14 Magnetic Tunnel Junctions

Dr. Anchal Srivastava

Contents of this Unit

1.      Introduction: Magnetic Tunnel Junctions.

2.      Tunnel Magnetoresistance.

3.      Basic phenomena in MTJs.

4.      Developments of MTJs and key ingredients for TMR.

5.      Additional Layer.

6.      Coherent spin tunnelling.

7.      Future perspective.

8.      Summary.

Learning Outcomes

After studying this module, you shall be able to

1. Learn about the basics of Magnetic tunnel junctions or MTJs are nanostructured devices within the field of magneto-electronics or spin electronics and spintronic.
2. Learn about the Tunnel magnetoresistance (TMR) and how this is a magneto-resistive effect that occurs in a magnetic tunnel junction (MTJ).
3. Learn about the basic phenomena in MTJs and learn about the developments of MTJs.
4. Learn about coherent spin tunnelling and what the future perspectives of MTJs are.

1.  INTRODUCTION: MAGNETIC TUNNEL JUNCTIONS (MTJs): –

In the early 1990s, high magnetoresistance (MR) was discovered for magnetic tunnel junction (MTJ) material. MTJ material is made of at least two magnetic layers separated by an insulating tunnel barrier. The current flows perpendicular to the film plane. The best results have been achieved with aluminium-oxide tunnel barriers. Since the initial experimental discovery of MTJ material with promising MR, the technique of producing these materials, as well as key properties, has been dramatically improved. Tunnelling MR values are in the 20–50 % range.

Instead of using inorganic insulating barrier layers like aluminium oxide, attempts are also made to insert organic insulating layers to make MTJ devices. This can also lead to the fabrication of flexible organic devices in future.

Around 1959, James Moor predicted that there would be a reduction in the transistor size with time. Ever since its depiction the trend in miniaturization of electronic devices has faithfully followed what is known as Moor’s law. It implies that every 18 months the reduction in the size doubles or the number of transistors on a chip doubles or the processing power of computers doubles. However, following this law, the devices have reached now a lowest size of 100 nm and deviation from the law has begun. It is not only increasingly difficult to achieve smaller and smaller sizes less than 100 nm but also difficult to retain linear nature of the graph that Moor had predicted. Below a size of 100 nm we know that besides the ‘surface effect’, materials also have size-dependent properties. Therefore, Nano devices using active or passive nanocomponents cannot be expected to behave like those of large (micrometre) size devices and components. Interestingly this very size-dependent nature can be used to obtain some novel devices, which were not imagined earlier. For example, single electron transistor (SET) is a completely new device due to unique properties of quantum dots. Magnetic Spin Valve and Magnetic Tunnel Junction (MTJ) using nanomaterials are some other high speed devices which are the products of nanotechnology.

The electronic devices with typical dimensions of few nano-meters in either of three directions display not just the miniaturization but unique properties not known over last 5–6 decades since the beginning of solid state devices. Single Electron Transistor (SET), spin valves, and Magnetic Tunnel Junctions (MTJ) are conceptually new devices based on nanotechnology. Such devices are fast, compact, relatively cheap and finding their way to market. Spin valve type devices are already being used in personal computers to ‘read’ disk which have enabled to increase data storage capacity of hard disks.

Interestingly, spin valve and MTJ are based on a concept which itself is growing into an area in itself known as spintronics or spin based electronics or magneto-electronics. It is well understood that an electron (or hole) has both charge and spin. However, electronics has so far used only the charge property of electron (or hole) and spin has been neglected. It has been now realized in recent years that if spin of an electron (or hole) is taken into account, properly fabricated devices would lead to some superior devices. Using an external magnetic field, spin transport can be controlled. Advantage with spin is that it cannot be easily destroyed by scattering from collisions with other charges, impurities or defects.

Many spin-based devices like Spin-FET, Spin-LED, Spin-RTD, optical switches with THz frequency, modulators, encoders, decoders, and q-bits for quantum computers are on the hot list of scientists and the technologists. We consider here devices based on Giant Magneto Resistance (GMR), spin valve, Magnetic Tunnel Junction (MTJ) and Spin Field Effect Transistor (SFET).

Fig: 1 – MgO-Based MTJ Sensor.

Magnetic tunnel junctions or MTJs are nanostructured devices within the field of magneto-electronics or spin electronics, hereafter called spintronic. In this area, the experimental observation of sizable and tunable magnetoresistance (change of materials resistance due to external fields) is intimately related to the exploitation of not only charge of the electrons but also its spin. The discovery of giant magnetoresistance (GMR; Barthelemy et al. 1999; see Giant Magnetoresistance) in multilayered ferromagnetic films separated by thin metallic spacers has initiated an enormous research interest, particularly also for a wealth of potential applications, e.g., in data-storage devices.

Fueled by these developments and earlier efforts in tunneling devices (Tedrow and Meservey 1971, Julliere 1975), Moodera et al. (1995) and Miyazaki and Tezuka (1995) have discovered that the tunneling current between two ferromagnetic films separated by a thin oxide layer strongly depends on an external magnetic field, an effect now known as tunnel magnetoresistance (TMR). Since then, the impact of MTJs on the field of spintronics has hugely expanded, particularly due to the enormous magnitude of the observed magnetoresistances at room temperature and its impact on potential applications (Chappert et al. 2007). Experiments using crystalline MgO barriers have dramatically improved the magnitude of TMR how the electronic structure of the complete tunneling junction may lead to enormous spin-selectivity.

2. TUNNEL MAGNETORESISTANCE (TMR):

Tunnel magnetoresistance (TMR) is a magneto-resistive effect that occurs in a magnetic tunnel junction (MTJ), which is a component consisting of two ferromagnets separated by a thin insulator. If the insulating layer is thin enough (typically a few nanometers), electrons can tunnel from one ferromagnet into the other. Since this process is forbidden in classical physics, the tunnel magnetoresistance is a strictly quantum mechanical phenomenon.

Magnetic tunnel junctions are manufactured in thin film technology. On an industrial scale the film deposition is done by magnetron sputter deposition; on a laboratory scale molecular beam epitaxy, pulsed laser deposition and electron beam physical vapor deposition are also utilized. The junctions are prepared by photolithography.

3. BASIC PHENOMENA IN MTJs:

When electrons are tunneling between two ferromagnetic metals, the magnitude of the tunneling current depends on the relative orientation of the magnetization of both electrodes.

This can be understood from a few elementary arguments:

(i) The tunneling current is, in first order, proportional to the product of the electrode density of states

(DOS) at the Fermi level;

(ii) In ferromagnetic materials, the ground-state energy bands in the vicinity of the Fermi level are shifted in

energy, yielding separate majority and minority bands for electrons with opposite spins; and

(iii) Assuming spin conservation for the tunneling electrons, there are two parallel currents of spin-up and

spin-down character.

As a result of these aspects, the current between electrodes with the same magnetization direction should be higher than those with opposite magnetization.

The change in resistance between antiparallel and parallel magnetization (normalized to the parallel resistance) is given by

where P1,2 are the so-called tunneling spin polarizations determined by the relative difference in DOS at the Fermi level. It is crucial to realize that not all electrons present at the Fermi level can efficiently tunnel through the barrier, and that this simple equation is not able to capture the physics behind a number of observations in MTJs. In many cases the spherically symmetric s-like electrons, which have a much lower DOS at the Fermi level, dominantly tunnel through the barrier, and the interface between the insulating tunnel barrier and the ferromagnets plays an essential role.

Fig. 2 – It illustrates the mechanism of TMR. Up; when the magnetization if parallel and below; when the magnetization is anti-parallel, leading to reduction of total tunneling current.

Nonetheless, this expression clearly demonstrates the presence of a magnetoresistance effect and the relevance of the magnetic character for the spin polarization of the tunneling electrons. Moreover, it shows that so-called half-metallic metals with only one of the two spin species available at the Fermi level (De Groot et al. 1983) may, in principle, engender infinitely high TMR. Indications for such behavior are indeed observed, for instance, in La₂/3Sr1/3MnO₃/SrTiO₃/La2/3Sr1/3MnO₃Sr (Bowen et al. 2003) and Co2FeAl0.5Si0.5 (Tezuka et al. 2007).

An important aspect for the presence of TMR is the ability to independently manipulate the direction of the magnetization of the electrodes. This can be accomplished by several methods which include the (sometimes combined) use of intrinsic differences in magnetic hysteresis of the ferromagnetic materials, exchange biasing with antiferromagnetic thin films (Coehoorn 2003), and antiferromagnetic interlayer coupling across nonmagnetic metallic films. The room-temperature resistance changes for a MJT with a MgO barrier. Two soft-magnetic CoFeB electrodes with different coercivities are used to create a clear distinction between the resistance levels in parallel and antiparallel alignment of the magnetization.

Fig.3. Resistance change in a magnetic tunnel junction consisting of (Co25Fe75)80B20/2.1 nm MgO/(Co25Fe75)80B20. The data are taken at room temperature. The arrows indicate the orientation of the CoFeB magnetization. Adapted from Lee Y M, Hayakawa J, Ikeda S, Matsukura F, Ohno H 2007 Effect of electrode composition on the tunnel magnetoresistance of pseudospin-valve magnetic tunnel junction with a MgO tunnel barrier. Appl. Phys. Lett. 90 (3), 212507.

4. DEVELOPMENTS OF MTJs AND KEY INGREDIENTS FOR TMR:

Tedrow and Meservey (1971) report the first experiments on spin tunneling. In their case, only one electrode is ferromagnetic (Ni), the other being a superconductor (Al). They have found that though minority electrons dominate the DOS at the Fermi level of Ni, majority electrons are most efficiently tunneling through the thin Al barrier. Later, it is suggested by Hertz and Aoi (1973) and Stearns (1977) that, although the dominant species of electrons at the Fermi level of transition metal ferromagnets are minority d-electrons, they do not couple well with the states over the barrier. Instead, highly dispersive majority s-like electrons have a much larger overlap integral with states in the barrier which leads to a larger transmission probability for these electrons. Moreover, the interaction between the s- and d-electrons (s–d hybridization) leads to a suppression of the s-DOS in regions of large d-DOS, which is also the case at the Fermi level of a 3d transition metal ferromagnet. Consequently, this induces a spin polarization of the s-DOS at the Fermi energy.

After these seminal papers on ferromagnetic tunneling, including the first prediction of a TMR effect by Julliere (1975), it took around two decades to do the same experiment with two ferromagnetic electrodes, as mentioned in the introduction (Moodera et al.1995, Miyazaki and Tezuka 1995). It should be noted that in all these experiments Al2O3 is preferred as barrier material, primarily since it allows an easy growth of a pinhole-free thin barrier by natural, thermal, or plasma oxidation of Al thin films.

We know that TMR is directly related to the tunneling spin polarization (P) induced by the ferromagnetic DOS. One may imagine that P is not constant over the whole Fermi surface, and varies depending on which direction in k-space one probes, that is, on the crystallographic orientation of the electrode at the interface with the tunnel barrier. The demonstration of such crystal anisotropy of the TMR is given by Yuasa et al. (2000), who have shown that the use of single-crystalline Fe electrodes of different orientations in MTJs resulted in a substantially different TMR.

5. ADDITIONAL LAYER:

Inserting an additional layer at the barrier–ferromagnet interface has been investigated to rigorously probe the origin of tunneling spin polarization P. LeClair et al. (2000) show that inserting one monolayer of Cu between the bottom Co electrode and the Al2O3 barrier leads to a strong reduction of TMR.

Yuasa et al. (2002) have further developed these experiments by achieving sharp interfaces between single crystalline Co (001) and Cu (001) using molecular beam epitaxy. They explain that majority electrons tunneling from NiFe into Co would transmit easily as compared to minority electrons which have a higher probability to be reflected at the Co–Cu interface. If multiple scattering occurs between the Co–Cu and Cu–Al2O3 interfaces, the minority electrons would form resonant quantum well states in the Cu layer, resulting in the oscillatory behavior of TMR.

Although most ferromagnets display a positive P in conjunction with Al2O3, Kaiser et al. (2005a) have reported that Co–Gd alloys can exhibit both positive and negative P systematically depending on the alloy composition. It is known that in these alloys the Co and Gd ferromagnetic subnetwork magnetization is aligned antiparallel with respect to each other, which may significantly influence the tunneling spin polarization. Now the sign of P depends on the orientation of the respective subnetwork magnetization with respect to the applied field.

The P from either of these subnetworks will be positive when its magnetization is aligned with the applied magnetic field, in contrast to the moments of the other subnetwork. Kaiser et al.(2005a) found that the measured P is the sum of independent spin-polarized tunneling currents from the Co and Gd subnetworks, resulting in a sign change of P with alloy composition. When combined with traditional ferromagnetic materials with positive P in an MTJ, this leads to positive or negative TMR, depending on the sign of the Co–Gd polarization.

6. COHERENT SPIN TUNNELLING:

One aspect which is highly unlikely in tunneling through an amorphous barrier is k conservation of the electron wave vector. On the contrary, in a crystalline barrier, k conservation (also known as coherent tunneling) is a distinct possibility.

This also implies that a wave vector selected at one interface, efficiently couples to a corresponding wave vector at the other interface.

Keeping in mind that P is not constant over the whole Fermi surface, one could imagine that using a certain electrode–barrier interface in a certain crystallographic orientation which would result in efficient electron tunneling for wave functions and which also have specific symmetries.

This in turn could lead to a very large tunneling spin polarization, even though the averaged DOS at the Fermi level of the ferromagnet is only moderately polarized.

7. FUTURE PERSPECTIVE:

In the field of spintronics, MTJs display magnetoresistance effects due to the spin dependence of the tunneling current when dealing with ferromagnetic electrodes. The physics behind TMR has been experimentally and theoretically explored by introducing novel concepts and engineered material combinations. This has dramatically increased our knowledge of tunneling between ferromagnetic materials.

See also: Giant Magnetoresistance; Half-metallic Magnetism; Magnetic Recording Systems: Spin Electronics;

Magnetic Recording Systems: Spin Valves; Magnetoresistive Heads: Physical Phenomena; Magnetic Tunnel

Transistor; Multilayers: Interlayer Coupling; Spin-polarized Scanning Tunneling Microscopy.

8. SUMMARY:

1.  Magnetic tunnel junctions or MTJs are nanostructured devices within the field of magneto-electronics or

spin electronics, hereafter called spintronic. In this area, the experimental observation of sizable and tunable

magnetoresistance (change of materials resistance due to external fields) is intimately related to the

exploitation of not only charge of the electrons but also its spin.

2. It has been now realized in recent years that if spin of an electron (or hole) is taken into account, properly

fabricated devices would lead to some superior devices. Using an external magnetic field, spin transport can

be controlled. Advantage with spin is that it cannot be easily destroyed by scattering from collisions with

other charges, impurities or defects.

3. Tunnel magnetoresistance (TMR) is a magneto-resistive effect that occurs in a magnetic tunnel junction

(MTJ), which is a component consisting of two ferromagnets separated by a thin insulator. If the insulating

layer is thin enough (typically a few nanometers), electrons can tunnel from one ferromagnet into the other.

Since this process is forbidden in classical physics, the tunnel magnetoresistance is a strictly quantum

mechanical phenomenon.

4.   When electrons are tunneling between two ferromagnetic metals, the magnitude of the tunneling

current depends on the relative orientation of the magnetization of both electrodes.

This can be understood from a few elementary arguments:

(i)  the tunneling current is, in first order, proportional to the product of the electrode density of states

(DOS) at the Fermi level;

(ii)   in ferromagnetic materials, the ground-state energy bands in the vicinity of the Fermi level are shifted

in energy, yielding separate majority and minority bands for electrons with opposite spins; and

(iii)  assuming spin conservation for the tunneling electrons, there are two parallel currents of spin-up and

spin-down character.

As a result of these aspects, the current between electrodes with the same magnetization direction should be higher than those with opposite magnetization.

5.  An important aspect for the presence of TMR is the ability to independently manipulate the direction of

the magnetization of the electrodes. This can be accomplished by several methods which include the

(sometimes combined) use of intrinsic differences in magnetic hysteresis of the ferromagnetic materials,

exchange biasing with antiferromagnetic thin films (Coehoorn 2003), and antiferromagnetic interlayer

coupling across nonmagnetic metallic films.

6.  Tedrow and Meservey (1971) report the first experiments on spin tunneling. In their case, only one

electrode is ferromagnetic (Ni), the other being a superconductor (Al). They have found that though minority

electrons dominate the DOS at the Fermi level of Ni, majority electrons are most efficiently tunneling

through the thin Al barrier.

7.  The demonstration of such crystal anisotropy of the TMR is given by Yuasa et al. (2000), who have shown

that the use of single-crystalline Fe electrodes of different orientations in MTJs resulted in a substantially

different TMR.

8.  Inserting an additional layer at the barrier–ferromagnet interface has been investigated to rigorously

probe the origin of tunneling spin polarization P. LeClair et al. (2000) show that inserting one monolayer of

Cu between the bottom Co electrode and the Al2O3 barrier leads to a strong reduction of TMR.

9.  One aspect which is highly unlikely in tunneling through an amorphous barrier is k conservation of the

electron wave vector. On the contrary, in a crystalline barrier, k conservation (also known as coherent

tunneling) is a distinct possibility.

10. In the field of spintronics, MTJs display magnetoresistance effects due to the spin dependence of the

tunneling  current  when  dealing  with   ferromagnetic  electrodes.  The   physics  behind  TMR   has   been experimentally and theoretically explored by introducing novel concepts and engineered material combinations.

This has dramatically increased our knowledge of tunneling between ferromagnetic materials.

Suggested Readings

1. Charles Poole, The Introduction of Nanotechnology.
2. W.I. Fletcher – An Engineering approach to Digital Design, PHI, 3Edition 3, Hand book of nanotechnology by B. Bhushan.
3. Chapter:  –  Nanostructured  Materials  for  Engineering  Applications,  pp  23-39,  Nanomagnetic Materials. Authors: – Silvana Da Dalt, Priscila Chaves Panta, Juliano Cantarelli Toniolo.
4. .Nanomagnetic Materials: A review, Authors: T.K.S.P. Gupta and S.S. Verma. Technical Proceedings of the 2003 Nanotechnology Conference and Trade Show, Volume 3, Nano Composites Chapter 6.