Periodic Report Summary 1 - HARFIR (Heusler Alloy Replacement for Iridium)
Project Context and Objectives:
Spin electronics is expected to displace volatile silicon memory technology within the next decade and is already in existence in the read head of hard disk drives (HDDs). The demand/supply disruption of the rare metal Iridium is already under way and, as spin electronics becomes more ubiquitous, the level of disruption and higher cost will only increase. The price of Iridium has risen by a factor of 4 in the last five years and by more than a factor of 10 in the last decade. It is expected to soar perhaps by a factor of 100 due to its wider applications.
Our research programme impacts this key material directly by providing an improved understanding of a wide ranging class of ternary alloys and we will seek to find new materials and new compositions of Heusler Alloys (HAs) to resolve this issue. Specifically, we intend to develop antiferromagnetic (AF) HA films to replace the AF alloy Iridium Manganese (IrMn).
HAs are alloys of the transition metals Iron (Fe), Cobalt (Co) or Manganese (Mn) with materials such as Silicon (Si) or Aluminium (Al). All these materials are abundant on Earth and hence success in our project would eliminate the problem of the soaring cost of Iridium and perhaps its availability in the future. As shown in the attached chart, Iridium is one of the world’s most scarce elements – three times as rare as Gold and Platinum, for example – with an average occurrence in igneous rock of only 0.001 parts per million.
We combine our expertise in ab initio calculations and HA film growth techniques to seek highly anisotropic AF HA films. These films will be characterised both structurally and magnetically using synchrotron beamlines, high-resolution (scanning) transmission electron microscopy and highly sensitive electrical and magnetic measurement facilities available within the consortium. We will demonstrate a device concept with the developed AF HA films at the end of this project, showing an exchange bias (EB) greater than 1 kOe in sheet form and a blocking temperature greater than 300K.
The developed AF HA films will be patented with the royalties shared equally among the partners in the consortium. To our knowledge we are the first group to realise the criticality of the position with regard to the supply of Ir. The innovation within HARFIR is therefore extremely high.
Project Results:
Four HA compounds – Fe2VAl, Ni2MnAl, Ru2MnAl and Cr2MnSb – were selected for the initial studies, based on theoretical calculations and crystallisation temperature measurements. We found that crystallising Fe2VAl requires 500C – very close to the limit for practical applications – whereas Ni2MnAl crystallises readily below 500C and Ru2MnAl crystallises into the B2 structure at 400C.
Alloys with the form X2YZ (full HA) belonging to families of the following compositions were prioritised for further studies:
• Fe2VAl: predicted to have the greatest potential for AF ordering.
• Mn2VAl: replaces the Fe2 with the very high moment Mn2.
• (CoMn)2VAl: analogous to the previous two except that X2 uses two high moment atoms.
• Ni2MnAl: has a high moment, with Mn replacing V.
• Ru2MnAl: has the potential advantage of both X and Y being high moment atoms, and is known to exhibit AF ordering.
We varied growth and annealing conditions of both polycrystalline and epitaxial Fe2VAl and Ni2MnAl films to determine optimised conditions. Using first-principles calculations and advanced classical spin-models, we investigated the magnetic state and its dependence on the atomic disorder and chemical composition for selected AFM and ferrimagnetic HAs.
We established the presence of AF for most of the selected bulk HAs. For Ni2MnAl, we demonstrated that chemical disorder leads to magnetic phase transition and the compensated AFM state occurs only in a fully disordered B2 phase – otherwise uncompensated AF or ferrimagnetism was observed. In the case of Ru-based AF Heusler alloys, calculations of the spin-spin correlation functions predicted complex AFM ordering for all compositions. We pointed out that the Néel temperature in Ru2MnSi and Ru2MnGe can be increased by improving the degree of L21 order.
We also initialised a database for calculated relativistic tensorial exchange interactions at interfaces between AF HAs and various FM media (pure FM metals or HAs). We achieved preliminary results for ground state magnetic configurations and hysteresis loops of Ni2MnAl/Fe, as obtained from spin-dynamics simulations, by using the database of first principles tensorial exchange interactions. For perfect Ni2MnAl (B2)/Fe bilayers, we found no (in-plane) EB effect.
We characterised the AF behaviour of the HA films using world-leading synchrotron light sources and a state-of-the-art HR-(S)TEM. The neutron-, muon-, soft X-ray- and electron-beams were also used for these characterisations. An XMCD analysis of epitaxially grown Ru2MnGe covered with 2nm Fe showed an AF coupling of the Manganese moments in this compound – i.e. for the correct phase, the XMCD signal vanishes.
Initial structural studies using HR-TEM were made to assess the crystalline quality of the HA films and the interfaces, which are crucial for the presence of EB. We used highly sensitive measurements to evaluate the electrical and magnetic properties of the HA AF films, in order to reveal the AF properties of the films on a microscopic scale. We established a method to measure the coercive field and EB that also allows experiments at low temperature.
All samples grown have been characterised, but no room temperature AF has been observed yet. Magnetisation measurements with an Alternating Gradient Magnetometer and the Magnetooptical Kerr Effect have confirmed a vanishing magnetic moment of the Ru2MnGe.
A Press Release about the project was distributed to relevant news outlets, and open workshops in the EU and in Japan have attracted many external researchers. We have participated in a number of Critical Raw Materials meetings, and given presentations at public scientific forums. We have also written articles about our work for several journals and have exchanged staff between the EU and Japanese collaborators.
Potential Impact:
We expect that the outcome of the project may be inexpensive HAs capable of replacing expensive (and rare) IrMn in spin electronic devices, and we will create GMR and TMR devices which use optimised AF or CF Heusler alloy films in order to demonstrate their performance.
It is widely recognised that spin electronic technologies will displace semiconductor technology within the next decade, yielding many areas where the project’s results can be applied. These range from existing technologies such as the read head in HDDs, to solid-state storage such as MRAM and thermally assisted MRAM (TA-MRAM) – a system proposed as a replacement for the common silicon-based dynamic random access memory (DRAM). The start-up time when switching on a computer or similar device would then be reduced dramatically because the storage of the information would be permanent rather than charge based or capacitive – where recharging the system is required and data has to be reloaded from a permanent storage system such as a hard disk. The advantages of such a system are clear.
The other major advantage of proposed MRAM technologies is that they are faster and capable of storage at a higher density than, for example, flash memory. They are also much less complex than flash and have the advantage of increased reliability, with no limit to the read-write cycles that can be performed. This is thanks to the infinite reversibility of a magnetic material’s orientation without damage to its structure – something not true of the charge-based systems that compose flash memory. It is also the case that flash memory has to be rewritten as an entire block when its data is modified. For large-scale storage or archiving, this power cost, with consequences for CO2 emissions, becomes considerable. It is worth noting that MRAM has recently appeared on the Intel roadmap for implementation as embedded memory on a three-to-five year timescale.
In addition to the storage capabilities of MRAM, there is also the distinct potential for magnetic logic systems based on similar technologies. These are known to have the capability to be faster and have higher aerial densities than is possible with current Silicon technologies. The decrease in complexity and the increase in reliability mean that magnetic logic systems would have major cost advantages over current Silicon techniques.
List of Websites:
http://www.harfir.eu/
Spin electronics is expected to displace volatile silicon memory technology within the next decade and is already in existence in the read head of hard disk drives (HDDs). The demand/supply disruption of the rare metal Iridium is already under way and, as spin electronics becomes more ubiquitous, the level of disruption and higher cost will only increase. The price of Iridium has risen by a factor of 4 in the last five years and by more than a factor of 10 in the last decade. It is expected to soar perhaps by a factor of 100 due to its wider applications.
Our research programme impacts this key material directly by providing an improved understanding of a wide ranging class of ternary alloys and we will seek to find new materials and new compositions of Heusler Alloys (HAs) to resolve this issue. Specifically, we intend to develop antiferromagnetic (AF) HA films to replace the AF alloy Iridium Manganese (IrMn).
HAs are alloys of the transition metals Iron (Fe), Cobalt (Co) or Manganese (Mn) with materials such as Silicon (Si) or Aluminium (Al). All these materials are abundant on Earth and hence success in our project would eliminate the problem of the soaring cost of Iridium and perhaps its availability in the future. As shown in the attached chart, Iridium is one of the world’s most scarce elements – three times as rare as Gold and Platinum, for example – with an average occurrence in igneous rock of only 0.001 parts per million.
We combine our expertise in ab initio calculations and HA film growth techniques to seek highly anisotropic AF HA films. These films will be characterised both structurally and magnetically using synchrotron beamlines, high-resolution (scanning) transmission electron microscopy and highly sensitive electrical and magnetic measurement facilities available within the consortium. We will demonstrate a device concept with the developed AF HA films at the end of this project, showing an exchange bias (EB) greater than 1 kOe in sheet form and a blocking temperature greater than 300K.
The developed AF HA films will be patented with the royalties shared equally among the partners in the consortium. To our knowledge we are the first group to realise the criticality of the position with regard to the supply of Ir. The innovation within HARFIR is therefore extremely high.
Project Results:
Four HA compounds – Fe2VAl, Ni2MnAl, Ru2MnAl and Cr2MnSb – were selected for the initial studies, based on theoretical calculations and crystallisation temperature measurements. We found that crystallising Fe2VAl requires 500C – very close to the limit for practical applications – whereas Ni2MnAl crystallises readily below 500C and Ru2MnAl crystallises into the B2 structure at 400C.
Alloys with the form X2YZ (full HA) belonging to families of the following compositions were prioritised for further studies:
• Fe2VAl: predicted to have the greatest potential for AF ordering.
• Mn2VAl: replaces the Fe2 with the very high moment Mn2.
• (CoMn)2VAl: analogous to the previous two except that X2 uses two high moment atoms.
• Ni2MnAl: has a high moment, with Mn replacing V.
• Ru2MnAl: has the potential advantage of both X and Y being high moment atoms, and is known to exhibit AF ordering.
We varied growth and annealing conditions of both polycrystalline and epitaxial Fe2VAl and Ni2MnAl films to determine optimised conditions. Using first-principles calculations and advanced classical spin-models, we investigated the magnetic state and its dependence on the atomic disorder and chemical composition for selected AFM and ferrimagnetic HAs.
We established the presence of AF for most of the selected bulk HAs. For Ni2MnAl, we demonstrated that chemical disorder leads to magnetic phase transition and the compensated AFM state occurs only in a fully disordered B2 phase – otherwise uncompensated AF or ferrimagnetism was observed. In the case of Ru-based AF Heusler alloys, calculations of the spin-spin correlation functions predicted complex AFM ordering for all compositions. We pointed out that the Néel temperature in Ru2MnSi and Ru2MnGe can be increased by improving the degree of L21 order.
We also initialised a database for calculated relativistic tensorial exchange interactions at interfaces between AF HAs and various FM media (pure FM metals or HAs). We achieved preliminary results for ground state magnetic configurations and hysteresis loops of Ni2MnAl/Fe, as obtained from spin-dynamics simulations, by using the database of first principles tensorial exchange interactions. For perfect Ni2MnAl (B2)/Fe bilayers, we found no (in-plane) EB effect.
We characterised the AF behaviour of the HA films using world-leading synchrotron light sources and a state-of-the-art HR-(S)TEM. The neutron-, muon-, soft X-ray- and electron-beams were also used for these characterisations. An XMCD analysis of epitaxially grown Ru2MnGe covered with 2nm Fe showed an AF coupling of the Manganese moments in this compound – i.e. for the correct phase, the XMCD signal vanishes.
Initial structural studies using HR-TEM were made to assess the crystalline quality of the HA films and the interfaces, which are crucial for the presence of EB. We used highly sensitive measurements to evaluate the electrical and magnetic properties of the HA AF films, in order to reveal the AF properties of the films on a microscopic scale. We established a method to measure the coercive field and EB that also allows experiments at low temperature.
All samples grown have been characterised, but no room temperature AF has been observed yet. Magnetisation measurements with an Alternating Gradient Magnetometer and the Magnetooptical Kerr Effect have confirmed a vanishing magnetic moment of the Ru2MnGe.
A Press Release about the project was distributed to relevant news outlets, and open workshops in the EU and in Japan have attracted many external researchers. We have participated in a number of Critical Raw Materials meetings, and given presentations at public scientific forums. We have also written articles about our work for several journals and have exchanged staff between the EU and Japanese collaborators.
Potential Impact:
We expect that the outcome of the project may be inexpensive HAs capable of replacing expensive (and rare) IrMn in spin electronic devices, and we will create GMR and TMR devices which use optimised AF or CF Heusler alloy films in order to demonstrate their performance.
It is widely recognised that spin electronic technologies will displace semiconductor technology within the next decade, yielding many areas where the project’s results can be applied. These range from existing technologies such as the read head in HDDs, to solid-state storage such as MRAM and thermally assisted MRAM (TA-MRAM) – a system proposed as a replacement for the common silicon-based dynamic random access memory (DRAM). The start-up time when switching on a computer or similar device would then be reduced dramatically because the storage of the information would be permanent rather than charge based or capacitive – where recharging the system is required and data has to be reloaded from a permanent storage system such as a hard disk. The advantages of such a system are clear.
The other major advantage of proposed MRAM technologies is that they are faster and capable of storage at a higher density than, for example, flash memory. They are also much less complex than flash and have the advantage of increased reliability, with no limit to the read-write cycles that can be performed. This is thanks to the infinite reversibility of a magnetic material’s orientation without damage to its structure – something not true of the charge-based systems that compose flash memory. It is also the case that flash memory has to be rewritten as an entire block when its data is modified. For large-scale storage or archiving, this power cost, with consequences for CO2 emissions, becomes considerable. It is worth noting that MRAM has recently appeared on the Intel roadmap for implementation as embedded memory on a three-to-five year timescale.
In addition to the storage capabilities of MRAM, there is also the distinct potential for magnetic logic systems based on similar technologies. These are known to have the capability to be faster and have higher aerial densities than is possible with current Silicon technologies. The decrease in complexity and the increase in reliability mean that magnetic logic systems would have major cost advantages over current Silicon techniques.
List of Websites:
http://www.harfir.eu/
