Skip to main content
European Commission logo print header

Solid State Energy Efficient Cooling

Final Report Summary - SSEEC (Solid state energy efficient cooling)

Executive summary:

SSEEC (Solid State Energy Efficient Cooling) has been a 36-month project dedicated to the development of magnetic refrigerant materials and the technology required to integrate them in an end-user application: a heat pump. The aim has been to lower the economic barrier to entry of magnetic cooling as a high efficiency cooling technology. Conventional cooling relies on the expansion and compression of a volatile liquid, typically an HFC. It therefore harnesses the latent heat that is evolved during a gas/liquid transformation, and is driven by changes in applied pressure. However, the volatility of HFCs is one factor that drives research into solid state refrigerants. A recent study predicted that HFCs could account for between 28 and 45 percent (CO2-equivalent basis) of projected global CO2 emissions by 2050.

Magnetic cooling is a solid state cooling method that promises high system efficiency. It utilises the temperature change that occurs when a magnetic material is driven through a change of state by an applied magnetic field. Magnetic refrigerants have magnetic phase transitions around room temperature, just as HFCs evaporate readily at room temperature.

Our project has been the first of its kind to examine all stages in the production of a magnetic cooling application solution, from the synthesis, fabrication and shaping of refrigerants, to measurement of their physical properties using the latest bespoke characterisation tools and the integration of refrigerants in cooling engines designed specifically according to models that optimise cost and efficiency. This technology chain has been iterated with three prototype cooling engines, the ultimate being an integrated heat pump system. We have produced a roadmap for magnetic cooling technology, addressing the domestic refrigeration sector, and also suggested further necessary steps in exploring: material processing; fundamental magnetism; standardisation of refrigerant characterisation and more unusual magnetic cooling effects in the future. Our major achievements in SSEEC are now listed. We have successfully:

Single phase magnetic refrigerants and novel cooling mechanisms
- Produced single phase refrigerants in small thicknesses suitable for heat exchange, with tuned transition temperatures
- Developed new refrigerant synthesis processes to address refrigerant machinability
- Investigated the fundamental magnetism of metamagnets that improve our understanding of novel magnetic cooling mechanisms based on magnetic order-order transitions.

Novel cooling mechanisms
- Investigated and syntheised micro- and nano-scale refrigerants that harness anisotropy rather than the degree of magnetic ordering
- Synthesised single crystal anisotropic refrigerants and textured polycrystalline materials.

Advanced characterisation
- Provided new plots to provide a simple means by which to compare refrigerant performance.
- Conducted round-robin activities that highlight the importance of sample preparation and comparative measurement techniques. These have also verified the accuracy of our methods.
- Developed advanced characterisation tools across the partnership which have been integrated into our materials development work, enabling greater flow of ideas and materials between partners.

Theoretical modelling
- Modelled the static, thermodyamic properties of MCE materials with continuous phase transitions
- Predicted of the effect of dynamics in sharp first order transitions, defining a range for the increase in magnetic field requirement, depending on transformation kinetics
- Established new magnetic cooling mechanisms due to 'artificial spin reorientation' in a hard ferromagnetic multilayer system

- Constructed 3 prototype magnetic cooling engines using Co-doped La-Fe-Si
- Met temperature span targets set for prototypes I and II
- Integrated with heat exchangers to build a heat pump in prototype III
- Developed a technology roadmap for magnetic cooling (see Section 4.1)

We hope that our results will provide the springboard for European development of magnetic cooling as a high efficiency, green and cost effective form of cooling.

Project Context and Objectives:

2. Project context and objectives
2.1 Context
The ultimate goal of our project was to lower the economic barrier to entry of magnetic cooling as a high efficiency cooling technology. Our efforts in this direction were centred around the development of a heat pump based on a highly efficient magnetic refrigeration cycle. The devices feature a magnetic refrigerant being magnetised and demagnetised by a permanent magnet.

Magnetic cooling: basic formalism
Our project has been the first of its kind to integrate all elements in the production of a magnetic cooling engine. Magnetic refrigerants have been synthesised, shaped, characterised, modelled and three prototype cooling engines have been built using the optimal materials that we have produced. For the description of the scientific and technological highlights that follow, it is useful to provide a brief overview of the evaluation of magnetic refrigerant materials and the nature of the magnetic cooling cycle used in prototypes.

How we evaluate magnetic refrigerants
In what follows we use the term 'magnetic refrigerant' and 'magnetocaloric' interchangeably. This terminology is because a magnetic refrigerant relies upon the magnetocaloric effect (MCE), the change of temperature that a material undergoes when exposed to a change in applied magnetic field. In non-magnetic materials it is very small (hundredths of a Kelvin) but in materials that have a magnetic phase transition it can be as large as 3-4 K in a 1 Tesla field – a field easily achievable with a permanent magnet. The applied field can trigger the change of magnetic phase, and with it bring about a release or uptake of heat by/from the material. The amount of heat exchanged with the surroundings can be sufficient to build a cooling engine; in the best MCE materials, around 1 kW cooling power per kilogram of refrigerant might be expected.

What we require of a magnetic refrigerant

Magnetic refrigerants therefore need to have the following properties:
1. A phase transition temperature that is tunable over the range of temperature required, so that the useful MCE response of the refrigerant is large over the required application temperature range. The phase transition has to be able to be triggered by an applied field.
2. Large heat release/uptake or temperature change when the field is applied (depending on the conditions of applying the field)
3. Non toxic, inexpensive, and shapeable

These conditions imply a set of more complicated requirements for magnetic properties, heat capacity and even thermal conductivity in application. However, the primary concern is that, when the magnetic field is applied condition 2 is satisfied. This means that:
- When the magnetic field is applied isothermally (at a constant temperature) a large heat is released or taken in by the material (approximately 10-20 J kg-1K-1). We quantify this in terms of isothermal entropy change, DeltaS.
- When the magnetic field is applied adiabaticaly (at constant entropy, or under conditions of no heat exchange with the material) a large temperature change is observed (approximately 2 Kelvin in a 1 Tesla field change). We call this quantity the adiabatic temperature change, DeltaTad.

How to make a magnetic cooling engine: the thermodynamic cycle
If a material satisfies the above requirements it can potentially be used in a magnetic cooling device. The magnetocaloric material is magnetised, resulting in a temperature change. Then heat can be removed by a secondary heat exchange fluid, and released by a radiator, such as exists at the back of a domestic refrigerator. On removal of the applied magnetic field, the material temperature changes, but now it is lower than the original temperature since heat was already removed to the radiator in the previous step. Therefore heat can be put into the material from the object(s) to be cooled, using another heat exchange fluid. This finally raises the temperature of the material back to the starting value, or close to it. The cycle is analagous to that of conventional gas compression, with the application and removal of a magnetic field corresponding to the compression and evaporation strokes respectively.

How to build a temperature span: the regnerator
If the change of temperature of a material on application of a 1 Tesla field is 2 Kelvin, how do we build a device that can span the 30-40 K that might exist between the hot end and cold end of an application? The answer lies in the use of a regenerator. This component allows different parts of a refrigerant block or stack to operate around different starting temperatures, thus meaning that a temperature gradient can be built along the length of the refrigerant. Since the MCE response of a magnetic refrigerant is not limited to a single temperature, but is instead spread over a range of temperatures around the phase transition temperature, heat can be exchanged with the refrigerant over a range of temperature along the length of an individual composition. If a wider temperature range of response is required, different material compositions are added. Composition typically shifts the transition temperature, and with it, the window of response of the magnetocaloric material.

In light of the above we often display plots of either isothermal entropy change (DeltaS) or adiabatic temperature change (DeltaTad) as a function of temperature for several compositions in a series. These plots inform how composition affects the functionality of a material via how the MCE response varies over a temperature range.

In summary, magnetic refrigerants and permanent magnets are the two essential materials components of a magnetic cooling engine. The SSEEC project was structured not only to innovate in fundamental materials research, producing new magnetic refrigerant nano-architectures, but also to pull through from those materials fundamentals to delivery of a technology where considerable improvements in energy efficiency in the cooling and heat pumping markets can be offered. In order to achieve these objectives, the consortium was made of four materials research institutions, an SME capable of developing prototype cooling engines, a medium scale materials manufacturer that is able to design and produce magnetocaloric materials as well as permanent magnets and a major systems end-user that provides industrially-guided feedback on system design and performance.

2.2 Objectives
Our project was supported under NMP-2007-2.2-3 Advanced material architectures for energy conversion. As outlined above, the magnetic cooling engine is a high efficiency energy converter for the conversion of mechanical work into cooling power, and represents the only solid state cooling technology capable of achieving a system efficiency greater than that provided by conventional gas compression technology. Although already used in research to achieve milliKelvin temperatures, economic viability of magnetic cooling engines for future room temperature applications will rely on two key materials factors:

(1) The availability of low cost, low hysteresis magnetic refrigerants, and
(2) The ability of these refrigerants to provide substantial cooling power in magnetising fields low enough to be provided by permanent, rather than superconducting- or electro-magnets;

and two cooling engine factors:
(3) Optimisation of heat exchange with the refrigerant, and
(4) Minimisation of refrigerant and permanent magnet volume

Before SSEEC started, prototype magnetic cooling engines were limited by all four factors, and relied on expensive refrigerants in a large and heavy cooling engine design. We aimed for SSEEC to produce refrigerant material architectures with new compositions, new morphologies and new underlying physical mechanisms (addressing items 1 and 2 above), whilst reducing the necessary volume of refrigerant and thereby both the cost of refrigerant and permanent magnet components in a cooling engine (item 4). Optimisation of heat exchange with the refrigerant (item 3) was to be an integral component of the prototype design. Refrigerant costs would also be reduced by improved synthesis processes. We thus aimed to lower the economic barrier to entry for high efficiency magnetic cooling.

Our materials methodology harnessed the combined expertise of the partners in materials synthesis, modeling and fundamental characterisation to produce a series of improved single phase refrigerant materials with a focus guided by items (3) and (4) above. Importantly, we also set out to explore a completely new avenue of magnetic refrigerant research – anisotropic nano-composite refrigerants - in an effort to widen the range of magnetic refrigerant materials open to the research and development community. The family of single-phase materials was to be implemented in a series of prototype heat pumps by our industrial partners. We specifically addressed item (4) above by exploring of prototype operation at high cycle frequency – a world first.

In summary, our scientific objectives were therefore:
a) To synthesise, characterise and optimise a family of improved single phase magnetic refrigerant materials
b) To model and develop a family of new anisotropic nano-composite refrigerants;
c) To design and produce low cost permanent magnet arrays;
d) To integrate the materials from (a), (b) and (c) into viable prototype heat pumps and air conditioners operating at high cycle frequency.

By the project completion date we assessed the material architectures developed within the project and integrated them with systems components such as the heat exchangers, regenerators and magnetic arrays required to produce an efficient heat pump. We also created a roadmap for the future deployment of magnetic cooling technology.

2.3 Structure
Having outlined the basic principles of MCE materials and their use, we move in the next section to a presentation of research highlights, organised in work package (WP) order. For clarity, the WPs were:

WP1: Synthesis of single phase La-Fe-Si and Co-Mn-Si materials
WP2: Exploiting magnetic anisotropy energy (intrinsic and artificial spin reorientation)
WP3: Advanced characterisation of materials
WP4: Theoretical modelling
WP5: Prototyping (the production of three prototypes labelled I, II and III)
WP6: Management

Project Results:
3. Main Science and Technology (S &T) results and foregrounds
3.1 WP1: Synthesis of single phase La-Fe-Si and Co-Mn-Si materials

- To establish economically viable routes to fabrication of optimised single phase magnetic refrigerant materials with workable entropy changes, low hysteresis and tunable critical temperatures
- Elimination of significant volume of parasitic second phases that dilute useful properties within the working volume of material
- Routine characterisation of these materials in terms of structure (XRD,SEM) and magnetisation M(T,H).

Summary of activities
Our efforts here have focussed on a better understanding of two families of single-phase alloys, LaFeSi and CoMnSi, that have magnetic field-induced phase transitions around room temperature.

The LaFeSi material system is one that promises to be cost effective and deliver significant cooling power. Before we started SSEEC it had not been used in an end user device. Our project examined all steps of material synthesis, fabrication and shaping required to bring about this goal. We have deployed LaFeSi in all of our prototype cooling engines. The activities of the two principle partners in this effort, IFW and VAC, are detailed in section 3.1.1. We synthesised this material by two different routes – powder metallurgy and melt-spinning – and examined the dependence of its magnetic properties on synthesis conditions, machining, and microstructure. In addition we examined the material's thermal and structural properties. The importance of thermal properties to magnetic cooling is relatively obvious; structural properties perhaps less so. However, in many potential magnetic refrigerants there is a change of volume associated with the magnetic field-induced change of refrigerant entropy. In a solid such changes of volume can lead to irreversibilities, cracking, and even in difficulties in machining the material if the associated magnetic transition is around room temperature, which it has to be for our applications.

A variety of synthesis routes and conditions have been tested on the La-Fe-Si system. This system contains an abrupt, 'first order' magnetic transition that can be tuned to room temperature by the addition of cobalt (which reduces the first order nature) or by the absorption of hydrogen (which maintains the first order nature and leads to a larger MCE). In the first 18 months Co-doped materials with smooth, continuous magnetic transitions have been studied from the point of view of applications. These have the advantage of good mechanical and machining properties, but are not the materials in the La-Fe-Si with the highest MCEs. In the second 18 months, the focus shifted to the role of hydrogenation. Fully hydrogenated La-Fe-Si has a Curie temperature well above room temperature. Partially hydrogenated material can have a Curie temperature in the useful range for room temperature applications (-20°C to 40°C) but can be unstable with respect to either hydrogen redistribution within the material or to dehydrogenation at elevated temperatures.

The partners, led in this effort by IFW and VAC built an understanding of the mechanisms of dehydrogenation, and of phase (in)stability in the ternary (La,Fe,Si) phase diagram. Importantly, substitutants were explored that allow the first order nature and high MCE of fully hydrogenated material to be harnessed at room temperature, by suppressing the Curie temperature of the non-hydrogenated starting compound by an appropriate amount.

Material machinability was explored so that structural changes did not interfere with the end product. Two new processes were developed and used by VAC: (i) the thermal decomposition and recombination (TDR) process for the production of plates of Co-doped material and (ii) a solid hydrogenation process (SH) used to hydrogenate materials in their desired shape, by avoiding cracking during hydrogenation. We have thereby examined link between the magnitude of structural changes and magnetothermal properties. We also successfully made materials with various Curie temperatures, thus extending the temperature range of the magnetocaloric effect compared to a material of a single composition. The delivery of single phase, cost-effective materials to span the temperature ranges of all three prototypes has thus been achieved, although the mechanical tolerances of these materials are a problem that is addressed in section 3.5.2 below. A range of analyses of these materials was undertaken, to establish the relation between magnetic, structural, compositional and magnetocaloric properties, and to explore the opportunities for optimisation of the latter.

Our second material family, based on Co-Mn-Si is used as a tool for gaining knowledge about the mechanisms of or driving forces behind 'metamagnetic' changes of state. These changes are less frequently observed in magnetic materials but can have large entropic effects associated with them. In this system, rather than a magnetic order-disorder (Curie), a transition between different types of magnetic order is studied. While this system shows less commercial promise, it has given rise to a number of fundamental insights in the metamagnetism of Mn-based alloys of the same (Pnma) space group and has resulted in three high profile publications. This work, performed mainly by Cambridge and Imperial, is detailed in section 3.1.2.

3.1.1 Highlights of La-Fe-Si activities
The highlights presented in this section are arranged under three headings:
- Preparation of single phase material with Curie transition tuned by Co, H and Mn
- Stability of hydrogenated materials and its improvement with carbon
- Machinability of materials improved by two new processes

Preparation of single phase material with Curie transition tuned by Co, H and Mn

La(Fe,Si)13 is an interesting MCE material because its magnetic transition – a Curie transition, Tc, at which temperature the material magnetically disorders – is sharp, or 'first order'. This means that, on application of a magnetic field, a large coordinated jump in entropy is observed as macroscopic magnetic ordering is induced. Without the addition of other elements, the Curie temperature, and therefore the temperature range of maximum MCE is about 190 K. In order to optimise the material for use as a room temperature magnetic refrigerant, its Curie temperature must be raised.

The two most common ways are either to use Co-substitution for Fe, or by introducing interstitial hydrogen. Whereas the former is associated with a significant decrease of the isothermal entropy change the latter yields alloys of the general composition LaFe13-xSixHz which, for z approximately 1.6 have a Curie temperature of about 350 K and a large isothermal entropy change of about 20 J/kgK for a magnetic field change of 1.5 Tesla. In principle it is possible to vary the hydrogen content, z, and obtain any Curie temperature between 190 K and 350 K while retaining the large isothermal entropy change.

Before considering the action of either method of Curie temperature tuning in detail, we first present results aimed at isolating single phase La-Fe-Si. Normally this material is metastable, forming from the melt by a peritectic reaction, and contains a parasitic alpha-Fe second phase, which does not contribute to the MCE. IFW has compared the structure and properties of melt-spun materials with those produced by conventional means (e.g. arc-melting and induction melting). The annealing temperature is of crucial importance for maximising the amount of 1:13 phase in LaFe13-xSix. However, little knowledge about the ternary phase diagram of La(Fe,Si)13 is available. The first aim was to optimize the annealing temperature for La(Fe,Si)13 bulk alloys with a first-order Curie transition.

Stability of hydrogenated materials and its improvement with carbon (IFW)
As explained above, one way to tune the Curie temperature of LaFe13-xSix alloys is to partially hydrogenate them and obtain transition temperatures between 190 K and 350 K. While this route is attractive because of the large entropy change of such alloys it has been found that magnetocaloric properties of partially hydrogenated LaFe13-xSixHz are not stable. Partially hydrogenated LaFe13-xSixHz alloys exhibit a phase separation effect by which a material with one well defined Curie temperature degrades into a material with two broader transitions if stored at the Curie temperature. Details of the mechanism of this 'peak splitting' are unknown but it is very likely that a partial hydrogen concentration causes the problem. This assumption is reasonable because alloys which are fully saturated with hydrogen do not exhibit 'peak splitting'.

We studied the use of Carbon (C) in combination with H to improve the thermal stability of our alloys. Based on microscopy and magnetisation data (below) we took the most promising C-containing samples – namely x = 0.1 and 0.2 with seven days annealing – and fully hydrided them. The endothermic peak shows that desorption of H starts at 460 K for the carbon-free alloy (black curve). In the carbon-containing samples, the desorption temperature is increased to 500 K (x = 0.1) and 540 K (x = 0.2). This shows that C stabilizes the H in LaFe11.6Si1.4CxHy alloys and is especially important as the materials should be stable over extended periods of time; even small losses in H content translate to large decrease in the working temperature as set by Tc.

Machinability of materials and its improvement via two new processes (VAC)

One characteristic of the magnetic phase transition in La-Fe-Si is the large magnetovolume effect associated with the change of the magnetic state. During machining of La-Fe-Co-Si, local heating leads to a change of the magnetic state in a small volume of the material and the material cracks. In order to be able to machine La-Fe-Co-Si with Curie temperatures around room temperature a new process was developed.

For this process the instability of the magnetocalorically active 1:13 phase was utilized. If La-Fe-Co-Si is annealed at temperatures below about 1000°C the 1:13 phase decomposes into alpha-Fe and La-rich phases. These phases do not exhibit anomalous thermal expansion and hence machining of the material is possible in this magnetocalorically inactive state. In this state the material contains up to about 75% alpha-Fe. After machining in the decomposed state the alpha-Fe reacts with the La-rich phases to form the desired 1:13 phase again. This homogenization process is conducted at about 1050°C. The material is then rapidly cooled to room temperature to preserve the magnetocalorically active state. The whole process is called Thermal Decomposition and Recombination (TDR).

We also developed a way to avoid cracking during hydrogenation. The key is the temperature at which hydrogen is offered to react with the alloy (the starting temperature). A second important factor is the cooling from the hydrogenation temperature to room temperature. Tests that were carried out show that disintegration can be avoided if hydrogenation of parts is conducted by first heating to 500°C under an inert atmosphere, e.g. argon. Then the inert gas has to be replaced with hydrogen followed by a dwell at 500°C for about one hour. The parts have to be cooled down to room temperature slowly (about eight to ten hours). From there it is seen that if the starting temperature is below 400°C the product of the hydrogenation is powder. The lower the starting temperature the finer the powder which is received after hydrogenation.

At a starting temperature of 500°C only little hydrogen is taken up by the magnetocaloric material. Since this temperature is relatively high hydrogen can diffuse easily in the material and generate a homogenous concentration throughout the specimen. The slow cooling guarantees the gradual and homogenous increase of the hydrogen concentration until the alloy is fully saturated with hydrogen at room temperature. If hydrogenation is carried at lower temperatures, a competition between diffusion (slow) and hydrogenation (fast) takes place, resulting in a hydrogen gradient through the sample. The resultant strain leads to disintegration of the parts.

3.1.2 Highlights of Co-Mn-Si activities
The CoMnSi system offers an alternative possibility for exploring magnetocaloric effects. As outlined in section 3.1.1 CoMnSi is antiferromagnetic, rather than ferromagnetic, and a magnetic field induces a 'metamagnetic' transition to a state of high magnetisation. All thermal and entropic features are of opposite sign, as compared to those of a ferromagnet such as La-Fe-Si at its Curie transition. For this reason, metamagnetic systems are known as 'inverse' MCE materials.

Nevertheless, the potential for application of metamagnets and the conditions on their use in magnetic cooling systems are the same as for ferromagnets. We wish to obtain a large (inverse) MCE in a small field, using as much of the material in phase transition process as possible (no second phase). Hysteresis, too, should be kept to a minimum. CoMnSi has an interesting feature in this regard; it can exhibit a continuous field-induced phase transition in low fields, or a first order, step-like phase transition at large field strengths, all in a single sample. The material therefore offers a so-called tricritical point, where the onset of first order behaviour, and associated large MCE is brought about with minimal hysteresis .

Our goals have been to:
1. Identify whether suitable alloying could bring the tricritical point in CoMnSi to room temperature. In the stoichiometric compound it occurs at a field of 3 Tesla, at 280 K.
2. Establish the link between sample synthesis and phase transition temperature. Literature values of the metamagnetic transition temperature in small fields vary by 200 degrees.

3.1.3 Conclusions of WP1
In summary, the main results of work in WP1 are:
- single phase La-Fe-Si (LFS) materials produced with small thicknesses
- Production of LFS specimens with tuned compositions and working temperatures
- Two new synthesis processes: thermal decomposition and reaction, 'TDR' and solid hydrogenation 'SH', particular to this material system have been developed.
- Optimal TDR and SH conditions have been identified
- Carbon has been shown to stabilise hydrogenated La-Fe-Si materials.
- Machinability has been addressed through TDR and SH syntheses as well as via other processes (see detailed annexe)
- Neutron investigation of Mn-based metamagnets, discovery of giant magnetoelastic coupling and theoretical extrapolation of metamagnetism in other Mn-containing materials

3.2 WP2: Anisotropic materials including intrinsic and artificial spin reorientation

- To set up a wet-chemistry route suitable for laboratory scale synthesis of nanoparticles of hexagonal barium ferrites and derived compounds.
- To investigate the effect of calcination temperature and other production parameters on the phase composition, mean particle size and magnetic properties of the above.
- To produce anisotropic polycrystalline W-type hexaferrites by pressing the synthesized powder in applied magnetic field with subsequent conventional sintering (VAC).
- To grow W-type hexaferrite single crystals from a mixture of oxides by the flux method (CNRS).
- To fabricate aligned NdCo5 intermetallic compacts using the following route: alloy preparation -greater than rapid solidification -greater than ball milling -greater than spark plasma sintering (CNRS and IFW).
- To confirm the presence of crystallographic anisotropy and spin reorientation transitions in the obtained materials (CNRS) and to measure the magnetocaloric effect in isothermal and adiabatic conditions as a function of temperature and magnetic field (INRIM and IFW).

Summary of activities
This part of the project was devoted to the development of a less-explored source of generating magnetocaloric effects. In the materials studied in WP2 the magnetocaloric effect is associated with changes in magnetic anisotropy rather than in the degree of magnetic order. The most common of these are so-called 'spin reorientation' transitions (SRT) at which the ordering direction of the magnetisation changes. This is due to a strong temperature dependence of the magnetocrystalline anisotropy constants (symbol: K ) which are the properties of a magnet that determine the orientation of the magnetisation relative to the crystalline axes, and how easy it is to change that orientation.

3.2.1 Highlights of hexaferrite and Nd-Co activities
The highlights presented in this section are arranged under three headings:
- Preparation of M-ferrite and sub-micron W-ferrite
- Magnetically oriented polycrystalline W-ferrite, magnetic anisotropy and SRT
- Single crystal W-ferrite, magnetic anisotropy and calorimetry around the SRT
- Fabrication of aligned NdCo5 intermetallic compacts

Preparation of M-ferrite and sub-micron W-ferrite
The production of metallic nanoparticles by cryogenic melting technique is a well-established procedure in CNRS-ICMPE. Our first goal was to implement the capacity for bottom-up synthesis of nanostructured ferrites. M-type hexagonal ferrite BaFe12O19 presents easy axis anisotropy at all temperatures below Curie point. And is easier to produce and require lower temperature of synthesis. However, W-ferrites are reported to display better magnetocaloric properies. W-type hexagonal ferrites with the general formula BaM2Fe16O27 (where M = Mg, Fe, Co, Ni, Zn, …) are also known to undergo spin reorientation due to changes in the anisotropy configuration. In the substituted BaCoxZn2-xFe16O27 compound with x = 0.7-0.8 the transitions between easy plane (EP), easy cone (EC) and easy axis (EA) magnetization occur in the vicinity of room temperature . For both hexaferrite types EP less than-greater than EC is expected to be a first order (sharp) transition whereas EC less than-greater than EA is generally continuous.

Magnetically oriented polycrystalline W-ferrite, magnetic anisotropy and SRT
Textured polycrystalline W-type BaCoxZn1–xFe16O27 (x = 0.7) hexaferrites have been prepared using powder magnetic alignment as follows:
- pressing of hexaferrite powder in simultaneously applied magnetic field;
- sintering of the crystallographically oriented compact at high temperature.

Single crystal W-ferrite, magnetic anisotropy and calorimetry around the SRT
Single crystals of BaCoxZn1–xFe16O27 (x = 0.62 and x = 0.7) compositions were grown using a flux method. The starting mixture was prepared from BaCO3, CoO, ZnO, Fe2O3 and Na2CO3 powders so that estimated hexaferrite to flux (NaFeO2) mass ratio was 2:1. XRD control of single crystals was performed at CNRS. X-ray diffraction confirmed the single crystalline nature of the samples, and their W-type hexaferrite phase composition. Magnetisation measurements (CNRS) of W-type BaCoxZn1–xFe16O27 (x = 0.62) hexaferrite single crystals are presented show typical easy axis anisotropy behaviour at room temperature. The crystal orientations determined from X-ray diffraction and magnetometry match. DSC data seems to confirm that easy plane -greater than easy cone -greater than easy axis spin reorientation in W-type hexaferrites proceeds through two continuous (not first order) phase transitions.

Fabrication of aligned NdCo5 intermetallic compacts
The purpose of this work was similar to that on polycrystalline hexaferrite; to take a material with known SRT (and larger associated magnetocaloric effect than the hexaferrite) and make an oriented compact from polycrystalline powder. Such a material should exhibit a spin reorientation MCE, since it consists of aligned, rather than randomised particles, but is much simpler to produce than a single crystal.

3.2.2 Conclusions of WP2
- Oriented polycrystalline W-type hexaferrites have been obtained by powder pressing in magnetic field followed by conventional sintering. XRD and VSM confirm crystal texture and anisotropy of magnetic properties.
- W-type hexaferrite single crystals have been synthesized using a flux method. XRD and VSM confirm correct crystal structure and orientation.
- Spark plasma sintering is proposed as a technique to produce textured NdCo5 and related compounds without application of magnetic field.
- Isothermal entropy change and adiabatic temperature change resulted from SRT have been measured in W-type hexaferrite single crystals and anisotropic polycrystalline samples, respectively, as a function of temperature and magnetic field.
- A simple anisotropy model has been proposed for the description of SRT-driven MCE in hard hexaferrites.

3.3 WP3: Advanced Characterisation

- Standardised methods of measurement and definition of a guide to refrigerant performance
- Characterise a set of refrigerants for prototype I
- Characterise a set of refrigerants with increased temperature range of operation (for prototype II)
- Develop new techniques of measurement as required for comprehensive characterisation of thermodynamic and/or other relevant refrigerant properties

Summary of activities
Work package 3 lies at the heart of the SSEEC project as it provides information to WP5 prototype design, modelling of performance WP4, characterisation of materials WP1 and WP2 and fundamental study of materials in WP3. The aim of the work package is to unify measurement method across the partners (using round robin samples), provide advanced characterisation and report on refrigerant capacity of all the materials produced in the program.

3.3.1 Guide to refrigerant performance
At month 18, a report on the notion of a 'Figure of Merit' was delivered. This found that, rather than a single figure of merit , a graphical comparison of material performance is most instructive. In particular, the cooling engine analysis in WP5 has highlighted the importance of the adiabatic temperature change DeltaTad. A plot of this quantity against peak isothermal entropy change in a fixed field of 2 Tesla (chosen due to the availability of literature data) was chosen as the clearest single plot with which to compare materials, although it is acknowledged that other 'dimensions' to this plot are important, for example operable temperature span of the material. But the graphical approach, similar to that of an Ashby Map in materials engineering, is recommended as a general route.
A comparison of literature materials
There are few straightforward comparisons of DeltaTad for different materials in the literature. The most significant recent reviews focus on entropy change. Indeed a plot of isothermal entropy change vs. Curie temperature for a large range of material systems and compositions reveals no clear features.

However, if we instead plot of DeltaTad vs DeltaS for materials with magnetocaloric effects in the room temperature range (270 K to 320 K) there are noteworthy differences between materials. In such a plot, the ideal material will occupy the upper right area of the figure, with high values of both DeltaTad and DeltaS. Current materials are scattered across the plot and several clues as to how to compare popular materials emerge. MnAs scores very poorly, as its DeltaTad is very low below 2.5 Tesla, due to the large transition hysteresis. The second order ferromagnets, Gd and La(Fe,Co)Si appear to the left of the plot, their entropy change values limited by the lack of sharpness of their phase transition. The first order materials occupy the right hand side of the plot, with Fe-Rh a notable exception in that its DeltaTad is remarkably high. The theoretical limit of DeltaTad has been explored as part of this WP and is described in the accompanying technical annexe.

3.3.2 Round robin exercise: the need for accurate measurement comparison
As the MCE research field grows the community is faced with a specific problem related to the reliability of measurement of the key physical properties M, DeltaS, Cp and DeltaTad. In order to address this problem we have conducted a second round robin exercise, passing a series of La(Fe,Si,Mn)HDelta samples between four partners (denoted here as IC, IFW, INRIM, and VAC), equipped with different commercial and bespoke measurement facilities in order to examine variability of measurement and analysis methods and the resulting measurements. We found that absolute agreement between measurements is a challenge, primarily due to variation in sample shape and measurement protocols used in different laboratories. Most notably commercial instrumentation can lead to the greatest discrepancies. The work highlights the need for a well characterised set of standards in order to unify methods of measurement across the community working in this field.

3.3.3 Development of specialised characterisation tools
A number of advanced characterisation tools have been developed across the partner nodes, now described.

Thermal Conductivity (IFW)
Although not written into the original Technical Annex, the measurement of thermal conductivity was set up at IFW and samples from WP1 were studied.

Combined thermal and magnetisation measurement (Imperial)
A probe developed at Imperial combines thermometry and magnetometry. M(H) and sample temperature measurements were performed using a VSM in combination with a Pt100 platinum resistance thermometer (sensitivity ±0.01 K) mounted on an alumina block stuck to the free surface of the plate using thin layer of GE varnish. The other surface of the sample plate was attached to the polymer sample holder. All samples were oriented with respect to the field direction to minimise demagnetizing effects. The probe fits into a cryostat with up to 8 Tesla magnetic field. The probe allows the study of temperature rise during the magnetic M-H loop cycle and demonstrates a break down in the isothermal conditions that depends on the sweep rate of the magnetic field (for some samples, sample size dependent).

Direct DeltaT in pulsed magnetic field (IFW)
An ultra sensitive ultra fast DeltaT measurement system has been developed at IFW allowing a proper adiabatic measure of the temperature rise in pulsed high magnetic field. In this way dynamic trends can be studied. It was found that under adiabatic conditions, the hysteresis is not field sweep rate dependent. The system has been used to examine many of the materials produced by the partnership; results are referred to throughout this report.

Direct Calorimetry over extended temperature range (INRIM)
In order to be able to perform measurements of the magnetocaloric effect of LaFeCoSi samples with low Co content, the temperature range of the existing setup had to be extended to lower temperatures. To this aim a second, low temperature, calorimeter was developed. With this setup can be measured: a) the isothermal entropy change due to the magnetic field and b) the specific heat as a function of temperature in isofield. The temperature can be changed in the range from 77K to 300K. The measurement is based on the evaluation of heat flux from the sample to a thermal bath measured by Peltier cells. The magnetic field is applied by an electromagnet (Hmax=1.8 T).

Thermal imaging (VAC)
Vacuumschmelze developed an apparatus for fast, direct measurement of the temperature change of a magnetocaloric material as a function of applied magnetic field. In this apparatus a circular magnet system with four different field strengths (0.4 0.8 1.2 and 1.6 T) along the arcs of the circle is rotated over a sample of a size of about 10 x 7 x 1 mm3. The temperature of the sample is continuously monitored using a contactless infrared thermometer. The development of this device was performed outside the budget of the SSEEC project but it has been used within the project.

3.3.4 Conclusions of WP3
- Plots of DeltaS vs. DeltaTad provide a simple means by which to compare refrigerant performance.
- Two round-robin activities have highlighted the importance of sample preparation and comparative measurement techniques. The have also verified the accuracy of our methods.
- A number of advanced characterisation tools have been developed across the partnership and have been integrated into our materials development work, enabling greater flow of ideas and materials between partners.

3.4 WP4: Theoretical modelling

- to formulate a micromagnetic model of exchange coupling in nano-composites
- to formulate a thermodynamic model of La-Fe-Si materials
- to write a report on the modeling of magnetic cooling cycles using La-Fe-Si data
- to formulate a kinetic-thermodynamic model of magnetocaloric materials that have a first order phase transition

Summary of activities
Modelling has been essential to SSEEC. Our work aimed to take the first truly integrated approach to magnetic cooling for an end-user application. This entails characterisation of materials, integration of those into prototype cooling engines, and integration of those engines into an end-user (air conditioning) system. The modelling required for the engineering of the prototype is contained in WP5. WP4, on the other hand, deals primarily with the magneto-thermodynamics of both single phase, hysteretic ('real') materials and of nano-scale exchange-coupled systems, that are also less explored experimentally. Theoretical work helps to give us invaluable feedback on experimental development, reducing the time that would otherwise be required to make multiple measurements or a full phase space of physical samples.

Modelling thermodynamic refrigerant properties
As a test material we selected the La1(FeCoSi)13 alloy family, used in all 3 prototypes, as prepared by VAC. Previous experiments and theories had focused on the size and origin of the available entropy, rather than the impact of hysteresis on this quantity. The particular alloy selected for theoretical investigation has a first order transition at around 206 K. The basics of the model used are described elsewhere.18,19 The starting idea is that in a first order transition two phases coexist and the resulting hysteresis is the consequence of the small energy barriers that hinder the transformation of one phase into the other. In the model this is done assuming an expression for the Gibbs free energy of each 'pure' phase and an independent description of the energy barriers. For a transition without discontinuities the energy barrier becomes a distribution of bistable units in which both the coercivity and the mean value are distributed. The resulting distribution takes the name of Preisach distribution because the resulting hysteresis scheme is equivalent to the Preisach model of hysteresis.

Thermodynamic cycles
INRIM also considered how to compute the thermodynamic cycle of the cross section of an active magnetic regenerator (AMR), the refrigeration device that uses a fluid to exchange heat between the magnetic material and the thermal reservoirs – see Section 2.1. We take a given section of the AMR and we suppose a perfect thermal contact between the material and the fluid and no thermal diffusion along the length. A realistic thermodynamic cycle of two adiabats and two iso-field transformations has been successfully computed, and shows little hysteresis in the case of Co-substituted (non first order) refrigerant materials. In AMR transformations the body is in contact with the fluid and the total entropy (body plus fluid) is then the sum of the two contributions. The example has been computed by solving the equation for adiabatic transformations as given elsewhere.19

Prediction of refrigerant parameters based on the kinetics of transformation of pure materials
In order to reach a desired temperature difference in a refrigerator based on a magnetocaloric refrigerant, an AMR cycle has to be run cyclically. It is therefore of crucial importance determine the maximum cycling frequency in order to achieve the maximum cooling power. The upper limit for the operation frequency of a magnetic refrigeration depends on both the intrinsic properties of the magnetic refrigerant and on the extrinsic effects related to the heat flow with the exchange fluid. Among the intrinsic properties there is the characteristic time for the material to achieve the entropy change. In materials with a first order phase transition, this time constant may be large as the transformation kinetics is governed by the energy barriers between the two phases. In this activity we have investigated the problem of the transformation kinetics from the theoretical viewpoint.

Magneto-elastic effects
In parallel with the hysteresis models which are essentially phenomenological in nature, we have investigated the microscopic origin of a first order phase transformation by analyzing the coupling between magnetism and structure. For this purpose one has to develop microscopic models of the magnetic and of the structural constituents. These models have been already studied in the MCE literature but have yielded discrepancies when compared with experimental data. A proper reconsideration was needed. The point of our work has been to describe the properties of the structural lattice in the simplest possible way, but taking into account the compatibility with the thermodynamic requirements constituted by the Maxwell relations.

3.4.2 Exchange coupling in nano-structured refrigerants
We have investigated the possibility to enhance the MCE by considering a composite material with a nanoscale magnetic ('exchange') coupling between a two phases. There are two examples:
1. A spin reorientation (SRT) phase of known MCE, with a soft ferromagnetic phase of high magnetic moment but with no relevant MCE. The aim is to discover whether the second ferromagnetic phase can enhance the properties of another MCE material.
2. Two hard ferromagnets, with perpendicular easy axes. The aim is to examine the possibility for generating a MCE from two phases that alone have no significant MCE. The mechanism of MCE in the composite should be an 'artificial' spin reorientation, as we expect a spontaneous change of magnetisation direction of one or both phases as a function of temperature.

Composite of SRT material and soft ferromagnet
Here the main MCE phase is the Co-Zn substituted W-type barium ferrite, see Section 3.2.1. With increasing temperature, this material displays a change in the orientation of the magnetization from easy plane to easy axis at T=230 K. The theoretical analysis (based on the theory originally developed by the group of Asti ) reveals that there is an entropy change associated with the reorientation of the spins and its magnitude is related to the temperature dependence of the anisotropy constants. It is then immediately clear that the best material for MCE is the one with the highest change of anisotropy.

Composite of two hard ferromagnets
In this subtask we considered how to exploit the MCE associated with the temperature dependence of the magnetocrystalline anisotropy energy K1(T). In the expression of the entropy of a magnetic system one finds: i) one term proportional to dMs/dT, which describes the contribution associated with ordering or disordering of magnetic spins, and ii) one term proportional to dK1/dT , which is an additional contribution related to the magnetic anisotropy energy. The microscopic origin of the term proportional to dK1/dT is the anisotropic magnetization mechanism : i.e. the spin system is more disordered (high entropy state) when the magnetization points along an hard direction. The interest in the anisotropy degree of freedom is in the fact that the entropy change can be achieved by a rotating field rather than by an alternating one, with possible simplification in refrigeration devices.

3.4.3 Conclusions of WP4
- Successful modelling of the static, thermodyamic properties of MCE materials with continuous phase transitions
- Prediction of the effect of dynamics in first order transitions, defining a range for the increase in magnetic field requirement, depending on transformation kinetics
- Thermodynamic cycle modelling
- Determination of the MCE of a composite of spin reorientation material and a soft ferromagnet (reduced entropy change, enhanced temperature range)
- Established new MCE due to 'artificial spin reorientation' in a hard ferromagnetic multilayer system

3.5 WP5: Prototyping

- Specification of end-user application (power, temperature span etc.)
- Refrigerant characterisation (heat capacity and shape of refrigerant plates)
- Delivery of 3 prototypes
a zero-load, 15 K span cooling engine (prototype I)
35 K span, multi-refrigerant system (prototype II)
an integrated heat pump (prototype III) span and field as in prototype II with complete integration with heat pump/air conditioner application; heat exchangers, control systems and environment

Summary of activities
WP5 is the technology stream of SSEEC, aiming to exploit the properties of magnetocaloric materials to deliver a cooling system capable of powering a domestic heat pump appliance. The resulting cooling system must be capable of being cost-competitive with current gas compressor technology. In this regard the single most important system variable is frequency of operation, as highlighted in our initial proposal; maximisation of operating frequency minimizes the amount (and thus the cost) of refrigerant and permanent magnetic material required for a given quantity of cooling power. It is therefore an objective for the technology delivered in this work package to operate at the maximum possible frequency.

3.5.1 Application specification
The specific application selected by Clivet within the first 6 months of the project was the Elfo air-air exchanger, a system for air renewal and purification in buildings. A key element of the Elfo system is a heat pump technology that actively recovers heat from extracted interior air which is used to pre-heat fresh filtered replacement air from the external environment.

As part of the WP5 effort, a comprehensive cost-benefit analysis of magnetic cooling at different levels of cooling power was undertaken, and reported at the 12 month point. The results of this analysis, which motivate future work in the domestic cooling sector, are given in Section 4.1.1.

3.5.2 Refrigerant characterisation (heat capacity and shape of refrigerant plates)
Heat capacity
Heat capacity is the critical measurement for designing and optimising the shape for active magnetic regenerators; from this single measurement most of the relevant thermodynamic properties of the refrigerant materials can be derived. A new piece of equipment (developed by Ozcan and Burdett at Cambridge) has facilitated the measurement of heat capacity and entropy of first and second order magnetic materials in a magnetic field using the Quantum Design PPMS system – this unique piece of kit has now been adopted by Quantum Design and is available from them.

Shaping of Co-substituted LaFeSi
The simplest parallel plate refrigerant geometry was selected as it seemed the easiest structure to fabricate in LaFeSiCo using the techniques available within the consortium. For a material such as LaFeSiCo the 'ideal' plate geometry for 10 Hz operation required plates with flatness tolerance of maximum 3%. This was according to the ideal geometry specified by Camfridge to VAC near the beginning of the project.

3.5.3 Delivery of three prototypes
The final deliverable, a fully integrated system, using the highest possible operating frequencies required careful design from the outset. The operating frequency focus was critical to achieving size, weight and power targets of an end-user application. If 1 kg of refrigerant can deliver 100 J of cooling per cycle, then when cycling at 10 Hz a total of 1 kW of cooling could be achieved. Thus operating frequency is a key driver to scaling cooling power upwards, or the size, weight or cost of a magnetic system downwards.

Operating span
Prototype 1 yielded a 15 K zero-load span using a single Co-substituted La-Fe-Si material thereby satisfying the target associated with the relevant deliverable. Prototype 2 demonstrated that an operating span of 35 K is achievable; (right, again achieving the target set out in the project plan). For prototype 3, VAC produced bulk materials with more finely tuned Curie temperatures. These materials have even more accurate Curie temperatures than those produced for prototype 2, thus enhancing operating span.

Final prototype (prototype III)
In terms of absolute size and weight the design for prototype 3 meets Clivet's requirements. The magnetic cooling engine is to be plugged into a pair of characterised heat exchangers simulating the ducts in the Elfo system. In associated work, Camfridge has developed a novel magnetically flow control system that:
- Easily integrates the magnetic cooling engine with Clivet heat exchangers
- Fully supports regenerative AMR cycles
- Minimises dead volume
- Eliminates the need for sliding seals
- Avoids fluid mixing in heat exchangers
- Scales to an arbitrary number of regenerator pairs
- Enables simple integration with appliances
- Utilizes a single hot and cold exchanger
- Requires a single pump to drive the system
- Does not need secondary loops
- Provides control mechanisms for defrosts and temperature control

Cooling power and efficiency
The biggest challenge for the prototype 3 is the cooling power. The only way to achieve a kW of cooling within weight and volume constrained design is to increase operating frequency. However, as discussed earlier, the maximum frequency of operation is limited by material processing (shorter length scales and better flatness) and the result achieved enables 5 Hz operation. Prototype 3 should achieve at least approximately 100 W cooling power over a 35K operating range, below the Clivet application target but achieving the original research proposal target in terms of operating temperature range and driving field. We note that this power range is also more in line with the cost-benefit optimum presented in Section 4.1.1.

3.5.4 Conclusions of WP5
WP5 has achieved most of its goals
- 3 prototype magnetic cooling engines have been constructed using Co-doped La-Fe-Si
- Temperature span targets have been met in prototypes I and II
- End user integration has been performed in prototype III
- A technology roadmap that best fits the characteristics of magnetic cooling has been developed (see Section 4.1.1)

The SSEEC project has rapidly advanced the development of magnetic cooling technology. Materials, produced for the first time to specification from a production process, have been integrated into a series of prototype systems targeted at achieving technical and commercial specifications for the technology.

Potential Impact:
4. Potential impact and main dissemination activities
4.1 Potential impact
Today, approximately 1/3 of the energy consumed in EU countries is destined to comfort in buildings (heating, air cooling and air exchange). As far as energy consumption of buildings is concerned, EC initiatives have mainly focused on the development of renewable energies (with associated long times and elevate set-up costs) and to the reduction of energy requirements (building energy efficiency certification). Nevertheless in the short term, significant results can be achieved by the adoption of low energy consumption systems.

4.1.1 Magnetic cooling: the future
The WP5 team considered concurrently what sectors would most benefit from switching to a magnetic cooling engine, in order to achieve the maximum possible synergy of the partners' skill sets, and to maximise the project's benefits for society beyond its 3-year life. At the 12-month point, a careful system-wide cost and efficiency analysis revealed the benefit of magnetic cooling at low powers (less than 500 Watt). This led to a re-evaluation of the long-term road map for the technology.

An analysis of the efficiency shows that the efficiency of low power gas compressors falls dramatically – by 50%. This means there is a significant opportunity to improve lower power cooling appliances – such as the domestic fridge.

4.1.2 Magnetic phase transition research
Our work has investigated continuous and first order phase transitions from many perspectives. We have aimed to advance: theoretical understanding, modelling of static and dynamic phenomena, tailoring of materials from theoretical work, bespoke experimental probes for the measurement of key materials properties, synthesis of micro-scale and nano-scale materials and single phase metastable alloys, shaping of refrigerant plates and non-destructive hydrogenation.

4.2 Main dissemination / exploitation of results
We have exploited many different forms of dissemination, a list of which is given in section A, 'Use and dissemination of foreground' of our SESAM submission accompanying this document. A total of 146 separate acts of dissemination were carried out, ranging from journal publications, oral and poster presentations at conferences, invited seminars at various institutions across the world, media interviews, magazine articles and articles in the popular press. These help to demonstrate that magnetic cooling is receiving increased attention within the physics, materials science and engineering research communities and from refrigerations industries and the general public. A summary under relevant headings is now presented.

Journal articles
During the course of the project (Oct 2008 – Sept 2011) we have published 16 peer-reviewed journal articles in high impact journals such as Physics Review Letters, Acta Materialia and Applied Physics Letters. Many of these involve co-authorship by two or more project partners, highlighting the degree of interaction within our consortium. Since the conclusion of the project another 8 peer-reviewed journal articles have followed, including one in Nature Materials. The total of 24 journal articles thus far will continue to grow in the coming months.

Conference presentations
The visibility of magnetocaloric materials and magnetic cooling research has greatly increased during the project, in no small way due to the activities of the consortium partners. Members of the consortium gave over 50 invited or contributed presentations at national or international conferences. Symposia on, or featuring, magnetocalorics and magnetic cooling were organised either wholly or in part at several conferences by Oliver Gutfleisch (IFW Dresden) and Karl Sandeman (SSEEC coordinator, Imperial), thus adding to the opportunities for dissemination of our results and interaction with other researchers. These symposia were held at Euromat 2009 (Glasgow, UK), MRS Fall 2010 (Boston, USA), TMS 2011 (San Diego, USA) and Euromat 2011 (Montpellier, France).

Invited presentations
During the course of the project a total of 18 invited presentations were given by partners at local meetings, workshops or as seminars at research institutions. These invitations were worldwide; seminars were given across Europe, the US, Canada, Australia and in Japan. The coordinator was invited to speak about SSEEC at the inaugural ICYRAM meeting in Singapore, after the project's conclusion.

Media interviews and articles in the popular press
The application of magnetic cooling to different market sectors continues to attract interest from general media and industry-led journals. Camfridge gave 3 BBC interviews during the project and featured in articles in New Scientist, The Daily Telegraph and 3 industry-led magazines. The coordinator wrote an invited article on gas-free refrigeration featuring the work of SSEEC for the launch edition of Magnetics Technology International in 2011.

IEEE Distinguished Lecturer Series 2011
Oliver Gutfleisch (IFW Dresden) was one of the 2011 IEEE Magnetics Society's distinguished lecturers, and gave 19 lectures across Europe, USA, China and Japan on the subject of 'Magnetic materials in sustainable energy', including the work of SSEEC, during the period January-September 2011. This dissemination activity continued after the project with a further 16 lectures across Europe and the USA during Oct-Dec 2011, yielding a total of 35 lectures delivered worldwide.

List of Websites: