetic analysis based on the bolometric detection [30a] of microwave resonant absorption and x-ray magnetic circular dichroism experiments [30b] was developed and employed. Intrinsic properties like orbital and spin magnetic moments were determined. Single particle magnetic anisotropy was measured and the contributions from shape and magnetocrystalline anisotropy disentangled.
5. With respect upscaling and consolidation
5.1 With respect to upscaling, the polyol process was scaled up to a few tens or hundreds of grams in order to be able to use the optimized materials for consolidation
5.2 It was he first time ever to use state-of-the-art consolidation techniques for nanoparticles to consolidate the high-aspect-ratio nanostructures for the fabrication of the novel class of rare-earth free permanent magnets.
6. Applications
We are among the first groups worldwide to prepare and consolidate nanowires-nanorods into macroscopic permanent magnets with energy so high energy products, indicating that our consortium had achieved one of its main objectives.
Worth to mention here, as far as we know, we are the first consortium in the world to fabricate and test in “real conditions” MnBi based PMs with very promising results.
We have developed novel high-throughput techniques to search for novel magnetic materials with the appropriate properties for PM, paving the way for future approaches of rapid-screening for other classes of materials.
7. Life cycle Impact analysis (LCIA) and sustainability
7.1 It is , as far as we know, the first time that an LCIA study, for the proposed novel class of nanostructured permanent magnets, will be undertaken. The state-of-the art protocol for nanoparticles will be used (ISO 14040:2006) with respect to the life cycle, durability, recyclability, impact on health and environmental impacts of the new magnets based on high-aspect-ratio nanostructures.
WP1: Ab-initio and Micromagnetic Modelling
Main Objective: The objective of this task was to establish computational design tools that will guide material development through the course of the program.
Ab-initio calculation of anisotropy and magnetization (UPPSALA)
Models used
Electronic model Density functional theory (DFT) for crystalline anisotropy
Mesocale model Micromagnetics dynamics for shape anisotropy
Simulation software and type of numerics used
WIEN2k, Elk, muffin-tin orbitals (EMTO) pseudopotential approaches with plane waves (VASP) or numerical local orbitals (SIESTA, OpenMX) for DFT calculations.Most of these codes provide MPI-parallel versions suitable for large-scale computations. MAGPAR, FEMME will be used for micromagnetics together with finite element discretisation codes (GID, Salome) (See Del 1.1 and 1.2)
Achievements of the model beyond experiments
The theory can pre-identify interesting materials prior to the actual synthesis and experiments and thus save time and resources.
Summary of the modelling
Two different kinds of magnetic anisotropies, namely the magneto-crystalline anisotropy and shape anisotropy are investigated. Due to their inherently different nature, they will also be studied with two different approaches. The magneto-crystalline anisotropy will be studied by first-principles electronic structure calculations with DFT using both full-potential and pseudopotentials methods. The project will use various basis sets, e.g. augmented plane waves with local orbitals. High-throughput algorithms will be employed to search the parameter-space of ternary compounds. An additional alloying element will be added to these selected systems to obtain additional degrees of freedom and the modeling will determine whether this improves the system. For alloys, some of the DFT codes implement the coherent potential approximation, which was shown to be the best single-site approximation for alloys.
Highlight clearly significant results;
Tetragonally distorted FexCo1-x alloys doped with B, C or N interstitial atoms
This work was motivated by a publication of Burkert et al, who has predicted that the MAE of tetragonally distorted Fe/Co alloys can reach large values at certain concentration of its constituents and at a particular tetragonal distortion. Since such structure is not stable, a tetragonally distorted Fe/Co alloy can only be realized as a thin layer that is epitaxially grown on a suitable substrate. We have explored an alternative way of stabilizing the tetragonal distortion, in particular, by adding another element, which is an atom with a sufficiently small atomic radius in order to occupy an interstitial site
RESULTS
(FexCo1-x)2B alloys
For (Fe/Co)2B alloys it is experimentally known that they have a natural stable tetragonal structure for the whole range of x=0 to 1. These compounds have been experimentally studied by Iga. Computational studies have never been performed. By means of coherent potential approximation (CPA) and magnetic torque method we have evaluated the MAE throughout the whole range of Fe and Co concentrations. Results are summarized in the figures below. The end-compounds consisting of boron and only iron or only cobalt have an easy magnetization plane. However for cobalt concentrations between ~10-65% there is an easy-axis magnetization with a maximum MAE of the order of 0.5MJ/m3. An optimum is reached at about 30% of cobalt content, at which the saturation magnetization reaches approximately 0.8T. Optimal cobalt concentration is at even lower value, compared to the carbon-doped Fe/Co alloys. Yet, the MAE obtained here are not very high.
Fe5SixP1-xB2 quaternary alloys
We have performed high-throughput calculations of MAE of the whole range of iron-based compounds Fe5SixP1-xB2, which have appeared as promising candidate materials due to their tetragonal layered crystal structure without need of any strain or interstitial elements. We have performed a full structural relaxation of these compounds as a function of silicon concentration. Free structural parameters have been also optimized. Since the Si and P are neighbors in the periodic table, the alloy was treated using the virtual crystal approximation. Resulting MAE is summarized in the figure on the right. As can be seen, the MAE is relatively low and it is positive only for low concentrations of silicon. At Si concentrations more than 20% the material has an easy plane of magnetization.
3d-based L10 structures
L10 structures span a wide range of layered binary compounds, which have a structure related to face-centered cubic. A tetragonal distortion, typically rather small, is achieved by alternating the elements in subsequent atomic layers. As a result, L10 structures are a limiting case of thin multilayers, where the layers are just one atom thick. Some of these structures, such as FePt or CoPt already found their way to applications, for example in perpendicular magnetic recording. Here, we focus on L10 compounds, which consist of cheap and abundant elements. The table summarizing results here is rather preliminary and we intend to study alloying these structures with third element and its influence on the MAE. Results so far obtained on binary compounds show a particularly large MAE for Mn-based L10 structures.
Tetragonal alloys (Fe/Co)2B for a certain range of compositions they have a rather large MAE and sufficiently high Curie temperature and saturation magnetization. In cooperation with Vienna and Darmstadt groups we have theoretically and experimentally explored these materials, including a possibility to enhance their anisotropy with small amounts of substitutional 5d elements, inspired by results of [6]. Our results are summarized in [P5]. We have found that a small amount of W or Re elements can efficiently double the MAE. The latter was also confirmed by experiment. The anisotropy reached by 5% doping with Re was 1.4 MJ/m3.
Finally, we have explored layered tetragonal materials based on Fe5SiB2 and their alloys with Co (in place of Fe) and P (in place of Si). Results of this work are subject of a manuscript in preparation. Pure Fe-based compounds showed only rather small value of MAE, despite their favorable highly anisotropic crystal structure. MAE of about 1.1 MJ/m3 can be reached by about 30% substitution of Fe with Co.
Micromagnetic modeling
Analytic and micromagnetic calculations on single cylindrical nanorods have been performed in order to understand and magnetocrystalline and shape anisotropy contributions to coercivity. Assuming coherent rotation, the coercive field can be calculated with
〖 H〗_(c,coh)= (2K_1)/〖μ_0 M〗_s +(N_⊥-N_∥ ) M_s
The formula contains the anisotropy field H_A=(2K_1)/〖μ_0 M〗_s and the shape anisotropy contribution 〖(N〗_⊥-N_∥)M_s with the demagnetizing factors N_⊥ and N_∥. This value acts as upper limit for the coercivity of the real nanorod. Although a higher saturation magnetization increases the shape anisotropy contribution to coercivity, it also lowers the anisotropy field. For that reason materials with a high uniaxial anisotropy (K1 > 400 kJ/m³) and moderate saturation magnetization such as Co should be preferred over materials with low magnetocrystalline anisotropy and high saturation such as Fe or CoFe (see fig. 4).
Fig. 4. Hysteresis of a single Co, Fe and CoFe
There are multiple analytic approximations for non-coherent magnetization reversal mechanisms such as curling or buckling, but in reality the magnetization reversal is a mixture of multiple mechanisms that can only be calculated numerically. The micromagnetic simulations have been performed with the FEMME package which solves the Landau-Lifshitz-Gilbert equation on a finite element discretization of the magnetic structure. The simulations make it possible to systematically examine the contributions of shape and magnetocrystalline anisotropy as well as the losses due to inherent magnetization reversal modes that even occur on high-aspect ratio, high-anisotropy structures (see fig. 5)
Fig 5.. Nucleation and incoherent magnetization reversal in a high-aspect-ratio Co nanorod. The demagnetizing field at the tips causes a curling-like vortex structure. The magnetization reversal is an incoherent process involving the movement of two domain walls from the tips to the center of the nanorod
The model of coherent rotation predicts a saturation of the shape anisotropy effect at aspect ratios n > 10 (fig. 6a).The micromagnetic simulations however, show different coercivity values for different cylinders with the same aspect ratio H/D but different height H and diameter D. In fact, the micromagnetic simulations show a strong relationship between coercivity and diameter (fig 6b) because higher diameters favor the formation of a vortex nucleation at the tips. At low diameters (D<20nm) a moderate aspect ratio of 1:5 is enough to maximize the shape anisotropy contribution.
Fig. 6. a) Theoretical upper limit of coercivity of Co nanorods as function of the aspect ratio H/D with simulation results. b) Simulated coercivity as function of aspect ratio with lines of constant height and diameter.
Besides the optimization of the aspect ratio (mainly through decreasing the nanorod diameter) it is also possible to optimize the shape of the nanorods. Micromagnetic simulations indicate that the coercivity of a nanorod can by increased by 10% by rounding the tips to a half-spherical shape. The rounded tips decrease the stray field and suppress a nucleation at the tips. Another approach would be the reduction of the saturation magnetization Ms to increase the anisotropy field but this would also decrease the potential energy density product.
2. Regular Arrangements of Nanorods
The packing density p of regular arrangements of cylindrical nanorods can be calculated by dividing the area of the nanorod's base circle by the area of the surrounding hexagonal or square cell. Fig 7 shows calculated packing densities of nanorods with D = 10 nm with varying distance d or Δ.
Fig. 7. Hexagonal and quadratic packing of nanorods with diameter D, center-center distance d and surface-surface-distance (wall thickness) Δ.
Micromagnetic Simulations on 10x10 grids with hexagonal and quadratic packing with varying distances and packing densities have been performed. The calculations showed that the coercivity of multiple nanorods is dominated by the surface-surface distance Δ and not by the overall packing density or packing lattice. Therefore, it is possible to project the results from two nanorods to a regular grid with the same distance in order to save computational time.
Fig. 6a shows the coercive field of two nanorods over the distance Δ. Due to numerical difficulties the value at Δ = 0 nm has to be extrapolated from the other results and is printed with a dashed line. In Fig. 6b the same coercive field data is transformed to a function of packing density using the distance-to-packing-density mapping in Fig 5. The energy density product (BH)max can approximately be calculated by
〖(BH)〗_max={■(〖B_r〗^2/4µ_0&if µ_0 H_c≥ B_r/2@(B_r-µ_0 H_c ) H_c&if µ_0 H_c≤B_r/2)┤
for square hysteresis loops, where Br = p Js is the magnetic flux density through the structure at the remanent state and Hc is the coercive field. The energy density curve in Fig. XX shows perfect quadratic behavior with respect to the packing density p, because the coercivity µ0Hc is in any case higher than Br/2.
Fig. 8. a) Coercivity of two parallel interacting nanorods with distance Δ. The coercivity of a single nanorod acts as upper coercivity limit with indefinitely high distance. b) Coercivity values in (a) transformed as function of packing density and the resulting energy density product.
For realistic packing densities around 50% (corresponds to a wall width of 3nm) the predicted energy density product is around 200 kJ/m³ [P2] which lies well above the project target of 60–160 kJ/m³.
3. Irregularly Packed Nanorods
We introduced an algorithm based on the bullet physics library to generate realistically packed structures of nanorods. The nanorods are modelled as ideal, stiff cylinders and their motion is based on gravity and collision forces between nanorods and the walls as well as between nanorods themselves. To obtain nanorod structures with different misalignments, we introduced a tuneable torque on each nanorod which mimics the alignment of magnetic dipoles in a homogenous external field. The misalignment is measured by the standard deviation of angles ϕ between the external field (along the x-axis) and the orientation of the nanorods, denoted as σϕ. Using 3200 nanorods with diameter D=10nm and height H=70nm we obtained structures with σϕ between 10.0° and 71.7° and the packing density varies between 37% and 41%. An example structure is shown in figure 9a.
Fig 9. a) Finite element model of a Co structure with 3200 nanorods. b) Calculated hysteresis properties according to the three simulation methods: Full micromagnetics (µmag) give the most accurate results, macro spin assumes coherent reversal and Hmag takes only crystal anisotropy into account. These simulation modes make it possible to calculate the loss due to incoherent reversal modes (Δ1) and the gain due to shape anisotropy (Δ2)
Remanence and coercivity are decreasing with increasing misalignment σϕ. In our observed range of misalignment (10.0°–71.7°) the remanence Br varies between 0.50T and 0.67T and the coercivity field µ0Hc varies between 0.53T and 0.61T (see figure 9). The values for the remanence are corrected by the packing density, because we are interested in the macroscopic flux density and not in the flux density in the magnetic material alone. The coercivities in are calculated by the µmag method (and are therefore only available for the small structures with 120-140 nanorods) and the remanence values are taken from macro spin calculations (including full structures as well as subsets).
Fig 10. Dependence of remanence (diamonds) and coercivity fields (squares) on misalignment σϕ. The data shows a linear decrease of both remanence and coercivity with increasing misalignment.
The energy density product (BH)max can be calculated directly from the hysteresis loop. Considering the packing density of 40%, we obtain an energy density of 83kJ/m³. If it would be possible to increase the packing density to 45% the energy density would increase to 103kJ/m³ assuming that the overall shape of the hysteresis loop does not change. By variation of material parameters in the simulation it is possible to give minimum criteria for K1 and Js to reach certain energy density products.
(BH)max min. Js min. K1
70 kJ/m³ 1.50 T 150 kJ/m³
90 kJ/m³ 1.75 T 250 kJ/m³
120 kJ/m³ 2.00 T 350 kJ/m³
160 kJ/m³ 2.25 T 450 kJ/m³
Table 2. Lower limits for material parameters to achieve certain energy density products of a packed nanorod structure with a packing density of 40%. Higher densities require lower saturation polarization and higher magnetocrystalline anisotropy for the same energy density product.
4. Antiferromagnetic Capping Layers
Until now, only materials with high uniaxial anisotropy such as hcp Co have been considered as nanorod material. bcc Fe would be a cheaper option with a higher saturation magnetization but bcc Fe has a low cubic anisotropy leading to a low coercivity compared to Co nanorods with the same dimensions (see fig. 1). The idea is to seal the tips of the Fe nanorods with an antiferromagnetically coupled layer with high uniaxial anisotropy. This capping layer suppresses the formation of the nucleation at the tips and increase coercivity [P4].
Micromagnetic simulations on cylindrical bcc Fe nanowires with 20nm diameter and 100nm length have been performed. The nanowires are assumed to have a cubic magnetocrystalline anisotropy with fourth and sixth order constants K1,NW = 46 kJ/m³ and K2,NW = 1.5 kJ/m³, saturation polarization Js = µ0Ms = 2.15 T and exchange stiffness ANW = 25 pJ/m. The cap thickness tcap has been varied between 3 nm and 7 nm and have high uniaxial magnetocrystalline anisotropy an anisotropy constant K1,cap. The interface exchange stiffness Aint between nanowire and cap has been varied between -1% and -10% of the bulk exchange ANW, the negative sign indicates antiferromagnetic exchange.
Fig. 9a shows the dependence of coercive field Hc and exchange bias field Hex on the magnetocrystalline anisotropy in the cap for different cap thicknesses at -2.5% and -5% interface exchange. The cap thickness does not increase the maximum achievable coercivity but thicker caps move the Hc peak to lower K1,cap. This means that the coercivity can be maximized by experimentally tuning the thickness of the capping layer [P4]. Increasing the magnetocrystalline anisotropy in the cap increases the coercive field of the nanowire until a critical point where the coercivity increase is lost and a bias field is introduced. This transition has also been predicted by analytical calculations and a condition for the formation of an exchange bias can be given by:
t_AFM∙K_(1,cap)>J_int (=A_int/a)
where Jint is the interface exchange energy a is the lattice constant. If the caps are hard enough to overcome the interface exchange, the micromagnetic spins in nanowire and capping layer rotate independently from each other and cause the exchange bias field. If Jint > tcap K1,cap, the exchange forces the spins in nanowire and cap to align and the hysteresis loop is unbiased and symmetric. Fig. 9b shows coercivity and bias field as function of the interface exchange stiffness Aint. As long the above condition is fulfilled, increasing Aint only increases the bias field. If the interface exchange is high enough, the exchange bias field Hex becomes zero and the coercivity increases with increasing Aint. Harder caps yield potentially higher coercivity increases but also exhibit higher interface exchange thresholds until the unbiased state is reached.
The simulations on 40nm nanowires consistently show a lower coercivity without caps but a higher increase with the antiferromagnetic caps. This indicates that the coercivity loss due to inhomogeneous magnetization reversal processes is higher for thicker nanowires. The antiferromagnetic capping layer is able to suppress the nucleation at the nanowire tips and compensates the coercivity loss due to nucleation.
Fig. 11. Coercive and bias field as function of a) magnetocrystalline anisotropy of the capping layer b) exchange between nanorod and capping layer.
Co nanowires. Micromagnetic modeling was performed firstly for Co nanowire arrays with different geometry characteristics (length and diameter) and crystalline structure. The general objective of the study included particularly the determination of the role played by the crystalline and shape anisotropies together with magnetostatic interactions for arrays of Co nanowires.
As starting point, the balance between magnetocrystalline and shape anisotropies has been simulated for arrays of Co nanopillars 120 nm height and varying diameter from 35 (mostly fcc cubic crystal phase) to 75 nm (mostly hcp hexagonal symmetry) for which experiment and modeling show crucial changes of hysteresis loops when the diameter is increased. Magpar package [1] has been used with finite element discretization for typical magnetic parameters for Co with hcp and fcc crystal phases. Meshed tetrahedra (average 4 nm sizes) were randomly assigned to corresponding textured phases.
Fig12.- Simulated remanence magnetization for arrays of 7 nanopillars with diameter D= 35nm (a) and 75 nm (b). Arrows indicate the magnetization direction and colors the longitudinal Mz (up) and transverse Mx magnetization on 3D images (up).
The simulations published in confirm the primary role of magnetocrystalline anisotropy and the importance of magnetostatic interactions between nanopillars. The simulations also reveal the change of remanent state when the diameter is increased: from the vortex at the flat end surfaces and single domain in the rest of the nanowire (small diameter of around 35 nm) to the vortex state spanning the total nanopillar (above 50 nm diameter). In the former case, the reversal mode corresponds to the propagation of the vortex domain walls from the flat surfaces, while in the latter it proceeds with the quasi-curling mode taking place in the whole nanopillar.
CoFeCu nanowires.
The modeling was then focused on CoFe magnetic nanowire arrays. In this case, the magnetocrystalline anisotropy plays a less relevant role, and the study addresses the magnetization process and particularly the increase in coercivity observed for the sample annealed with the temperature (see report WP4). For the simulations, a polycrystalline structure of CoFe nanowires with [110] textured bcc phase in agreement with XRD studies was considered. For a CoFe nanowire with given composition, typical exchange and magnetocrystalline constants are respectively A = 10.7x10-12 J/m, K = 104 J/m3 , while the values of the saturation magnetization, Ms, as a function of annealing temperature have been extracted from the experimental measurements. We have assumed that the wires microstructure consists of grains with [110] direction parallel to the nanowires axis and random in-plane easy axis components.
Fig 13. Experimental (filled symbols) RT evolution of coercivity with the annealing temperature for CoFeCu nanowires of indicated diameter, Dw, and corresponding simulations for a single nanowire (lines) and for an array of 7 nanowires (dashed line).
Particular attention has been paid to CoFeCu nanowire arrays with small Cu content where important effect of thermal treatments was observed. After the magnetic field is applied parallel to the nanowire axis, at the remanence state, the simulations indicate the presence of a single domain structure practically along the whole nanowires length with longitudinal magnetization for nanowires with diameter in the range Dw = 8-22 nm. At both nanowire ends open vortex structures appear that minimize the magnetostatic energy. The magnetization reversal occurs by propagation of a domain wall (DW) which structure, either transverse (TDW) or vortex (VDW) mode, appears to be strongly dependent on the saturation magnetization value, for fixed wire diameter. Upon decrease of the saturation magnetization value in the range from 2.0 T (Ms for CoFeCu nanowires in as prepared state) to 1.7 T (Ms after annealing), a significant effect occurs namely, a transition from VDW to TDW mode owing to the change of the magnetic correlation length.
The coercivity value for the transverse wall reversal mode is known to be always larger than that for the vortex wall mode, and thus, during the transition between the two DW types, the coercivity increases. Consequently, the experimentally observed enhanced coercivity after annealing can be understood in terms of the transition between VDW propagation for as prepared nanowires and TDW propagation for the annealed nanowires
In most recent work, relevant parameters including nanowire diameter and crystalline structure have been collected for these micromagnetic simulations.
Other Phases
Tetragonal alloys (Fe/Co)2B for a certain range of compositions they have a rather large MAE and sufficiently high Curie temperature and saturation magnetization. In cooperation with Vienna and Darmstadt groups we have theoretically and experimentally explored these materials, including a possibility to enhance their anisotropy with small amounts of substitutional 5d elements. We have found that a small amount of W or Re elements can efficiently double the MAE. The latter was also confirmed by experiment. The anisotropy reached by 5% doping with Re was 1.4 MJ/m3.
Finally, we have explored layered tetragonal materials based on Fe5SiB2 and their alloys with Co (in place of Fe) and P (in place of Si). Results of this work are subject of a manuscript in preparation [P6]. Pure Fe-based compounds showed only rather small value of MAE, despite their favorable highly anisotropic crystal structure. MAE of about 1.1 MJ/m3 can be reached by about 30% substitution of Fe with Co.
WP2: Preparation of high-aspect-ratio nanostructures (INSAT, CSIC)
Three different methods were developed for high aspect ratio nanostructures synthesis: chemical, electrochemical and mechano-chemical syntheses of polymetallic nanowires and nanoflakes.
Chemical synthesis
“Chemical Synthesis of {Fe-Co}-X-Y Nanostructures” aims on the synthesis of nanorods (NRs) and nanowires (NWs) by the polyol process, suitable for an up-scale for several grams of powder (WP5). Cobalt nanorods with high coercivity (µ0Hc > 0.4 T for a random orientation and µ0Hc > 0.7 T for partially oriented assemblies) were obtained by this method. Core shell Co-Fe anisotropic particles were synthesized through the growth of iron on cobalt rods (D.2.2). The iron shell was varied from 2-10 nm. This shell increased the magnetization but decreased the coercivity, in agreement with the micromagnetic modeling carried out by partner 2 (MS2). In order to target BHmax =160 kJ.m-3 following the conclusions of MS2, the future efforts will be focused on Co and Co-Fe(2 nm) NWs.
Fig 14. TEM images of cobalt nanorods synthesized by the polyol process (a) and (b) with classical heating mantle, (c) and (d) with a microwave oven (Scale bar denotes 200 nm). Mean diameter, Dm, and mean length, Lm: (a) Dm = 18 nm, Lm = 280 nm; (b) Dm = 16 nm, Lm = 160 nm; (c) Dm = 7.5 nm, Lm = 28 nm; Dm = 8 nm, Lm = 42 nm. Sample (a) was prepared with 2.5% of hydrated RuCl3 ref. Sigma Aldrich 463779; (b) with 2.5 % of hydrated RuCl3 ref. Sigma Aldrich 84050; (c) and (d) with 2.5% and 2 % of anhydrous RuCl3 ref. Sigma Aldrich 208523 as nucleating agent.
“Electrochemical synthesis of CoFe rich nanowires”
Magnetic nanowires were prepared by electroplating filling of self-assembled pores in anodic alumina oxide (AAO) templates. The ordered AAO membranes were synthesized by two-step anodization process on 99.999% Al foils in oxalic acid electrolyte by applying a constant voltage of 40V and keeping the temperature between 3-4oC. The first anodization was performed for 24 hours and the second one was performed for 20 hours to assure a thickness of nanoporous alumina template of about 40 μm. The final hexagonal self-assembling of pores with diameter of around 35 nm and interpore distance of 105 nm is achieved. The diameter of pores was subsequently enlarged using H3PO4 (5 % wt). In this way the pores can be enlarged up to 80nm in diameters (Figs 15a-b).
Fig15. Top view SEM images of AAO membranes with pores of: (a) 40nm in diameter and 105 interpore distance, (b) 80nm in diameter and 105nm interpore distance and (c)18nm and 50nm interpore distance.
Well hexagonally ordered templates with much smaller diameters of the pores (16-27nm) are obtained using sulfuric acid as electrolyte (Fig 15c). The geometrical parameters (length, diameters and interpore distance) can be further tailored by changing the anodization voltage and the concentration of H2SO4 electrolyte (see Fig. 16, where A=3% wt. H2SO4, B=10% wt. H2SO4, and C=20% wt. H2SO4).
Fig 16. Dependence of pore diameter for different electrolyte concentrations (A, B and C series) on applied voltages.
Afterwards, Al was chemically etched from the bottom of the membrane and an Au nanolayer was sputtered to serve later as an electrode for the final electroplating of nanowires. CoxFe100-x (0 < x < 100) and FeCoCu nanowires were deposited into AAO membranes at room-temperature by DC electrodeposition from sulfate-based electrolytes containing CoSO4∙7H2O (5-45g/l)+ FeSO4∙7H2O (5-45g/l)+H3BO3 (10g/l) + ascorbic acid (10g/l) (CoxFe100-x nanowires) and CoSO4∙7H2O (35g/l)+ FeSO4∙7H2O (15g/l)+CuSO4∙7H2O (2g/l) +H3BO3 (10g/l) + ascorbic acid (10g/l) (FeCoCu). The pH values of the electrolytes were maintained constant at about 3.0.
Fig.17. Cross-section SEM images of a) CoFe nanowires with diameters of 80nm, b) CoFeCu nanowires with diameters of 20nm, c) CoFeCu nanowires with modulated diameters and d) CoFe/Au multilayered nanowires with 40 nm in diameter. The insert shows a close-up image of modulated nanowires in (c).
Figure 17 shows some examples of CoFe nanowires produced by the method described above. Apart from the already mentioned nanowires with uniform diameters (Fig 17a-b), modulated in diameter pores (i.e the nanowire) along the length can be obtained by combining mild and hard anodization, and specific parameters. CoFeCu nanowires with modulated diameters (22 and 35nm) were obtained (Fig 17c and insert).
Physical Methods-Nanoflakes (TUDA)
The ball milling of Fe-Co-X-Y compounds has been carried out in a planetary ball milling (Pulverisette 6) device. As it is mentioned in the DOW we focused on the composition where the Fe-Co-X-Y compounds show the highest magneto-crystalline anisotropy. It has been reported that the (Fe1-xCox)2B system shows both uniaxial and in-plane anisotropy as a function of chemical composition. Initially, we carried out magnetic measurements to determine the optimum composition of the Fe-Co ratio. Our detailed studies show that, the x = 0.25 composition shows the highest anisotropy constant K1.
Fig.18. SEM images of ball milled Fe1.5Co0.5B samples for (a) dry milling, (b) wet milling under ethanol and (c) wet milling under heptane and oleic acid (surfactant assisted). For comparison of the particle size the same rotational speed and milling time was used for all samples.
After the determination of the optimum chemical composition for the (Fe,Co)2B system we carried out ball milling on the (Fe0.75Co0.25)2B to increase the extrinsic magnetic properties. Initially we started with dry milling and the milling procedure is carried out for 18h at 250 RPM under protective argon atmosphere. Ball to powder ratio is selected as 10:1 which ensures a good energy product. After dry milling, observed average particle size was between 4-6 μm. To reduce the particle size further, we carried out wet milling on Fe1.5Co0.5B. For the milling in ambient we used ethanol and heptane with oleic acid as a surfactant. Figure 18 shows the secondary electron images of (Fe0.75Co0.25)2B samples milled under different ball milling mediums.
After the determination of the crystal structure and approximate particle size distribution we investigate the magnetic properties of the ball milled Fe1.5Co0.5B samples. The room temperature VSM measurement results of the ball milled Fe1.5Co0.5B samples are shown in Table 1.
Task 2.3 Surface modification and recycling
The objective of this task was to grow a thin layer at the surface of magnetic high aspect ratio nanostructures to improve their thermal stability and to avoid the coalescence during a consolidation process.
a) Coating of cobalt nanorods with low melting point metals (Co@M, M = Sn, Sb)
Cobalt nanorods were prepared by the polyol method, washed with absolute ethanol and dispersed in a solution of tetradecanediol in oleylamine containing the tin precursor (tin acetate) or the antimony precursor (antimony acetate). Different ratio M/Co were targeted from 1/8 to 1/2. The suspensions were heated for 30 min at high temperature: 300°C for the tin coating and 250°C for the antimony coating (see more experimental details in D2.3 and D4.9). The final powders were analysed by XRD, TEM and local EDS spectroscopy. X-ray diffraction showed that the reaction of tin acetate with the Co NRs at 300°C lead to the growth of the intermetallic CoSn phase. The TEM images showed composite fibers containing the Co NWs and the CoSn phase. The local EDS spectroscopy showed the presence of tin at the surface of the Co rods (Fig. 19).
Fig. 19. (a) Bright field TEM image of a core-shell Co/CoSb nanorod obtained by the decomposition of antimony acetate in oleylamine at 250°C (Co:Sb ratio = 4:1); (b) STEM-HAADF image (left) and corresponding EDX analysis (right) of a single rod ( Co map in red, Sb map in green and combined map of Co and Sb).
In conclusion, the CoSb shell does not modify the magnetic anisotropy of the cobalt rods and improve their thermal stability. The Co@CoSb anisotropic particles are stable up to 400°C when they are annealed in a H2/Ar atmosphere. These particles are promising for further consolidations to get isotropic permanent magnets.
Highlight clearly significant results;
The electrochemical process was extended to the growth of magnetic NWs with smaller diameter, higher crystallinity and optimized chemical composition for the objectives of the project, high magnetization and high coercivity.
Values of coercivity larger than 0.27 T and squareness ratio above 0.9 were obtained for the Co-rich samples. Very high magnetization, Ms=2.3 T (1830 emu/cm3), were obtained with the Co40Fe60 composition. CoFeCu NWs with complex nanostructures and high coercivity were also obtained by this method. These NWs annealed in good conditions combine high magnetization and high coercivity and are promising for microdevices.
The polyol process was extended to the synthesis of Co NWs with very high coercivities (µ0Hc >0.7 T). The important point is that this method can be extended to the production of several grams of powder at the lab scale (see WP5).
We succeeded in the growth of a thin layer of carbon or CoSn alloy at the surface of the Co NWs without degradation of their magnetic properties. The thin layer can act as a barrier to avoid the Co NWs sintering during the consolidation process.
Sn and Sb are good coatings for the nanowires before compaction and do not degrade their magnetic properties
WP3: High-Throughput synthesis of Rare-Earth Free permanent magnet materials libraries
The main objective of the addition of a third element to the Fe-Co lattice is to maintain a tetragonal distortion. The latter may be set by applying a suitable film substrate or buffer layer. High magnetic anisotropies are expected for tetragonally strained Fe-Co. However, a relaxation of the strained lattice was observed in binary Fe-Co). For ternary Fe-Co-C, a recent publication from Partner 4 predicts ternary phases with spontaneous (‘stable’) lattice strain and remarkable values of magnetic anisotropy. This paper reports on the progress to set this approach into practice and to extend it to ternary systems Fe-Co-X, which include different elements of X (B, C, N, Nb).
Developing “Bain library”, with tailored lattice parameters
According to theoretical calculations FeCo in tetragonal distorted phase could exhibit high magnetocrystalline anisotropy making it a promising material for permanent magnets applications, since it exhibits the highest magnetic moment among any other ferromagnetic material included Fe. Since the equilibrium state of FeCo is the cubic phase, a stabilization of tetragonal distorted FeCo is a demanding task. The highest magnetic anisotropy is expected for c/a lattice constant ratios between 1.2 and 1.25.
Combinatorial experimentation is a systematic way to fabricate and characterize a batch of samples on one substrate, covering broad regions of the compositional diagram and therefore, called materials libraries, under identical conditions in one experiment. Using combinatorial fabrication with techniques like co-sputtering or wedge–type depositions we are able to fabricate material libraries with well-defined composition and thickness gradients. Moreover high-throughput characterization methods (magneto-optical Kerr magnetometry, X-ray diffraction, EDX) allow us not only to verify the theoretical predictions but also to study extrinsic magnetic properties as coercivity (Hc) or remanent magnetization (Mr) which are almost impossible to be predicted by calculations and are crucial factors for high performance permanent magnets.
AuxCu1-x Library (Buffer layer)
The main idea is that the in-plane lattice parameters of FeCo on AuCu will adopt those of AuCu and the out-of-plane lattice parameter c is expected to adapt accordingly, since the unit cell volume should remain constant. The AuxCu1-x library was made by co-sputtering of Au and Cu elemental targets, at room temperature in confocal deposition geometry (K2 sputtering system – RUB facilities). The range of Au content starts from 13% < x < 46% (Fig.20). The epitaxial growth of the Au-Cu buffers was confirmed by XRD texture measurements in a four-circle set-up (IFW facilities) and can also be seen from XRD Bragg-Brentano measurements (Fig. 21 with the high (002) reflection around 2θ = 66°). Moreover we grew successfully the same buffer layer AuxCu1-x again by HT-sputtering techniques on MgO substrates at 300 oC, improving further the degree of texture and making it appropriate for higher temperatures (<300 oC) depositions which are needed for the crystallization not only of B-doped FeCo compounds but also of the binary system.
Fig 20. Compositional spread of Au content x(%) in AuxCu1-x buffer layer library
High-throughput deposition of Fe-Co-X on Bain library
After the deposition of the AuxCu1-x buffer layer library, the wafer was transferred into K1 sputtering system (RUB facilities) for the growth of a thickness gradient FeCo library. K1 is more adequate for thickness-dependent depositions than K2 sputtering system due to the movable deposition shutters of which the velocity can be adjusted from 1 mm/sec to 5 mm/sec. Initially the wafer is completely covered by the shutter which is retracted in a constant velocity during the deposition process, thus making a thickness gradient. The wedge-type thickness FeCo layer was deposited perpendicular to the AuxCu1-x composition spread axis A FeCo compound target (99.95% purity) was used for the deposition of FeCo layers. Under these conditions, a nominal thickness gradient ranging from 3.2 nm to 10.3 nm was achieved. According to the literature the relaxation volume of FeCo on Rh (c/a=1.24) is 12 monolayers (ML) or 1.14 nm. It is obvious that we are above the relaxation volume of FeCo, but this is the goal of this task: the study of magnetic properties of thicker strained FeCo films. More details about the magnetic properties are given in the next section. The material library was characterized structurally by a high-throughput x-ray diffractometer (RUB facilities) which allows screening of 301 measurement regions using a highly sensitive x-ray detector.
Figure 21. X-ray spectra for a 30 nm thick FeCo thin film deposited on a AuxCu1-x/Pd/Cu/Si substrate.
Following our methodology we can also estimate the c/a ratio of FeCo as a function of the composition of AuxCu1-x buffer layer (Fig.22).
Figure 22. c/a ratio of FeCo as a function of the Cu content in the AuxCu1-x buffer layer.
X=C: Fe-Co-C films
In most of the prepared films of Fe-Co-X type, we aimed at the compositions which promised the best performing in the calculations: (FezCo1-z)16X with z close to 0.4. The addition of approx. 2 at% carbon into the Fe-Co lattice stabilises a tetragonal strain of c/a = 1.03. We suggest that this strain originates from the formation of a spontaneously strained Fe-Co-C phase. The RHEED results recorded during growth of the (Fe0.4Co0.6)0.98C0.02 films (Fig. 23a) revealed a reduced driving force for strain relaxation. Together with the observed c/a ratios at thicker films (up to 100 nm thickness) containing 2 at% C (Fig. 23 b), these are strong indications for a lattice softening due to the C and a spontaneously strained Fe-Co-C phase which contains only 2 at% C and has 3% tetragonal distortion.
Fig 23: a) Comparison of ternary (Fe0.4Co0.6)0.98C0.02 with binary Fe-Co strain relaxation during growth detected with in situ RHEED. b) c/a ratios of tetragonal distortion from texture XRD measurements of (Fe0.4Co0.6)0.98C0.02 films in dependence on film thickness.
X=B: Fe-Co-B films
A larger content series with B contents up to 9.6 at% was performed with PLD prepared films. For a B content of 4 at%, the detected c/a ratio is 1.045. A further increase of B content, however, led to a decreased c/a ratio. A likely reason is the higher fraction of amorphous phase in the alloy as indicated by the reduced x-ray diffraction intensity. The magnetocrystalline anisotropy is also dependent on the B content. The estimated magnetic anisotropy energy MAE reveals a maximum close to the maximum of c/a. The reason for the shift of the maximum to a lower B content are the opposing effects of lattice strain and magnetic saturation: While a higher B content first increases the lattice strain, it reduces the magnetic moment of the alloy at the same time. In consequence, maximum MAE is shifted to B contents of approx. 2 at%. The highest MAE are observed in ultrathin films (d < 5 nm) due to the strong contribution of coherent strain at the Au-Cu interface. At higher thicknesses, the lattice relaxation becomes relevant and MAE decreases. For the binary films, it nearly vanishes. Due to the stabilised strain in the ternary films, a residual MAE around 0.4 MJ/m³ remains.
X=Nb: Fe-Co-Nb films
A [Fe/Co/Nb]24 multilayer materials library (ML) was fabricated by a wedge-type multilayer approach using a combinatorial magnetron sputtering system..In XRD analysis, we observe (110) oriented Fe-Co and
Figure 24: Hysteresis loops of Fe-Co-Nb for average Nb contents from 0 to 21 at. %
(FeCo)3Nb with various orientations. The higher the Nb content, the more was the fraction of (FeCo)3Nb. Together with TEM measurements (including EELS) the formation of a nanocomposite structure of the Fe-Co-Nb thin films, consisting of a matrix based on cubic FeCo and (FeCo)3Nb precipitates, is suggested.
The magnetization reversal of the composite system (Fig. 24) occurs at two discrete fields indicating that the new compound is also ferromagnetic but possesses significantly different magnetic properties compared with the FeCo-based matrix. The magnetic measurements revealed that the system behaves as a hard-soft magnetic composite. Particularly, hysteresis loop measurements at various angles showed that the system dominates by a domain wall (DW) pinning switching mechanism and (FeCo)3Nb inclusions act as pinning sites due to the anisotropy difference between the two phases. A rough estimation of the anisotropy constant based on DW pinning model showed that it is the same order of magnitude of that of ferrites. Furthermore, the energy product, estimated only by using the magnetic properties of this particular phase and not of the composite, is also comparable with the ferrites energy product, pointing that (FeCo)3Nb is a promising compound for permanent magnet applications.
Search for new phases in Heusler families
Theoretical studies have predicted the existence of tetragonal phases in Fe2YZ alloys with high magnetic moments. Thin films with Y = Ni, Cu, Co and Z = Sn, Ga were grown on thermally oxidized Si(100) substrates with 500 nm thick SiO2 layer, and epitaxial substrates, with a Mantis UHV magnetron sputtering system with three sources in a confocal geometry. Adjusting the power of the sources was the method that controlled the composition. The Fe-Y-Z films were studied with EDX, X-ray reflectivity measurements, X-ray diffraction, TEM and magnetically characterized by means of a VSM.
Fe-Ni-Sn films grown on BaF (001), independently of film thickness and exact composition, are crystallized in the hexagonal structure P63/mmc(194) showing high magnetic moment (450 emu/cc). Phase separation or amorphous phases were observed in the other systems.
Fe2CuGa films were also studied using combinatorial approach. The XRD pattern was close to the theoretically expected corresponding to tetragonal half-Heusler alloy structure. The effect of stoichiometry was crucial, since for different compositions than the nominal, another crystallographic orientation was revealed. The structural properties were also strongly depended on deposition temperature, while the optimum results observed at 300 °C.
Fig 25: Calculated X-ray diffraction pattern for the tetragonal Heusler alloy (up) and the X-ray diffraction pattern of Fe2CuGa on Si (111) (below) sample are shown. The peaks of the regular cubic structure are presented in blue color.
Another approach was the modification of various stoichiometric Heusler alloys by substituting –in bulk or melt spun materials– specific 3d elements with Fe and Co, in order to tune the magnetic properties. Main emphasis was given to Mn-Ni-Sn family and Mn-Ga family. The adopted strategy was to introduce Fe in substitution for Mn. Fe2-xMnxNiSn compound presented high saturation magnetization (80.0 Am2/kg), significantly larger even from 5 K values reported for Mn2NiSn. Mn0.4Fe0.3Ga0.3 material presented overall interesting values for its magnetic properties, with saturation magnetization equal to 51.3 Am2/kg and coercive field equal to 3 kOe in room temperature.
Other approaches for inducing tetragonal distortion
A way to induce uniaxial anisotropy in FeCo-based alloys is by the formation of new ternary phases. The spirit of this approach is similar to that of epitaxial growth: in both cases the cubic symmetry of the parent FeCo alloys should be broken. Bulk (Fe0.7Co0.3)2B can be stabilized in a tetragonal phase which exhibits high uniaxial anisotropy. We fabricated and studied Nb-doped FeCo library. Nb was selected to replace rare earth (RE) elements due to its low cost and high abundance. Moreover, binary FeNb and CoNb systems form hexagonal structures fulfilling the criterion of broken cubic symmetry. Also, Nb is heavier than Fe and Co which means that possesses higher spin-orbit coupling which is one of the most important factors for achieving high magnetocrystalline anisotropy. A new hexagonal ferromagnetic phase (FeCo)3Nb was detected. The fabrication of pure (FeCo)3Nb films which will allow the study of the magnetic properties of this new compound.
WP4: Characterization –structural and magnetic- of high-aspect-ratio Nanostructures and RE-free intermetallics
Various compositions with different Co and Fe content have been studied. Cast ingots of different compositions have been synthesized by induction melting in inert atmosphere using master alloys of Co85B15 and Fe80B20. In addition, melt-spun ribbons were prepared from the cast ingots by a pressure difference of He between gas chamber and vacuum chamber of 200 mbar and at different wheel speeds. Hysteresis curve of the samples was measured by VSM at room temperature. Melt spun ribbons of selected compositions were annealed at 500 °C for 1 hour. Selected alloys were investigated by TEM and XRD. A significant dominance of the crystal structure Fe2B is observed.
Table 3: Overview of various Fe-Co-B compositions
For (Fe0.7Co0.3)71B29 a coercivity of μ0Hc = 0.05 T was obtained with a wheel speed of 26.7 m/s and a pressure difference of 200 mbar.
Nanoanalytical TEM characterisation of post annealed Fe/Co/Nb multilayers
Different Fe/Co/Nb multilayer specimens were prepared for TEM investigations by the focussed ion beam lift-out specimen preparation technique. The specimen thickness is in the range of 120 – 175 nm. The Fe/Co/Nb multilayers exhibit a granular microstructure with a grain size of about 50 to 100 nm in diameter. Thickness of the Fe/Co/Nb magnetic layer is 315 nm, which is embedded between two SiO2 layers. The thickness of the top SiO2 layer is 12.4 nm. Two regions (A and B) were chosen for elemental characterization by energy loss spectroscopy (EELS). In region A besides Fe (48.5 at %) and Co (42.4 at %) also Nb (9.1 at%) was identified. No Nb was found in region B, indicating that Nb is inhomogeneously distributed within the magnetic layer.
Fig 26: Z-contrast image of the cross section (left) and bright field image with indexed pattern showing a granular microstructure (right) of Fe/Co/Nb multilayers
High-aspect-ratio nanostructures
Antimony coating Co nanorods (Co@CoSb) from partner 10 (INSAT) (see WP2) were used for scaling up and consolidation of Co-based nanorods with high densities for permanent magnets. Coating the Co nanorods can significantly help to avoid them oxidation and reach high density packing but preventing the strong dipolar coupling. Systematic study of structural microstructural and magnetic properties of CoSb coated Co nanorods with different shall thicknesses and at different temperatures has been performed. The morphology (shape) of these systems in both cases is preserved up to the end of reaction of Co nanorods with antimony acetate at 250 °C for 30 min.
Systematic XRD measurements of prepared systems have revealed that the presence of a CoSb shell around the Co rods delays their oxidation. The coercivity is in the range 3-3.6 kOe, rather constant, showing that no strong modification of the magnetic Co core occurred with the coating. The remanence to saturation magnetization ratio Mr/Ms of 0.5 the expected value for a powder in which particles are randomly oriented in the applied magnetic field. After annealing no apparent changes in the coercivity of the nanorods was noted, that marks the enhanced stability of these nanorods at 400°C.
Figure 27: Magnetization loop at room temperature of the raw Co nanorods (a); Co@CoSb with Co/Sb=8 (b); Co@CoSb with Co/Sb= 4 (c); Co@CoSb with Co/Sb= 2 (d).
Conventional Structural, microstructural and magnetic properties of nanostructures-nanowires
The role of Co on the structure and magnetic properties of CoxFe100-x nanowires with two selected diameters, 40 and 20 nm, respectively, was studied. The X-ray diffraction results show cubic symmetry in the whole range of compositions that evolves from body-centered cubic, bcc, structure for pure Fe and small Co content nanowires towards face-centered cubic, fcc, as Co content increases. For high Co content, including pure Co nanowires, the only stable phase is the γ phase, fcc structure with a [110] preferred orientation. A similar compositional evolution is found for nanowires with 20 nm in diameter: i) bcc cubic crystals with (110) texture for more reduced Co content until pure Fe nanowires, ii) fcc cubic crystals with (110) texture for high (around 90%) Co content, and iii) hcp hexagonal symmetry structure with (100) texture for pure Co nanowires. By increasing the Co content, the coercivity reaches a maximum at about 70-80% Co, and then decreases. Values of coercivity larger than 0.27 T and squareness ratio above 0.9 were obtained for the Co-rich samples. The maximum saturation magnetization of about Ms=2.3 T is obtained for the Co40Fe60 alloy nanowires. Room temperature oercivity of Co67Fe28Cu5 nanowires with diameters of 18 and 22 nm and the length of about 8 mm increases with annealing temperature up to 500 °C. By further increase of the annealing temperature it decreases. Squareness of all annealed samples was around 0.95. Saturation magnetization takes values of 2.0 and 1.7 T for as-deposited and 500 °C annealed samples respectively. This reduction has been ascribed to a partial oxidation of the nanowires. Note that increased coercivity is obtained after thermal treatment in spite of the reduction of saturation magnetization and the corresponding shape anisotropy.
Fig 28: Parallel and perpendicular hysteresis loops for the 18nm diameter nanowire array in as-deposited (a) and 500 ºC annealed (b). The EMD lies along the nanowire axis
Then, we have focused on the systematic study of the influence of thermal annealing on the structure and magnetic characteristics. XRD and TEM structural studies indicate that, after annealing, CoFeCu (with intermediate Co content) structure remains cubic bcc structure with reduced defects and sharper crystallite bounds. After suitable annealing the <111> growth direction is identified, together with a gap between wire and surrounding alumina that could contribute to the observed magnetic hardening after annealing.
Fig 29: (c) HRTEM images of CoFeCu nanowire 40 nm in diameter annealed at 700 ºC showing <111> direction as the nanowire growth direction. (d) Fourier transform of the area inside the box depicted in (c). (e) Close up showing the lattice image. (f) A gap (marked by the two arrows) between the nanowire and the alumina is observed.
CoFe and CoFeCu nanowire arrays, were annealed at different temperatures for 2 hours in Ar atmosphere. After annealing, a magnetic hardening is observed for all, more effective in the Cu ones. During the annealing, the presence of Cu is more effective to induce a noticeable magnetic hardening in comparison to nanowires without Cu content. In addition, the annealing temperature at which optimal magnetic hardening is achieved in CoFe and CoFeCu depends on geometry characteristics.
Fig 30: Temperature dependence of coercivity for several CoFe and CoFeCu nanowires (DNW=18 nm) in their as-prepared state and after optimal annealing.
Thin films approach
The multilayer approach (preparing [AuCu/Fe45Co55(C)]n=3 multilayers) supports to reach the large magnetocrystalline anisotropy, which is competing with the shape anisotropy of FeCo films. Using FMR we found this anisotropy to be of the order of 1 MJ/m3. From the FMR we also found that C doping supports the increase of MAE in FeCo sublayers. Element-specific XMCD measurements revealed the increase of the mL/mS ratio at the Fe-site, showing that the modification of the Fe electron band structure in our multilayers can be reason of obtained large MAE.Magnetic – nonmagnetic [AuCu/Fe45Co55]n=3 and [AuCu/ Fe45Co55C]n=3 multilayers were prepared using magnetron sputter deposition. Magnetic properties of samples were measured by SQUID, FMR and XMCD.
Hysteresis loops with external magnetic field parallel and perpendicular to the film plane for both samples are shown in Fig.8. Saturation magnetisation Ms of sample [AuCu/Fe45Co55]n=3 (RE-71) was found to be 1530 emu/cm3, whereas as sample doped by C (RE-70) has Ms of 1700 emu/cm3. These facts point toward the presence of large perpendicular anisotropy fields, comparable to the demagnetizing fields of 4πMs = 19.2 kOe and 21.3 kOe respectively.
Fig 31. SQUID measurements of [AuCu/Fe45Co55(C)]n=3 samples.
Advanced High-Through-put Characterization techniques (structural, microstructural & magnetic). A Fe-Co-Cr-Ni thin film library was fabricated using the sputtering system, which has been used for the deposition of the Fe-Co-Cr ternary library. After the deposition the library was annealed at 550 oC in vacuum for 80 min. As in the case of Fe-Co-Cr only a few very weak peaks were observable in XRD patterns, indicating the formation of a nanostructured or amorphous material. Screening of the magnetic properties by HT-MOKE revealed that the maximum coercive field as in the case of the ternary system was achieved for Cr rich regions (> 50%). Hysteresis loops measured by VSM did not show any significant enhancement of the coercivity (90 mT) compared with the values of the ternary system. The ratio between the remanence magnetization Mr and the saturation magnetization remains low (~0.58). The study of the ternary FeCoCr and quaternary FeCoCrNi libraries showed that a moderate coercive field can be achieved which is higher than that of commercial FeCoCr compounds.
Other local probe advanced techniques for single sample characterization
The Procedure to correlate Texture maps and microstructure was achieved with the upgrade of the available 4 circle neutron diffractometer to SUPER-6T2 version. A specially designed microwave shortcut can be driven at 2 to about 24 GHz continuously. The device is mounted in a quartz tube which is attached to a conventional Ultrahigh Vacuum System. Powder samples can be introduced through a load-lock which also includes a plasma cleaning stage for controlled removal of Carbon or Oxygen species on the surface of colloidal nanostructures supplied by the partners in the consortium. Auger spectroscopy is used to verify the purity of the nanostructures which then can be moved in-situ to the FMR detection stage. The load-lock mechanism and the reducing action of the plasma treatment have been confirmed. The detection limit of the multifrequency in-situ FMR device has been successfully tested with a 10 nm Fe film on GaAs(100).
Highlights of most significant results
Systematic study and detailed analysis of FeCo nanowires has been performed by partners 7 and 5. The performance of the nanowires could be significantly improved for permanent magnet applications. Particularly the magnetic hardening (increase of remanent magnetization and coercivity) was achieved by depositing hard magnetic (or antiferromagnetic) materials at the tips of nanowires and by the small addition of Cu with subsequent annealing at 400 0C.
The SANS measurements of powders of Co-based nanorods by partners 9 and 4 have supported the optimization of the washing procedures for studied systems. The optimization of the washing procedure had strong impact on the alignment of nanorods during the consolidation. This optimization was in significance for achievement of the large energy product in Co-based consolidated nanorods for permanent magnets reported in WP5. Additionally, it has been shown by Partner 4 that CoSn shell doesn’t influence on structural and magnetic properties of the Co core and this shell can be used as a protector from degradation of compressed Co nanorods in permanent magnets.
New hard Fe-Co based magnetic phases have been discovered by partners 6 and 2 using the high-throughput characterization techniques and the melt-spinning facility, respectively. Superior magnetic properties of hard magnetic Mn70Ga30 phases has been obtained by partner 13 while optimasing the texturing process. These are important ingridients for consolidation of the semi-hard (Fe-Co-B or Fe-Co-Nb) and hard (Mn70Ga30) magnetic phases in order to obtain an exchange spring magnet with high energy product.
The SQUID, FMR and XMCD measurements performed by partners 1 and 5 have reveald an exceptionaly (for the Fe-Co films) large magnetocrystalline energy density of oabout 1 MJ/m3 in the [AuCu/Fe45Co55]n=3 and [AuCu/ Fe45Co55C]n=3 multilayers prepared by partner 1. This energy density is comparable with MAE of SmCo alloys. This finding suggests the possibility to use these systems as PM in write-read heads of magnetic memory devises.
It has been found by partner 8 that with increasing the (B+C) content, the tetragonal strain in the Fe-Co lattice can be increased up to approx. 5%. This is remarkable since this strain is higher than the so far observed strains in ternary films of the type Fe-Co-B or Fe-Co-C where the highest strain was approx. 4%. However, they also have found that the crystal size is also important with regard to the magnetocrystalline anisotropy: if the appropriate crystal size is not maintained in the Fe-Co films, the anisotropy constant decreases strongly. The XMCD study performed by partner 5 confirms unambiguously that the observed enhanced MAE in spontaneously strained FeCo films due to B (C) doping is related to the electronic structure of studied systems and not to the possible grain structure or morphology of prepared films.
WP5: Up-scaling and Consolidation Techniques of nanocomposites-sintered magnets
While most of the preparation routes described up to now will be directly applicable for the growing market of magnetic MEMS, the large volume market requires novel routes to prepare rare earth free bulk permanent magnets. In this WP we will examine the feasibility of conventional and novel up- scaling techniques. We consider this WP as absolutely crucial when moving from optimized intrinsic to extrinsic magnetic properties or from micro to macro scale. This WP certainly represents a high risk endeavor as most of the following approaches will be applied for these materials for the first time. The challenge is to develop novel up-scaling and consolidation techniques (or to apply existing ones in an innovative manner) which yield densities as high as possible, without degrading the nanostructure while avoiding magnetic dipolar interactions and maintaining the high-aspect-ratio character of the precursor materials. Finally routes towards near net shape permanent magnets with unique magnetic properties in the chosen temperature window for application are developed. The bottom-up approaches described in the previous work packages will allow the preparation of high-aspect-ratio nano structures in addition to the theoretical predictions and high-throughput synthesis of thin films. As the market of permanent magnets requires bulk material, in this work package we report on the feasibility of up-scaling processes and consolidation techniques.
Up-scaling processes
The up-scaling processes task aims on supplying sufficient material for the tasks of consolidation and/or towards application s.
Upscaling of high aspect ratio nanostructures
The cobalt nanorods (Co NRs) were found to be the most promising high aspect ratio nanoparticles for
Figure 32: (a) TEM image of cobalt nanorods prepared at a scale of 10 g in 2L of butanediol using a 3L jacketed reactor; (b) magnetization curve of an assembly of the rods measured with the applied field parallel to the rod long axis.
\further consolidations. Pure cobalt rods indeed are the best compromise to get high magnetization and high coercivity. Coercivities higher than 7 kOe were measured on cobalt rods prepared at small scale. Moreover the polyol method is better suited to a scale-up than organometallic chemistry. Textured assemblies of nanorods were prepared by drying suspensions applying an external magnetic field. Co NRs were washed twice with ethanol, chloroform, then were sonicated in chloroform. The suspensions were placed into a mould in air gap of an electromagnet. The drying of the suspension and the rod alignment was achieved under a field of 1T. Large scale alignments were obtained (Fig.33) with a good correlation distance as inferred from small angle neutron scattering (SANS) characterizations.
Figure 33: (a) wafer of cobalt nanorods obtained by drying a suspension in chloroform under an external magnetic field of 1T; (b) Small angle neutron intensity profile scattered by a cobalt nanorod assembly, perpendicular (black square) and parallel (red diamond) to the rod alignment. Inset: corresponding 2D SANS pattern. The two correlation spots scattered perpendicularly to the rods show a very good rod ordering in the assembly.
From thin to thick films and foils
The study of the magnetic properties of Fe-Co-X single layers in dependence of the applied in-plane lattice parameter of the buffer material revealed a large contribution of a surface related anisotropy Ksurface around 1.8 mJ/mm², which includes contributions of both, the buffer interface and the free surface (see D3.5). Although the main fraction may be related to the free surface, the role of the interfaces is also remarkable. An additional contribution may originate from an induced strain due to the applied underlayers. However, the studies on single layers did not reveal significant differences, when comparing different Au-Cu buffer compositions i.e. very different in-plane lattice parameters (see also D3.5). Multilayers, which have a repetitive architecture of Fe-Co(-X) layers with interlayers of the Au-Cu buffer material (Fig. 34), could thus increase the positive contribution of the interfaces to the overall magnetocrystalline anisotropy KU in film architectures of higher thickness, i.e. upscaled films. For the Fe-Co- layers, we chose the composition (Fe0.4Co0.6)0.98B0.02 since this gave the highest MCAs according to previous studies on single layers (see D3.3).
Figure 34: Schematic architecture of the multilayers. The overall multilayer thickness is 100 nm for reasons of comparability.
X-ray diffraction and TEM measurements of multilayers with dAu-Cu = dFe-Co-B = 5 nm revealed that a further decrease of the Fe-Co-B layer thickness is necessary, since the structural and magnetic properties in terms of tetragonal strain and MCA do not fundamentally exceed the values of single layers. However, in situ RHEED measurements during the deposition of multilayers with dFe Co-B = 2 nm showed that a interlayer thickness of 4 nm is necessary to maintain the epitaxial growth up to an overall thickness of 100 nm. Such multilayers exhibit KU values around 0.95 MJ/m³ in architectures with at least 20 repetions. Their MCA is between two and three times the value of a single layered film with equivalent thickness . Very similar KU is also observed in multilayers based on C doped Fe-Co with Au-Cu interlayers, but a smaller overall thickness
In conclusion, the multilayer approach may be a suitable way to prepare film architectures with both, high overall thicknesses and high KU with respect to Fe-Co(-X). With further improvement, the shape anisotropy might be overcome and these multilayers could be an option for small scale systems, where the material cost of the interlayer alloy is not crucial. However, regarding possible bulk magnets, one also has to consider the magnetic saturation MS which is reduced substantially compared to single layers, when the whole multilayer volume is taken into account (see D5.7). In consequence, the possible energy product BHmax, which scales with MS2, is drastically reduced to about the factor of 1/10. Table 4 summarises this point as it compares the properties of a multilayer with 17 repetitions and the best performing 100 nm thick (Fe0.4Co0.6)0.98C0.02single layer.
Table41: Comparison of the properties of single and multilayers.
KU / MJ/m³ MS / T BHth,max /kJ/m³
(Fe0.4Co0.6)0.98C0.02single layer 0,44 2,1 880
(2nm (Fe40Co60)98B2/4nm Au46Cu54) multilayer (n = 17) 0,98 0,7 90
Inducing metastable Fe-Co-X-Y phases
The goal of this task is to stabilize tetragonal distortion in the Fe-Co-X-Y family. For this task in addition to the Fe-Co-X-Y systems we carried out investigations on a Mn-based (Mn-Ga) system.
Fe-Co-X-Y:
Theoretical predictions show that ternary (Fe,Co)2B and (Fe,Co)3B systems are promising candidates for tetragonal distorted systems. After the optimization of the composition of (Fe,Co)2B system we carried out quite a number of additions to the (Fe0.75Co0.25)2B system. To observe the effect we add Ta, Ti, Sc, Y, Cr, W, Re, Ir, Zr, Mn, Ni, P, Ga, Al, Si, In, Sn, Sb elements to the pure (Fe0.75Co0.25)2B system. The magnetic measurements show that only Ta, Zr, Ti and Re shows promising results. The Ta and Zr substituted samples show higher coercivity in the bulk form compared to the pure system. The Ti substituted sample shows lower coercive field with slightly increased saturation magnetization. The hysteresis measurements of the Ti, Ta and Zr samples are shown in Figure 5.6 in comparison to the pure (Fe0.75Co0.25)2B system.
The observed increase of the anisotropy energy (or anisotropy constant) for Re and decrease for Ir is confirmed by the first principle calculations. Theory predicts and increase for W and Re samples and a decrease is observed for other 5d transition metals including the Ir. Calculated magneto-crystalline anisotropy energies are shown in Figure 5.8.
Figure 35: MAE for various elements X in (Fe0.675Co0.3X0.025)2B and (Fe0.675Co0.275X0.05)2B. The dotted line indicates the MAE of (Fe0.7Co0.3)2B for comparison.
The Mn-Ga system
The goal of the work on the Mn-Ga system was to find an up scalable processing route of Heusler like DO22 structure of Mn3-δGa system. Different processing routes were investigated for the optimization of the DO22 phase. These results indicate that the optimized process route to obtain the DO22 phase is to arc melt the bulk alloy, cold rolling it and subsequent field assisted annealing at 730 K for 24 h. Using these parameters, we are able to produce samples of superior properties to those reported on bulk samples in the literature. Table 5 shows a comparison of this study with the existing literature.
Table 5: Comparison of magnetic properties and sample preparation routines of Mn3-δGa.
Composition Remanence
(Am2kg-1) Coercive field (T) Type of the sample Preparation details Reference
Mn70Ga30 20.3 0.58 Bulk (rolled) 2h at 700K Sample 8b
Mn70Ga30 23.6 0.72 Bulk (rolled) 2h at 700K under 1T Sample 9b
Mn70Ga30 26.1 1.24 Bulk (rolled) 24h at 700K under 1T Sample 10
Mn75Ga25 14.4 0.43 Bulk (arc melted) 336h at 623K J.Winterlik et al., Phys. Rev. B 77, 054406
Mn75Ga25 16.1 0.36 Bulk (arc melted) 168h at 673K B. Balke et al.,
Appl. Phys. Letters
90, 152504
Mn70Ga30 25.0 1.35 Bulk (arc melted) 24h at 1173K + several days at 673K H. Niida et al.,
J. Appl. Phys. 79, 5946
Mn70Ga30 10.5 0.57 Melt spun ribbon 1h at 973K T. Saito et al.,
J. Appl. Phys. 112, 083901
Mn75Ga25 28.0 0.65 Melt spun ribbon 50h at 723K Y. Huh et al.,
IEEE Trans. Magn. 49, 3277
Mn67Ga33 19.5 2.05 Thin film Sputtered at 673K C.L.Zha et al.,
J. Appl. Phys. 110, 093902
Mn75Ga25 17.0 1.19 Thin film Sputtered at 623K K.Rode et al.,
Phys. Rev. B 87, 184429
In conclusion, an optimized production routine for the metastable phase of Mn70Ga30 system was found. Our study indicates that 24 hours of annealing in 1T at 730 K is sufficient to stabilize the metastable DO22 phase in cold rolled samples. The advantage of the preparation routes is up-scalability in bulk.
Consolidation techniques
Texturing of nanowires by magnetic fields
The goal was to align the nanorods in a macroscopic powder with the following characteristics:
a high magnetic nanorods volume fraction in order to get a high magnetization per volume;
a perfect alignment to get the most square possible magnetization loop;
a preserved high aspect ratio character to maintain a high coercivity.
The SEM images of magnetically alligned cobalt needles are shown in Figure 5.12 for two different magnification. For the production of the needles, concentrated Co nanowires suspension in chloroform or in toluene were dried in an electromagent with an applied field within the range of 0.2-1.5 T. The optimum alingment was observed for the external field of 1 T. The lower magnetic fields leads to higher misalingment and bad magnetic properties.
It was the first time that it is clearly showed that an energy product higher than 160 kJ/m3 can be reached with high aspect ratio magnetic particles.
Figure 36: Hysteresis loop (a) and second quadrant of the corresponding B(H) loop (b) of three rod alignments exhibiting the same magnetic volume fraction: (blue) dm = 22 nm, volume fraction = 48.7%, BHmax = 126 kJ.m-3 ; (black) dm = 24 nm, volume fraction = 48.7%, BHmax = 82.5 kJ.m-3; (red) dm = 28 nm, volume fraction = 48.8%, BHmax = 51 kJ.m-3.
Figure 37: Second quadrant of the B(H) loop of two assemblies of the same cobalt nanorods (dm = 22 nm) with a magnetic volume fraction VM = 48.7% and a BHmax = 126 kJ.m-3 (blue) and a magnetic volume fraction VM = 54.4% and a BHmax = 167 kJ.m-3 (magenta).
Hot pressing (and cold compaction)
The goal was to obtain hot compacted dense permanent magnet materials by using high-aspect-ratio nanostructures. The hot pressing studies on high-aspect ratio nanostructures were carried out for different pressures and temperatures. To ensure the hexagonal crystal structure of the Co nanowires all studies were carried out below the structural transition temperature of the cobalt. Prior to the compaction, the Co nanowires were aligned under external magnetic field of 1.7 T to ensure a good texture. After texturing these “green compacts” were used for hot pressing. The results (Figure 5.15) show that it is not possible to reach to the 75% texture by using the hot pressing which is due to the used pressure values. The best sample is obtained for the pressure for 40 MPa.
Figure 38: Room temperature pulse magnetometer measurements of hot compacted Co nanowires.
Magnetization measurements of hot compacted sample shows remanence to saturation ratio around 62% which is slightly better than the polymer bonded Co nanowires. The room temperature (BH)max value for the hot compacted Co nanowire magnet is calculated as 15.7 kJ/m3. The reason of this low energy product can be explained by the low packing density of 44%.
In addition to the pure Co nanowires we carried out hot pressing on the antimony coated Co NWs. Same hot pressing conditions are used for Co@Sb nanowires and the magnetization results are given in Figure 5.16.
Compaction studies for Mn-Bi samples were carried out in a polymer press equipped with a electromagnet. For the compaction of these magnets a special design press form was used. Fig 39 shows the images of the press form and a final product magnet as an example.
Fig. 39 Images of the press die (a) and the final magnet (b)
Spark-plasma sintering (SPS)
The goal of the spark-plasma-sintering studies were to obtain dense permanent magnets by using high-aspect-ratio nanostructures. Similar to the hot compacted samples, the Co nanowires were pre-aligned with an external magnetic field of 1.7 and the consolidation studies were carried out on these pre-aligned “green compacts”. The SPS studies on Co nanowire powder leads to similar results (texture ratio of 60-65% after consolidation) as the hot compacted samples. This observation shows that the hot pressing and SPS has almost the same effect on the consolidation of the Co nanowires. The studies show that using Co needles leads to better magnetic properties compared to Co nanowire powders. Due to this reason we carried out SPS studies on Co needles which are consist of aligned co nanowires. Figure 5.20 shows an image of the Co needles/platelets.
Figure 40: Examples of pre-aligned needles/platelets obtained in the 5mm width parallelepiped mold.
The SPS studies which are carried on the pre-aligned Co needles show promising results even after compaction. Figure 5.21 shows a room temperature hysteresis curve of a SPS compacted Co nanowires. Hysteresis measurement of spark plasma sintered Co NWs show remanence to saturation ratio of 78%. Observed (BH)max value for the SPS Co NWs is 27.5 kJ/m3 at room temperature.
Figure 41: Room temperature pulse magnetometer measurements of Co nanowire needles after SPS.
In addition to the different consolidation methods the effect of shape of the magnet is investigated in detail. The room temperature hysteresis measurements confirm that the most promising shape for the application of Co NWs is the needle shape. For further increase of the magnetic properties one can carry out the consolidation study under the existence of the external magnetic field. In our experiments the samples were pre-aligned under 1.6 T external magnetic field and the pre-aligned powder was compacted without field. Due to the setups we have it is only possible to carry out cold compaction studies under external magnetic field.
Our studies show that it is possible to achieve the goal of the REFREEPERMAG project (developing new generation novel materials for high performance permanent magnets with energy product 60 kJ/m3 < (BH)max < 160 kJ/m3 without any rare earth or platinum). Our studies show that highest achievable (BH)max value for consolidated magnet is 64.9 kJ/m3 after hot compaction.
WP6: Towards Applications
The aim of this work package is to demonstrate the feasibility of our approaches in making prototype magnets that can directly be tested by the industry and outperform existing families of PMs in terms of the price/performance ratio. In addition we will explore the possibilities to extend the knowledge developed for thin/thick films during the project towards “emulation” of stress-strain in core-shell nanoparticles and eventually in bulk metallurgy, which will lead to large-scale applications.
From some of the candidate materials, for example MnGa, MnAl, (Fe0.65Co0.3Re0.05)2B (Fe0.4Co0.6)32B and MnBi, which have large enough magnetization, Curie temperature and magnetocrystalline anisotropy energy (MAE) we have selected and mastered after the month 24rth the MnBi system.
Fig. 42 Various candidate structures of Mn-Based alloys
Detailed description
The characteristic properties of MnBi materials are ( Ames)
Saturation Magnetization (RT) 60-75 Am2/kg
Hc ~ 2T
(BH)max (theoretical) 17.6 MGOe
(BH)max (Our MAGNETS) <5 MGOe
Magnetocrystaline Anisotropy 1.6 x 106 Jm-3
Operating limit 473 K
The magnets already used by WCM are of the NdFeB bonded and the dimensions are shown in Fig.44
We have prepared a mold made from special steel using wire cutting in order to use for magnets casting e.g insert powder, alligh in magnetic field and compress it , properly remove it and sinter in inert atmosphere at ~ 260-270 for 40-60 minutes.
a) Precursor alloy melting ( 5-10 % more Mn) in the alloy Mn-Bi
b) Powder production using ball-milling for about 50-6- minutes
c) Compression in a magnetic field of ~ 1 T
d) Anneal at lower temperature 250-270 C
We prepared various batches of alloys and characterize them with VSM and SQUID magnetometry and data are given in the following figures.
Fig. 43 The Optimized magnets showing uniformity and their characteristic properties.
Doing so we have obtained sintered magnets which are uniform and have magnetization of 0.4 T and coercivity ~ 0,5 KOe, as shown in Fig. 6.6
Fig. 44 MnBi and NdFe B magnets glued on the test-motor
Operation and evaluation of its performance. Comparison with the state-of-the-art micromotor.
The magnets are bonded in the rotor with alternating direction of magnetization as shown in Fig. 44
Fig. 45 Schematic of a 8-pole motor with PM mounted on the rotor different magnetization orientation.
We run the motor at ~ 700 and 1000 rpm and the results are given below
We obtained excellent performance , with no extra resonances.
The measured output is:
MnBi
N= 737 rpm V=0,738 V N= 1000 rpm V= 1,001 V
NdFeB
N= 667 V= 2,717 V N=1000 rpm V=4,073 V
This output scales rougly to the (BH) max of MnBi vs the NdFeB magnets used.
Work is in progress to fully characterize the motor performance, measuring torque, stability etc.
We are one of the first groups in the word to use MnBi –based permanent magnets in a 8-pole motor susccesfully. If we can improve the fabrication process and increase the energy product to the one obtained by other groups (10 MGOe) and make an estimation of the cost of the process it might be an alternative non-rare earth permanent magnet- for some applications.
Highlight clearly significant results;
MnBi PM were synthesized, tested and bonded on a rotor with alternating direction of magnetization
We are one of the few labs in the world to test these magnets in a motor with accepted performance, a statement on the use of resources, in particular highlighting and explaining deviations between actual and planned person-months per work package and per beneficiary in Annex 1 (Description of Work);
We have already informed the consortium that NCSR D will undertake the work of magnet preparation, characterization, finishing, bonding atc and WCM will perform in their testing labs the performance, shifting their 6-PM to NCSR D.
WP7: Life Cycle Impact Assessment (LCIA)
WP7 deals with the evaluation of the environmental impact caused by the nanomaterials used in the fabrication of novel PMs developed in this project. To achieve this main goal, the following specific objectives are proposed:
To design a new framework for the life cycle impact assessments of nanomaterials based on the existing model like USEtox or ReCiPe.
To perform the LCIA analysis of nanomaterials following the framework previously described
To develop the correct algorithms for the impact analysis of the nanomaterials in the LCIA analysis
The LCA is aimed to integrate results generated in REFREEPERMAG on toxicity of new FeCo-X nanostructured materials and their functionality in new magnet compacts developed in order to have a global knowledge of the environmental performance of these products in all life cycle stages. The main objectives of the study were to identify which processes pose greater impacts, how these impacts are distributed along the different life cycle stages, and their potential impact on environment and human health. The characteristics of the system defined for FeCo-X nanomaterials are detailed, defining the scope and boundaries of the system, the functional unit, modelling rules and hypothesis. Preliminary information available for inventory is presented in this deliverable as well.
Goal and scope definition
The first phase of the LCA study is the goal and scope definition, which defines the general context for the study. In the goal definition, parameters such as the functional unit, the intended application, the reasons for carrying out the study, the target audience or the limitations and assumptions were identified.
Inventory analysis
In the inventory data, each life stage was analysed to determine the relevant inputs and outputs of the system. This information was gathered from REFREEPERMAG partners, literature and also from other projects related primary data .
Impact assessment
The impact assessment calculated the environmental potential impacts associated with inventory flows. The base methodology chosen for the present study was the ReCiPe Mid/Endpoint method. This methodology combines both midpoints and endpoints in a consistent framework (cause-effect chain) which allow a step-by-step interpretation and revision of the results.
Interpretation of results
In the final step of the LCA, the interpretation of results, a critical revision of the results was done in order to verify its reliability. In this step the completeness, sensitivity and consistency of data gathered and results obtained were done in order to guarantee their representativeness and suitability to be incorporated in created processes for datasets and impact assessment methods.
The methodology and steps followed are given in Fig. 46.
Figure 46 . Methodology and steps for LCA studies according to ISO:14044:2006 should be prioritized and secondary data from databases and literature will be used when needed.
The Preliminary scope of the study on FeCo-X based materials systems is given in Fig, 47
Figure 47 . Preliminary scope of FeCo-X systems
Description of the studied system
The system selected for the study is based on the synthesis, production, manufacturing, use and end of life of nanomaterials (FeCo-X, nanowires and nanofilms) for magnetic compact, as presented in Figure 3 and Table 1.
The first life stage considered is the synthesis of nanomaterials. Different partners have synthesised FeCo-X nanomaterials using different methods. The methods considered in this study require extraction of raw material and manufacturing of Fe, Co, and X, as well as pyrol synthesis and electrochemical synthesis of nanostructures, sputtering of FeCo-Ni and ball milling.
Once the FeCo-X nanowires and films are synthesized, bulk structured nanomaterials in pellet forms can be obtained and transported in order to produce prototype modules, such as permanent magnets for sensors.
In the use stage not important environmental impacts are expected to occur.
For the end-of-life phase, different scenarios were defined according to the prototype characteristics that will determine the recyclability of the piece once the useful life will be over. Current European Statistics on waste treatment for End-of-Life vehicles will be used to define the possible waste treatments to be assessed.
The general allocation criterion is the mass parameter, which means that the reference flow related to the functional unit is taken to allocate environmental loads in each stage. For the allocation of environment impact the product along all life stages, the following rules have been determined.
In synthesis, modification and structuring of nanomaterials, all the inputs and outputs needed to process nanomaterials have been included.
For manufacturing, all inputs and outputs have been taken into account.
In the stage of use, not important environmental impacts are expected to occur..
Fig. 48. System flowchart for Applications
Table 7. Life Cycle Stages and information required for nanomaterials
Life Cycle Stage Information Needed Primary data sources information from Partners Secondary data sources
Nanomaterials Synthesis processing
Nanomaterials Structure Processing Processes description
Inputs Outputs Data provided by manufacturers partners
Previous studies (literature), LCI databases
Prototype module manufacturing • Processes description
• Inputs and outputs Data provided by manufacturers partners
Previous studies (literature), LCI databases
Functional module-based system manufacturing • Module description (parts, materials)
• Processes description
• Inputs and outputs Partners Involved Previous studies (literature), LCI databases
During module-based system use (Sensor application and Motor applications) • Performance during use
• Energy recuperation/saving
• Maintenance/Substitution parts
• Inputs needed
• Outputs generated
• Useful life duration
Partners Involved Previous studies (literature), LCI databases
Disposal processes • Waste generated (types, weight, treatment) Data provided by REFREEPERMAG Previous studies (literature), LCI databases
For recycling processes, only the waste treatment of the nanomaterials will be considered. At the end-of-life stage, impacts generated during the treatment of materials will be allocated into the system until the moment where the nanomaterials reach the state of end-of-waste.
Human toxicity of nanomaterials along their life cycle
The human toxicity of the FeCo-X nanomaterials can be assumed to be similar to that of its main constituents: cobalt and and iron. In addition, in the case of the presence of nickel this can also contribute to their toxicity, in particular for people sensitized to this element. The manipulation nanomaterials should be done taking the necessary protective measures to reduce inhalation and the subsequent risk of lung diseases, respiratory sensitization and carcinogenicity. Similarly, the frequent manipulation of FeCo-X nanomaterials should be done with gloves to minimize contact with cobalt and reduce the risk of sensitization. Sn or Sb nanocoatings also in small amounts have the same effect as their metals.
Among the life cycle steps evaluated for these materials, the manipulation of the nanomaterials prior to the compactation step and also the recycling phase after their use, are the most important phases in terms of toxicological risk and appropriate measures to minimize worker exposure should be taken.
Ecotoxicity of nanomaterials along their life cycle
To achieve the entire comprehension of the impact of new materials in the environment, the full set of impact categories to be considered at the midpoint level. At the endpoint level, three major endpoint categories were considered:
damage to human health (HH),
damage to ecosystem diversity (ED),
damage to resource availability (RA).
Table 3. Overview of the endpoint categories, indicators and characterization factors.
Area of protection Impact category Indicator
Name Abr Name Unit
Human Health Damage to Health HH Disability-adjusted loss of life years DALY
Ecosystems Damage to ecosystem diversity ED Loss of species during a year Species yr
Resources Damage to resource availability RA increased cost $
Man-made environment NA - NA -
Conclusions and discussions
The ecotoxicity of the FeCo-X nanowires, films and sintered PMs can be assumed to be similar to that of its main constituents: cobalt, iron, antimony, boron, tin etc. The manipulation of the nanowire/nanorods and to a lesser extend the thin films/multilayers should be done with the necessary protective meaures to reduce inhalation and the subsequent risk of lung deseases, respiratory sensitization and carcinogenicity. For the finished products the risk is lower and the same preacutions should be taken. Coating of PMs can prevent some contamination and it is strongly recommended.
WP8: Management
We used our previous experience in managing EU and National projects to meet the challenges for managing such an excellent consortium and did not faced any problem.
WP9: Scientific Coordination
All the meetings were well organized and minutes were collected
WP10: Dissemination and exploitation plan
A solid exploitation plan and dissemination has been developed
The potential impact (including the socio-economic impact and the wider societal implications of the project so far) and the main dissemination activities and exploitation of results (not exceeding 10 pages).
Exploitation
On completion of the project, the following main deliverables deserve special attention for exploitation
System Software
New improved alloys
New high performance grades of finished PM’s
New processes or materials (e.g. nanocomposites) for production of PM’s.
Novel high-throughput techniques to search for novel magnetic materials with the appropriate properties for PM
System Software
The partners TUW, UPPSALA, CSIC, INSAT and CEA have developed software packages that are suitable for exploitation.
The software packages used are WIEN2k, Elk, muffin-tin orbitals (EMTO) pseudopotential approaches with plane waves (VASP) or numerical local orbitals (SIESTA, OpenMX) for DFT calculations. Most of these codes provide MPI-parallel versions suitable for large-scale computations. MAGPAR, FEMME will be used for micromagnetics together with finite element discretisation codes (GID, Salome).
UPPSALA is concentrated on combinatorial calculations using standard codes as modules and had successfully TUW, CSIC and INSAT and CEA for micromagnetic simulations. Their software packages present an excellent opportunity for exploitation, but also as an important issue and strength of the partner in future submissions to the EU.
New improved alloys (NCSR D, INSAT, RUB, IFW, TUDA)
Chemical and electrochemical approach for the preparation of nanorods, nanowires of the type Co, Fe-Co-X
Bulk synthesis (melt spinning, ball milling, arc-melting, compaction)
New high performance grades of finished PM’s
Two classes of finished PMs were developed along the duration of the REFREEPERMAG project.
Co-based nanowire-nanorod PMs by blending and compaction
MnBi- based alloys by cold compaction and sintering
Co-based nanowire-nanorod PMs by blending and compaction
We have achieved the benchmark set in the proposal of (BH)max = 160 KJ/m3 in compacted Co-nanorods with an aspect ratio of ~ 5. Magnets have been prepared by INSAT by blending with epoxy and by the compaction methods developed by TUDA.
We are among the first groups worldwide to prepare and consolidate nanowires-nanorods into macroscopic permanent magnets with energy so high energy products, indicating that our consortium had achieved one of its main objectives.
MnBi- based alloys by cold compaction and sintering
We have succeeded to fabricate Bulk PMs with very good properties and energy product in sintered (NCSR D) and epoxy bonded magnets (MagnetFabrick). The energy product for the sintered magnet was ~ 40 KJ/m3 and the epoxy bonded ~ 5 KJ/m3. A degradation of the original magnetic properties was observed during sintering, but the final PM was good enough to be used for incorporation in a micromotor (WCM) and run a test performance.
The test performance has given the following results compared to the currently used bonded NdFeB magnets
MnBi
N= 737 rpm Vout =0,738 V N= 1000 rpm V out = 1,001 V
NdFeB
N= 667 V out = 2,717 V N=1000 rpm V out =4,073 V
Given that the output Voltage of the motor is proportional to the flux Φ = (Β.S) and from the measured values of both original NdFeB and our MnBi the voltage output of our- non optimized MnBi magnets- scales as ¼ of the voltage output of the NdFeB magnets ( 1.1 T is the remanence magnetization for NdFeB compared to the ~ 0,35 T for our sintered magnets). There is room for improvement of the sintering process of the MnBi PMs in order to preserve the very good properties of the powder ( Br ~ 60 emu/gr and Hc ~ 1 T).
Worth to mention here, as far as we know, we are the first consortium in the world to fabricate and test in “real conditions” MnBi based PMs with very promising results.
Economics of PMs
Co-based nanowire-nanorod PMs by blending and compaction
There are three preparation routes for the polyol mediated synthesis of cobalt nanorods:
1. Microwave oven – which offers the possibility to heat quickly a large volume of polyol and experiment within a big range of heating rates. High heating rates favors the formation of cobalt nanorods with small diameter, presenting the best magnetic properties. However, the strong drawback of the microwave is the heterogeneous heating that creates a metallic mirror on the flasks of the spherical glassware. The particles formed in the bulk of the suspension were found homogeneous with interesting magnetic properties, but the particles formed on the walls exhibited poor magnetic properties.
2. Heating mantle – which offers a better homogeneity of temperature inside the flask leading to
monodispersed cobalt nanorods. This set-up provided nanorods with the best magnetic
properties. Its limitation is the narrow range of the heating rate that can’t exceed 10 °C/min.
3. Jacket reactor – is expected to combine all the advantages of the above: a very homogeneous
temperature inside the solution, a good stirring and a high heating rate. The efficiency,
scalability and reproducibility of the first two approaches were examined.
These three set-ups are used at industrial scale to produce chemicals and including inorganic particles. In the table 1 the prices of the different chemicals used in the different steps of the “standard” synthesis of cobalt nanorods at the laboratory scale are given. Of course these prices are much higher than that expected for an industrial supply. Savings can be made on the cobalt precursor. The cobalt acetate is clearly not the cheapest precursor. Nevertheless it is noteworthy that 80% of the price comes from the butanediol.
Chemical reactions
The synthesis of the cobalt nanorods involves three steps:
1. The synthesis of the sodium laurate from lauric acid and sodium hydroxide
2. The cobalt laurate synthesis from cobalt chloride and cobalt acetate
3. The cobalt nanorods synthesis from the reduction of the cobalt laurate in a sodium hydroxide
solution in 1,2 butanediol containing 2.5 % of ruthenium chloride.
In the table the prices of the different chemicals used in the different steps of the “standard” synthesis of cobalt nanorods at the laboratory scale are summarized. Of course these prices are much higher than that expected for an industrial supply. Savings can be made on the cobalt precursor. The cobalt acetate is clearly not the cheapest precursor. Nevertheless it is noteworthy that 80% of the price comes from the butanediol.
The concentration of cobalt in the “standard” synthesis is 0.08 mol.L-1 that allows the synthesis of 5 g in one batch. Current price excluding personnel cost is 7€/g. Can be reduced to 2.5 €/g.
Several syntheses were done in the microwave oven with higher cobalt concentrations: 0.12 M and 0.24M. In every case the anhydrous ruthenium chloride was used as nucleating agent. These concentrations also showed that it is possible to prepare anisotropic cobalt particles similar to those prepared with the standard protocol at reduced cost.
In conclusion nanoparticles of Co- cost very much and we are in doubt that this technique will be used for mass production of permanent magnets (see tables below). Only niche markets are suitable for exploitation.
b) 1. MnBi sintered- based PMs
In order to evaluate the possibility for exploitation of PMs based on MnBi- alloys we show a recent study in ROMEO project funded by EU by P. McGuinness et al, considering that from the producer and consumer point of view, the cost of a permanent magnet with certain magnetic properties is one of the most important factors.
Their study is based on the cost of raw materials as of 2014 for alloy production by different methods for the most common and the best performing PMs based on rare-Earths in €/Kg Fig.1 for comparison, and in the following figure the cost in €/Kg for the emerging substitutes . Among them is MnBi at a cost of 13 €/Kg.
Fig.49. Raw material prices of several RE-based PMs alloys (after P. McGuinness, 2015)
Fig. 50 RE-Free alloys estimated cost for prices No 2013-April 2014 (after P. McGuinness, 2015).
In red are the prices for MnBi and for Co-nanowires ( 7500 €/Kg and using different solvent (to replace the 1,2 butanediol by the 1,2 propanediol.) according to the INSAT in the framework of the REFREEPERMAG project.
Considering the performance of the MnBi magnets, WCM feels that improvement of performance can position MnBi next to Ferrite magnets at cost 3-20 €/Kg and similar characteristics.
2. MnBi-Bonded Magnets ( MFB)
Magnetic powders of MnBi were mixed with epoxy to achieve as high density as possible and were measured at MagnetFabrick . Typical curves obtained are shown in Fig. 10.3
Fig. 51 Typical BH-loops of MnBi-bonded magnets
In the table below the characteristics of various bonded magnets are given
Density
g/cm3 Br
kA/m Hcb
kA/m Hcj
kA/m (BH)max
kJ/m3
MnBi-22 6,898 0,151 100,7 539,7 3,83
MnBi-11-b3 6,690 0,163 108,5 884,9 4,17
MnBi-11 6,584 0,133 91,88 841,1 3,06
MnBi-p06 6,288 0,133 96,87 777 3,22
Result:
Polymer bonded magnets have been prepared from different powders delivered by the consortium partners. As an interesting shape for PM motors and acuators a disc shaped magnet have been choosen with diameter 12mm in different length. Demagnetisation curves have been measured on the bonded magnets with a permagraph from Magnetphysik Steingroever. All samples prepared showed permanent magnetic properties in accordance with the expectations from the powder properties.
Summary:
It could be shown that the production of a near net shape magnet is feasible with the powders coming from different production routes. The magnet production performed can be scaled up for a series process. Powder quantities received from partners were sufficient to produce epoxy bonded magnets. As expected from the beginning of the project the powder preparation possibilities of our partners could not supply sufficient quantities to produce magntes based on thermoplastic binders as our preproduction of compounds would require some Kg masses of powders.
However, it has been deduced from the tests and produced magnets based on thermosettings in a compaction molding route that also a injection molding process will be applicable.
Novel high-throughput techniques to search for novel magnetic materials with the appropriate properties for PM
Combinatorial and high-throughput (CHT) Materials Science and Engineering techniques represent a powerful approach to rapidly screen new compounds and accelerate materials development. CHT methods are based on: i) materials modelling using novel computational tools ii) fabrication of materials libraries (multi-element compositional spreads), iii) high-throughput and spatially resolved characterization techniques, and iv) data management and analysis for processing high volume data.
CHT methods reduce the cost and time of materials Research & Development, effectively addressing the pressing issues arising on Materials Science and Engineering. A major societal challenge with global implications, climate change, demands the development of clean energy and energy efficiency technologies. As emphasized by the U.S. Department of Energy , scientific breakthroughs in new materials will foster technological advancements providing sustainable energy solutions. CHT methods ensure accelerated materials development for coping with present-day environmental issues.
High-Throughput techniques were employed to exploit strain-induced epitaxial growth of distorted cubic to tetragonal high magnetization alloys, that resulted in the high anisotropy which is required for the materialization of novel permanent magnets e.g. Fe-Co based and Heusler alloys, leading to the identification and characterization of new magnetic multifunctional materials of scientific and technological interest.
IFW and RUB used high-throughput synthesis on various substrates following the Bain path and realizing a lateral variation of in-plane lattice parameters using combinatorial film deposition of epitaxial Cu-Au on a 4-inch Si wafer. This template gave the possibility to adjust the in-plane lattice parameter over a wide range from 0.365 nm up to 0.382 nm. Cu-Au buffer was used to stain and assist the epitaxial growth of FeCo magnetic layer. High-throughput X-ray diffraction (XRD) measurements were performed to determine the structure of the Cu-Au Bain library.
RUB used well developed combinatorial methodologies in order to fabricate and study strained wedge type Fe-Co on AuxCu1-x libraries. For the magnetic characterization of the wedge type FeCo libraries, a high-throughput magneto-optical Kerr magnetometer (MOKE – RUB facilities) was used. Rapid characterization techniques, which can automatically measure the whole material library for the specific properties of interest, where also used for the structural properties study. A third or fourth element addition was studied by co-sputtering along with FeCo, B, Cr or Cr-Ni. RUB also prepared wedge type layers of Fe-Co-Nb using a post-annealing procedure, resulting to a ternary library.
NCSR D employed combinatorial thin film synthesis in order to create libraries of Cu-Au buffer alloy, as well as of {Fe-Co}-C magnetic layer. In the first case a both tetragonal and cubic Cu-Au crystallographic phases were studied on a single 4 inch wafer using high-throughput structural characterization technique. Moreover a wide range of stoichiometries of Fe-Co alloy were produced and studied, along with the variation of Carbon percentage effect (up to 20% at.). For rapid high-throughput magnetic characterization, a 3d magnetic moment mapping system was developed.
High-Throughput combinatorial method was also used by NCSR D to screen possible compositions and synthesis conditions of Heusler alloys of the type X2YZ. Fe2CuGa was deposited on various substrates at different deposition temperature. A great amount of variable alloy stoichiometries were produced and studied structurally using rapid characterization combinatorial XRD on a Rigaku system. On a 2 inch Si wafer different crystallographic orientations were observed in one step procedure.
Our project is one of the very few in the word to use this approach for screening a) Theoretically and b) experimentally the search for new magnetic materials for permanent magnets and is well cited in meetings all-over the world.
After 36 months (the completion of the project), the exploitation activities will be continued aiming at manufacturing & marketing of the products in a cost-effective way. Support from Eureka, CRAFT and local schemes will be considered, encouraging a bottom-up approach to technological development and strengthening competitive position of project companies on the world market. SMEs would also benefit from loans from the European Investment Bank (EIB) or national/regional sources may be exploited for further commercialisation, in addition to the partners’ own investments. Such strategy of financial engineering is expected to ensure the sustainability of the activities and further improvement of the cooperation in the European Research Area.
Potential Impact:
The potential impact (including the socio-economic impact and the wider societal implications of the project so far) and the main dissemination activities and exploitation of results (not exceeding 10 pages).
Exploitation
On completion of the project, the following main deliverables deserve special attention for exploitation
• System Software
• New improved alloys
• New high performance grades of finished PM’s
• New processes or materials (e.g. nanocomposites) for production of PM’s.
• Novel high-throughput techniques to search for novel magnetic materials with the appropriate properties for PM
→ System Software
The partners TUW, UPPSALA, CSIC, INSAT and CEA have developed software packages that are suitable for exploitation.
The software packages used are WIEN2k, Elk, muffin-tin orbitals (EMTO) pseudopotential approaches with plane waves (VASP) or numerical local orbitals (SIESTA, OpenMX) for DFT calculations. Most of these codes provide MPI-parallel versions suitable for large-scale computations. MAGPAR, FEMME will be used for micromagnetics together with finite element discretisation codes (GID, Salome).
UPPSALA is concentrated on combinatorial calculations using standard codes as modules and had successfully TUW, CSIC and INSAT and CEA for micromagnetic simulations. Their software packages present an excellent opportunity for exploitation, but also as an important issue and strength of the partner in future submissions to the EU.
→ New improved alloys (NCSR D, INSAT, RUB, IFW, TUDA)
a) Chemical and electrochemical approach for the preparation of nanorods, nanowires of the type Co, Fe-Co-X
b) Bulk synthesis (melt spinning, ball milling, arc-melting, compaction)
→ New high performance grades of finished PM’s
Two classes of finished PMs were developed along the duration of the REFREEPERMAG projec
• Co-based nanowire-nanorod PMs by blending and compaction
• MnBi- based alloys by cold compaction and sintering
• Co-based nanowire-nanorod PMs by blending and compaction
We have achieved the benchmark set in the proposal of (BH)max = 160 KJ/m3 in compacted Co-nanorods with an aspect ratio of ~ 5. Magnets have been prepared by INSAT by blending with epoxy and by the compaction methods developed by TUDA.
We are among the first groups worldwide to prepare and consolidate nanowires-nanorods into macroscopic permanent magnets with energy so high energy products, indicating that our consortium had achieved one of its main objectives.
• MnBi- based alloys by cold compaction and sintering
We have succeeded to fabricate Bulk PMs with very good properties and energy product in sintered (NCSR D) and epoxy bonded magnets (MagnetFabrick). The energy product for the sintered magnet was ~ 40 KJ/m3 and the epoxy bonded ~ 5 KJ/m3. A degradation of the original magnetic properties was observed during sintering, but the final PM was good enough to be used for incorporation in a micromotor (WCM) and run a test performance.
The test performance has given the following results compared to the currently used bonded NdFeB magnets
MnBi
N= 737 rpm Vout =0,738 V N= 1000 rpm V out = 1,001 V
NdFeB
N= 667 V out = 2,717 V N=1000 rpm V out =4,073 V
Given that the output Voltage of the motor is proportional to the flux Φ = (Β.S) and from the measured values of both original NdFeB and our MnBi the voltage output of our- non optimized MnBi magnets- scales as ¼ of the voltage output of the NdFeB magnets ( 1.1 T is the remanence magnetization for NdFeB compared to the ~ 0,35 T for our sintered magnets). There is room for improvement of the sintering process of the MnBi PMs in order to preserve the very good properties of the powder ( Br ~ 60 emu/gr and Hc ~ 1 T).
Worth to mention here, as far as we know, we are the first consortium in the world to fabricate and test in “real conditions” MnBi based PMs with very promising results.
Economics of PMs
a) Co-based nanowire-nanorod PMs by blending and compaction
There are three preparation routes for the polyol mediated synthesis of cobalt nanorods:
1. Microwave oven – which offers the possibility to heat quickly a large volume of polyol and experiment within a big range of heating rates. High heating rates favors the formation of cobalt nanorods with small diameter, presenting the best magnetic properties. However, the strong drawback of the microwave is the heterogeneous heating that creates a metallic mirror on the flasks of the spherical glassware. The particles formed in the bulk of the suspension were found homogeneous with interesting magnetic properties, but the particles formed on the walls exhibited poor magnetic properties.
2. Heating mantle – which offers a better homogeneity of temperature inside the flask leading to
monodispersed cobalt nanorods. This set-up provided nanorods with the best magnetic
properties. Its limitation is the narrow range of the heating rate that can’t exceed 10 °C/min.
3. Jacket reactor – is expected to combine all the advantages of the above: a very homogeneous
temperature inside the solution, a good stirring and a high heating rate. The efficiency,
scalability and reproducibility of the first two approaches were examined.
These three set-ups are used at industrial scale to produce chemicals and including inorganic particles. In the table 1 the prices of the different chemicals used in the different steps of the “standard” synthesis of cobalt nanorods at the laboratory scale are given. Of course these prices are much higher than that expected for an industrial supply. Savings can be made on the cobalt precursor. The cobalt acetate is clearly not the cheapest precursor. Nevertheless it is noteworthy that 80% of the price comes from the butanediol.
Chemical reactions
The synthesis of the cobalt nanorods involves three steps:
1. The synthesis of the sodium laurate from lauric acid and sodium hydroxide
2. The cobalt laurate synthesis from cobalt chloride and cobalt acetate
3. The cobalt nanorods synthesis from the reduction of the cobalt laurate in a sodium hydroxide
solution in 1,2 butanediol containing 2.5 % of ruthenium chloride.
In the table the prices of the different chemicals used in the different steps of the “standard” synthesis of cobalt nanorods at the laboratory scale are summarized. Of course these prices are much higher than that expected for an industrial supply. Savings can be made on the cobalt precursor. The cobalt acetate is clearly not the cheapest precursor. Nevertheless it is noteworthy that 80% of the price comes from the butanediol.
The concentration of cobalt in the “standard” synthesis is 0.08 mol.L-1 that allows the synthesis of 5 g in one batch. Current price excluding personnel cost is 7€/g. Can be reduced to 2.5 €/g.
Several syntheses were done in the microwave oven with higher cobalt concentrations: 0.12 M and 0.24M. In every case the anhydrous ruthenium chloride was used as nucleating agent. These concentrations also showed that it is possible to prepare anisotropic cobalt particles similar to those prepared with the standard protocol at reduced cost.
In conclusion nanoparticles of Co- cost very much and we are in doubt that this technique will be used for mass production of permanent magnets (see tables below). Only niche markets are suitable for exploitation.
b) 1. MnBi sintered- based PMs
In order to evaluate the possibility for exploitation of PMs based on MnBi- alloys we show a recent study in ROMEO project funded by EU by P. McGuinness et al, considering that from the producer and consumer point of view, the cost of a permanent magnet with certain magnetic properties is one of the most important factors.
Their study is based on the cost of raw materials as of 2014 for alloy production by different methods for the most common and the best performing PMs based on rare-Earths in €/Kg Fig.1 for comparison, and in the following figure the cost in €/Kg for the emerging substitutes . Among them is MnBi at a cost of 13 €/Kg.
Fig.49. Raw material prices of several RE-based PMs alloys (after P. McGuinness, 2015)
Fig. 50 RE-Free alloys estimated cost for prices No 2013-April 2014 (after P. McGuinness, 2015).
In red are the prices for MnBi and for Co-nanowires ( 7500 €/Kg and using different solvent (to replace the 1,2 butanediol by the 1,2 propanediol.) according to the INSAT in the framework of the REFREEPERMAG project.
Considering the performance of the MnBi magnets, WCM feels that improvement of performance can position MnBi next to Ferrite magnets at cost 3-20 €/Kg and similar characteristics.
2. MnBi-Bonded Magnets ( MFB)
Magnetic powders of MnBi were mixed with epoxy to achieve as high density as possible and were measured at MagnetFabrick . Typical curves obtained are shown in Fig. 10.3
Fig. 51 Typical BH-loops of MnBi-bonded magnets
In the table below the characteristics of various bonded magnets are given
Density
g/cm3 Br
kA/m Hcb
kA/m Hcj
kA/m (BH)max
kJ/m3
MnBi-22 6,898 0,151 100,7 539,7 3,83
MnBi-11-b3 6,690 0,163 108,5 884,9 4,17
MnBi-11 6,584 0,133 91,88 841,1 3,06
MnBi-p06 6,288 0,133 96,87 777 3,22
Result:
Polymer bonded magnets have been prepared from different powders delivered by the consortium partners. As an interesting shape for PM motors and acuators a disc shaped magnet have been choosen with diameter 12mm in different length. Demagnetisation curves have been measured on the bonded magnets with a permagraph from Magnetphysik Steingroever. All samples prepared showed permanent magnetic properties in accordance with the expectations from the powder properties.
Summary:
It could be shown that the production of a near net shape magnet is feasible with the powders coming from different production routes. The magnet production performed can be scaled up for a series process. Powder quantities received from partners were sufficient to produce epoxy bonded magnets. As expected from the beginning of the project the powder preparation possibilities of our partners could not supply sufficient quantities to produce magntes based on thermoplastic binders as our preproduction of compounds would require some Kg masses of powders.
However, it has been deduced from the tests and produced magnets based on thermosettings in a compaction molding route that also a injection molding process will be applicable.
• Novel high-throughput techniques to search for novel magnetic materials with the appropriate properties for PM
Combinatorial and high-throughput (CHT) Materials Science and Engineering techniques represent a powerful approach to rapidly screen new compounds and accelerate materials development. CHT methods are based on: i) materials modelling using novel computational tools ii) fabrication of materials libraries (multi-element compositional spreads), iii) high-throughput and spatially resolved characterization techniques, and iv) data management and analysis for processing high volume data.
CHT methods reduce the cost and time of materials Research & Development, effectively addressing the pressing issues arising on Materials Science and Engineering. A major societal challenge with global implications, climate change, demands the development of clean energy and energy efficiency technologies. As emphasized by the U.S. Department of Energy , scientific breakthroughs in new materials will foster technological advancements providing sustainable energy solutions. CHT methods ensure accelerated materials development for coping with present-day environmental issues.
High-Throughput techniques were employed to exploit strain-induced epitaxial growth of distorted cubic to tetragonal high magnetization alloys, that resulted in the high anisotropy which is required for the materialization of novel permanent magnets e.g. Fe-Co based and Heusler alloys, leading to the identification and characterization of new magnetic multifunctional materials of scientific and technological interest.
IFW and RUB used high-throughput synthesis on various substrates following the Bain path and realizing a lateral variation of in-plane lattice parameters using combinatorial film deposition of epitaxial Cu-Au on a 4-inch Si wafer. This template gave the possibility to adjust the in-plane lattice parameter over a wide range from 0.365 nm up to 0.382 nm. Cu-Au buffer was used to stain and assist the epitaxial growth of FeCo magnetic layer. High-throughput X-ray diffraction (XRD) measurements were performed to determine the structure of the Cu-Au Bain library.
RUB used well developed combinatorial methodologies in order to fabricate and study strained wedge type Fe-Co on AuxCu1-x libraries. For the magnetic characterization of the wedge type FeCo libraries, a high-throughput magneto-optical Kerr magnetometer (MOKE – RUB facilities) was used. Rapid characterization techniques, which can automatically measure the whole material library for the specific properties of interest, where also used for the structural properties study. A third or fourth element addition was studied by co-sputtering along with FeCo, B, Cr or Cr-Ni. RUB also prepared wedge type layers of Fe-Co-Nb using a post-annealing procedure, resulting to a ternary library.
NCSR D employed combinatorial thin film synthesis in order to create libraries of Cu-Au buffer alloy, as well as of {Fe-Co}-C magnetic layer. In the first case a both tetragonal and cubic Cu-Au crystallographic phases were studied on a single 4 inch wafer using high-throughput structural characterization technique. Moreover a wide range of stoichiometries of Fe-Co alloy were produced and studied, along with the variation of Carbon percentage effect (up to 20% at.). For rapid high-throughput magnetic characterization, a 3d magnetic moment mapping system was developed.
High-Throughput combinatorial method was also used by NCSR D to screen possible compositions and synthesis conditions of Heusler alloys of the type X2YZ. Fe2CuGa was deposited on various substrates at different deposition temperature. A great amount of variable alloy stoichiometries were produced and studied structurally using rapid characterization combinatorial XRD on a Rigaku system. On a 2 inch Si wafer different crystallographic orientations were observed in one step procedure.
Our project is one of the very few in the word to use this approach for screening a) Theoretically and b) experimentally the search for new magnetic materials for permanent magnets and is well cited in meetings all-over the world.
After 36 months (the completion of the project), the exploitation activities will be continued aiming at manufacturing & marketing of the products in a cost-effective way. Support from Eureka, CRAFT and local schemes will be considered, encouraging a bottom-up approach to technological development and strengthening competitive position of project companies on the world market. SMEs would also benefit from loans from the European Investment Bank (EIB) or national/regional sources may be exploited for further commercialisation, in addition to the partners’ own investments. Such strategy of financial engineering is expected to ensure the sustainability of the activities and further improvement of the cooperation in the European Research Area.
Socioeconomic Impact
In a nutshell we moved from technology readiness level 2 (TRL 2, see table) to TRL 4..
Table 1- Extract from Part 18 - Commission Decision C(2013)8631 : Technology Readiness Level
TRL1 basic principles observed
-----------------------------------------------------------------------------------------------------------------------------------
TRL2 technology concept formulated REFREEPERMAG starting point
TRL3 experimental proof of concept
TRL4 technology validated in lab REFREEPERMAG ending
TRL5 technology validated in relevant environment (industriallyrelevant environment in the case of key enabling technologies)
TRL6 technology demonstrated in relevant environment (industrially relevant environment in the case of key enabling technologies)
-----------------------------------------------------------------------------------------------------------------------------------
TRL7 system prototype demonstration in operational environment
TRL8 system complete and qualified
TRL9 actual system proven in operational environment (competitive
Manufacturing in the case of enabling technologies; or in space
New Ab-initio calculation scheme: REFREEPERMAG project delivered a series of ab-initio simulations and micromagnetic calculations for the discovery of novel PMs without rare earths. Further theREFREEPERMAG project has successfully demonstrated that the combinatorial approach is the most suitable for the rapid search of novel materials meeting predefined criteria.
New materials: REFREEPERMAG project produced a plethora of FeCo-X-Y materials in the form of nanowires/nanorods, thin films and bulk that possess the required properties as candidate PMs.
New Infrastructure for combinatorial and specialized measurements:
The project developed new Combinatorial and specialized methods with sufficient accuracy to differentiate between materials with different compositions , structure , intrinsic properties for the search of novel magnetic phases. Namely a novel Combi-Kerr, a Combi-XRD, a 3D-Magscan for magnetic field mapping, a ultrasensitive FMR, a SANS (Small Angle Newtron Scattering) for aligned nanowires/nanorods magnets and many more , that eventually can be exploited, paving the way of a fast high-trroughput characterization approach.
New commercial knowledge: Both within Europe and elsewhere a number of parties are searching for rare-earth free permanent magnets ( ARPA-E in USA, JAPAN coordinated by Dr. S. Hirosawa 10-year due to the economics involved wwith respect to energy savings, Electric vehicles and wind energy. We have demonstrated that there is a fast-pathway to explore the landscape for nonel magnetic materials for PMs. We have demonstrated that even nanowires/nanorods , through chemical synthesis and upscaling , can be commercialized for specific applications. The knowledge developed for nanowires can have spill-overs to other sectors like LEDs, since the next generation of LEDs will be based on nanowires of GaN-type and on quantum dots, because the consume less energy. The attempt to study in-depth the Mn-based systems and the fabrication of MnBi PMs , for demonstration only, showns that a commercial knowledge has been developed as well. Who will exploit this knowledge is a matter of further discussions.
Environmental impact: In the car industry there is actually a huge focus on reducing CO2 emissions. This focus is basically generated by new laws and the need of a responsible and sustainable use of energy sources. However, most of the measurements which are analyzed actually have a high cost, which avoids the introduction of such new technology. Also, the carbon footprint of every new technology has to be considered. The main measure which is taken in count today to quantify the relation between CO2 reduction and their costs i.e. Euros per gram of CO2. In reality, most of the measures are currently around 0 and 150€ per gram of CO2.
Considering that a 5 % improvement in (BH)max or equivalently a reduction in price of the raw materials compared to the currently used rare-earth based materials, will facilitate the replacement of existing magnets and will contribute , along with other technologies, to a better and cleaner environment.
The REFREEPERMAG project simply was essential in order to be able to quantify the potential of the technology and to identify the biggest issues, which have to be resolved in order to get a novel PMs .All partners intend to continue the exploitation of these new capabilities to further develop their business.
List of Websites:
The address of the project public website, if applicable as well as relevant contact details.
http://refreepermag-fp7.eu/
Furthermore, project logo, diagrams or photographs illustrating and promoting the work of the project (including videos, etc...) as well as the list of all beneficiaries with the corresponding contact names can be submitted without any restriction.
4.1 Final publishable summary report
There exist a number of families of PMs, each one possessing specific characteristics. The most powerful PMs are the ones based on rare-earths and cannot be superseded by any transition- metal alloys. With respect to the energy product the NdFeB magnets are by far the best, however with a relative high cost and limited maximum operational temperature of 200 0C. For higher operational temperatures and at a much higher cost the Sm-Co 1:5 and 2:17 families are available. Ferrites are the most widely used by a very wide margin due to their very low price. Fe based alloys containing (Al,Ni,Co) that is AlNiCo magnets find their uses in electronic instruments, sensors, loudspeakers, TWT amplifiers, etc.
Export quotas posed recently by the rare-earth producers, mainly China which is currently the source of more than 95 % of the world’s supply on rare-earths, has resulted in shortages and a dramatic prize increase of these strategic materials. The rare- earth shortage has stimulated intense effort towards the search for new non-rare-earth containing magnets either by modifying the Alnico and Ferrite type magnets or discovering new materials containing no rare-earths. The search is driven not only by scientific interest but mainly from the applications sector’s needs, especially for the electric car and the wind energy industries.
This project aimed and succeeded at developing a new generation of high-performance permanent magnets (PM) without rare-earths. Our approach was based on:
a) novel production of high-aspect-ratio (>5) nanostructures (nanowires, nanoparticles, nanorods, nanoflakes) by exploiting the magnetic shape anisotropy of the constituents that can be produced via chemical nanosynthesis -polyol process or electrodeposition- in Co-based novel magnetic phases, which were consolidated with novel compaction process for a new generation of rare-earth free permanent magnets with energy product in the range of (BH)max > 160 KJ/m3 at room temperature.
b) using a high-throughput thin film synthesis and high-through-put characterization approach we have identified novel magnetic phases of the type Fe-Co-X (X= C, B) with very high anisotropy in excess of 0.5 MJ/m3 and coercive field of ~ 1 KOe stabilized in a tetragonal or hexagonal structure by epitaxial growth on Au1-xCux combinatorial searched suitable substrate, under various conditions of pressure, stoichiometry and temperature, and
c) Mn-X based alloys (X=Ga, Bi) as bulk processed permanent magnets
To obtain these remarkable results we were based our approach on combinatorial modelling using state of the art approaches and software such as WIEN2k, Elk, muffin-tin orbitals (EMTO) pseudopotential approaches with plane waves (VASP) or numerical local orbitals (SIESTA, OpenMX) for DFT calculations.Most of these codes provide MPI-parallel versions suitable for large-scale computations. MAGPAR, FEMME will be used for micromagnetics together with finite element discretisation codes. We have combined modeling with unique facilities for preparation and characterization that existed in the partner’s labs, such chemical or electrochemical synthesis of nanoparticles and nanowires, combinatorial sputtering for thin films and multilayers, ultrasensitive FMR for anisotropy evaluation and consolidation techniques for bulk PMs. We have developed new tools for high-throughput characterization such as M-O Kerr effect, combi- XRD , combi 3D-scanner , polarization set-up for thermal neutrons and a procedure to correlate texture maps and microstructure.
In addition, due to the largely unknown environmental and toxicological impact of nanomaterials, a comprehensive Life Cycle Assessment was done for chemically synthesized nanorods and sputtered thin films and the magnets produced. The potential environmental impacts of these modules over their entire life were quantified.
This activity has resulted in 43 journal publication, 61 conference publications , 2 book chapters and 3-PhD and more than fifty invited-oral presentation in conferences, summer schools or special workshops. The developed concepts were well accepted by the scientific community and frequently cited.
In addition, we are among the pioneers in developing permanent magnets based on nanorods with very high energy product and among the first groups that tested Mn-Bi based magnets in a motor with very promising performance. A techno economical study revealed that the nanorods-based magnets are for niche markets due to the high cost although the Mn-Bi magnets with some improvement can compete with ferrites. No patents have been filled until the end of project
Project Context and Objectives:
A summary description of project context and objectives (not exceeding 4 pages).
Main Ideas: The scientific base or our proposal
Magnetic anisotropy is the directional dependence of a material's magnetic properties. In the absence of an applied magnetic field, a magnetically isotropic material has no preferential direction for its magnetic moment while a magnetically anisotropic material will align its moment along one of the easy axes. Magnetic anisotropy is a prerequisite for hysteresis in ferromagnets: without it a ferromagnet is magnetically soft. Magnetocrystalline anisotropy arises from quantum-mechanical interactions between electrons and the lattice. The spin-orbit interaction couples the spin degrees of freedom to the lattice and leads to the alignment of the magnetization along certain preferred crystallographic orientations. There are different sources of magnetic anisotropy:
• Magnetocrystalline anisotropy: the atomic structure of a crystal introduces preferential directions for the magnetization (due to spin-orbit interaction)
• Shape anisotropy: when a particle is not perfectly spherical, the demagnetizing field will not be equal for all directions, creating one or more easy axes (due to intrinsic dipolar interaction).
• Magnetoelastic anisotropy: stress may alter magnetic behavior, leading to magnetic anisotropy.
• Surface/interface anisotropy: the lower coordination at interfaces and its associated reduced symmetry lead to strong modification of the magnetocrystalline anisotropy
• Exchange anisotropy: appears at the interface of magnetic phases with different character (i.e. antiferromagnetic and ferromagnetic materials)
• Externally imposed (induced) anisotropy
The total magnetic anisotropy can be expressed as the sum of the various contributions
HTotal = Hcr + Hshape + H stress + Hexc +Hother (Eq. 1)
where, Hcr = h ( 2K1/ Ms), Hshape = h (Na -Nc) Ms , Hstress = h (3λsi σ/Ms), Hexc = h(-J12 s1s2)
and Na , Nc are the demagnetizing coefficients parallel to a and c axes of an “ellipsoid”, λsi the saturation magnetostriction, σ is the stress and K1 the crystal anisotropy constant. The Na and Nc are expressed as a function of the c/a ratio. Assuming a coherent rotation magnetization reversal mechanism, the coercivity for an assembly of elongated particles can be found from the relation
Ηc = (1-p) (Νb – Νa) Μs (Eq.2)
where p is the packing fraction and Na and Nb the demagnetization factors along and perpendicular to the length of the particles. Assuming an Ms somewhere in the range of 1700 - 1900 G (for Fe or Fe-Co-X alloys), Nb - Na = 2 and p ~ 0.4-0.6 an Hc of 4-6 kOe is expected with (BH)max ~ 240 KJ/m3. Having this in mind we propose to exploit this possibility to prepare a novel series of magnets covering the range 60
KJ/m3 < (BH)max < 160 KJ/m3, depending on optimization of packing, magnetization and intrinsic properties of the high-aspect-ratio nanostructures.
The best example of exploitation of the shape anisotropy is the AlNico series. AlNico magnets are alloys of Fe, Co, Ni, and Al with small amounts of other elements (Cu, Nb, Ta, and Ti) which are used to optimize coercivity at the expense of magnetization. These magnets develop their coercivity after a carefully controlled heat treatment which leads to precipitation hardening. However, in this case the coercivity is due to the shape anisotropy of {Fe-Co}-rich precipitates ( ~30 nm lateral dimension) instead of magnetocrystalline anisotropy as in the case of 2:17 magnets. We propose to develop a new type of nano-composite material for the fabrication of permanent magnets by a bottom-up approach. The new material will be based on assembled metallic nanostructures with high aspect ratio (nanowires, nanoparticles, nanorods, nanoflakes). We will use state-of-the-art techniques to fabricate high-aspect-ratio nanostructures and consolidate them after alignment into bulk permanent magnets. This is a disruptive approach for new permanent magnets relying on shape anisotropy and not an incremental improvement of AlNiCo magnets. We propose to develop the synthesis of metallic Fe-Co based nanoparticles with high aspect ratio by a chemical organometallic synthesis, gas phase methods and ball milling. The fine control of the chemical processes will allow to control the morphology of the nano-objects (aspect ratio – diameter – composition) with high precision and reproducibility. These approaches can potentially be scaled up to large volume production. Nanostructures combine several advantages compared to existing permanent magnet materials. The anisotropy will be an extrinsic property based on the shape anisotropy. This property is rather insensitive to temperature effects contrary to magneto-crystalline anisotropy. We think that it will be possible to reach high energy products. Our new composite material could be used as a high performance replacement for permanent magnets, which are operating at high temperature (T>200°C).
The scientific base- the Bain Path- for the search of new non rare-earth containing phases with large magnetocrystalline anisotropy is based on the previous success of growing phases –with phase instability- by epitaxy on selected substrates. Strain effects in functional materials are of great interest currently for modifying material properties. By strained epitaxial film growth, physical properties can be controlled and improved, e.g. in semiconductors, multiferroic materials, and ferromagnets . Moreover, there are even materials which exhibit functional properties such as ferroelectricity only in strained films. During epitaxial growth, the film orientation is determined by the substrate onto which the material is deposited and the temperature used. The thin film adapts its in-plane lattice parameters to the substrate, even if its equilibrium lattice parameters differ considerably. The particular case in which the lattice parameters of the film material are strained such that they are identical to those of the substrate is called strained coherent film growth. In common rigid metals, straining of the crystal lattice by coherent film growth requires substantial elastic energy. As a consequence, already at strains of a few percent, coherent growth is limited to ultrathin films with thicknesses of up to several atomic layers. For the growth of coherent epitaxial films with high strains and large thicknesses, soft materials should be used, as suggested by van der Merwe. Such softening is observed in crystals with lattice instabilities, e.g. in materials with a martensitic transformation. As an example of how to exploit such a martensitic instability, Godlevsky and Rabe predicted the possibility of inducing a cubic to tetragonal distortion with c/a ratios from 0.95 to 1.25 in the magnetic shape memory material Ni2MnGa. In fact, in experiments Dong et al. demonstrated a considerable epitaxial strain of 3% in a Ni-Mn-Ga film. This shows that for improved tunability of crystal lattice and functional properties in thicker films, it is advantageous to exploit alloy lattice instabilities
In particular, the possibility of stabilizing intermediate lattice geometries-fct- and varying smoothly between stable phases along Bain transformation paths offers new routes to modify and investigate structure-property relations in functional materials. To achieve this cumbersome and difficult task we will employ a novel approach, which offers flexibility, multiplicity and high speed towards a first identification of materials with the desired properties.
While discoveries in materials science have made a major impact they are mostly based on simple materials - either elemental as Si or binaries, which are synthesized in a traditional way, where experience and intuition are the drivers. Serial processing is the norm due to time/manpower constraints. However, the high-tech industries demand more and more tailor-made materials with a complex list of requirements. Most often only optimized materials having four to five components are able to fulfil the requirements, which increases the number of required experiments by orders of magnitude. The solution to these problems and limitations are the corner stones of this proposal.
The Combinatorial Technique (CT) or High-Throughput (HT) approach is a revolutionary step forward in the development of new materials. It involves the development and
application of new tools for systematic and parallel synthesis focussed on the characterization to the most industry relevant and innovative of multi-functional classes of magnetic materials. Automatic well- controlled production and characterization of binary, ternary and quaternary systems yielding thousands of new materials becomes possible and it is the most effective way to search for new phases. Further development in bulk form is beyond this proposal. The high-throughput synthesis (HTS) approach involves the parallel production of huge number of samples in form of (arrays) of samples, also
called material libraries. For example in an array of 10x10 samples it is possible to vary composition parameter A in the X direction in 10 steps, and composition or fabrication parameter B (e.g. composition, substrate temperature) by 10 steps in Y direction. This produces a ‘library’ of 100 different but very well- comparable samples with two independently varied production parameters within one single experiment. The next step is to have rapid high-quality characterization techniques, which can, in one operation, measure the whole material library for the specific properties of interest. These are termed ‘High Throughput Characterization’ (HTC) measurements. The final stage is the appropriate informatics systems - data handling (visualization) and mining to aid in the analysis of the large quantities of data produced.
Coupling this (HT)approach with the possibility of epitaxial multicomponent thin film growth at various temperatures and on various substrates, to stabilize new types of structures in an analogous way for bulk materials- e.g. spinodal or similar type- using (HTS) techniques will permit us to explore a wide range of promising high-magnetization/ high anisotropy/Curie temperature alloys a)of the type {Fe-Co}- X-Y (X=other 3d or 4d metals and Y=B,C,P or N) and b) of Heusler alloys with the general formula X2YZ (where X is usually Fe, Co, Ni, Cu; Y other transition metals, most often Mn; and Z a group-B element (Al, Ga, Ge, Sn...).
1.1.3 Scientific and technical objectives
Through an integration of professional skills (chemists, physicists, engineers, materials scientists) from academic and industrial groups, the activity will approach different aspects of modelling of permanent magnet materials and devices, materials preparation and characterization of their electric, magnetic and other relevant properties e.g. porosity, strength, thermal conductivity, core fabrication and device assembly, performance evaluation and optimization of materials and processing steps to exploitation plans. The know-how on the different aspects will be transferred from the lab size to commercial size products aiming at filling the gap between research laboratories and industrial production. A prerequisite for achieving the above technological goal is an improved understanding of magnetic phenomena at the nanoscale (in thin films, nanoparticles, nanowires, nanoflakes), which thus represents the intrinsic strategic baseline of the project.
The strategic objectives of the project are:
a) use of the shape anisotropy as the dominant mechanism to develop novel high-aspect-ratio nanostructures as building blocks to fabricate novel non-rare-earth containing permanent magnets using a bottom–up approach, and b) use HT techniques to exploit strain-induced epitaxial growth of distorted cubic to tetragonal high magnetization alloys that will result in the high anisotropy that is required for the production of novel permanent magnets e.g. Fe-Co based alloys, leading to the identification and characterization of new magnetic multifunctional materials of scientific and technological interest, c) to use nanoscale building blocks to exploit their large surface/interface anisotropies at artificial grain boundaries, and d) to mix particles with different high-aspect shapes to create well defined pinning centers for domain wall motion and thereby enhance the coercive field.
Thus the project will focus on the following specific scientific objectives:
SO1 To develop novel modelling techniques for optimization of hard magnetic properties of high- aspect-ratio nanostructures (inducing strong non-linear edge effects)
SO2 To develop cutting-edge basic research into the intrinsic -atomically resolved- properties of hard magnetic high-aspect-ratio nanostructures.
SO3 To develop online monitoring techniques during and after synthesis of the starting materials before consolidation
SO4 To understand the correlation of shape, composition, structure and morphology with magnetic anisotropy, magnetic moment
SO5 To provide from atomistic approach calculation the magnetocrystalline anisotropy of novel distorted cubic materials of the type Fe-Co and Heusler alloys in a single and a high-
throughput way.
SO6 To understand the conditions of stabilization of the tetragonal –phase along the “Bain-path” in
the Fe-Co and in Heusler alloys
SO7 To identify best practice examples for the full life cycle of the materials from synthesis to re-
cycling taking into account sustainability and cost issues
Technical objectives
The technical objectives of REFREEPERMAG are derived from the scientific objectives and are expected to lead to feasible materials for the fabrication of novel permanent magnets suitable for replacement of some applications of -earth based products or the development of a novel class of permanent magnets, first through thin film combinatorial approach with a possible extension or suggestion of bulk tetragonally-distorted Fe-Co or Heusler Alloys. More specifically:
TO1 To develop new technologies for high-aspect-ratio nanostructures production of new doped
{Fe-Co}-X-Y intermetallics with high magnetization (Br >1 T) and anisotropy (Hc>1T) using novel environmentally friendly fabrication techniques.
TO2 Use novel processing techniques for the consolidation of nanowires into Permanent Magnet compacts, with a view to realising energy products 60 KJ/m3 <(BH)max < 160 KJ/m3 at room temperature , better than Alnico and Ferrites and approaching those of rare-earth bonded permanent magnets.
T03 To discover new hard magnetic phases using a HT approach of thin films of {Fe-Co}-X-Y or
Heusler alloys, deposited epitaxially on various substrates with K1 > 107 ergs/cm3
TO4 Develop and establish advanced high-throughput-characterization techniques for structural
and magnetic properties resulting in a database of complex alloys and composites
TO5 Produce thick films/foils of {Fe-Co}-X-Y or Heusler alloys using the interlayer libraries approach
TO6 Demonstrate as a proof-of-principle one demonstration of the novel PMs into a micromotor
TO7 Suggest DIN standards and quality as well as sustainability control procedures for these novel
PMs
Project Results:
A description of the main S&T results/foregrounds (not exceeding 25 pages)
The major S&T task of the REFREEPERMAG project were divided into seven workpackages that are listed below with extra three workpackages for management, scientific coordination and exploitation -dissemination:
WP
Number WP Title Type of Activity
WP1 Ab-initio and Micromagnetic Modelling RTD
WP2 Ab-initio and Micromagnetic Modelling RTD
WP3 High-Throughput synthesis of Rare-Earth Free permanent magnet materials libraries RTD
WP4 Characterization –structural and magnetic of high-aspect-ratio nanostructures and Rare-Earth-Free thin films/multilayers and bulk PMs RTD
WP5 Up-scaling and Consolidation Techniques of nanocomposites-sintered magnets RTD
WP6 Towards Applications RTD
WP7 Life Cycle Impact Assessment (LCIA) RTD
WP8 Management MGT
WP9 Scientific Coordination RTD
WP10 Exploitation and dissemination OTHER
Advances that this project brought about beyond the state-of-the-art are:
1. With respect to Modelling
Employed for the first time combinatorial atomistic modelling – an extension of the work of Partner # 3- to search for unexplored compositional spreads of candidate magnetic phases of the type {Fe- Co}-X-Y and Heusler alloys of the type X2YZ that were stabilized in lower than cubic symmetry in conjunction with our combinatorial and high-throughput experimental deposition and characterization techniques, that were employed for the first time into the search of novel candidate materials for permanent magnets. Finite element micromagnetic modelling and micromagnetic- Monte Carlo simulations were used to simulate the best geometrical configurations for high coercivity and thermal stability of the switching processes.
2. With respect to fabrication approaches
2.1 The polyol synthesis, an environmentally friendly process, was extended to systems beyond the Co and Co-Ni system to more complicated ones with larger magnetization such as Fe and Fe-Co. These systems are more interesting because they will have higher magnetization (fcc Fe has a 30% larger atomic magnetic moment than bcc Fe), and consequently higher anisotropy when employing the shape anisotropy. Partner # 1 has successfully fabricated air stable FeCo nanoparticles with high magnetization using a modified polyol process.
Organometallic chemistry was developed by partner # 4 for the synthesis of air-stable Fe and FeCo nanoparticles with spherical and cubic shapes. This method was extended to the synthesis of iron-based anisotropic particles. The growth of iron-based nanorods and nanowires by a chemical process was a real breakthrough
2.2 The fabrication of nanowires using electrodeposition (Partner #7) were optimized firstly with respect to the materials composition as above. Secondly, the different parameters of fabrication (e.g. pH of the bath, current density, or temperature) were systematically checked for a controlled tuning of crystalline phases of nanowires as well as the plating time that determine the length of nanowires. In addition, crystal size and new crystal phases were induced by appropriate annealing. Finally, the distance between self-assembled pores in the alumina template (thickness of walls between the holes) were tuned serving the purposes of achieving the maximum production of nanowires, and if the structure is right wereused as prepared (e.g. as a monolithic sensor) to minimize or control the effect of dipolar interactions.
2.3 For the first time we explored the possibilities of shape anisotropy of nanoflakes, were produced either by modifying the ratio and kind of solvents in the polyol process (Partner #4) or by intensive ball milling in the presence of surfactants as in the rare-earth containing magnets (Partner #13).
3. With respect to the non-equilibrium techniques for stabilizing the fct-phase along the Bain path of high-Ms materials
3.1 We advanced the search of novel magnetic materials to a state beyond the state-of-the-art by using a high-throughput synthesis (HTS) technique in a very systematic way never used before as far we know. (HTS) is required due to the large number of parameters which need to be varied in order to achieve the conditions for stabilization of the tetragonal –fct- phase along the Bain path. These parameters were the substrate lattice constant, temperature, third or fourth element addition to the {Fe-Co}–base alloy. The (HTS) approach was the best approach to achieve the goal to identify efficiently the range of parameters for further detailed studies. Combined with the (HT) computational modelling, we introduced the search for novel permanent magnets a fast and reliable process never tried before.
3.2 The high-throughput approach was ideal as a fast and reliable method to stabilize novel Heusler alloys, which in the bulk possess already high moment and high Curie temperature, but yet lack large anisotropy. We carried out an intensive composition spread study of selected phases extending our studies from thin to thick films as well and succeeded to stabilize a tetragonal-like Fe2GuGa –phase on Si(111) at 300 0C.
4. With respect to intrinsic and extrinsic properties characterization
4.1 The vast number of materials libraries produced with the high-throughput synthesis need adequate high-throughput characterization techniques. We used existing and develop new screening techniques as non-destructive techniques assessing chemical (automated EDX in SEM), structural and microstructural (automated XRD) and magnetization properties. We optimized especially the two techniques for magnetization and anisotropy measurements, using a high- throughput high-temperature SQUID magnetometer for magnetization measurements and high- throughput Kerr-effect set-up, one of them with up to 2T, for high-throughput anisotropy-coercivity measurements. A correlation with structural data w lead us to conventional techniques for structure, microstructure and magnetization measurements for selected samples with promising properties within the preselected ranges. It has to be noted that the combinatorial approach is also here beneficial as we have complete data sets of very comparable samples from the materials libraries. From this we can select efficiently the samples which were studied in more detail e.g. for investigating the nanostructure of a discovered interesting material we made targeted preparations of cross-sectional samples for TEM (transmission electron microscopy) from a materials library by using a FIB (focused ion beam).
4.2 Newly developed anisotropic nanomagnets with high aspect ratios need to be magnetically characterized on an individual basis to understand the correlation of surface chemistry [23] (ligands for example), shape, morphology crystalline structure and magnetic properties. Single particle magnetic analysis based on the bolometric detection [30a] of microwave resonant absorption and x-ray magnetic circular dichroism experiments [30b] was developed and employed. Intrinsic properties like orbital and spin magnetic moments were determined. Single particle magnetic anisotropy was measured and the contributions from shape and magnetocrystalline anisotropy disentangled.
5. With respect upscaling and consolidation
5.1 With respect to upscaling, the polyol process was scaled up to a few tens or hundreds of grams in order to be able to use the optimized materials for consolidation
5.2 It was he first time ever to use state-of-the-art consolidation techniques for nanoparticles to consolidate the high-aspect-ratio nanostructures for the fabrication of the novel class of rare-earth free permanent magnets.
6. Applications
We are among the first groups worldwide to prepare and consolidate nanowires-nanorods into macroscopic permanent magnets with energy so high energy products, indicating that our consortium had achieved one of its main objectives.
Worth to mention here, as far as we know, we are the first consortium in the world to fabricate and test in “real conditions” MnBi based PMs with very promising results.
We have developed novel high-throughput techniques to search for novel magnetic materials with the appropriate properties for PM, paving the way for future approaches of rapid-screening for other classes of materials.
7. Life cycle Impact analysis (LCIA) and sustainability
7.1 It is , as far as we know, the first time that an LCIA study, for the proposed novel class of nanostructured permanent magnets, will be undertaken. The state-of-the art protocol for nanoparticles will be used (ISO 14040:2006) with respect to the life cycle, durability, recyclability, impact on health and environmental impacts of the new magnets based on high-aspect-ratio nanostructures.
WP1: Ab-initio and Micromagnetic Modelling
Main Objective: The objective of this task was to establish computational design tools that will guide material development through the course of the program.
Ab-initio calculation of anisotropy and magnetization (UPPSALA)
Models used
Electronic model Density functional theory (DFT) for crystalline anisotropy
Mesocale model Micromagnetics dynamics for shape anisotropy
Simulation software and type of numerics used
WIEN2k, Elk, muffin-tin orbitals (EMTO) pseudopotential approaches with plane waves (VASP) or numerical local orbitals (SIESTA, OpenMX) for DFT calculations.Most of these codes provide MPI-parallel versions suitable for large-scale computations. MAGPAR, FEMME will be used for micromagnetics together with finite element discretisation codes (GID, Salome) (See Del 1.1 and 1.2)
Achievements of the model beyond experiments
The theory can pre-identify interesting materials prior to the actual synthesis and experiments and thus save time and resources.
Summary of the modelling
Two different kinds of magnetic anisotropies, namely the magneto-crystalline anisotropy and shape anisotropy are investigated. Due to their inherently different nature, they will also be studied with two different approaches. The magneto-crystalline anisotropy will be studied by first-principles electronic structure calculations with DFT using both full-potential and pseudopotentials methods. The project will use various basis sets, e.g. augmented plane waves with local orbitals. High-throughput algorithms will be employed to search the parameter-space of ternary compounds. An additional alloying element will be added to these selected systems to obtain additional degrees of freedom and the modeling will determine whether this improves the system. For alloys, some of the DFT codes implement the coherent potential approximation, which was shown to be the best single-site approximation for alloys.
Highlight clearly significant results;
Tetragonally distorted FexCo1-x alloys doped with B, C or N interstitial atoms
This work was motivated by a publication of Burkert et al, who has predicted that the MAE of tetragonally distorted Fe/Co alloys can reach large values at certain concentration of its constituents and at a particular tetragonal distortion. Since such structure is not stable, a tetragonally distorted Fe/Co alloy can only be realized as a thin layer that is epitaxially grown on a suitable substrate. We have explored an alternative way of stabilizing the tetragonal distortion, in particular, by adding another element, which is an atom with a sufficiently small atomic radius in order to occupy an interstitial site
RESULTS
(FexCo1-x)2B alloys
For (Fe/Co)2B alloys it is experimentally known that they have a natural stable tetragonal structure for the whole range of x=0 to 1. These compounds have been experimentally studied by Iga. Computational studies have never been performed. By means of coherent potential approximation (CPA) and magnetic torque method we have evaluated the MAE throughout the whole range of Fe and Co concentrations. Results are summarized in the figures below. The end-compounds consisting of boron and only iron or only cobalt have an easy magnetization plane. However for cobalt concentrations between ~10-65% there is an easy-axis magnetization with a maximum MAE of the order of 0.5MJ/m3. An optimum is reached at about 30% of cobalt content, at which the saturation magnetization reaches approximately 0.8T. Optimal cobalt concentration is at even lower value, compared to the carbon-doped Fe/Co alloys. Yet, the MAE obtained here are not very high.
Fe5SixP1-xB2 quaternary alloys
We have performed high-throughput calculations of MAE of the whole range of iron-based compounds Fe5SixP1-xB2, which have appeared as promising candidate materials due to their tetragonal layered crystal structure without need of any strain or interstitial elements. We have performed a full structural relaxation of these compounds as a function of silicon concentration. Free structural parameters have been also optimized. Since the Si and P are neighbors in the periodic table, the alloy was treated using the virtual crystal approximation. Resulting MAE is summarized in the figure on the right. As can be seen, the MAE is relatively low and it is positive only for low concentrations of silicon. At Si concentrations more than 20% the material has an easy plane of magnetization.
3d-based L10 structures
L10 structures span a wide range of layered binary compounds, which have a structure related to face-centered cubic. A tetragonal distortion, typically rather small, is achieved by alternating the elements in subsequent atomic layers. As a result, L10 structures are a limiting case of thin multilayers, where the layers are just one atom thick. Some of these structures, such as FePt or CoPt already found their way to applications, for example in perpendicular magnetic recording. Here, we focus on L10 compounds, which consist of cheap and abundant elements. The table summarizing results here is rather preliminary and we intend to study alloying these structures with third element and its influence on the MAE. Results so far obtained on binary compounds show a particularly large MAE for Mn-based L10 structures.
Tetragonal alloys (Fe/Co)2B for a certain range of compositions they have a rather large MAE and sufficiently high Curie temperature and saturation magnetization. In cooperation with Vienna and Darmstadt groups we have theoretically and experimentally explored these materials, including a possibility to enhance their anisotropy with small amounts of substitutional 5d elements, inspired by results of [6]. Our results are summarized in [P5]. We have found that a small amount of W or Re elements can efficiently double the MAE. The latter was also confirmed by experiment. The anisotropy reached by 5% doping with Re was 1.4 MJ/m3.
Finally, we have explored layered tetragonal materials based on Fe5SiB2 and their alloys with Co (in place of Fe) and P (in place of Si). Results of this work are subject of a manuscript in preparation. Pure Fe-based compounds showed only rather small value of MAE, despite their favorable highly anisotropic crystal structure. MAE of about 1.1 MJ/m3 can be reached by about 30% substitution of Fe with Co.
Micromagnetic modeling
Analytic and micromagnetic calculations on single cylindrical nanorods have been performed in order to understand and magnetocrystalline and shape anisotropy contributions to coercivity. Assuming coherent rotation, the coercive field can be calculated with
〖 H〗_(c,coh)= (2K_1)/〖μ_0 M〗_s +(N_⊥-N_∥ ) M_s
The formula contains the anisotropy field H_A=(2K_1)/〖μ_0 M〗_s and the shape anisotropy contribution 〖(N〗_⊥-N_∥)M_s with the demagnetizing factors N_⊥ and N_∥. This value acts as upper limit for the coercivity of the real nanorod. Although a higher saturation magnetization increases the shape anisotropy contribution to coercivity, it also lowers the anisotropy field. For that reason materials with a high uniaxial anisotropy (K1 > 400 kJ/m³) and moderate saturation magnetization such as Co should be preferred over materials with low magnetocrystalline anisotropy and high saturation such as Fe or CoFe (see fig. 4).
Fig. 4. Hysteresis of a single Co, Fe and CoFe
There are multiple analytic approximations for non-coherent magnetization reversal mechanisms such as curling or buckling, but in reality the magnetization reversal is a mixture of multiple mechanisms that can only be calculated numerically. The micromagnetic simulations have been performed with the FEMME package which solves the Landau-Lifshitz-Gilbert equation on a finite element discretization of the magnetic structure. The simulations make it possible to systematically examine the contributions of shape and magnetocrystalline anisotropy as well as the losses due to inherent magnetization reversal modes that even occur on high-aspect ratio, high-anisotropy structures (see fig. 5)
Fig 5.. Nucleation and incoherent magnetization reversal in a high-aspect-ratio Co nanorod. The demagnetizing field at the tips causes a curling-like vortex structure. The magnetization reversal is an incoherent process involving the movement of two domain walls from the tips to the center of the nanorod
The model of coherent rotation predicts a saturation of the shape anisotropy effect at aspect ratios n > 10 (fig. 6a).The micromagnetic simulations however, show different coercivity values for different cylinders with the same aspect ratio H/D but different height H and diameter D. In fact, the micromagnetic simulations show a strong relationship between coercivity and diameter (fig 6b) because higher diameters favor the formation of a vortex nucleation at the tips. At low diameters (D<20nm) a moderate aspect ratio of 1:5 is enough to maximize the shape anisotropy contribution.
Fig. 6. a) Theoretical upper limit of coercivity of Co nanorods as function of the aspect ratio H/D with simulation results. b) Simulated coercivity as function of aspect ratio with lines of constant height and diameter.
Besides the optimization of the aspect ratio (mainly through decreasing the nanorod diameter) it is also possible to optimize the shape of the nanorods. Micromagnetic simulations indicate that the coercivity of a nanorod can by increased by 10% by rounding the tips to a half-spherical shape. The rounded tips decrease the stray field and suppress a nucleation at the tips. Another approach would be the reduction of the saturation magnetization Ms to increase the anisotropy field but this would also decrease the potential energy density product.
2. Regular Arrangements of Nanorods
The packing density p of regular arrangements of cylindrical nanorods can be calculated by dividing the area of the nanorod's base circle by the area of the surrounding hexagonal or square cell. Fig 7 shows calculated packing densities of nanorods with D = 10 nm with varying distance d or Δ.
Fig. 7. Hexagonal and quadratic packing of nanorods with diameter D, center-center distance d and surface-surface-distance (wall thickness) Δ.
Micromagnetic Simulations on 10x10 grids with hexagonal and quadratic packing with varying distances and packing densities have been performed. The calculations showed that the coercivity of multiple nanorods is dominated by the surface-surface distance Δ and not by the overall packing density or packing lattice. Therefore, it is possible to project the results from two nanorods to a regular grid with the same distance in order to save computational time.
Fig. 6a shows the coercive field of two nanorods over the distance Δ. Due to numerical difficulties the value at Δ = 0 nm has to be extrapolated from the other results and is printed with a dashed line. In Fig. 6b the same coercive field data is transformed to a function of packing density using the distance-to-packing-density mapping in Fig 5. The energy density product (BH)max can approximately be calculated by
〖(BH)〗_max={■(〖B_r〗^2/4µ_0&if µ_0 H_c≥ B_r/2@(B_r-µ_0 H_c ) H_c&if µ_0 H_c≤B_r/2)┤
for square hysteresis loops, where Br = p Js is the magnetic flux density through the structure at the remanent state and Hc is the coercive field. The energy density curve in Fig. XX shows perfect quadratic behavior with respect to the packing density p, because the coercivity µ0Hc is in any case higher than Br/2.
Fig. 8. a) Coercivity of two parallel interacting nanorods with distance Δ. The coercivity of a single nanorod acts as upper coercivity limit with indefinitely high distance. b) Coercivity values in (a) transformed as function of packing density and the resulting energy density product.
For realistic packing densities around 50% (corresponds to a wall width of 3nm) the predicted energy density product is around 200 kJ/m³ [P2] which lies well above the project target of 60–160 kJ/m³.
3. Irregularly Packed Nanorods
We introduced an algorithm based on the bullet physics library to generate realistically packed structures of nanorods. The nanorods are modelled as ideal, stiff cylinders and their motion is based on gravity and collision forces between nanorods and the walls as well as between nanorods themselves. To obtain nanorod structures with different misalignments, we introduced a tuneable torque on each nanorod which mimics the alignment of magnetic dipoles in a homogenous external field. The misalignment is measured by the standard deviation of angles ϕ between the external field (along the x-axis) and the orientation of the nanorods, denoted as σϕ. Using 3200 nanorods with diameter D=10nm and height H=70nm we obtained structures with σϕ between 10.0° and 71.7° and the packing density varies between 37% and 41%. An example structure is shown in figure 9a.
Fig 9. a) Finite element model of a Co structure with 3200 nanorods. b) Calculated hysteresis properties according to the three simulation methods: Full micromagnetics (µmag) give the most accurate results, macro spin assumes coherent reversal and Hmag takes only crystal anisotropy into account. These simulation modes make it possible to calculate the loss due to incoherent reversal modes (Δ1) and the gain due to shape anisotropy (Δ2)
Remanence and coercivity are decreasing with increasing misalignment σϕ. In our observed range of misalignment (10.0°–71.7°) the remanence Br varies between 0.50T and 0.67T and the coercivity field µ0Hc varies between 0.53T and 0.61T (see figure 9). The values for the remanence are corrected by the packing density, because we are interested in the macroscopic flux density and not in the flux density in the magnetic material alone. The coercivities in are calculated by the µmag method (and are therefore only available for the small structures with 120-140 nanorods) and the remanence values are taken from macro spin calculations (including full structures as well as subsets).
Fig 10. Dependence of remanence (diamonds) and coercivity fields (squares) on misalignment σϕ. The data shows a linear decrease of both remanence and coercivity with increasing misalignment.
The energy density product (BH)max can be calculated directly from the hysteresis loop. Considering the packing density of 40%, we obtain an energy density of 83kJ/m³. If it would be possible to increase the packing density to 45% the energy density would increase to 103kJ/m³ assuming that the overall shape of the hysteresis loop does not change. By variation of material parameters in the simulation it is possible to give minimum criteria for K1 and Js to reach certain energy density products.
(BH)max min. Js min. K1
70 kJ/m³ 1.50 T 150 kJ/m³
90 kJ/m³ 1.75 T 250 kJ/m³
120 kJ/m³ 2.00 T 350 kJ/m³
160 kJ/m³ 2.25 T 450 kJ/m³
Table 2. Lower limits for material parameters to achieve certain energy density products of a packed nanorod structure with a packing density of 40%. Higher densities require lower saturation polarization and higher magnetocrystalline anisotropy for the same energy density product.
4. Antiferromagnetic Capping Layers
Until now, only materials with high uniaxial anisotropy such as hcp Co have been considered as nanorod material. bcc Fe would be a cheaper option with a higher saturation magnetization but bcc Fe has a low cubic anisotropy leading to a low coercivity compared to Co nanorods with the same dimensions (see fig. 1). The idea is to seal the tips of the Fe nanorods with an antiferromagnetically coupled layer with high uniaxial anisotropy. This capping layer suppresses the formation of the nucleation at the tips and increase coercivity [P4].
Micromagnetic simulations on cylindrical bcc Fe nanowires with 20nm diameter and 100nm length have been performed. The nanowires are assumed to have a cubic magnetocrystalline anisotropy with fourth and sixth order constants K1,NW = 46 kJ/m³ and K2,NW = 1.5 kJ/m³, saturation polarization Js = µ0Ms = 2.15 T and exchange stiffness ANW = 25 pJ/m. The cap thickness tcap has been varied between 3 nm and 7 nm and have high uniaxial magnetocrystalline anisotropy an anisotropy constant K1,cap. The interface exchange stiffness Aint between nanowire and cap has been varied between -1% and -10% of the bulk exchange ANW, the negative sign indicates antiferromagnetic exchange.
Fig. 9a shows the dependence of coercive field Hc and exchange bias field Hex on the magnetocrystalline anisotropy in the cap for different cap thicknesses at -2.5% and -5% interface exchange. The cap thickness does not increase the maximum achievable coercivity but thicker caps move the Hc peak to lower K1,cap. This means that the coercivity can be maximized by experimentally tuning the thickness of the capping layer [P4]. Increasing the magnetocrystalline anisotropy in the cap increases the coercive field of the nanowire until a critical point where the coercivity increase is lost and a bias field is introduced. This transition has also been predicted by analytical calculations and a condition for the formation of an exchange bias can be given by:
t_AFM∙K_(1,cap)>J_int (=A_int/a)
where Jint is the interface exchange energy a is the lattice constant. If the caps are hard enough to overcome the interface exchange, the micromagnetic spins in nanowire and capping layer rotate independently from each other and cause the exchange bias field. If Jint > tcap K1,cap, the exchange forces the spins in nanowire and cap to align and the hysteresis loop is unbiased and symmetric. Fig. 9b shows coercivity and bias field as function of the interface exchange stiffness Aint. As long the above condition is fulfilled, increasing Aint only increases the bias field. If the interface exchange is high enough, the exchange bias field Hex becomes zero and the coercivity increases with increasing Aint. Harder caps yield potentially higher coercivity increases but also exhibit higher interface exchange thresholds until the unbiased state is reached.
The simulations on 40nm nanowires consistently show a lower coercivity without caps but a higher increase with the antiferromagnetic caps. This indicates that the coercivity loss due to inhomogeneous magnetization reversal processes is higher for thicker nanowires. The antiferromagnetic capping layer is able to suppress the nucleation at the nanowire tips and compensates the coercivity loss due to nucleation.
Fig. 11. Coercive and bias field as function of a) magnetocrystalline anisotropy of the capping layer b) exchange between nanorod and capping layer.
Co nanowires. Micromagnetic modeling was performed firstly for Co nanowire arrays with different geometry characteristics (length and diameter) and crystalline structure. The general objective of the study included particularly the determination of the role played by the crystalline and shape anisotropies together with magnetostatic interactions for arrays of Co nanowires.
As starting point, the balance between magnetocrystalline and shape anisotropies has been simulated for arrays of Co nanopillars 120 nm height and varying diameter from 35 (mostly fcc cubic crystal phase) to 75 nm (mostly hcp hexagonal symmetry) for which experiment and modeling show crucial changes of hysteresis loops when the diameter is increased. Magpar package [1] has been used with finite element discretization for typical magnetic parameters for Co with hcp and fcc crystal phases. Meshed tetrahedra (average 4 nm sizes) were randomly assigned to corresponding textured phases.
Fig12.- Simulated remanence magnetization for arrays of 7 nanopillars with diameter D= 35nm (a) and 75 nm (b). Arrows indicate the magnetization direction and colors the longitudinal Mz (up) and transverse Mx magnetization on 3D images (up).
The simulations published in confirm the primary role of magnetocrystalline anisotropy and the importance of magnetostatic interactions between nanopillars. The simulations also reveal the change of remanent state when the diameter is increased: from the vortex at the flat end surfaces and single domain in the rest of the nanowire (small diameter of around 35 nm) to the vortex state spanning the total nanopillar (above 50 nm diameter). In the former case, the reversal mode corresponds to the propagation of the vortex domain walls from the flat surfaces, while in the latter it proceeds with the quasi-curling mode taking place in the whole nanopillar.
CoFeCu nanowires.
The modeling was then focused on CoFe magnetic nanowire arrays. In this case, the magnetocrystalline anisotropy plays a less relevant role, and the study addresses the magnetization process and particularly the increase in coercivity observed for the sample annealed with the temperature (see report WP4). For the simulations, a polycrystalline structure of CoFe nanowires with [110] textured bcc phase in agreement with XRD studies was considered. For a CoFe nanowire with given composition, typical exchange and magnetocrystalline constants are respectively A = 10.7x10-12 J/m, K = 104 J/m3 , while the values of the saturation magnetization, Ms, as a function of annealing temperature have been extracted from the experimental measurements. We have assumed that the wires microstructure consists of grains with [110] direction parallel to the nanowires axis and random in-plane easy axis components.
Fig 13. Experimental (filled symbols) RT evolution of coercivity with the annealing temperature for CoFeCu nanowires of indicated diameter, Dw, and corresponding simulations for a single nanowire (lines) and for an array of 7 nanowires (dashed line).
Particular attention has been paid to CoFeCu nanowire arrays with small Cu content where important effect of thermal treatments was observed. After the magnetic field is applied parallel to the nanowire axis, at the remanence state, the simulations indicate the presence of a single domain structure practically along the whole nanowires length with longitudinal magnetization for nanowires with diameter in the range Dw = 8-22 nm. At both nanowire ends open vortex structures appear that minimize the magnetostatic energy. The magnetization reversal occurs by propagation of a domain wall (DW) which structure, either transverse (TDW) or vortex (VDW) mode, appears to be strongly dependent on the saturation magnetization value, for fixed wire diameter. Upon decrease of the saturation magnetization value in the range from 2.0 T (Ms for CoFeCu nanowires in as prepared state) to 1.7 T (Ms after annealing), a significant effect occurs namely, a transition from VDW to TDW mode owing to the change of the magnetic correlation length.
The coercivity value for the transverse wall reversal mode is known to be always larger than that for the vortex wall mode, and thus, during the transition between the two DW types, the coercivity increases. Consequently, the experimentally observed enhanced coercivity after annealing can be understood in terms of the transition between VDW propagation for as prepared nanowires and TDW propagation for the annealed nanowires
In most recent work, relevant parameters including nanowire diameter and crystalline structure have been collected for these micromagnetic simulations.
Other Phases
Tetragonal alloys (Fe/Co)2B for a certain range of compositions they have a rather large MAE and sufficiently high Curie temperature and saturation magnetization. In cooperation with Vienna and Darmstadt groups we have theoretically and experimentally explored these materials, including a possibility to enhance their anisotropy with small amounts of substitutional 5d elements. We have found that a small amount of W or Re elements can efficiently double the MAE. The latter was also confirmed by experiment. The anisotropy reached by 5% doping with Re was 1.4 MJ/m3.
Finally, we have explored layered tetragonal materials based on Fe5SiB2 and their alloys with Co (in place of Fe) and P (in place of Si). Results of this work are subject of a manuscript in preparation [P6]. Pure Fe-based compounds showed only rather small value of MAE, despite their favorable highly anisotropic crystal structure. MAE of about 1.1 MJ/m3 can be reached by about 30% substitution of Fe with Co.
WP2: Preparation of high-aspect-ratio nanostructures (INSAT, CSIC)
Three different methods were developed for high aspect ratio nanostructures synthesis: chemical, electrochemical and mechano-chemical syntheses of polymetallic nanowires and nanoflakes.
Chemical synthesis
“Chemical Synthesis of {Fe-Co}-X-Y Nanostructures” aims on the synthesis of nanorods (NRs) and nanowires (NWs) by the polyol process, suitable for an up-scale for several grams of powder (WP5). Cobalt nanorods with high coercivity (µ0Hc > 0.4 T for a random orientation and µ0Hc > 0.7 T for partially oriented assemblies) were obtained by this method. Core shell Co-Fe anisotropic particles were synthesized through the growth of iron on cobalt rods (D.2.2). The iron shell was varied from 2-10 nm. This shell increased the magnetization but decreased the coercivity, in agreement with the micromagnetic modeling carried out by partner 2 (MS2). In order to target BHmax =160 kJ.m-3 following the conclusions of MS2, the future efforts will be focused on Co and Co-Fe(2 nm) NWs.
Fig 14. TEM images of cobalt nanorods synthesized by the polyol process (a) and (b) with classical heating mantle, (c) and (d) with a microwave oven (Scale bar denotes 200 nm). Mean diameter, Dm, and mean length, Lm: (a) Dm = 18 nm, Lm = 280 nm; (b) Dm = 16 nm, Lm = 160 nm; (c) Dm = 7.5 nm, Lm = 28 nm; Dm = 8 nm, Lm = 42 nm. Sample (a) was prepared with 2.5% of hydrated RuCl3 ref. Sigma Aldrich 463779; (b) with 2.5 % of hydrated RuCl3 ref. Sigma Aldrich 84050; (c) and (d) with 2.5% and 2 % of anhydrous RuCl3 ref. Sigma Aldrich 208523 as nucleating agent.
“Electrochemical synthesis of CoFe rich nanowires”
Magnetic nanowires were prepared by electroplating filling of self-assembled pores in anodic alumina oxide (AAO) templates. The ordered AAO membranes were synthesized by two-step anodization process on 99.999% Al foils in oxalic acid electrolyte by applying a constant voltage of 40V and keeping the temperature between 3-4oC. The first anodization was performed for 24 hours and the second one was performed for 20 hours to assure a thickness of nanoporous alumina template of about 40 μm. The final hexagonal self-assembling of pores with diameter of around 35 nm and interpore distance of 105 nm is achieved. The diameter of pores was subsequently enlarged using H3PO4 (5 % wt). In this way the pores can be enlarged up to 80nm in diameters (Figs 15a-b).
Fig15. Top view SEM images of AAO membranes with pores of: (a) 40nm in diameter and 105 interpore distance, (b) 80nm in diameter and 105nm interpore distance and (c)18nm and 50nm interpore distance.
Well hexagonally ordered templates with much smaller diameters of the pores (16-27nm) are obtained using sulfuric acid as electrolyte (Fig 15c). The geometrical parameters (length, diameters and interpore distance) can be further tailored by changing the anodization voltage and the concentration of H2SO4 electrolyte (see Fig. 16, where A=3% wt. H2SO4, B=10% wt. H2SO4, and C=20% wt. H2SO4).
Fig 16. Dependence of pore diameter for different electrolyte concentrations (A, B and C series) on applied voltages.
Afterwards, Al was chemically etched from the bottom of the membrane and an Au nanolayer was sputtered to serve later as an electrode for the final electroplating of nanowires. CoxFe100-x (0 < x < 100) and FeCoCu nanowires were deposited into AAO membranes at room-temperature by DC electrodeposition from sulfate-based electrolytes containing CoSO4∙7H2O (5-45g/l)+ FeSO4∙7H2O (5-45g/l)+H3BO3 (10g/l) + ascorbic acid (10g/l) (CoxFe100-x nanowires) and CoSO4∙7H2O (35g/l)+ FeSO4∙7H2O (15g/l)+CuSO4∙7H2O (2g/l) +H3BO3 (10g/l) + ascorbic acid (10g/l) (FeCoCu). The pH values of the electrolytes were maintained constant at about 3.0.
Fig.17. Cross-section SEM images of a) CoFe nanowires with diameters of 80nm, b) CoFeCu nanowires with diameters of 20nm, c) CoFeCu nanowires with modulated diameters and d) CoFe/Au multilayered nanowires with 40 nm in diameter. The insert shows a close-up image of modulated nanowires in (c).
Figure 17 shows some examples of CoFe nanowires produced by the method described above. Apart from the already mentioned nanowires with uniform diameters (Fig 17a-b), modulated in diameter pores (i.e the nanowire) along the length can be obtained by combining mild and hard anodization, and specific parameters. CoFeCu nanowires with modulated diameters (22 and 35nm) were obtained (Fig 17c and insert).
Physical Methods-Nanoflakes (TUDA)
The ball milling of Fe-Co-X-Y compounds has been carried out in a planetary ball milling (Pulverisette 6) device. As it is mentioned in the DOW we focused on the composition where the Fe-Co-X-Y compounds show the highest magneto-crystalline anisotropy. It has been reported that the (Fe1-xCox)2B system shows both uniaxial and in-plane anisotropy as a function of chemical composition. Initially, we carried out magnetic measurements to determine the optimum composition of the Fe-Co ratio. Our detailed studies show that, the x = 0.25 composition shows the highest anisotropy constant K1.
Fig.18. SEM images of ball milled Fe1.5Co0.5B samples for (a) dry milling, (b) wet milling under ethanol and (c) wet milling under heptane and oleic acid (surfactant assisted). For comparison of the particle size the same rotational speed and milling time was used for all samples.
After the determination of the optimum chemical composition for the (Fe,Co)2B system we carried out ball milling on the (Fe0.75Co0.25)2B to increase the extrinsic magnetic properties. Initially we started with dry milling and the milling procedure is carried out for 18h at 250 RPM under protective argon atmosphere. Ball to powder ratio is selected as 10:1 which ensures a good energy product. After dry milling, observed average particle size was between 4-6 μm. To reduce the particle size further, we carried out wet milling on Fe1.5Co0.5B. For the milling in ambient we used ethanol and heptane with oleic acid as a surfactant. Figure 18 shows the secondary electron images of (Fe0.75Co0.25)2B samples milled under different ball milling mediums.
After the determination of the crystal structure and approximate particle size distribution we investigate the magnetic properties of the ball milled Fe1.5Co0.5B samples. The room temperature VSM measurement results of the ball milled Fe1.5Co0.5B samples are shown in Table 1.
Task 2.3 Surface modification and recycling
The objective of this task was to grow a thin layer at the surface of magnetic high aspect ratio nanostructures to improve their thermal stability and to avoid the coalescence during a consolidation process.
a) Coating of cobalt nanorods with low melting point metals (Co@M, M = Sn, Sb)
Cobalt nanorods were prepared by the polyol method, washed with absolute ethanol and dispersed in a solution of tetradecanediol in oleylamine containing the tin precursor (tin acetate) or the antimony precursor (antimony acetate). Different ratio M/Co were targeted from 1/8 to 1/2. The suspensions were heated for 30 min at high temperature: 300°C for the tin coating and 250°C for the antimony coating (see more experimental details in D2.3 and D4.9). The final powders were analysed by XRD, TEM and local EDS spectroscopy. X-ray diffraction showed that the reaction of tin acetate with the Co NRs at 300°C lead to the growth of the intermetallic CoSn phase. The TEM images showed composite fibers containing the Co NWs and the CoSn phase. The local EDS spectroscopy showed the presence of tin at the surface of the Co rods (Fig. 19).
Fig. 19. (a) Bright field TEM image of a core-shell Co/CoSb nanorod obtained by the decomposition of antimony acetate in oleylamine at 250°C (Co:Sb ratio = 4:1); (b) STEM-HAADF image (left) and corresponding EDX analysis (right) of a single rod ( Co map in red, Sb map in green and combined map of Co and Sb).
In conclusion, the CoSb shell does not modify the magnetic anisotropy of the cobalt rods and improve their thermal stability. The Co@CoSb anisotropic particles are stable up to 400°C when they are annealed in a H2/Ar atmosphere. These particles are promising for further consolidations to get isotropic permanent magnets.
Highlight clearly significant results;
The electrochemical process was extended to the growth of magnetic NWs with smaller diameter, higher crystallinity and optimized chemical composition for the objectives of the project, high magnetization and high coercivity.
Values of coercivity larger than 0.27 T and squareness ratio above 0.9 were obtained for the Co-rich samples. Very high magnetization, Ms=2.3 T (1830 emu/cm3), were obtained with the Co40Fe60 composition. CoFeCu NWs with complex nanostructures and high coercivity were also obtained by this method. These NWs annealed in good conditions combine high magnetization and high coercivity and are promising for microdevices.
The polyol process was extended to the synthesis of Co NWs with very high coercivities (µ0Hc >0.7 T). The important point is that this method can be extended to the production of several grams of powder at the lab scale (see WP5).
We succeeded in the growth of a thin layer of carbon or CoSn alloy at the surface of the Co NWs without degradation of their magnetic properties. The thin layer can act as a barrier to avoid the Co NWs sintering during the consolidation process.
Sn and Sb are good coatings for the nanowires before compaction and do not degrade their magnetic properties
WP3: High-Throughput synthesis of Rare-Earth Free permanent magnet materials libraries
The main objective of the addition of a third element to the Fe-Co lattice is to maintain a tetragonal distortion. The latter may be set by applying a suitable film substrate or buffer layer. High magnetic anisotropies are expected for tetragonally strained Fe-Co. However, a relaxation of the strained lattice was observed in binary Fe-Co). For ternary Fe-Co-C, a recent publication from Partner 4 predicts ternary phases with spontaneous (‘stable’) lattice strain and remarkable values of magnetic anisotropy. This paper reports on the progress to set this approach into practice and to extend it to ternary systems Fe-Co-X, which include different elements of X (B, C, N, Nb).
Developing “Bain library”, with tailored lattice parameters
According to theoretical calculations FeCo in tetragonal distorted phase could exhibit high magnetocrystalline anisotropy making it a promising material for permanent magnets applications, since it exhibits the highest magnetic moment among any other ferromagnetic material included Fe. Since the equilibrium state of FeCo is the cubic phase, a stabilization of tetragonal distorted FeCo is a demanding task. The highest magnetic anisotropy is expected for c/a lattice constant ratios between 1.2 and 1.25.
Combinatorial experimentation is a systematic way to fabricate and characterize a batch of samples on one substrate, covering broad regions of the compositional diagram and therefore, called materials libraries, under identical conditions in one experiment. Using combinatorial fabrication with techniques like co-sputtering or wedge–type depositions we are able to fabricate material libraries with well-defined composition and thickness gradients. Moreover high-throughput characterization methods (magneto-optical Kerr magnetometry, X-ray diffraction, EDX) allow us not only to verify the theoretical predictions but also to study extrinsic magnetic properties as coercivity (Hc) or remanent magnetization (Mr) which are almost impossible to be predicted by calculations and are crucial factors for high performance permanent magnets.
AuxCu1-x Library (Buffer layer)
The main idea is that the in-plane lattice parameters of FeCo on AuCu will adopt those of AuCu and the out-of-plane lattice parameter c is expected to adapt accordingly, since the unit cell volume should remain constant. The AuxCu1-x library was made by co-sputtering of Au and Cu elemental targets, at room temperature in confocal deposition geometry (K2 sputtering system – RUB facilities). The range of Au content starts from 13% < x < 46% (Fig.20). The epitaxial growth of the Au-Cu buffers was confirmed by XRD texture measurements in a four-circle set-up (IFW facilities) and can also be seen from XRD Bragg-Brentano measurements (Fig. 21 with the high (002) reflection around 2θ = 66°). Moreover we grew successfully the same buffer layer AuxCu1-x again by HT-sputtering techniques on MgO substrates at 300 oC, improving further the degree of texture and making it appropriate for higher temperatures (<300 oC) depositions which are needed for the crystallization not only of B-doped FeCo compounds but also of the binary system.
Fig 20. Compositional spread of Au content x(%) in AuxCu1-x buffer layer library
High-throughput deposition of Fe-Co-X on Bain library
After the deposition of the AuxCu1-x buffer layer library, the wafer was transferred into K1 sputtering system (RUB facilities) for the growth of a thickness gradient FeCo library. K1 is more adequate for thickness-dependent depositions than K2 sputtering system due to the movable deposition shutters of which the velocity can be adjusted from 1 mm/sec to 5 mm/sec. Initially the wafer is completely covered by the shutter which is retracted in a constant velocity during the deposition process, thus making a thickness gradient. The wedge-type thickness FeCo layer was deposited perpendicular to the AuxCu1-x composition spread axis A FeCo compound target (99.95% purity) was used for the deposition of FeCo layers. Under these conditions, a nominal thickness gradient ranging from 3.2 nm to 10.3 nm was achieved. According to the literature the relaxation volume of FeCo on Rh (c/a=1.24) is 12 monolayers (ML) or 1.14 nm. It is obvious that we are above the relaxation volume of FeCo, but this is the goal of this task: the study of magnetic properties of thicker strained FeCo films. More details about the magnetic properties are given in the next section. The material library was characterized structurally by a high-throughput x-ray diffractometer (RUB facilities) which allows screening of 301 measurement regions using a highly sensitive x-ray detector.
Figure 21. X-ray spectra for a 30 nm thick FeCo thin film deposited on a AuxCu1-x/Pd/Cu/Si substrate.
Following our methodology we can also estimate the c/a ratio of FeCo as a function of the composition of AuxCu1-x buffer layer (Fig.22).
Figure 22. c/a ratio of FeCo as a function of the Cu content in the AuxCu1-x buffer layer.
X=C: Fe-Co-C films
In most of the prepared films of Fe-Co-X type, we aimed at the compositions which promised the best performing in the calculations: (FezCo1-z)16X with z close to 0.4. The addition of approx. 2 at% carbon into the Fe-Co lattice stabilises a tetragonal strain of c/a = 1.03. We suggest that this strain originates from the formation of a spontaneously strained Fe-Co-C phase. The RHEED results recorded during growth of the (Fe0.4Co0.6)0.98C0.02 films (Fig. 23a) revealed a reduced driving force for strain relaxation. Together with the observed c/a ratios at thicker films (up to 100 nm thickness) containing 2 at% C (Fig. 23 b), these are strong indications for a lattice softening due to the C and a spontaneously strained Fe-Co-C phase which contains only 2 at% C and has 3% tetragonal distortion.
Fig 23: a) Comparison of ternary (Fe0.4Co0.6)0.98C0.02 with binary Fe-Co strain relaxation during growth detected with in situ RHEED. b) c/a ratios of tetragonal distortion from texture XRD measurements of (Fe0.4Co0.6)0.98C0.02 films in dependence on film thickness.
X=B: Fe-Co-B films
A larger content series with B contents up to 9.6 at% was performed with PLD prepared films. For a B content of 4 at%, the detected c/a ratio is 1.045. A further increase of B content, however, led to a decreased c/a ratio. A likely reason is the higher fraction of amorphous phase in the alloy as indicated by the reduced x-ray diffraction intensity. The magnetocrystalline anisotropy is also dependent on the B content. The estimated magnetic anisotropy energy MAE reveals a maximum close to the maximum of c/a. The reason for the shift of the maximum to a lower B content are the opposing effects of lattice strain and magnetic saturation: While a higher B content first increases the lattice strain, it reduces the magnetic moment of the alloy at the same time. In consequence, maximum MAE is shifted to B contents of approx. 2 at%. The highest MAE are observed in ultrathin films (d < 5 nm) due to the strong contribution of coherent strain at the Au-Cu interface. At higher thicknesses, the lattice relaxation becomes relevant and MAE decreases. For the binary films, it nearly vanishes. Due to the stabilised strain in the ternary films, a residual MAE around 0.4 MJ/m³ remains.
X=Nb: Fe-Co-Nb films
A [Fe/Co/Nb]24 multilayer materials library (ML) was fabricated by a wedge-type multilayer approach using a combinatorial magnetron sputtering system..In XRD analysis, we observe (110) oriented Fe-Co and
Figure 24: Hysteresis loops of Fe-Co-Nb for average Nb contents from 0 to 21 at. %
(FeCo)3Nb with various orientations. The higher the Nb content, the more was the fraction of (FeCo)3Nb. Together with TEM measurements (including EELS) the formation of a nanocomposite structure of the Fe-Co-Nb thin films, consisting of a matrix based on cubic FeCo and (FeCo)3Nb precipitates, is suggested.
The magnetization reversal of the composite system (Fig. 24) occurs at two discrete fields indicating that the new compound is also ferromagnetic but possesses significantly different magnetic properties compared with the FeCo-based matrix. The magnetic measurements revealed that the system behaves as a hard-soft magnetic composite. Particularly, hysteresis loop measurements at various angles showed that the system dominates by a domain wall (DW) pinning switching mechanism and (FeCo)3Nb inclusions act as pinning sites due to the anisotropy difference between the two phases. A rough estimation of the anisotropy constant based on DW pinning model showed that it is the same order of magnitude of that of ferrites. Furthermore, the energy product, estimated only by using the magnetic properties of this particular phase and not of the composite, is also comparable with the ferrites energy product, pointing that (FeCo)3Nb is a promising compound for permanent magnet applications.
Search for new phases in Heusler families
Theoretical studies have predicted the existence of tetragonal phases in Fe2YZ alloys with high magnetic moments. Thin films with Y = Ni, Cu, Co and Z = Sn, Ga were grown on thermally oxidized Si(100) substrates with 500 nm thick SiO2 layer, and epitaxial substrates, with a Mantis UHV magnetron sputtering system with three sources in a confocal geometry. Adjusting the power of the sources was the method that controlled the composition. The Fe-Y-Z films were studied with EDX, X-ray reflectivity measurements, X-ray diffraction, TEM and magnetically characterized by means of a VSM.
Fe-Ni-Sn films grown on BaF (001), independently of film thickness and exact composition, are crystallized in the hexagonal structure P63/mmc(194) showing high magnetic moment (450 emu/cc). Phase separation or amorphous phases were observed in the other systems.
Fe2CuGa films were also studied using combinatorial approach. The XRD pattern was close to the theoretically expected corresponding to tetragonal half-Heusler alloy structure. The effect of stoichiometry was crucial, since for different compositions than the nominal, another crystallographic orientation was revealed. The structural properties were also strongly depended on deposition temperature, while the optimum results observed at 300 °C.
Fig 25: Calculated X-ray diffraction pattern for the tetragonal Heusler alloy (up) and the X-ray diffraction pattern of Fe2CuGa on Si (111) (below) sample are shown. The peaks of the regular cubic structure are presented in blue color.
Another approach was the modification of various stoichiometric Heusler alloys by substituting –in bulk or melt spun materials– specific 3d elements with Fe and Co, in order to tune the magnetic properties. Main emphasis was given to Mn-Ni-Sn family and Mn-Ga family. The adopted strategy was to introduce Fe in substitution for Mn. Fe2-xMnxNiSn compound presented high saturation magnetization (80.0 Am2/kg), significantly larger even from 5 K values reported for Mn2NiSn. Mn0.4Fe0.3Ga0.3 material presented overall interesting values for its magnetic properties, with saturation magnetization equal to 51.3 Am2/kg and coercive field equal to 3 kOe in room temperature.
Other approaches for inducing tetragonal distortion
A way to induce uniaxial anisotropy in FeCo-based alloys is by the formation of new ternary phases. The spirit of this approach is similar to that of epitaxial growth: in both cases the cubic symmetry of the parent FeCo alloys should be broken. Bulk (Fe0.7Co0.3)2B can be stabilized in a tetragonal phase which exhibits high uniaxial anisotropy. We fabricated and studied Nb-doped FeCo library. Nb was selected to replace rare earth (RE) elements due to its low cost and high abundance. Moreover, binary FeNb and CoNb systems form hexagonal structures fulfilling the criterion of broken cubic symmetry. Also, Nb is heavier than Fe and Co which means that possesses higher spin-orbit coupling which is one of the most important factors for achieving high magnetocrystalline anisotropy. A new hexagonal ferromagnetic phase (FeCo)3Nb was detected. The fabrication of pure (FeCo)3Nb films which will allow the study of the magnetic properties of this new compound.
WP4: Characterization –structural and magnetic- of high-aspect-ratio Nanostructures and RE-free intermetallics
Various compositions with different Co and Fe content have been studied. Cast ingots of different compositions have been synthesized by induction melting in inert atmosphere using master alloys of Co85B15 and Fe80B20. In addition, melt-spun ribbons were prepared from the cast ingots by a pressure difference of He between gas chamber and vacuum chamber of 200 mbar and at different wheel speeds. Hysteresis curve of the samples was measured by VSM at room temperature. Melt spun ribbons of selected compositions were annealed at 500 °C for 1 hour. Selected alloys were investigated by TEM and XRD. A significant dominance of the crystal structure Fe2B is observed.
Table 3: Overview of various Fe-Co-B compositions
For (Fe0.7Co0.3)71B29 a coercivity of μ0Hc = 0.05 T was obtained with a wheel speed of 26.7 m/s and a pressure difference of 200 mbar.
Nanoanalytical TEM characterisation of post annealed Fe/Co/Nb multilayers
Different Fe/Co/Nb multilayer specimens were prepared for TEM investigations by the focussed ion beam lift-out specimen preparation technique. The specimen thickness is in the range of 120 – 175 nm. The Fe/Co/Nb multilayers exhibit a granular microstructure with a grain size of about 50 to 100 nm in diameter. Thickness of the Fe/Co/Nb magnetic layer is 315 nm, which is embedded between two SiO2 layers. The thickness of the top SiO2 layer is 12.4 nm. Two regions (A and B) were chosen for elemental characterization by energy loss spectroscopy (EELS). In region A besides Fe (48.5 at %) and Co (42.4 at %) also Nb (9.1 at%) was identified. No Nb was found in region B, indicating that Nb is inhomogeneously distributed within the magnetic layer.
Fig 26: Z-contrast image of the cross section (left) and bright field image with indexed pattern showing a granular microstructure (right) of Fe/Co/Nb multilayers
High-aspect-ratio nanostructures
Antimony coating Co nanorods (Co@CoSb) from partner 10 (INSAT) (see WP2) were used for scaling up and consolidation of Co-based nanorods with high densities for permanent magnets. Coating the Co nanorods can significantly help to avoid them oxidation and reach high density packing but preventing the strong dipolar coupling. Systematic study of structural microstructural and magnetic properties of CoSb coated Co nanorods with different shall thicknesses and at different temperatures has been performed. The morphology (shape) of these systems in both cases is preserved up to the end of reaction of Co nanorods with antimony acetate at 250 °C for 30 min.
Systematic XRD measurements of prepared systems have revealed that the presence of a CoSb shell around the Co rods delays their oxidation. The coercivity is in the range 3-3.6 kOe, rather constant, showing that no strong modification of the magnetic Co core occurred with the coating. The remanence to saturation magnetization ratio Mr/Ms of 0.5 the expected value for a powder in which particles are randomly oriented in the applied magnetic field. After annealing no apparent changes in the coercivity of the nanorods was noted, that marks the enhanced stability of these nanorods at 400°C.
Figure 27: Magnetization loop at room temperature of the raw Co nanorods (a); Co@CoSb with Co/Sb=8 (b); Co@CoSb with Co/Sb= 4 (c); Co@CoSb with Co/Sb= 2 (d).
Conventional Structural, microstructural and magnetic properties of nanostructures-nanowires
The role of Co on the structure and magnetic properties of CoxFe100-x nanowires with two selected diameters, 40 and 20 nm, respectively, was studied. The X-ray diffraction results show cubic symmetry in the whole range of compositions that evolves from body-centered cubic, bcc, structure for pure Fe and small Co content nanowires towards face-centered cubic, fcc, as Co content increases. For high Co content, including pure Co nanowires, the only stable phase is the γ phase, fcc structure with a [110] preferred orientation. A similar compositional evolution is found for nanowires with 20 nm in diameter: i) bcc cubic crystals with (110) texture for more reduced Co content until pure Fe nanowires, ii) fcc cubic crystals with (110) texture for high (around 90%) Co content, and iii) hcp hexagonal symmetry structure with (100) texture for pure Co nanowires. By increasing the Co content, the coercivity reaches a maximum at about 70-80% Co, and then decreases. Values of coercivity larger than 0.27 T and squareness ratio above 0.9 were obtained for the Co-rich samples. The maximum saturation magnetization of about Ms=2.3 T is obtained for the Co40Fe60 alloy nanowires. Room temperature oercivity of Co67Fe28Cu5 nanowires with diameters of 18 and 22 nm and the length of about 8 mm increases with annealing temperature up to 500 °C. By further increase of the annealing temperature it decreases. Squareness of all annealed samples was around 0.95. Saturation magnetization takes values of 2.0 and 1.7 T for as-deposited and 500 °C annealed samples respectively. This reduction has been ascribed to a partial oxidation of the nanowires. Note that increased coercivity is obtained after thermal treatment in spite of the reduction of saturation magnetization and the corresponding shape anisotropy.
Fig 28: Parallel and perpendicular hysteresis loops for the 18nm diameter nanowire array in as-deposited (a) and 500 ºC annealed (b). The EMD lies along the nanowire axis
Then, we have focused on the systematic study of the influence of thermal annealing on the structure and magnetic characteristics. XRD and TEM structural studies indicate that, after annealing, CoFeCu (with intermediate Co content) structure remains cubic bcc structure with reduced defects and sharper crystallite bounds. After suitable annealing the <111> growth direction is identified, together with a gap between wire and surrounding alumina that could contribute to the observed magnetic hardening after annealing.
Fig 29: (c) HRTEM images of CoFeCu nanowire 40 nm in diameter annealed at 700 ºC showing <111> direction as the nanowire growth direction. (d) Fourier transform of the area inside the box depicted in (c). (e) Close up showing the lattice image. (f) A gap (marked by the two arrows) between the nanowire and the alumina is observed.
CoFe and CoFeCu nanowire arrays, were annealed at different temperatures for 2 hours in Ar atmosphere. After annealing, a magnetic hardening is observed for all, more effective in the Cu ones. During the annealing, the presence of Cu is more effective to induce a noticeable magnetic hardening in comparison to nanowires without Cu content. In addition, the annealing temperature at which optimal magnetic hardening is achieved in CoFe and CoFeCu depends on geometry characteristics.
Fig 30: Temperature dependence of coercivity for several CoFe and CoFeCu nanowires (DNW=18 nm) in their as-prepared state and after optimal annealing.
Thin films approach
The multilayer approach (preparing [AuCu/Fe45Co55(C)]n=3 multilayers) supports to reach the large magnetocrystalline anisotropy, which is competing with the shape anisotropy of FeCo films. Using FMR we found this anisotropy to be of the order of 1 MJ/m3. From the FMR we also found that C doping supports the increase of MAE in FeCo sublayers. Element-specific XMCD measurements revealed the increase of the mL/mS ratio at the Fe-site, showing that the modification of the Fe electron band structure in our multilayers can be reason of obtained large MAE.Magnetic – nonmagnetic [AuCu/Fe45Co55]n=3 and [AuCu/ Fe45Co55C]n=3 multilayers were prepared using magnetron sputter deposition. Magnetic properties of samples were measured by SQUID, FMR and XMCD.
Hysteresis loops with external magnetic field parallel and perpendicular to the film plane for both samples are shown in Fig.8. Saturation magnetisation Ms of sample [AuCu/Fe45Co55]n=3 (RE-71) was found to be 1530 emu/cm3, whereas as sample doped by C (RE-70) has Ms of 1700 emu/cm3. These facts point toward the presence of large perpendicular anisotropy fields, comparable to the demagnetizing fields of 4πMs = 19.2 kOe and 21.3 kOe respectively.
Fig 31. SQUID measurements of [AuCu/Fe45Co55(C)]n=3 samples.
Advanced High-Through-put Characterization techniques (structural, microstructural & magnetic). A Fe-Co-Cr-Ni thin film library was fabricated using the sputtering system, which has been used for the deposition of the Fe-Co-Cr ternary library. After the deposition the library was annealed at 550 oC in vacuum for 80 min. As in the case of Fe-Co-Cr only a few very weak peaks were observable in XRD patterns, indicating the formation of a nanostructured or amorphous material. Screening of the magnetic properties by HT-MOKE revealed that the maximum coercive field as in the case of the ternary system was achieved for Cr rich regions (> 50%). Hysteresis loops measured by VSM did not show any significant enhancement of the coercivity (90 mT) compared with the values of the ternary system. The ratio between the remanence magnetization Mr and the saturation magnetization remains low (~0.58). The study of the ternary FeCoCr and quaternary FeCoCrNi libraries showed that a moderate coercive field can be achieved which is higher than that of commercial FeCoCr compounds.
Other local probe advanced techniques for single sample characterization
The Procedure to correlate Texture maps and microstructure was achieved with the upgrade of the available 4 circle neutron diffractometer to SUPER-6T2 version. A specially designed microwave shortcut can be driven at 2 to about 24 GHz continuously. The device is mounted in a quartz tube which is attached to a conventional Ultrahigh Vacuum System. Powder samples can be introduced through a load-lock which also includes a plasma cleaning stage for controlled removal of Carbon or Oxygen species on the surface of colloidal nanostructures supplied by the partners in the consortium. Auger spectroscopy is used to verify the purity of the nanostructures which then can be moved in-situ to the FMR detection stage. The load-lock mechanism and the reducing action of the plasma treatment have been confirmed. The detection limit of the multifrequency in-situ FMR device has been successfully tested with a 10 nm Fe film on GaAs(100).
Highlights of most significant results
Systematic study and detailed analysis of FeCo nanowires has been performed by partners 7 and 5. The performance of the nanowires could be significantly improved for permanent magnet applications. Particularly the magnetic hardening (increase of remanent magnetization and coercivity) was achieved by depositing hard magnetic (or antiferromagnetic) materials at the tips of nanowires and by the small addition of Cu with subsequent annealing at 400 0C.
The SANS measurements of powders of Co-based nanorods by partners 9 and 4 have supported the optimization of the washing procedures for studied systems. The optimization of the washing procedure had strong impact on the alignment of nanorods during the consolidation. This optimization was in significance for achievement of the large energy product in Co-based consolidated nanorods for permanent magnets reported in WP5. Additionally, it has been shown by Partner 4 that CoSn shell doesn’t influence on structural and magnetic properties of the Co core and this shell can be used as a protector from degradation of compressed Co nanorods in permanent magnets.
New hard Fe-Co based magnetic phases have been discovered by partners 6 and 2 using the high-throughput characterization techniques and the melt-spinning facility, respectively. Superior magnetic properties of hard magnetic Mn70Ga30 phases has been obtained by partner 13 while optimasing the texturing process. These are important ingridients for consolidation of the semi-hard (Fe-Co-B or Fe-Co-Nb) and hard (Mn70Ga30) magnetic phases in order to obtain an exchange spring magnet with high energy product.
The SQUID, FMR and XMCD measurements performed by partners 1 and 5 have reveald an exceptionaly (for the Fe-Co films) large magnetocrystalline energy density of oabout 1 MJ/m3 in the [AuCu/Fe45Co55]n=3 and [AuCu/ Fe45Co55C]n=3 multilayers prepared by partner 1. This energy density is comparable with MAE of SmCo alloys. This finding suggests the possibility to use these systems as PM in write-read heads of magnetic memory devises.
It has been found by partner 8 that with increasing the (B+C) content, the tetragonal strain in the Fe-Co lattice can be increased up to approx. 5%. This is remarkable since this strain is higher than the so far observed strains in ternary films of the type Fe-Co-B or Fe-Co-C where the highest strain was approx. 4%. However, they also have found that the crystal size is also important with regard to the magnetocrystalline anisotropy: if the appropriate crystal size is not maintained in the Fe-Co films, the anisotropy constant decreases strongly. The XMCD study performed by partner 5 confirms unambiguously that the observed enhanced MAE in spontaneously strained FeCo films due to B (C) doping is related to the electronic structure of studied systems and not to the possible grain structure or morphology of prepared films.
WP5: Up-scaling and Consolidation Techniques of nanocomposites-sintered magnets
While most of the preparation routes described up to now will be directly applicable for the growing market of magnetic MEMS, the large volume market requires novel routes to prepare rare earth free bulk permanent magnets. In this WP we will examine the feasibility of conventional and novel up- scaling techniques. We consider this WP as absolutely crucial when moving from optimized intrinsic to extrinsic magnetic properties or from micro to macro scale. This WP certainly represents a high risk endeavor as most of the following approaches will be applied for these materials for the first time. The challenge is to develop novel up-scaling and consolidation techniques (or to apply existing ones in an innovative manner) which yield densities as high as possible, without degrading the nanostructure while avoiding magnetic dipolar interactions and maintaining the high-aspect-ratio character of the precursor materials. Finally routes towards near net shape permanent magnets with unique magnetic properties in the chosen temperature window for application are developed. The bottom-up approaches described in the previous work packages will allow the preparation of high-aspect-ratio nano structures in addition to the theoretical predictions and high-throughput synthesis of thin films. As the market of permanent magnets requires bulk material, in this work package we report on the feasibility of up-scaling processes and consolidation techniques.
Up-scaling processes
The up-scaling processes task aims on supplying sufficient material for the tasks of consolidation and/or towards application s.
Upscaling of high aspect ratio nanostructures
The cobalt nanorods (Co NRs) were found to be the most promising high aspect ratio nanoparticles for
Figure 32: (a) TEM image of cobalt nanorods prepared at a scale of 10 g in 2L of butanediol using a 3L jacketed reactor; (b) magnetization curve of an assembly of the rods measured with the applied field parallel to the rod long axis.
\further consolidations. Pure cobalt rods indeed are the best compromise to get high magnetization and high coercivity. Coercivities higher than 7 kOe were measured on cobalt rods prepared at small scale. Moreover the polyol method is better suited to a scale-up than organometallic chemistry. Textured assemblies of nanorods were prepared by drying suspensions applying an external magnetic field. Co NRs were washed twice with ethanol, chloroform, then were sonicated in chloroform. The suspensions were placed into a mould in air gap of an electromagnet. The drying of the suspension and the rod alignment was achieved under a field of 1T. Large scale alignments were obtained (Fig.33) with a good correlation distance as inferred from small angle neutron scattering (SANS) characterizations.
Figure 33: (a) wafer of cobalt nanorods obtained by drying a suspension in chloroform under an external magnetic field of 1T; (b) Small angle neutron intensity profile scattered by a cobalt nanorod assembly, perpendicular (black square) and parallel (red diamond) to the rod alignment. Inset: corresponding 2D SANS pattern. The two correlation spots scattered perpendicularly to the rods show a very good rod ordering in the assembly.
From thin to thick films and foils
The study of the magnetic properties of Fe-Co-X single layers in dependence of the applied in-plane lattice parameter of the buffer material revealed a large contribution of a surface related anisotropy Ksurface around 1.8 mJ/mm², which includes contributions of both, the buffer interface and the free surface (see D3.5). Although the main fraction may be related to the free surface, the role of the interfaces is also remarkable. An additional contribution may originate from an induced strain due to the applied underlayers. However, the studies on single layers did not reveal significant differences, when comparing different Au-Cu buffer compositions i.e. very different in-plane lattice parameters (see also D3.5). Multilayers, which have a repetitive architecture of Fe-Co(-X) layers with interlayers of the Au-Cu buffer material (Fig. 34), could thus increase the positive contribution of the interfaces to the overall magnetocrystalline anisotropy KU in film architectures of higher thickness, i.e. upscaled films. For the Fe-Co- layers, we chose the composition (Fe0.4Co0.6)0.98B0.02 since this gave the highest MCAs according to previous studies on single layers (see D3.3).
Figure 34: Schematic architecture of the multilayers. The overall multilayer thickness is 100 nm for reasons of comparability.
X-ray diffraction and TEM measurements of multilayers with dAu-Cu = dFe-Co-B = 5 nm revealed that a further decrease of the Fe-Co-B layer thickness is necessary, since the structural and magnetic properties in terms of tetragonal strain and MCA do not fundamentally exceed the values of single layers. However, in situ RHEED measurements during the deposition of multilayers with dFe Co-B = 2 nm showed that a interlayer thickness of 4 nm is necessary to maintain the epitaxial growth up to an overall thickness of 100 nm. Such multilayers exhibit KU values around 0.95 MJ/m³ in architectures with at least 20 repetions. Their MCA is between two and three times the value of a single layered film with equivalent thickness . Very similar KU is also observed in multilayers based on C doped Fe-Co with Au-Cu interlayers, but a smaller overall thickness
In conclusion, the multilayer approach may be a suitable way to prepare film architectures with both, high overall thicknesses and high KU with respect to Fe-Co(-X). With further improvement, the shape anisotropy might be overcome and these multilayers could be an option for small scale systems, where the material cost of the interlayer alloy is not crucial. However, regarding possible bulk magnets, one also has to consider the magnetic saturation MS which is reduced substantially compared to single layers, when the whole multilayer volume is taken into account (see D5.7). In consequence, the possible energy product BHmax, which scales with MS2, is drastically reduced to about the factor of 1/10. Table 4 summarises this point as it compares the properties of a multilayer with 17 repetitions and the best performing 100 nm thick (Fe0.4Co0.6)0.98C0.02single layer.
Table41: Comparison of the properties of single and multilayers.
KU / MJ/m³ MS / T BHth,max /kJ/m³
(Fe0.4Co0.6)0.98C0.02single layer 0,44 2,1 880
(2nm (Fe40Co60)98B2/4nm Au46Cu54) multilayer (n = 17) 0,98 0,7 90
Inducing metastable Fe-Co-X-Y phases
The goal of this task is to stabilize tetragonal distortion in the Fe-Co-X-Y family. For this task in addition to the Fe-Co-X-Y systems we carried out investigations on a Mn-based (Mn-Ga) system.
Fe-Co-X-Y:
Theoretical predictions show that ternary (Fe,Co)2B and (Fe,Co)3B systems are promising candidates for tetragonal distorted systems. After the optimization of the composition of (Fe,Co)2B system we carried out quite a number of additions to the (Fe0.75Co0.25)2B system. To observe the effect we add Ta, Ti, Sc, Y, Cr, W, Re, Ir, Zr, Mn, Ni, P, Ga, Al, Si, In, Sn, Sb elements to the pure (Fe0.75Co0.25)2B system. The magnetic measurements show that only Ta, Zr, Ti and Re shows promising results. The Ta and Zr substituted samples show higher coercivity in the bulk form compared to the pure system. The Ti substituted sample shows lower coercive field with slightly increased saturation magnetization. The hysteresis measurements of the Ti, Ta and Zr samples are shown in Figure 5.6 in comparison to the pure (Fe0.75Co0.25)2B system.
The observed increase of the anisotropy energy (or anisotropy constant) for Re and decrease for Ir is confirmed by the first principle calculations. Theory predicts and increase for W and Re samples and a decrease is observed for other 5d transition metals including the Ir. Calculated magneto-crystalline anisotropy energies are shown in Figure 5.8.
Figure 35: MAE for various elements X in (Fe0.675Co0.3X0.025)2B and (Fe0.675Co0.275X0.05)2B. The dotted line indicates the MAE of (Fe0.7Co0.3)2B for comparison.
The Mn-Ga system
The goal of the work on the Mn-Ga system was to find an up scalable processing route of Heusler like DO22 structure of Mn3-δGa system. Different processing routes were investigated for the optimization of the DO22 phase. These results indicate that the optimized process route to obtain the DO22 phase is to arc melt the bulk alloy, cold rolling it and subsequent field assisted annealing at 730 K for 24 h. Using these parameters, we are able to produce samples of superior properties to those reported on bulk samples in the literature. Table 5 shows a comparison of this study with the existing literature.
Table 5: Comparison of magnetic properties and sample preparation routines of Mn3-δGa.
Composition Remanence
(Am2kg-1) Coercive field (T) Type of the sample Preparation details Reference
Mn70Ga30 20.3 0.58 Bulk (rolled) 2h at 700K Sample 8b
Mn70Ga30 23.6 0.72 Bulk (rolled) 2h at 700K under 1T Sample 9b
Mn70Ga30 26.1 1.24 Bulk (rolled) 24h at 700K under 1T Sample 10
Mn75Ga25 14.4 0.43 Bulk (arc melted) 336h at 623K J.Winterlik et al., Phys. Rev. B 77, 054406
Mn75Ga25 16.1 0.36 Bulk (arc melted) 168h at 673K B. Balke et al.,
Appl. Phys. Letters
90, 152504
Mn70Ga30 25.0 1.35 Bulk (arc melted) 24h at 1173K + several days at 673K H. Niida et al.,
J. Appl. Phys. 79, 5946
Mn70Ga30 10.5 0.57 Melt spun ribbon 1h at 973K T. Saito et al.,
J. Appl. Phys. 112, 083901
Mn75Ga25 28.0 0.65 Melt spun ribbon 50h at 723K Y. Huh et al.,
IEEE Trans. Magn. 49, 3277
Mn67Ga33 19.5 2.05 Thin film Sputtered at 673K C.L.Zha et al.,
J. Appl. Phys. 110, 093902
Mn75Ga25 17.0 1.19 Thin film Sputtered at 623K K.Rode et al.,
Phys. Rev. B 87, 184429
In conclusion, an optimized production routine for the metastable phase of Mn70Ga30 system was found. Our study indicates that 24 hours of annealing in 1T at 730 K is sufficient to stabilize the metastable DO22 phase in cold rolled samples. The advantage of the preparation routes is up-scalability in bulk.
Consolidation techniques
Texturing of nanowires by magnetic fields
The goal was to align the nanorods in a macroscopic powder with the following characteristics:
a high magnetic nanorods volume fraction in order to get a high magnetization per volume;
a perfect alignment to get the most square possible magnetization loop;
a preserved high aspect ratio character to maintain a high coercivity.
The SEM images of magnetically alligned cobalt needles are shown in Figure 5.12 for two different magnification. For the production of the needles, concentrated Co nanowires suspension in chloroform or in toluene were dried in an electromagent with an applied field within the range of 0.2-1.5 T. The optimum alingment was observed for the external field of 1 T. The lower magnetic fields leads to higher misalingment and bad magnetic properties.
It was the first time that it is clearly showed that an energy product higher than 160 kJ/m3 can be reached with high aspect ratio magnetic particles.
Figure 36: Hysteresis loop (a) and second quadrant of the corresponding B(H) loop (b) of three rod alignments exhibiting the same magnetic volume fraction: (blue) dm = 22 nm, volume fraction = 48.7%, BHmax = 126 kJ.m-3 ; (black) dm = 24 nm, volume fraction = 48.7%, BHmax = 82.5 kJ.m-3; (red) dm = 28 nm, volume fraction = 48.8%, BHmax = 51 kJ.m-3.
Figure 37: Second quadrant of the B(H) loop of two assemblies of the same cobalt nanorods (dm = 22 nm) with a magnetic volume fraction VM = 48.7% and a BHmax = 126 kJ.m-3 (blue) and a magnetic volume fraction VM = 54.4% and a BHmax = 167 kJ.m-3 (magenta).
Hot pressing (and cold compaction)
The goal was to obtain hot compacted dense permanent magnet materials by using high-aspect-ratio nanostructures. The hot pressing studies on high-aspect ratio nanostructures were carried out for different pressures and temperatures. To ensure the hexagonal crystal structure of the Co nanowires all studies were carried out below the structural transition temperature of the cobalt. Prior to the compaction, the Co nanowires were aligned under external magnetic field of 1.7 T to ensure a good texture. After texturing these “green compacts” were used for hot pressing. The results (Figure 5.15) show that it is not possible to reach to the 75% texture by using the hot pressing which is due to the used pressure values. The best sample is obtained for the pressure for 40 MPa.
Figure 38: Room temperature pulse magnetometer measurements of hot compacted Co nanowires.
Magnetization measurements of hot compacted sample shows remanence to saturation ratio around 62% which is slightly better than the polymer bonded Co nanowires. The room temperature (BH)max value for the hot compacted Co nanowire magnet is calculated as 15.7 kJ/m3. The reason of this low energy product can be explained by the low packing density of 44%.
In addition to the pure Co nanowires we carried out hot pressing on the antimony coated Co NWs. Same hot pressing conditions are used for Co@Sb nanowires and the magnetization results are given in Figure 5.16.
Compaction studies for Mn-Bi samples were carried out in a polymer press equipped with a electromagnet. For the compaction of these magnets a special design press form was used. Fig 39 shows the images of the press form and a final product magnet as an example.
Fig. 39 Images of the press die (a) and the final magnet (b)
Spark-plasma sintering (SPS)
The goal of the spark-plasma-sintering studies were to obtain dense permanent magnets by using high-aspect-ratio nanostructures. Similar to the hot compacted samples, the Co nanowires were pre-aligned with an external magnetic field of 1.7 and the consolidation studies were carried out on these pre-aligned “green compacts”. The SPS studies on Co nanowire powder leads to similar results (texture ratio of 60-65% after consolidation) as the hot compacted samples. This observation shows that the hot pressing and SPS has almost the same effect on the consolidation of the Co nanowires. The studies show that using Co needles leads to better magnetic properties compared to Co nanowire powders. Due to this reason we carried out SPS studies on Co needles which are consist of aligned co nanowires. Figure 5.20 shows an image of the Co needles/platelets.
Figure 40: Examples of pre-aligned needles/platelets obtained in the 5mm width parallelepiped mold.
The SPS studies which are carried on the pre-aligned Co needles show promising results even after compaction. Figure 5.21 shows a room temperature hysteresis curve of a SPS compacted Co nanowires. Hysteresis measurement of spark plasma sintered Co NWs show remanence to saturation ratio of 78%. Observed (BH)max value for the SPS Co NWs is 27.5 kJ/m3 at room temperature.
Figure 41: Room temperature pulse magnetometer measurements of Co nanowire needles after SPS.
In addition to the different consolidation methods the effect of shape of the magnet is investigated in detail. The room temperature hysteresis measurements confirm that the most promising shape for the application of Co NWs is the needle shape. For further increase of the magnetic properties one can carry out the consolidation study under the existence of the external magnetic field. In our experiments the samples were pre-aligned under 1.6 T external magnetic field and the pre-aligned powder was compacted without field. Due to the setups we have it is only possible to carry out cold compaction studies under external magnetic field.
Our studies show that it is possible to achieve the goal of the REFREEPERMAG project (developing new generation novel materials for high performance permanent magnets with energy product 60 kJ/m3 < (BH)max < 160 kJ/m3 without any rare earth or platinum). Our studies show that highest achievable (BH)max value for consolidated magnet is 64.9 kJ/m3 after hot compaction.
WP6: Towards Applications
The aim of this work package is to demonstrate the feasibility of our approaches in making prototype magnets that can directly be tested by the industry and outperform existing families of PMs in terms of the price/performance ratio. In addition we will explore the possibilities to extend the knowledge developed for thin/thick films during the project towards “emulation” of stress-strain in core-shell nanoparticles and eventually in bulk metallurgy, which will lead to large-scale applications.
From some of the candidate materials, for example MnGa, MnAl, (Fe0.65Co0.3Re0.05)2B (Fe0.4Co0.6)32B and MnBi, which have large enough magnetization, Curie temperature and magnetocrystalline anisotropy energy (MAE) we have selected and mastered after the month 24rth the MnBi system.
Fig. 42 Various candidate structures of Mn-Based alloys
Detailed description
The characteristic properties of MnBi materials are ( Ames)
Saturation Magnetization (RT) 60-75 Am2/kg
Hc ~ 2T
(BH)max (theoretical) 17.6 MGOe
(BH)max (Our MAGNETS) <5 MGOe
Magnetocrystaline Anisotropy 1.6 x 106 Jm-3
Operating limit 473 K
The magnets already used by WCM are of the NdFeB bonded and the dimensions are shown in Fig.44
We have prepared a mold made from special steel using wire cutting in order to use for magnets casting e.g insert powder, alligh in magnetic field and compress it , properly remove it and sinter in inert atmosphere at ~ 260-270 for 40-60 minutes.
a) Precursor alloy melting ( 5-10 % more Mn) in the alloy Mn-Bi
b) Powder production using ball-milling for about 50-6- minutes
c) Compression in a magnetic field of ~ 1 T
d) Anneal at lower temperature 250-270 C
We prepared various batches of alloys and characterize them with VSM and SQUID magnetometry and data are given in the following figures.
Fig. 43 The Optimized magnets showing uniformity and their characteristic properties.
Doing so we have obtained sintered magnets which are uniform and have magnetization of 0.4 T and coercivity ~ 0,5 KOe, as shown in Fig. 6.6
Fig. 44 MnBi and NdFe B magnets glued on the test-motor
Operation and evaluation of its performance. Comparison with the state-of-the-art micromotor.
The magnets are bonded in the rotor with alternating direction of magnetization as shown in Fig. 44
Fig. 45 Schematic of a 8-pole motor with PM mounted on the rotor different magnetization orientation.
We run the motor at ~ 700 and 1000 rpm and the results are given below
We obtained excellent performance , with no extra resonances.
The measured output is:
MnBi
N= 737 rpm V=0,738 V N= 1000 rpm V= 1,001 V
NdFeB
N= 667 V= 2,717 V N=1000 rpm V=4,073 V
This output scales rougly to the (BH) max of MnBi vs the NdFeB magnets used.
Work is in progress to fully characterize the motor performance, measuring torque, stability etc.
We are one of the first groups in the word to use MnBi –based permanent magnets in a 8-pole motor susccesfully. If we can improve the fabrication process and increase the energy product to the one obtained by other groups (10 MGOe) and make an estimation of the cost of the process it might be an alternative non-rare earth permanent magnet- for some applications.
Highlight clearly significant results;
MnBi PM were synthesized, tested and bonded on a rotor with alternating direction of magnetization
We are one of the few labs in the world to test these magnets in a motor with accepted performance, a statement on the use of resources, in particular highlighting and explaining deviations between actual and planned person-months per work package and per beneficiary in Annex 1 (Description of Work);
We have already informed the consortium that NCSR D will undertake the work of magnet preparation, characterization, finishing, bonding atc and WCM will perform in their testing labs the performance, shifting their 6-PM to NCSR D.
WP7: Life Cycle Impact Assessment (LCIA)
WP7 deals with the evaluation of the environmental impact caused by the nanomaterials used in the fabrication of novel PMs developed in this project. To achieve this main goal, the following specific objectives are proposed:
To design a new framework for the life cycle impact assessments of nanomaterials based on the existing model like USEtox or ReCiPe.
To perform the LCIA analysis of nanomaterials following the framework previously described
To develop the correct algorithms for the impact analysis of the nanomaterials in the LCIA analysis
The LCA is aimed to integrate results generated in REFREEPERMAG on toxicity of new FeCo-X nanostructured materials and their functionality in new magnet compacts developed in order to have a global knowledge of the environmental performance of these products in all life cycle stages. The main objectives of the study were to identify which processes pose greater impacts, how these impacts are distributed along the different life cycle stages, and their potential impact on environment and human health. The characteristics of the system defined for FeCo-X nanomaterials are detailed, defining the scope and boundaries of the system, the functional unit, modelling rules and hypothesis. Preliminary information available for inventory is presented in this deliverable as well.
Goal and scope definition
The first phase of the LCA study is the goal and scope definition, which defines the general context for the study. In the goal definition, parameters such as the functional unit, the intended application, the reasons for carrying out the study, the target audience or the limitations and assumptions were identified.
Inventory analysis
In the inventory data, each life stage was analysed to determine the relevant inputs and outputs of the system. This information was gathered from REFREEPERMAG partners, literature and also from other projects related primary data .
Impact assessment
The impact assessment calculated the environmental potential impacts associated with inventory flows. The base methodology chosen for the present study was the ReCiPe Mid/Endpoint method. This methodology combines both midpoints and endpoints in a consistent framework (cause-effect chain) which allow a step-by-step interpretation and revision of the results.
Interpretation of results
In the final step of the LCA, the interpretation of results, a critical revision of the results was done in order to verify its reliability. In this step the completeness, sensitivity and consistency of data gathered and results obtained were done in order to guarantee their representativeness and suitability to be incorporated in created processes for datasets and impact assessment methods.
The methodology and steps followed are given in Fig. 46.
Figure 46 . Methodology and steps for LCA studies according to ISO:14044:2006 should be prioritized and secondary data from databases and literature will be used when needed.
The Preliminary scope of the study on FeCo-X based materials systems is given in Fig, 47
Figure 47 . Preliminary scope of FeCo-X systems
Description of the studied system
The system selected for the study is based on the synthesis, production, manufacturing, use and end of life of nanomaterials (FeCo-X, nanowires and nanofilms) for magnetic compact, as presented in Figure 3 and Table 1.
The first life stage considered is the synthesis of nanomaterials. Different partners have synthesised FeCo-X nanomaterials using different methods. The methods considered in this study require extraction of raw material and manufacturing of Fe, Co, and X, as well as pyrol synthesis and electrochemical synthesis of nanostructures, sputtering of FeCo-Ni and ball milling.
Once the FeCo-X nanowires and films are synthesized, bulk structured nanomaterials in pellet forms can be obtained and transported in order to produce prototype modules, such as permanent magnets for sensors.
In the use stage not important environmental impacts are expected to occur.
For the end-of-life phase, different scenarios were defined according to the prototype characteristics that will determine the recyclability of the piece once the useful life will be over. Current European Statistics on waste treatment for End-of-Life vehicles will be used to define the possible waste treatments to be assessed.
The general allocation criterion is the mass parameter, which means that the reference flow related to the functional unit is taken to allocate environmental loads in each stage. For the allocation of environment impact the product along all life stages, the following rules have been determined.
In synthesis, modification and structuring of nanomaterials, all the inputs and outputs needed to process nanomaterials have been included.
For manufacturing, all inputs and outputs have been taken into account.
In the stage of use, not important environmental impacts are expected to occur..
Fig. 48. System flowchart for Applications
Table 7. Life Cycle Stages and information required for nanomaterials
Life Cycle Stage Information Needed Primary data sources information from Partners Secondary data sources
Nanomaterials Synthesis processing
Nanomaterials Structure Processing Processes description
Inputs Outputs Data provided by manufacturers partners
Previous studies (literature), LCI databases
Prototype module manufacturing • Processes description
• Inputs and outputs Data provided by manufacturers partners
Previous studies (literature), LCI databases
Functional module-based system manufacturing • Module description (parts, materials)
• Processes description
• Inputs and outputs Partners Involved Previous studies (literature), LCI databases
During module-based system use (Sensor application and Motor applications) • Performance during use
• Energy recuperation/saving
• Maintenance/Substitution parts
• Inputs needed
• Outputs generated
• Useful life duration
Partners Involved Previous studies (literature), LCI databases
Disposal processes • Waste generated (types, weight, treatment) Data provided by REFREEPERMAG Previous studies (literature), LCI databases
For recycling processes, only the waste treatment of the nanomaterials will be considered. At the end-of-life stage, impacts generated during the treatment of materials will be allocated into the system until the moment where the nanomaterials reach the state of end-of-waste.
Human toxicity of nanomaterials along their life cycle
The human toxicity of the FeCo-X nanomaterials can be assumed to be similar to that of its main constituents: cobalt and and iron. In addition, in the case of the presence of nickel this can also contribute to their toxicity, in particular for people sensitized to this element. The manipulation nanomaterials should be done taking the necessary protective measures to reduce inhalation and the subsequent risk of lung diseases, respiratory sensitization and carcinogenicity. Similarly, the frequent manipulation of FeCo-X nanomaterials should be done with gloves to minimize contact with cobalt and reduce the risk of sensitization. Sn or Sb nanocoatings also in small amounts have the same effect as their metals.
Among the life cycle steps evaluated for these materials, the manipulation of the nanomaterials prior to the compactation step and also the recycling phase after their use, are the most important phases in terms of toxicological risk and appropriate measures to minimize worker exposure should be taken.
Ecotoxicity of nanomaterials along their life cycle
To achieve the entire comprehension of the impact of new materials in the environment, the full set of impact categories to be considered at the midpoint level. At the endpoint level, three major endpoint categories were considered:
damage to human health (HH),
damage to ecosystem diversity (ED),
damage to resource availability (RA).
Table 3. Overview of the endpoint categories, indicators and characterization factors.
Area of protection Impact category Indicator
Name Abr Name Unit
Human Health Damage to Health HH Disability-adjusted loss of life years DALY
Ecosystems Damage to ecosystem diversity ED Loss of species during a year Species yr
Resources Damage to resource availability RA increased cost $
Man-made environment NA - NA -
Conclusions and discussions
The ecotoxicity of the FeCo-X nanowires, films and sintered PMs can be assumed to be similar to that of its main constituents: cobalt, iron, antimony, boron, tin etc. The manipulation of the nanowire/nanorods and to a lesser extend the thin films/multilayers should be done with the necessary protective meaures to reduce inhalation and the subsequent risk of lung deseases, respiratory sensitization and carcinogenicity. For the finished products the risk is lower and the same preacutions should be taken. Coating of PMs can prevent some contamination and it is strongly recommended.
WP8: Management
We used our previous experience in managing EU and National projects to meet the challenges for managing such an excellent consortium and did not faced any problem.
WP9: Scientific Coordination
All the meetings were well organized and minutes were collected
WP10: Dissemination and exploitation plan
A solid exploitation plan and dissemination has been developed
The potential impact (including the socio-economic impact and the wider societal implications of the project so far) and the main dissemination activities and exploitation of results (not exceeding 10 pages).
Exploitation
On completion of the project, the following main deliverables deserve special attention for exploitation
System Software
New improved alloys
New high performance grades of finished PM’s
New processes or materials (e.g. nanocomposites) for production of PM’s.
Novel high-throughput techniques to search for novel magnetic materials with the appropriate properties for PM
System Software
The partners TUW, UPPSALA, CSIC, INSAT and CEA have developed software packages that are suitable for exploitation.
The software packages used are WIEN2k, Elk, muffin-tin orbitals (EMTO) pseudopotential approaches with plane waves (VASP) or numerical local orbitals (SIESTA, OpenMX) for DFT calculations. Most of these codes provide MPI-parallel versions suitable for large-scale computations. MAGPAR, FEMME will be used for micromagnetics together with finite element discretisation codes (GID, Salome).
UPPSALA is concentrated on combinatorial calculations using standard codes as modules and had successfully TUW, CSIC and INSAT and CEA for micromagnetic simulations. Their software packages present an excellent opportunity for exploitation, but also as an important issue and strength of the partner in future submissions to the EU.
New improved alloys (NCSR D, INSAT, RUB, IFW, TUDA)
Chemical and electrochemical approach for the preparation of nanorods, nanowires of the type Co, Fe-Co-X
Bulk synthesis (melt spinning, ball milling, arc-melting, compaction)
New high performance grades of finished PM’s
Two classes of finished PMs were developed along the duration of the REFREEPERMAG project.
Co-based nanowire-nanorod PMs by blending and compaction
MnBi- based alloys by cold compaction and sintering
Co-based nanowire-nanorod PMs by blending and compaction
We have achieved the benchmark set in the proposal of (BH)max = 160 KJ/m3 in compacted Co-nanorods with an aspect ratio of ~ 5. Magnets have been prepared by INSAT by blending with epoxy and by the compaction methods developed by TUDA.
We are among the first groups worldwide to prepare and consolidate nanowires-nanorods into macroscopic permanent magnets with energy so high energy products, indicating that our consortium had achieved one of its main objectives.
MnBi- based alloys by cold compaction and sintering
We have succeeded to fabricate Bulk PMs with very good properties and energy product in sintered (NCSR D) and epoxy bonded magnets (MagnetFabrick). The energy product for the sintered magnet was ~ 40 KJ/m3 and the epoxy bonded ~ 5 KJ/m3. A degradation of the original magnetic properties was observed during sintering, but the final PM was good enough to be used for incorporation in a micromotor (WCM) and run a test performance.
The test performance has given the following results compared to the currently used bonded NdFeB magnets
MnBi
N= 737 rpm Vout =0,738 V N= 1000 rpm V out = 1,001 V
NdFeB
N= 667 V out = 2,717 V N=1000 rpm V out =4,073 V
Given that the output Voltage of the motor is proportional to the flux Φ = (Β.S) and from the measured values of both original NdFeB and our MnBi the voltage output of our- non optimized MnBi magnets- scales as ¼ of the voltage output of the NdFeB magnets ( 1.1 T is the remanence magnetization for NdFeB compared to the ~ 0,35 T for our sintered magnets). There is room for improvement of the sintering process of the MnBi PMs in order to preserve the very good properties of the powder ( Br ~ 60 emu/gr and Hc ~ 1 T).
Worth to mention here, as far as we know, we are the first consortium in the world to fabricate and test in “real conditions” MnBi based PMs with very promising results.
Economics of PMs
Co-based nanowire-nanorod PMs by blending and compaction
There are three preparation routes for the polyol mediated synthesis of cobalt nanorods:
1. Microwave oven – which offers the possibility to heat quickly a large volume of polyol and experiment within a big range of heating rates. High heating rates favors the formation of cobalt nanorods with small diameter, presenting the best magnetic properties. However, the strong drawback of the microwave is the heterogeneous heating that creates a metallic mirror on the flasks of the spherical glassware. The particles formed in the bulk of the suspension were found homogeneous with interesting magnetic properties, but the particles formed on the walls exhibited poor magnetic properties.
2. Heating mantle – which offers a better homogeneity of temperature inside the flask leading to
monodispersed cobalt nanorods. This set-up provided nanorods with the best magnetic
properties. Its limitation is the narrow range of the heating rate that can’t exceed 10 °C/min.
3. Jacket reactor – is expected to combine all the advantages of the above: a very homogeneous
temperature inside the solution, a good stirring and a high heating rate. The efficiency,
scalability and reproducibility of the first two approaches were examined.
These three set-ups are used at industrial scale to produce chemicals and including inorganic particles. In the table 1 the prices of the different chemicals used in the different steps of the “standard” synthesis of cobalt nanorods at the laboratory scale are given. Of course these prices are much higher than that expected for an industrial supply. Savings can be made on the cobalt precursor. The cobalt acetate is clearly not the cheapest precursor. Nevertheless it is noteworthy that 80% of the price comes from the butanediol.
Chemical reactions
The synthesis of the cobalt nanorods involves three steps:
1. The synthesis of the sodium laurate from lauric acid and sodium hydroxide
2. The cobalt laurate synthesis from cobalt chloride and cobalt acetate
3. The cobalt nanorods synthesis from the reduction of the cobalt laurate in a sodium hydroxide
solution in 1,2 butanediol containing 2.5 % of ruthenium chloride.
In the table the prices of the different chemicals used in the different steps of the “standard” synthesis of cobalt nanorods at the laboratory scale are summarized. Of course these prices are much higher than that expected for an industrial supply. Savings can be made on the cobalt precursor. The cobalt acetate is clearly not the cheapest precursor. Nevertheless it is noteworthy that 80% of the price comes from the butanediol.
The concentration of cobalt in the “standard” synthesis is 0.08 mol.L-1 that allows the synthesis of 5 g in one batch. Current price excluding personnel cost is 7€/g. Can be reduced to 2.5 €/g.
Several syntheses were done in the microwave oven with higher cobalt concentrations: 0.12 M and 0.24M. In every case the anhydrous ruthenium chloride was used as nucleating agent. These concentrations also showed that it is possible to prepare anisotropic cobalt particles similar to those prepared with the standard protocol at reduced cost.
In conclusion nanoparticles of Co- cost very much and we are in doubt that this technique will be used for mass production of permanent magnets (see tables below). Only niche markets are suitable for exploitation.
b) 1. MnBi sintered- based PMs
In order to evaluate the possibility for exploitation of PMs based on MnBi- alloys we show a recent study in ROMEO project funded by EU by P. McGuinness et al, considering that from the producer and consumer point of view, the cost of a permanent magnet with certain magnetic properties is one of the most important factors.
Their study is based on the cost of raw materials as of 2014 for alloy production by different methods for the most common and the best performing PMs based on rare-Earths in €/Kg Fig.1 for comparison, and in the following figure the cost in €/Kg for the emerging substitutes . Among them is MnBi at a cost of 13 €/Kg.
Fig.49. Raw material prices of several RE-based PMs alloys (after P. McGuinness, 2015)
Fig. 50 RE-Free alloys estimated cost for prices No 2013-April 2014 (after P. McGuinness, 2015).
In red are the prices for MnBi and for Co-nanowires ( 7500 €/Kg and using different solvent (to replace the 1,2 butanediol by the 1,2 propanediol.) according to the INSAT in the framework of the REFREEPERMAG project.
Considering the performance of the MnBi magnets, WCM feels that improvement of performance can position MnBi next to Ferrite magnets at cost 3-20 €/Kg and similar characteristics.
2. MnBi-Bonded Magnets ( MFB)
Magnetic powders of MnBi were mixed with epoxy to achieve as high density as possible and were measured at MagnetFabrick . Typical curves obtained are shown in Fig. 10.3
Fig. 51 Typical BH-loops of MnBi-bonded magnets
In the table below the characteristics of various bonded magnets are given
Density
g/cm3 Br
kA/m Hcb
kA/m Hcj
kA/m (BH)max
kJ/m3
MnBi-22 6,898 0,151 100,7 539,7 3,83
MnBi-11-b3 6,690 0,163 108,5 884,9 4,17
MnBi-11 6,584 0,133 91,88 841,1 3,06
MnBi-p06 6,288 0,133 96,87 777 3,22
Result:
Polymer bonded magnets have been prepared from different powders delivered by the consortium partners. As an interesting shape for PM motors and acuators a disc shaped magnet have been choosen with diameter 12mm in different length. Demagnetisation curves have been measured on the bonded magnets with a permagraph from Magnetphysik Steingroever. All samples prepared showed permanent magnetic properties in accordance with the expectations from the powder properties.
Summary:
It could be shown that the production of a near net shape magnet is feasible with the powders coming from different production routes. The magnet production performed can be scaled up for a series process. Powder quantities received from partners were sufficient to produce epoxy bonded magnets. As expected from the beginning of the project the powder preparation possibilities of our partners could not supply sufficient quantities to produce magntes based on thermoplastic binders as our preproduction of compounds would require some Kg masses of powders.
However, it has been deduced from the tests and produced magnets based on thermosettings in a compaction molding route that also a injection molding process will be applicable.
Novel high-throughput techniques to search for novel magnetic materials with the appropriate properties for PM
Combinatorial and high-throughput (CHT) Materials Science and Engineering techniques represent a powerful approach to rapidly screen new compounds and accelerate materials development. CHT methods are based on: i) materials modelling using novel computational tools ii) fabrication of materials libraries (multi-element compositional spreads), iii) high-throughput and spatially resolved characterization techniques, and iv) data management and analysis for processing high volume data.
CHT methods reduce the cost and time of materials Research & Development, effectively addressing the pressing issues arising on Materials Science and Engineering. A major societal challenge with global implications, climate change, demands the development of clean energy and energy efficiency technologies. As emphasized by the U.S. Department of Energy , scientific breakthroughs in new materials will foster technological advancements providing sustainable energy solutions. CHT methods ensure accelerated materials development for coping with present-day environmental issues.
High-Throughput techniques were employed to exploit strain-induced epitaxial growth of distorted cubic to tetragonal high magnetization alloys, that resulted in the high anisotropy which is required for the materialization of novel permanent magnets e.g. Fe-Co based and Heusler alloys, leading to the identification and characterization of new magnetic multifunctional materials of scientific and technological interest.
IFW and RUB used high-throughput synthesis on various substrates following the Bain path and realizing a lateral variation of in-plane lattice parameters using combinatorial film deposition of epitaxial Cu-Au on a 4-inch Si wafer. This template gave the possibility to adjust the in-plane lattice parameter over a wide range from 0.365 nm up to 0.382 nm. Cu-Au buffer was used to stain and assist the epitaxial growth of FeCo magnetic layer. High-throughput X-ray diffraction (XRD) measurements were performed to determine the structure of the Cu-Au Bain library.
RUB used well developed combinatorial methodologies in order to fabricate and study strained wedge type Fe-Co on AuxCu1-x libraries. For the magnetic characterization of the wedge type FeCo libraries, a high-throughput magneto-optical Kerr magnetometer (MOKE – RUB facilities) was used. Rapid characterization techniques, which can automatically measure the whole material library for the specific properties of interest, where also used for the structural properties study. A third or fourth element addition was studied by co-sputtering along with FeCo, B, Cr or Cr-Ni. RUB also prepared wedge type layers of Fe-Co-Nb using a post-annealing procedure, resulting to a ternary library.
NCSR D employed combinatorial thin film synthesis in order to create libraries of Cu-Au buffer alloy, as well as of {Fe-Co}-C magnetic layer. In the first case a both tetragonal and cubic Cu-Au crystallographic phases were studied on a single 4 inch wafer using high-throughput structural characterization technique. Moreover a wide range of stoichiometries of Fe-Co alloy were produced and studied, along with the variation of Carbon percentage effect (up to 20% at.). For rapid high-throughput magnetic characterization, a 3d magnetic moment mapping system was developed.
High-Throughput combinatorial method was also used by NCSR D to screen possible compositions and synthesis conditions of Heusler alloys of the type X2YZ. Fe2CuGa was deposited on various substrates at different deposition temperature. A great amount of variable alloy stoichiometries were produced and studied structurally using rapid characterization combinatorial XRD on a Rigaku system. On a 2 inch Si wafer different crystallographic orientations were observed in one step procedure.
Our project is one of the very few in the word to use this approach for screening a) Theoretically and b) experimentally the search for new magnetic materials for permanent magnets and is well cited in meetings all-over the world.
After 36 months (the completion of the project), the exploitation activities will be continued aiming at manufacturing & marketing of the products in a cost-effective way. Support from Eureka, CRAFT and local schemes will be considered, encouraging a bottom-up approach to technological development and strengthening competitive position of project companies on the world market. SMEs would also benefit from loans from the European Investment Bank (EIB) or national/regional sources may be exploited for further commercialisation, in addition to the partners’ own investments. Such strategy of financial engineering is expected to ensure the sustainability of the activities and further improvement of the cooperation in the European Research Area.
Potential Impact:
The potential impact (including the socio-economic impact and the wider societal implications of the project so far) and the main dissemination activities and exploitation of results (not exceeding 10 pages).
Exploitation
On completion of the project, the following main deliverables deserve special attention for exploitation
• System Software
• New improved alloys
• New high performance grades of finished PM’s
• New processes or materials (e.g. nanocomposites) for production of PM’s.
• Novel high-throughput techniques to search for novel magnetic materials with the appropriate properties for PM
→ System Software
The partners TUW, UPPSALA, CSIC, INSAT and CEA have developed software packages that are suitable for exploitation.
The software packages used are WIEN2k, Elk, muffin-tin orbitals (EMTO) pseudopotential approaches with plane waves (VASP) or numerical local orbitals (SIESTA, OpenMX) for DFT calculations. Most of these codes provide MPI-parallel versions suitable for large-scale computations. MAGPAR, FEMME will be used for micromagnetics together with finite element discretisation codes (GID, Salome).
UPPSALA is concentrated on combinatorial calculations using standard codes as modules and had successfully TUW, CSIC and INSAT and CEA for micromagnetic simulations. Their software packages present an excellent opportunity for exploitation, but also as an important issue and strength of the partner in future submissions to the EU.
→ New improved alloys (NCSR D, INSAT, RUB, IFW, TUDA)
a) Chemical and electrochemical approach for the preparation of nanorods, nanowires of the type Co, Fe-Co-X
b) Bulk synthesis (melt spinning, ball milling, arc-melting, compaction)
→ New high performance grades of finished PM’s
Two classes of finished PMs were developed along the duration of the REFREEPERMAG projec
• Co-based nanowire-nanorod PMs by blending and compaction
• MnBi- based alloys by cold compaction and sintering
• Co-based nanowire-nanorod PMs by blending and compaction
We have achieved the benchmark set in the proposal of (BH)max = 160 KJ/m3 in compacted Co-nanorods with an aspect ratio of ~ 5. Magnets have been prepared by INSAT by blending with epoxy and by the compaction methods developed by TUDA.
We are among the first groups worldwide to prepare and consolidate nanowires-nanorods into macroscopic permanent magnets with energy so high energy products, indicating that our consortium had achieved one of its main objectives.
• MnBi- based alloys by cold compaction and sintering
We have succeeded to fabricate Bulk PMs with very good properties and energy product in sintered (NCSR D) and epoxy bonded magnets (MagnetFabrick). The energy product for the sintered magnet was ~ 40 KJ/m3 and the epoxy bonded ~ 5 KJ/m3. A degradation of the original magnetic properties was observed during sintering, but the final PM was good enough to be used for incorporation in a micromotor (WCM) and run a test performance.
The test performance has given the following results compared to the currently used bonded NdFeB magnets
MnBi
N= 737 rpm Vout =0,738 V N= 1000 rpm V out = 1,001 V
NdFeB
N= 667 V out = 2,717 V N=1000 rpm V out =4,073 V
Given that the output Voltage of the motor is proportional to the flux Φ = (Β.S) and from the measured values of both original NdFeB and our MnBi the voltage output of our- non optimized MnBi magnets- scales as ¼ of the voltage output of the NdFeB magnets ( 1.1 T is the remanence magnetization for NdFeB compared to the ~ 0,35 T for our sintered magnets). There is room for improvement of the sintering process of the MnBi PMs in order to preserve the very good properties of the powder ( Br ~ 60 emu/gr and Hc ~ 1 T).
Worth to mention here, as far as we know, we are the first consortium in the world to fabricate and test in “real conditions” MnBi based PMs with very promising results.
Economics of PMs
a) Co-based nanowire-nanorod PMs by blending and compaction
There are three preparation routes for the polyol mediated synthesis of cobalt nanorods:
1. Microwave oven – which offers the possibility to heat quickly a large volume of polyol and experiment within a big range of heating rates. High heating rates favors the formation of cobalt nanorods with small diameter, presenting the best magnetic properties. However, the strong drawback of the microwave is the heterogeneous heating that creates a metallic mirror on the flasks of the spherical glassware. The particles formed in the bulk of the suspension were found homogeneous with interesting magnetic properties, but the particles formed on the walls exhibited poor magnetic properties.
2. Heating mantle – which offers a better homogeneity of temperature inside the flask leading to
monodispersed cobalt nanorods. This set-up provided nanorods with the best magnetic
properties. Its limitation is the narrow range of the heating rate that can’t exceed 10 °C/min.
3. Jacket reactor – is expected to combine all the advantages of the above: a very homogeneous
temperature inside the solution, a good stirring and a high heating rate. The efficiency,
scalability and reproducibility of the first two approaches were examined.
These three set-ups are used at industrial scale to produce chemicals and including inorganic particles. In the table 1 the prices of the different chemicals used in the different steps of the “standard” synthesis of cobalt nanorods at the laboratory scale are given. Of course these prices are much higher than that expected for an industrial supply. Savings can be made on the cobalt precursor. The cobalt acetate is clearly not the cheapest precursor. Nevertheless it is noteworthy that 80% of the price comes from the butanediol.
Chemical reactions
The synthesis of the cobalt nanorods involves three steps:
1. The synthesis of the sodium laurate from lauric acid and sodium hydroxide
2. The cobalt laurate synthesis from cobalt chloride and cobalt acetate
3. The cobalt nanorods synthesis from the reduction of the cobalt laurate in a sodium hydroxide
solution in 1,2 butanediol containing 2.5 % of ruthenium chloride.
In the table the prices of the different chemicals used in the different steps of the “standard” synthesis of cobalt nanorods at the laboratory scale are summarized. Of course these prices are much higher than that expected for an industrial supply. Savings can be made on the cobalt precursor. The cobalt acetate is clearly not the cheapest precursor. Nevertheless it is noteworthy that 80% of the price comes from the butanediol.
The concentration of cobalt in the “standard” synthesis is 0.08 mol.L-1 that allows the synthesis of 5 g in one batch. Current price excluding personnel cost is 7€/g. Can be reduced to 2.5 €/g.
Several syntheses were done in the microwave oven with higher cobalt concentrations: 0.12 M and 0.24M. In every case the anhydrous ruthenium chloride was used as nucleating agent. These concentrations also showed that it is possible to prepare anisotropic cobalt particles similar to those prepared with the standard protocol at reduced cost.
In conclusion nanoparticles of Co- cost very much and we are in doubt that this technique will be used for mass production of permanent magnets (see tables below). Only niche markets are suitable for exploitation.
b) 1. MnBi sintered- based PMs
In order to evaluate the possibility for exploitation of PMs based on MnBi- alloys we show a recent study in ROMEO project funded by EU by P. McGuinness et al, considering that from the producer and consumer point of view, the cost of a permanent magnet with certain magnetic properties is one of the most important factors.
Their study is based on the cost of raw materials as of 2014 for alloy production by different methods for the most common and the best performing PMs based on rare-Earths in €/Kg Fig.1 for comparison, and in the following figure the cost in €/Kg for the emerging substitutes . Among them is MnBi at a cost of 13 €/Kg.
Fig.49. Raw material prices of several RE-based PMs alloys (after P. McGuinness, 2015)
Fig. 50 RE-Free alloys estimated cost for prices No 2013-April 2014 (after P. McGuinness, 2015).
In red are the prices for MnBi and for Co-nanowires ( 7500 €/Kg and using different solvent (to replace the 1,2 butanediol by the 1,2 propanediol.) according to the INSAT in the framework of the REFREEPERMAG project.
Considering the performance of the MnBi magnets, WCM feels that improvement of performance can position MnBi next to Ferrite magnets at cost 3-20 €/Kg and similar characteristics.
2. MnBi-Bonded Magnets ( MFB)
Magnetic powders of MnBi were mixed with epoxy to achieve as high density as possible and were measured at MagnetFabrick . Typical curves obtained are shown in Fig. 10.3
Fig. 51 Typical BH-loops of MnBi-bonded magnets
In the table below the characteristics of various bonded magnets are given
Density
g/cm3 Br
kA/m Hcb
kA/m Hcj
kA/m (BH)max
kJ/m3
MnBi-22 6,898 0,151 100,7 539,7 3,83
MnBi-11-b3 6,690 0,163 108,5 884,9 4,17
MnBi-11 6,584 0,133 91,88 841,1 3,06
MnBi-p06 6,288 0,133 96,87 777 3,22
Result:
Polymer bonded magnets have been prepared from different powders delivered by the consortium partners. As an interesting shape for PM motors and acuators a disc shaped magnet have been choosen with diameter 12mm in different length. Demagnetisation curves have been measured on the bonded magnets with a permagraph from Magnetphysik Steingroever. All samples prepared showed permanent magnetic properties in accordance with the expectations from the powder properties.
Summary:
It could be shown that the production of a near net shape magnet is feasible with the powders coming from different production routes. The magnet production performed can be scaled up for a series process. Powder quantities received from partners were sufficient to produce epoxy bonded magnets. As expected from the beginning of the project the powder preparation possibilities of our partners could not supply sufficient quantities to produce magntes based on thermoplastic binders as our preproduction of compounds would require some Kg masses of powders.
However, it has been deduced from the tests and produced magnets based on thermosettings in a compaction molding route that also a injection molding process will be applicable.
• Novel high-throughput techniques to search for novel magnetic materials with the appropriate properties for PM
Combinatorial and high-throughput (CHT) Materials Science and Engineering techniques represent a powerful approach to rapidly screen new compounds and accelerate materials development. CHT methods are based on: i) materials modelling using novel computational tools ii) fabrication of materials libraries (multi-element compositional spreads), iii) high-throughput and spatially resolved characterization techniques, and iv) data management and analysis for processing high volume data.
CHT methods reduce the cost and time of materials Research & Development, effectively addressing the pressing issues arising on Materials Science and Engineering. A major societal challenge with global implications, climate change, demands the development of clean energy and energy efficiency technologies. As emphasized by the U.S. Department of Energy , scientific breakthroughs in new materials will foster technological advancements providing sustainable energy solutions. CHT methods ensure accelerated materials development for coping with present-day environmental issues.
High-Throughput techniques were employed to exploit strain-induced epitaxial growth of distorted cubic to tetragonal high magnetization alloys, that resulted in the high anisotropy which is required for the materialization of novel permanent magnets e.g. Fe-Co based and Heusler alloys, leading to the identification and characterization of new magnetic multifunctional materials of scientific and technological interest.
IFW and RUB used high-throughput synthesis on various substrates following the Bain path and realizing a lateral variation of in-plane lattice parameters using combinatorial film deposition of epitaxial Cu-Au on a 4-inch Si wafer. This template gave the possibility to adjust the in-plane lattice parameter over a wide range from 0.365 nm up to 0.382 nm. Cu-Au buffer was used to stain and assist the epitaxial growth of FeCo magnetic layer. High-throughput X-ray diffraction (XRD) measurements were performed to determine the structure of the Cu-Au Bain library.
RUB used well developed combinatorial methodologies in order to fabricate and study strained wedge type Fe-Co on AuxCu1-x libraries. For the magnetic characterization of the wedge type FeCo libraries, a high-throughput magneto-optical Kerr magnetometer (MOKE – RUB facilities) was used. Rapid characterization techniques, which can automatically measure the whole material library for the specific properties of interest, where also used for the structural properties study. A third or fourth element addition was studied by co-sputtering along with FeCo, B, Cr or Cr-Ni. RUB also prepared wedge type layers of Fe-Co-Nb using a post-annealing procedure, resulting to a ternary library.
NCSR D employed combinatorial thin film synthesis in order to create libraries of Cu-Au buffer alloy, as well as of {Fe-Co}-C magnetic layer. In the first case a both tetragonal and cubic Cu-Au crystallographic phases were studied on a single 4 inch wafer using high-throughput structural characterization technique. Moreover a wide range of stoichiometries of Fe-Co alloy were produced and studied, along with the variation of Carbon percentage effect (up to 20% at.). For rapid high-throughput magnetic characterization, a 3d magnetic moment mapping system was developed.
High-Throughput combinatorial method was also used by NCSR D to screen possible compositions and synthesis conditions of Heusler alloys of the type X2YZ. Fe2CuGa was deposited on various substrates at different deposition temperature. A great amount of variable alloy stoichiometries were produced and studied structurally using rapid characterization combinatorial XRD on a Rigaku system. On a 2 inch Si wafer different crystallographic orientations were observed in one step procedure.
Our project is one of the very few in the word to use this approach for screening a) Theoretically and b) experimentally the search for new magnetic materials for permanent magnets and is well cited in meetings all-over the world.
After 36 months (the completion of the project), the exploitation activities will be continued aiming at manufacturing & marketing of the products in a cost-effective way. Support from Eureka, CRAFT and local schemes will be considered, encouraging a bottom-up approach to technological development and strengthening competitive position of project companies on the world market. SMEs would also benefit from loans from the European Investment Bank (EIB) or national/regional sources may be exploited for further commercialisation, in addition to the partners’ own investments. Such strategy of financial engineering is expected to ensure the sustainability of the activities and further improvement of the cooperation in the European Research Area.
Socioeconomic Impact
In a nutshell we moved from technology readiness level 2 (TRL 2, see table) to TRL 4..
Table 1- Extract from Part 18 - Commission Decision C(2013)8631 : Technology Readiness Level
TRL1 basic principles observed
-----------------------------------------------------------------------------------------------------------------------------------
TRL2 technology concept formulated REFREEPERMAG starting point
TRL3 experimental proof of concept
TRL4 technology validated in lab REFREEPERMAG ending
TRL5 technology validated in relevant environment (industriallyrelevant environment in the case of key enabling technologies)
TRL6 technology demonstrated in relevant environment (industrially relevant environment in the case of key enabling technologies)
-----------------------------------------------------------------------------------------------------------------------------------
TRL7 system prototype demonstration in operational environment
TRL8 system complete and qualified
TRL9 actual system proven in operational environment (competitive
Manufacturing in the case of enabling technologies; or in space
New Ab-initio calculation scheme: REFREEPERMAG project delivered a series of ab-initio simulations and micromagnetic calculations for the discovery of novel PMs without rare earths. Further theREFREEPERMAG project has successfully demonstrated that the combinatorial approach is the most suitable for the rapid search of novel materials meeting predefined criteria.
New materials: REFREEPERMAG project produced a plethora of FeCo-X-Y materials in the form of nanowires/nanorods, thin films and bulk that possess the required properties as candidate PMs.
New Infrastructure for combinatorial and specialized measurements:
The project developed new Combinatorial and specialized methods with sufficient accuracy to differentiate between materials with different compositions , structure , intrinsic properties for the search of novel magnetic phases. Namely a novel Combi-Kerr, a Combi-XRD, a 3D-Magscan for magnetic field mapping, a ultrasensitive FMR, a SANS (Small Angle Newtron Scattering) for aligned nanowires/nanorods magnets and many more , that eventually can be exploited, paving the way of a fast high-trroughput characterization approach.
New commercial knowledge: Both within Europe and elsewhere a number of parties are searching for rare-earth free permanent magnets ( ARPA-E in USA, JAPAN coordinated by Dr. S. Hirosawa 10-year due to the economics involved wwith respect to energy savings, Electric vehicles and wind energy. We have demonstrated that there is a fast-pathway to explore the landscape for nonel magnetic materials for PMs. We have demonstrated that even nanowires/nanorods , through chemical synthesis and upscaling , can be commercialized for specific applications. The knowledge developed for nanowires can have spill-overs to other sectors like LEDs, since the next generation of LEDs will be based on nanowires of GaN-type and on quantum dots, because the consume less energy. The attempt to study in-depth the Mn-based systems and the fabrication of MnBi PMs , for demonstration only, showns that a commercial knowledge has been developed as well. Who will exploit this knowledge is a matter of further discussions.
Environmental impact: In the car industry there is actually a huge focus on reducing CO2 emissions. This focus is basically generated by new laws and the need of a responsible and sustainable use of energy sources. However, most of the measurements which are analyzed actually have a high cost, which avoids the introduction of such new technology. Also, the carbon footprint of every new technology has to be considered. The main measure which is taken in count today to quantify the relation between CO2 reduction and their costs i.e. Euros per gram of CO2. In reality, most of the measures are currently around 0 and 150€ per gram of CO2.
Considering that a 5 % improvement in (BH)max or equivalently a reduction in price of the raw materials compared to the currently used rare-earth based materials, will facilitate the replacement of existing magnets and will contribute , along with other technologies, to a better and cleaner environment.
The REFREEPERMAG project simply was essential in order to be able to quantify the potential of the technology and to identify the biggest issues, which have to be resolved in order to get a novel PMs .All partners intend to continue the exploitation of these new capabilities to further develop their business.
List of Websites:
The address of the project public website, if applicable as well as relevant contact details.
http://refreepermag-fp7.eu/
Furthermore, project logo, diagrams or photographs illustrating and promoting the work of the project (including videos, etc...) as well as the list of all beneficiaries with the corresponding contact names can be submitted without any restriction.
