Forskjell mellom versjoner av «Giant magnetoresistance»

Revisjonen fra 26. mar. 2009 kl. 00:46

In production by Einar, Fabio og Fabio, please do not edit unless you are one of us.

When magnetic fields are appplied to metals, most show an increase in resistance, known as magnetoresisitance (MR), but GMR exploits the fact that when these junctions is under a critical length the magnitude of this magnetoresistance is greatly amplified. This has been used as a way of packing data extremely dense, and most modern hard-drives (HDD) uses this technology.

Theoretical background

History

Discovery and commercialization of Magnetoresistance (MR), a change in the electrical resistance of a conductor caused by an applied magnetic field, was first observed by the physicist William Thomson (Lord Kelvin) in 1857, wheras the physics underlying electron spin — which is the ultimate source of magnetism in most materials — dates back to the work of Paul Dirac, Wolfgang Pauli and others in the golden era of quantum mechanics. The effect was quite small, typically a few percent, but it was still large enough to be exploited in read heads for magnetic disks and sensors for detecting magnetic fields. However, that all changed with the discovery of GMR in 1988. Grünberg and his team at the Jülich Research Center in Germany made their discovery — a change of about 10% in electrical resistance in the presence of a magnetic field — in a structure containing a 1-nm thick layer of chromium (which is not magnetic) sandwiched between two thicker layers of iron (which is magnetic). Meanwhile Fert and co-workers at the University of Paris-Sud and Thomson CSF observed an even larger effect (a change of about 50%) in more complex structures containing about 60 alternating layers of chromium and iron. Both teams used molecular beam epitaxy (MBE), a central if often-overlooked research tool in the history of nanotechnology, to make their multilayer samples. Although the French team coined the term ‘giant magnetoresistance’, it was Grünberg who recognized that this effect could be used to detect faint magnetic fields, so he filed for a patent as his group wrote up its results. However, the two group leaders agreed to share the credit for the discovery. GMR also represented the first example of a new kind of technology called ‘spintronics’, so-called because it exploits the spin of the electron, as well as its electric charge, to store and process information.

Ordinary magnetoresistance

In the absence of an external field electrons travel through a solid in straight lines in between scattering events. For a free electron gas, the same is true even in the presence of an applied field. Althoug the applied field exerts a force (the Lorence force) on the electrons which deflects them from their path, the electric field created by the displaced electrons exactly balances the Lorentz force, and at equilibrium the electrons follow the same straight-line path.This is known as the Hall effect. However in a real metal the conduction electrons have different average velocities, and although on average the transverse Hall electric field exactly balances the magnetic field, individual electrons travel in a curved path and because of this they scatter more. Consequently the resistance in the presence of the field is larger than the resistance in the absence of the field. This effect is, however, very small, and doesn't have technological applications.

Anisotropic magnetoresistance

Larger magnetoresistive effects, of around 2%, are observed in ferromagnetic metals and their alloys. The phenomenon is called anisotropic magnetoresistance (AMR) because the change in resistance when a field is applied parallel to the current direction is different from that when a field is perpendicular to the curren direction. In fact, the resistance for current flowing parallel to the field direction increases when a field is applied, whereas the resistance for current flowing perpendicular to the field direction decreases by approximately the same amount. The effect is significant even in small fields, and his origin lies in the spin-orbit coupling of the electrons. From about 1993 until the late 1990s, anisotropic magnetoresistive materials were used almost exclusively as the read elements in recording heads.

Giant magnetoresistance

The magnetoresistive component in modern read heads operates on the giant magnetoresistive (GMR) effect. In GMR materials, thin layers of of magnetic materials are separated by layers of non-magnetic materials. Depending on the thickness of the non magnetic layers, the magnetic layers couple either ferromagnetically or antiferromagnetically. Giant magnetoresistance occurs when the thicknesses are chosen such that the adjacent layers are antiferromegnetic in zero applied field. This results in a high-resistance state as up-spins electrons are scattered by regions of down spin magnetization and vice versa. Then the GMR effect works by changing the relative magnetization directions between adjacent magnetic layers. A low-resistance state is obtained when a magnetic field strong enough to overcome the antiferromagnetic coupling is applied and the magnetization of the layers is rotated to a ferromagnetic configuration. When the magnetic layers are ferromagnetically aligned, conduction electrons of compatible spin-type are able to move through the heterostructure with minimal scattering, and the overall resistance of the material is lowered.

Although the highest values of GMR were measured in antiferromagnetically-coupled magnetic multilayers, these multilayers are not the best materials for technological applications. This is due to the large magnetic fields, which are required to saturate the magnetization of the multilayers and to obtain a sizeable change in the resistance. For example, as is evident from Fig.6, the saturation fields in the Fe/Cr multilayers are of the order of 10-20 kG which is three orders of magnitude higher than the fields required for applications. The sensitivity, which is defined as ΔR/R per unit magnetic field, is much too small.

By the way, antiparallel alignment can also be obtained by introducing different coercivities of the successive ferromagnetic layers. In this case (pseudo spin valve structure) the magnetic moments of the soft and hard magnetic layers switch at different values of the applied magnetic field providing a field range in which they are antiparallel and the resistance is higher.

Another structure which gives much better performances from the point of view if applications is the spin valve. In the spin valve the magnetization of one ferromagnetic layer is pinned by the exchange coupling with an adjacent antiferromagnetic layer, whereas the magnetization of the other ferromagnetic layer is free to rotate with the applied magnetic field. Although the measured values of GMR are higher in magnetic multilayers, spin valves are more attractive from the point of view of applications, because only small magnetic fields need to be applied to change the resistance.

Magnetic granular solids represent another system, which displays the GMR effect. They consist of a non-magnetic metal alloyed with a ferromagnetic metal, which precipitates into granules. Although the granules can be magnetically coupled, in the absence of the applied field their magnetic moments are randomly-oriented. Applying the magnetic field aligns the moments of the granules, which results in the resistance drop. The saturation fields, which are required to align the moments, are as high as in the antiferromagnetically-coupled multilayers, i.e. of the order of 10kG. This fact makes the applicability of granular materials fairly limited. In addition, the magnitude of GMR in granular materials at room temperature is strongly reduced due to superparamagnetic relaxation, which originates from thermal fluctuations of the magnetic moments of the granules.

Mott's model and spin dependent transport

GMR can be qualitatively understood using the Mott model, which was introduced as early as 1936 to explain the sudden increase in resistivity of ferromagnetic metals as they are heated above the Curie temperature. There are two main points proposed by Mott. First, the electrical conductivity in metals can be described in terms of two largely independent conducting channels, corresponding to the up-spin and down-spin electrons, which are distinguished according to the projection of their spins along the quantization axis. The probability of spin-flip scattering processes in metals is normally small as compared to the probability of the scattering processes in which the spin is conserved. This means that the up-spin and down-spin electrons do not mix over long distances and, therefore, the electrical conduction occurs in parallel for the two spin channels. Second, in ferromagnetic metals the scattering rates of the up-spin and down-spin electrons are quite different, whatever the nature of the scattering centers is. According to Mott, the electric current is primarily carried by electrons from the valence sp bands due to their low effective mass and high mobility. The d bands play an important role in providing final states for the scattering of the sp electrons. In ferromagnets the d bands are exchange-split, so that the density of states is not the same for the upspin and down-spin electrons at the Fermi energy. The probability of scattering into these states is proportional to their density, so that the scattering rates are spin-dependent, i.e. are different for the two conduction channels. Using Mott’s arguments it is straightforward to explain GMR in magnetic multilayers. We consider collinear magnetic configurations, and assume that the scattering is strong for electrons with spin antiparallel to the magnetization direction, and is weak for electrons with spin parallel to the magnetization direction. This is supposed to reflect the asymmetry in the density of states at the Fermi level, in accordance with Mott’s second argument.

For the parallel-aligned magnetic layers, the up-spin electrons pass through the structure almost without scattering, because their spin is parallel to the magnetization of the layers. On the contrary, the down-spin electrons are scattered strongly within both ferromagnetic layers, because their spin is antiparallel to the magnetization of the layers. Since conduction occurs in parallel for the two spin channels, the total resistivity of the multilayer is determined mainly by the highly-conductive up-spin electrons and appears to be low. For the antiparallel-aligned multilayer, both the up-spin and down-spin electrons are scattered strongly within one of the ferromagnetic layers, because within the one of the layers the spin is antiparallel to the magnetization direction. Therefore, in this case the total resistivity of the multilayer is high.

Band structure of ferromagnetic materials

Review article

Multilayer stack configuration (cut and paste from the review article we printed. Do with it as you want!)

With increasing number of FM/NM bilayers within a multilayer the value of GMR monotonically increases until it reaches saturation. This is due to the diffuse scattering at the outer boundaries of the multilayer. If the longest mean free path is larger than the total thickness of the multilayer, then the diffuse outer-boundary scattering reduces the conductivity of the “good” spin channel and hence effects GMR negatively. The magnetoresistance ratio becomes independent of the number of bilayers when the total thickness of the multilayer is much larger than the longest mean free path.

Most experiments on GMR are performed by measuring the electric current in the plane of the multilayer, i.e. within the current-in-the-plane (CIP) geometry. This geometry is currently used for the industrial applications of GMR. Measuring the current perpendicular to the multilayer plane, i.e. within the current-perpendicular-to-the-plane (CPP) geometry, is much more difficult. This is due to the very small thickness of the multilayer and consequently the very low CPP resistance, which is not easy to detect.

Materials and synthesis methods

Material properties

Importance of band structure

In most metals, such as for instance copper, there is an equal number of spin-up and spin-down electrons. This causes the atom to be spin neutral. The ferromagnetic materials used most frequently in GMR applications are metals of the first transition series. For these elements it is energetically favourable for the d-electrons to have parallel spins in order to reduce their Coulomb repulsion. This is known as exchange splitting, and it gives rise to a net magnetic moment which makes the metal ferromagnetic. Exchange splitting leads to a difference in conductivity of spin-up and spin-down electrons.

At a junction between a ferromagnet and a non-magnetic material the transmission of electric current will depend on how well the band structure of the up- and down-spin electrons in the ferromagnet match with the nonmagnet. In order to achieve high transmission of spin-up and not spin-down, for instance, it is necessary to select a nonmagnet whose band structure closely resembles the band structure of spin-up electrons in the ferromagnet.

Importance of lattice matching

When building a superlattice of two or more crystalline materials it is important to consider the lattice parameters of the materials. If the lattice parameters differ significantly there would be significant strain at and near the interface, causing dislocations and imperfections which would scatter both the spin-up and spin-down electrons equally. This would be derogative to the total GMR effect. Hence, it is imperative that the lattice parameters are very closely matched.

Importance of purity

The high impurity or high internal strains in a ferromagnet increases its coercivity. Recall that a high coercivity means that a high magnetic field must be applied to change the direction of magnetization. Good purity control is therefore essential in synthesis to tailor the properties.<ref name="Kittel">Kittel, page 352 </ref>

In order to maximize the GMR effect, we generally want to minimize the spin-independent scattering events and maximize the spin-dependent scattering. Previously, we have discussed only the most fundamental type of spin-dependent scattering, namely interface scattering. However, bulk impurities such as dislocations, stacking faults and grain boundaries can also show spin-dependent scattering. This can be a severe disadvantage or a powerful tool, depending on whether those imperfections are present by design or by random. <ref name="Gurney et.al.">B.A. Gurney, V.S. Speriosu, J.P. Nozi~res, H. Lefakis, D.R. Wilhoit and O.U. Need, Phys. Rev. Lett 71 (1993) 4023.</ref>

Successful materials and examples of structures

Synthesis

Thickness control

In the antiferromagnetic superlattice structure, a certain thickness of the layers must be present in order to achieve any GMR at all. [fig exchangecoupling] The following discussion is about multilayer superlattice structures (not spin valves).

Ferromagnetic layer thickness. As the ferromagnetic layer thickness is reduced, there will obviously be less bulk scattering events. The critical thickness is at the mean free path of the strongly scattered electrons. If the layer is any thinner than this, the GMR effect drops dramatically. Experiments show that interface scattering also depends on the layer thickness. If the layer is thinner than approximately 2 monolayers, the scattering is not optimal. <ref name="Parkin"> S.S.P. Parkin, Phys. Rev. Lett. 71 (1993) 1641</ref>

As the ferromagnetic layer thickness is increased, current will be shunted towards the center of the channel. This leads to a higher current density, with a resultant inrease in resistance and a drop of GMR. <ref name="Dieny"> B. Dieny, Journal of Magnetism and Magnetic Materials 136 (1994) 335-359</ref>

Nonferromagnetic layer thickness. Again, as the nonferromagnetic layer thickness is increased, current will be shunted towards the center of the channel. This leads to a higher current density, with a resultant inrease in resistance and a drop of GMR. Moreover, a thicker layer leads to more bulk scattering events. Hence the GMR effect drops monotonically with increasing laer thickness.<ref name="Dieny"> B. Dieny, Journal of Magnetism and Magnetic Materials 136 (1994) 335-359</ref>

If the nonmagnetic layer gets too thin, pinholes will cause a strong ferromagnetic coupling of the magnetic layers, which reduces the GMR effect dramatically. Normally, therefore, the nonmagnetic layer must be at least 15Å thick. <ref name="Dieny"> B. Dieny, Journal of Magnetism and Magnetic Materials 136 (1994) 335-359</ref>

Extensive research has gone into investigating how the overall GMR effect is changed if one introduces thin layers (~ 2Å) of various impurities at or near interfaces within the structure. The observed effects are complex and highly dependent on the particular structure under investigation, but in some cases the value of the GMR effect can be doubled. The key point here is that it's imperative to choose a synthesis method which allows excellent control of the layer thickness and position. <ref name="Main review article 47 pages"> Main review article</ref>

Methods in use

The thin film industry has developed substantially in the past few decades, and numerous synthesis methods are available and frequently used to make superlattices. In this context, the most important metal deposition techniques of crystalline metal films will be discussed briefly.

Sputtering

Sputtering involves a plasma reactor where the plasma ions are accelerated towards a metal target, knocking loose metal atoms in the process which subsequently migrate towards the wafer and deposit to form a uniform thin film. By controlling the plasma density, vacuum conditons and acceleration potential for the plasma, high thickness precision is achieved.<ref name="Quirk and Serda">Quirk and Serda, p. 314 </ref>

Molecular beam epitaxy

In molecular beam epitaxy (MBE), a metal target is evaporated by an electron gun. The metal atoms then migrate to the wafer and deposit. The growth rate of MBE is about one monolayer per second, which gives excellent control of the layer thickness. Moreover, the process occurs at relatively low temperatures, which minimizes diffusion across interfaces. <ref name="Jenkins">Jenkins, p. 74 </ref> If the temperature is too high, the atoms will eventually migrate across the interface and form clusters. If the ferromagnetic material does this, ferromagnetic interaction between these particles reduces the GMR effect. <ref name="cocusystem2">Article, abbr. cocosystem2</ref>

Electrochemical deposition

(Rephrase) Electrochemical deposition ECD, is an attractive alternative technique because of its simplicity, cost effectiveness, and high deposition rates in comparison to vacuum techniques. However, electrochemically deposited multilayers usually show a lower GMR than those prepared by physical vapor deposition techniques and, therefore, are most commonly used only in low-end devices.<ref name="cocusystem">Article abbr. "cocusystem" </ref>

GMR in hard-drives

Requirements for the magnetic materials in the hard drive itself:

• Soft magnet. Low coercivity, meaning that you don't need a strong magnetic field to magnetize or demagnetize (1 or 0), and the process doesn't require a lot of energy.
• Hysterisis loop must be narrow (--> low energy losses for (de)magnetization) and square shaped (--> swift change from 1 to 0

For a nice graphical presentation, click here (IBM pages).

How it works

Hard disk drives (HDD)

In GMR Hard disk drives (HDD) data is stored on magnetically coated platters made of metal or glass that revolve at several thousand rotations per minute (RPM), driven by an electric motor. It is extremely important that these platters are completely flat (less than a few nanometers in height-difference on the entire disk), because the read-write head is scanning the surface at a relative speed of 10-30 m/s at ~15 nm distance. This is also the reason why you should never move a HDD in use, because the read-write heads might come in contact with the disks and scratch the surface, which would leave the disk useless. The capacity of the drive depends on the number of platters, typically from 4 to 8 and the type of magnetic coating. The magnetic coating is divided into small areas called bits, where one bit represent either a 1 or a 0, depending on the direction of the magnetization. When the disks spin the data is read and written by read-write heads located on the "arms". The position of the read-write heads is controlled by an actuator operating the arm.

Read-write head

A GMR read-write head is located on the arm sweeping over the revolving platters typically separated from the platters by a mere 15 nm. The read head measures the resistance, and depending on the magnetization the resistance is measured as either high or low, corresponding to 1's and 0's which is provided back to the computer by the means of a control circuit. The write part of the read-write head consist of an electro-magnet which is able to shift the polarization of the mentioned bits and thereby writing the 1's and 0's.

Problems

Superparamagnetic effect (SPE)

The energy needed to flip between the different magnetization directions is related both to the physical size of the bit and the properties of the material. And simply described, SPE is a physical phenomenon that occurs in data storage when the energy that holds the magnetic spin in the atoms making up a bit either a 0 or 1 becomes comparable to the ambient thermal energy. When this happens bits become subject to random flipping between 0’s and 1’s, corrupting the information they represent. One way of reducing this problem is by keeping the bit-size large, but since the scaling tends to go toward smaller bit-sizes and higher densities, manufacturers try to increase the flip-energy per area by making different microstructures, such as the Current-Perpendicular to plane (CPP) or Tunnel Magnetoresistance (TMR). Even though low coercivity is preffered in GMR devices, materials with higher coercivity might be used to be able to decrease the bit size without experiencing SPE.

Coming technologies

Tunneling magnetoresistance (TMR)

TMR occurs when current flows through two ferromagnets separated by a thin (~1nm) insulator. The resistance is dramatically changed from the parallel to the anti-parallel alignment of the two ferromagnets, which gives higher signal-to-noise ratio for the TMR than GMR and possibility of higher density. The increase in resistance for TMR is in the orders of 200% and often much more. Today, almost all HDD's produced is based on TMR.

Current-perpendicular-to-plane GMR (CPP-GMR)

According to newer research <ref name="Keiichi ">Keiichi Nagasaka, CPP-GMR technology for magnetic read heads of future high-density recording systems, Journal of Magnetism and Magnetic Materials Volume 321, Issue 6, March 2009, Pages 508-511 </ref> TMR is approaching an upper density-limit of 300–400 Gbit/in2. This means that if new technologies are not invented, the ongoing evolution of capacity increase will come to a halt. The CPP-GMR, also called vertical bits because of the reading heads sending a sensing current perpendicular to the disks, promises higher density because of the smaller surface area needed for a single bit.

References

<references/>

• The Discovery of Gigant Magnetoresistance, Scientific background for the 2007 Nobel prize in physics.