more Chapters on this topic:IntroductionTransport Eqs.Spin Proximity/ Spin InjectionSpin DetectionBoltzmann Eqs.Band currentScattering currentMean-free pathCurrent near InterfaceOrdinary Hall effectAnomalous Hall effect, AMR effectSpin-Orbit interactionSpin Hall effectNon-local Spin DetectionLandau -Lifshitz equationExchange interactionsp-d exchange interactionCoercive fieldPerpendicular magnetic anisotropy (PMA)Voltage- controlled magnetism (VCMA effect)All-metal transistorSpin-orbit torque (SO torque)What is a hole?spin polarizationCharge accumulationMgO-based MTJMagneto-opticsSpin vs Orbital momentWhat is the Spin?model comparisonQuestions & AnswersEB nanotechnologyReticle 11
more Chapters on this topic:IntroductionTransport Eqs.Spin Proximity/ Spin InjectionSpin DetectionBoltzmann Eqs.Band currentScattering currentMean-free pathCurrent near InterfaceOrdinary Hall effectAnomalous Hall effect, AMR effectSpin-Orbit interactionSpin Hall effectNon-local Spin DetectionLandau -Lifshitz equationExchange interactionsp-d exchange interactionCoercive fieldPerpendicular magnetic anisotropy (PMA)Voltage- controlled magnetism (VCMA effect)All-metal transistorSpin-orbit torque (SO torque)What is a hole?spin polarizationCharge accumulationMgO-based MTJMagneto-opticsSpin vs Orbital momentWhat is the Spin?model comparisonQuestions & AnswersEB nanotechnologyReticle 11
|
Spin-Orbit Torque (SOT) Spin and Charge TransportAbstract: An electrical current generates the spin-polarized electrons due to the Spin Hall effect. These spin-polarized changes the magnetic properties of the nanowire. This effect is called the effect of spin-orbit torque (SOT). The change of magnetic properties by an electrical current may be substantial. For example, the magnetization can be reversed. The SOT effects describes the dependence of a magnetic properties of a magnetic nanowire on polarity and magnitude of electrical current.Magnetic parameters, which depend on polarity of current in a nanomagnet:
|
|
|||||||||||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
(fact): There are 3 possible mechanism of magnetization reversal: 1: spin-injection into the bulk of nanomagnet; 2: parametric reversal; 3: thermally-activated reversal | |||||||||||||||||||||||||
|
|||||||||||||||||||||||||
click on image to enlarge it |
|
|||||||||||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
![]() |
|||||||||||||||||||||||||
|
|||||||||||||||||||||||||
click on image to enlarge it |
3 terminal MTJ memory cell |
---|
![]() |
Fig.1 . The writing and reading circuits are separated in this design. The reading current flows though the tunnel barrier. The writing current flows through the non-magnetic metal. The spin current is generated at free-layer/ non-magnetic-metal interface, which reverses the magnetization of the free-layer due to SOT effect and a data is recorded. |
click on image to enlarge it |
It improves: (1) memory durability; (2) operational speed;
reading circuit: The reading voltage is applied between "free" and "pinned" layers
The resistivity of the MTJ is lower, when the magnetizations of "free" and "pinned" layers are parallel.
The resistivity of the MTJ is higher, when the magnetizations of "free" and "pinned" layers are anti parallel.
writing circuit: The writing voltage is applied between sides of the non-magnetic metal
The spin current is generated at free-layer/ non-magnetic-metal interface , which induces the torque on the "free" layer due to the SOT effect.
The SOT torque is opposite for the opposite polarities of the writing current and it reverses the magnetization into two opposite directions.
Effect of Spin-orbit torque (SOT) |
---|
There are two spin pumps. (spin pump 1): Localized d-electrons, which constantly creates spin-up conduction electrons. (spin pump 2): Due to Spin Hall effect, spin-left electrons is created. |
![]() |
Big ball shows a large number of spin-polarized electrons of electron gas. The small balls shows direction of injected spin-polarized electrons from two spin pumps. |
Spin direction of the front spin pump is toward left. Spin direction of the backside spin pump is toward up. |
arrows shows the spin-direction and the volume of balls is proportional to the number of the spin polarized electrons |
Details on Spin Torque are here |
click on image to enlarge it |
The creation (origin) of the spin-orbit torque can be divided into two steps. At the first step, the spin-polarized electrons are created by an electrical current. At the second step, the created spin-polarized electrons affects the magnetic properties of the nanomagnet
The creation of the spin-polarized conduction electrons by an electrical current is called the Spin Hall effect (See here for details).
The major mechanism of creation of spin-polarized electrons is the spin-dependent scatterings (See Spin Hall effect for more details)
When there are spin-dependent scatterings, the spin-polarized electrons are accumulated at side edges of an electrical wire (Spin Hall effect). For example, if initially the conduction electrons are not spin-polarized, the probability of a scattering of a spin-up electron is larger to the left with respect to current direction and the probability of a scattering of a spin-down electron is larger to the right, then there are more spin-up electrons at the left side of wire and more spin-down electrons at the right side of the wire.
The spin-polarization is created due to the spin-dependent scatterings across an interface. Typically spin-dependent scattering occurs at an interface between a non-magnetic heavy metal (like Pt, Ta, W) and ferromagnetic metal (like Fe,Co, FeCoB).
The spin-polarization is created due to the spin-dependent scatterings in the bulk of ferromagnetic metal. Typically the spin-dependent scattering occurs in a ferromagnetic metal containing a heavy metal (like FeBTb)
The SOT effect is usually observed in a ferromagnetic metal, where there are two groups of conduction electrons: (group 1) spin-unpolarized electrons and (group 2) spin-polarized electrons (see here). Correspondingly, there are two origins for creations of new spin-polarized electrons.
Due to spin-dependent scattering, some spin-unpolarized electrons becomes spin-polarized. The spin-polarization of these created spin-polarized electrons are different on opposite sides of the wire.
A spin-dependent scattering of already-existed spin-polarized electrons creates the spin-polarized electrons of different spin direction. As a result, there are two groups of spin-polarized electrons of different spin directions: (group 1) large group of "already-existed" spin-polarized electrons and (group 2) tiny group of "newly-created" of spin-polarized electrons. These two groups quickly interact with each other (See Spin Torque)
(influence 1) Spin torque
It is the case when the spin direction of "newly-created" spin-polarized electrons is different from the spin direction of "already-existed" spin-polarized electrons. In this case the Spin Torque is created. As a result, the spin direction of a large number of "already-existed" spin-polarized electrons rotates toward the spin-direction of a tiny number of "newly-created" spin-polarized electrons. This effect is called the Spin Torque.
Depending on the spin direction of "newly-created" spin-polarized electrons and the corresponded direction the Spin Torque., two torque torque can be distinguished: "damp-like" torque and "field-like" torque.
(influence 2) Change of size of nucleation domain for the magnetization switching
The electrical current induces the spin-transfer torque (it is the mechanism of the current induced magnetization reversal in a MTJ). Under influence of the spin-transfer torque, the domain wall of the nucleation domain for magnetization switching may may. As a result,
(influence 3) Change of the spin polarization
The electrical current creates the spin polarized electrons, which added to "already-existed" spin-polarized electrons. Depending on the polarity of the electrical current, the spin direction of "newly-created" spin-polarized electrons is either along or opposite to the spin direction of the "already-existed" spin-polarized electrons. As a result, the total spin polarization either decrease or increases for two opposite directions of the electrical current.
This influence makes current-dependent all magnetic properties, which depend on the spin polarization.
(influence 4) Change of the PMA energy
For a reason, which has not been understood yet, the PMA energy EPMA is changed by the electrical current. It leads to current-dependency of anisotropy field Hanis, coercive field Hc and delta Δ.
Transformation of Hysteresis loop due SOT effect |
||
|
||
The x-axis is applied out-plane magnetic field. Click on image to enlarge it. |
This SOT effect is similar to the effect produced by an usual magnetic filed HDL , which is applied perpendicularly to the electrical current and perpendicularly to the magnetization.
The direction of the magnetic field HFL depends on the magnetization direction. When magnetization rotates along the z-axis. The magnetic field HDL rotates as well.
This SOT effect is similar to the effect produced by an usual magnetic filed HFL , which is applied along the electrical current.
The magnetic field HFL does not depend on the magnetization direction.
The bias current generates a spin-polarized electrons. The spin-polarized electrons at may affect the magnetization near film interface and consequently the the strength of the perpendicular magnetic anisotropy (PMA)
Change of magnetization switching time under current due to SOTdependence is opposite for spin-down to up and spin-down to up switching |
||||||
|
||||||
The magnetization switching time at a different current density. Sample ud30 Volt53B Ta(2.5):FeCoB(1):MgO Nanowire width: 1000 nm, length 200 nm. Measurements date is 10. 2018 | ||||||
Click on image to enlarge it |
Under a bias current, the hysteresis loop are shifted from its center position (See Fig). It looks similar as an additional magnetic field applied perpendicularly to the film.
The switching field from spin-up to down state became different from switching field from spin-down to up state
The Δ and retention time characterize stability of the magnetization against a thermally- activated reversal.
The modulation of the Δ changes the probability thermally-activated magnetization switching
The Meff is magnetization of first magnetic domain (nucleation domain), which triggers the magnetization reversal.
The bias current may move domain wall due to the spin-transfer torque. As a result the size of the nucleation domain becomes smaller or larger. Consequently, the Meff becomes smaller or larger.
The Hall angle or the Hall resistance depends on the magnetization of the ferromagnetic metal, spin-polarization of the conduction electrons and the strength of the spin-orbit (SO) interaction. The bias current generates a spin-polarized electrons. As a result, the spin polarization of electron gas and its distribution across film changes. It causes the change of the Hall angle.
Experiment |
||
|
||
Click on image to enlarge it. |
All SOT measurements were done using the Anomalous Hall Effect (AHE).
Fabrication of FeB, FeCoB and FeTbB nanomagnets connected to a Hall probe
The FeB, FeCoB and FeTbB films were grown on a Si/SiO2 substrate by sputtering. A Ta layer was used as a non-magnetic adhesion layer. The thickness of the Ta was between 2 and 10 nanometers and wafers of different Ta thickness were tested. A nanowire of different width between 100 and 1000 nm with a Hall probe was fabricated by the argon milling. The width of the Hall probe is 50 nm. The FeB and FeCoB layers were etched out from top of the nanowire except a small region of the nanomagnet, which was aligned to the Hall probe. The nanomagnets of different lengths between 100 nm to 1000 nm were fabricated.
The SOT becomes substantial at the current of about 10-100 mA/um2. The heating of nanowire is substantial at this current. It is hard to remove the heating even when a pulse mode is used. For example, in my standard measurements an electrical pulse of 300 ms following 5 s cooling is used. However, there is still a substantial heating in this pulse mode (see below).
A. The SOT effect is linearly proportional to current, but heating ~I2, at a relatively-small current the SOT dominates, but at a higher current the heating dominates.
1) Sweep polarity of current
Usually (but not always) the SOT changes its polarity when the polarity of current is reversed. The heating does not dependent on the current polarity
2) use a narrower and shorter nanowire.
The dissipation of heating is more effective in this case.
Measurement method of current-induced in- plane magnetic field |
|||
|
|||
Dependence of in-plane component of magnetization on in-plane plane magnetic field. The magnetic field is applied in-plane and along the current. HFL is the the offset magnetic field, which is proportional to the current. The HFL is evaluated by linear fitting of the dependence, which measured in Hall configuration. | |||
Click on image to enlarge it |
(fact) In order to explain rather- complex measurement data of the 2nd harmonic method and the symmetrical asymmetrical contributions to FMR resonance, the "damp-like" torque and the "field-like" torque were incorrectly introduced.
Why have the "damp-like" torque and the "field-like" torque been introduced?
The current-induced spin- dynamics has many contributions (see above) and, therefore, is complex. It was incorrectly believed that the usage of two independent subjects as two independent torques could resolve all that complexity.
(The torque as a subject of the Classical Physics to resolve problem of Quantum Mechanic): vs.
The torque is a subject of the Classical Mechanics describing how a force changes rotation of an object. The spin does not describe any rotation (See here). In the Quantum Mechanic, the orbital moment describes the object rotation (See here). The spin describes the properties of the broken time inverse symmetry, according to which the spin can either precess or align along or opposite to an external magnetic field. Any introduction of the classical torque should fit to the fundamental properties of the time inverse symmetry
This method is similar to the method of measurement of anisotropy field Hanis (See here). In this measurement the in-plane component of the magnetization is measured as a function of an external magnetic field Hext. The dependence is linear (See here). Therefore, it can be measured with a high precision. The anisotropy field Hanis is defined as the in-plane magnetic field, at which initially-perpendicular magnetization turns completely into the in-plane direction.
Under a sufficient bias current, additionally to Hext., there are two more additional magnetic fields: 1) effective field of "field-like" torque HFL in direction of current and 2) effective field of "damp-like" torque HDL in direction perpendicularly to the current. Therefore, the magnetization experiences two magnetic field:
In-plane magnetic field along bias current (the x-axis) Htotal,x=Hext,x+HFL ,
In-plane magnetic field perpendicularly to bias current (the y-axis) Htotal,y=Hext,y+HDL ,
Non-existent field like torque |
![]() |
(fact): Field- like torque contradicts with the laws of Quantum mechanics |
(fact): Field- like torque violates the important conservation law associated with the time-inverse symmetry |
Click on image to enlarge it |
As a result, the dependence M vs Hext is shifted on HFL (Hext is applied along x-axis) or HDL (Hext is applied along y-axis). From the shift the HFL are HDL are evaluated.
Note: The same measurements can be done by two method
Method 1. The direct measurement using a nano voltmeter (See Fig.2 above)
Method 2. The 2nd-harmonic measurement using a lock-in amplifier (See here)
Both measurements give the same values of HFL and HDL and nearly the measurement precision. However, the usage of the direct measurement is preferable for the following reasons. From the direct measurement, the dependence of HFL and HDL on bias current can be evaluated.
From measured dependence of M vs H, the contributions of the "field-like" and "damp-like" torque can be separated in the case when they are in the same direction. In contrast to the 2nd-harmonic measurement, in the direct measurement the undesired contribution due to sample heating can be removed.
"Damping-like" torque |
![]() |
The electrical current creates a magnetic field HDL, which is directed along current and perpendicularly to the magnetization M.. Due to HDL, the magnetization M is inclined to the front direction. When external magnetic field Hext is applied, the magnetization M turns in-plane. Following M, HDL turns as well. From measurements of magnetization M vs Hext, HDL can be evaluated. |
Click on image to enlarge it |
This SOT effect is similar to the effect produced by an usual magnetic filed HDL , which is applied perpendicularly to the electrical current and perpendicularly to the magnetization.
The "damping-like" torque is described as
where the effective magnetic field of the "damping-like" torque is defined as
Measurement method of "damping -like" torque HDL |
|||
|
|||
Dependence of in-plane component of magnetization on in-plane plane magnetic field. The magnetic field is applied in-plane and perpendicularly to the current. The center of line is shifted, but its ends are at the same position. The HDL is the the offset magnetic field, which is proportional to the current. The HDL is evaluated by linear fitting of the dependence, which measured in Hall configuration. | |||
Click on image to enlarge it |
It can be measured by the same measurement, which is used to measure the anisotropy field (See here). The in-plane component of magnetization is measured as a function of in-plane magnetic field. The in-plane magnetic field, which is applied perpendicularly to the electrical current.
The HDL gives the field offset for such measurement (See right Fig). From a linear fitting of measured dependence, the HDL is evaluated.
In the case of the "damp-like" torque the dependence M vs H is not linear. Even in the case the fitting gives a high precision.
Effective magnetic field HDL of the "damping -like" torque |
|||||||||
|
|||||||||
Measurements date is 10. 2018 | |||||||||
Click on image to enlarge it |
Effective magnetic field of "damp-like" and "anti damp-like" torque |
||||||
|
||||||
Red arrow shows the spin (the magnetization). Blue arrow show the magnetic field Hext. Green arrow shows the effective magnetic field of the damp Hdamp (anti damp Hanti damp) torque. | ||||||
Click on image to enlarge it |
Difference between "damp-like" torque and "field-like" torque
Properties of effective magnetic field of "damp-like" torque
Its direction changes, when the spin direction (magnetization direction) changes.
Its magnitude changes, when the spin direction (magnetization direction) changes. The magnitude is the largest, when the spin is perpendicular to Hext and the magnitude is the smallest (equals to 0), when the spin is parallel to Hext.
What is the direction of the "damp-like" torque?
3 components of the "damp-like" torque can be distinguished. They are labeled as Hdamp,x , Hdamp,y and Hdamp,z.
Since the direction and magnitude of the effective magnetic field of "damp-like" torque changes when the magnetization direction is changed, the following definition is used:
Hdamp,x aligns magnetization along the x-axis (along bias current)
Hdamp,y aligns magnetization along the y-axis. (in-plane and perpendicularly to bias current)
Hdamp,z aligns magnetization along the z-axis. (perpendicularly to plane)
Direction of effective magnetic field of "damp-like" torque |
||||||
|
||||||
Red Arrow :M is the magnetization (the spin). |
||||||
Click on image to enlarge it |
most probable direction of the "damp-like" torque is Hdamp,x
It is because of the following reason: The bias current breaks the time-reversal symmetry along the x-axis. Similarly, the time-reversal symmetry breaks in this direction, when a magnetic field is applied along the x-axis. Then, damp-like" torque is Hdamp,x aligns the magnetization along this field
Note: The existence of Hdamp,y and Hdamp,z is also allowed by the symmetry.
How Hdamp,x , Hdamp,y and Hdamp,z change their magnitude and direction when magnetization is rotated in the yz-plane and the xz-plane
It is important because from measurements of such rotation both is "field-like" torque and "damp-like" torque are evaluated.
"Field like" torque |
![]() |
The electrical current creates a magnetic field HFL, which is directed along current. Due to HFL, the magnetization M is inclined to the right direction. When external magnetic field Hext is applied, the magnetization M turns fully in-plane at a smaller field Hext= Hanisotropy - HFL,in the left direction. The magnetization M turns field in-plane at a larger field Hext= Hanisotropy + HFL,in the right direction. Measuring the difference between two field, HFL can be evaluated |
Click on image to enlarge it |
This SOT effect is similar to the effect produced by an usual magnetic filed HFL , which is applied along the electrical current.
The "field-like" torque is described as
where the effective magnetic field of the "field-like" torque is defined as
Measurement method of "field -like" torque HFL |
|||
|
|||
Dependence of in-plane component of magnetization on in-plane plane magnetic field. The magnetic field is applied in-plane and along the current. HFL is the the offset magnetic field, which is proportional to the current. The HFL is evaluated by linear fitting of the dependence, which measured in Hall configuration. | |||
Click on image to enlarge it |
It can be measured by the same measurement, which is used to measure the anisotropy field (See here). The in-plane component of magnetization is measured as a function of in-plane magnetic field. The in-plane magnetic field is applied along the electrical current.
The HFL gives the field offset for such measurement (See right Fig). From a linear fitting of measured dependence, the HFL is evaluated.
The dependence of HFL on the current is rather linear. All FeB and FeCoB samples, which I have measured by Nov. 2018, shows the same sign (positive) of HFL.
Effective magnetic field HFL of the "field like" torque induced by SOT |
|||||||||
|
|||||||||
Measurements date is 10. 2018 | |||||||||
Click on image to enlarge it |
Change of anisotropy field Hanis & PMA energy due to the SOT effect |
||||||||||||||||
|
||||||||||||||||
Measurements date is 10. 2018 | ||||||||||||||||
Click on image to enlarge it |
Measurement method of anisotropy field |
||
Dependence of in-plane component of magnetization on in-plane plane magnetic field. The slope depends on the electrical current. The anisotropy field is evaluated from the slope of the line. | ||
Click on image to enlarge it |
Change of coercive field due to the SOT effect (case 1:stronger heating) |
||||||
|
||||||
Data was measured using method described here, which gives measurement precision of coercive field better than 0.1 Oe. | ||||||
Sample Ta(5)/FeB(0.9)/ MgO(6)/ Ta(1)/Ru(5) (Volt55 free44). Measurements date is 06. 2018 | ||||||
Click on image to enlarge it |
When current increases, two effects occur:
1.Heating
Even though the measurements of the coercive field are done in pulse mode, it is difficult completely avoid heating.
Due to the heating coercive field decreases. However, the decrease of the switching field between spin-down to up and switching field between spin-up to down states are absolutely identical and symmetrical (See here)
2. SOT effect
Change of coercive field due to the SOT effect (case 2: weaker heating) |
||||||
|
||||||
Sample ud30 Volt53B Ta(2.5):FeCoB(1):MgO Nanowire width: 1000 nm, length 200 nm. | ||||||
Click on image to enlarge it |
Sample distribution of ΔHc |
![]() |
Measured sample distribution of the current- modulation of the coercive field ΔHc in FeB and FeCoB samples. |
Click on image to enlarge it |
|
||||||
Sample ud30 Volt53B Ta(2.5):FeCoB(1):MgO Nanowire width: 1000 nm, length 200 nm | ||||||
Click on image to enlarge it |
modulation of effective size of nucleation domain |
||||
---|---|---|---|---|
|
||||
Dependence of size of nucleation domain on the electrical current. The black line shows for the case of magnetization switching from spin-down to spin-up state. The red line show for the case of switching from spin-up to spin-down state. | ||||
Click on image to enlarge it |
SOT modulation of Hall angle |
![]() |
Difference of Hall for two opposite polarities of the bias current vs absolute value of current. The proximity of MgO modifies significantly |
Sample: Ta(5):FeCoB ( 1 nm, x=0.3):MgO(7) Volt58A (L58B); nanowire width is 3000 nm, nanowire length is 25 um, length of um etched section is 3 um. For measurements of different magnetic properties of this sample click here |
Click on image to enlarge it |
Dependence of spin polarization on polarity of bias current (SOT effect) |
||||||
|
||||||
click on image to enlarge it |
The "damp-like" torque and "field-like" torque can be measured by following techniques:
1) from measurement of anisotropy field; 2) by 2f-lockin technique; 3) from ST-FMR measurements
Difference between "Damp-like" torque and "Field-like" torque |
|||||||||
|
|||||||||
Click on image to enlarge it |
There are two substantially different types of the Spin-Orbit Torque.
One type does not depend on the magnetization direction of the ferromagnetic metal. It is only depend on the direction of the current. Such torque is called the field-like torque. The second
The second type does depend on the magnetization direction of the ferromagnetic metal. Such torque is called the anti-damping torque.
Even without any current, there are spin-polarized conduction electrons in a ferromagnetic metal. All conduction electrons in a ferromagnetic metal can be divided into two groups: group of spin-polarized electrons and group of spin-unpolarized electrons (See here for details). As was mentioned above, the spin-dependent scatterings originate the Spin-Orbit Torque. Since the properties of the spin-unpolarized electrons does not depend on the magnetization direction, the scattering of these electrons creates the anti-damping torque.. Spin direction of spin-polarized conduction electrons is parallel to the magnetization (See here for details). Therefore, the scattering of the spin-polarized creates the field-like torque
The field-like torque
1.It does not depend on the magnetization direction of the ferromagnetic metal.
2. It is created due to scattering of the spin-unpolarized electrons
The anti-damping torque
1.It does depend on the magnetization direction of the ferromagnetic metal.
2. It is created due to scattering of the spin-polarized electrons
Dependence of SOT on the interface material |
||||||||
|
||||||||
Sample: Ta(5):FeCoB ( 1 nm, x=0.3):MgO(7) Volt58A (L58B); nanowire width is 3000 nm, nanowire length is 25 um, length of um etched section is 3 um. For measurements of different magnetic properties of this sample click here | ||||||||
Click on image to enlarge it |
Measurement of the Spin-orbit torque by 2d harmonic lock-in technique |
||||||
|
||||||
In this method the 2d harmonic of Hall voltage is measured as a function of an in-plane magnetic field | ||||||
Click on image to enlarge it |
The Hall voltage is proportional to the current, the perpendicular components of magnetization and the spin polarization of conduction electrons. When current is modulated with frequency ω, the magnetization direction, the magnetization magnitude and spin-polarization of the conduction electrons may be modulated due to the effect of the SO torque. As a result, the Hall voltage is modulated with frequency 2ω. Measuring this 2d harmonic the amplitude of the SO torque can be estimated.
1. "Field-like" torque
It can be evaluated from the asymmetric (even) component of dependence of the 2d-harmonic voltage vs the in-plane magnetic field, when the magnetic field is applied along the current.
2. "Damp-like" torque
It can be evaluated from the asymmetric (even) component of dependence of the 2d-harmonic voltage vs the in-plane magnetic field, when the magnetic field is applied perpendicularly the current.
3. Current-modulation of anisotropy field
It can be evaluated from the symmetric (odd) component of dependence of the 2d-harmonic voltage vs the in-plane magnetic field. The result is the same whether the magnetic field applied along or perpendicularly to the current.
The fields of the the spin-orbit torque can be calculated from the following dependance of the 2d-harmonic voltage VHall,2ω vs applied in-plane magnetic field Hx :
where ΔHanis,ω is the current induced change of the anisotropy field Hanis; ΔHoff,ω is the effective magnetic field HFL,ω of the "field-like" torque, when Hx is applied along electrical current; and ΔHoff,ω is the effective magnetic field HDL,ω of the "damp-like" torque, when Hx is perpendicularly to the electrical current;
the odd and even components can be calculated as
Rwire is the is the ohmic resistance of the wire; RHall,0 is the is the Hall resistance, when a in-plane magnetic field is not applied ;
Hanisot is the anisotropy field, which can be measured directly (See here) with a high precision or from 1st harmonic with a moderate precision.
The Hanisot can be evaluated from the following dependance of the 1d-harmonic voltage VHall,ω vs applied in-plane magnetic field Hx :
The 2d harmonic lock-in technique measures the current- modulation of the effective magnetic field HDL of "damp- like" torque, the effective magnetic field HFL of "field- like" torque and the anisotropy field ΔHanis
Without electrical current, the in-plane component of the magnetization Mx depends on the applied external in-plane magnetic field Hx as (see here)
where Hanis is the anisotropy field
As was demonstrated above, the spin-orbit torque (SOT) produces the offset magnetic field ΔHoff and changes the anisotropy field Hanis on ΔHanis. As a result, the Eq.(4.1) is modified as
where ΔHoff equals to HFL when the in-plane magnetic field is applied along current and ΔHoff equals to HDL when the in-plane magnetic field is applied in-plane and perpendicularly to the current
In the case when
Eq.(4.2) can be simplified as
or
where
from (4.2) we have .
The z- component of the magnetization Mz can be calculated as
or
The Hall voltage is calculated as
when magnetization is not perpendicular to plane, the Hall voltage is calculated as
where Mz is the perpendicular-to-film component of magnetization, RHall,0 is the Hall resistance when the magnetization is perpendicular to the film (Mz =M).
When the current is modulated with frequency ω ,
both the ΔHoff and ΔHanis are modulated as well:
Using a trigonometric relation
and substituting Eqs (4.7) (4.10),(4.11) into Eqs. (4.7) gives the Hall voltage VHall,2ω of the 2d harmonic (the coefficient at cos(2ωt)) as
In a lock-in measurement it is convenient to use the reference voltage Vω rather than reference current Iω
where Rwire is the resistance of metallic nanowire.
Substituting Eqs. (4.2) and (4.14) into Eq. (4.13) gives
or
The voltage of the second harmonic has two component. The first component is proportional to ΔHoff and is an odd function in the respect to Hx. The first component is proportional to ΔHanis and is an even function in the respect to Hx. Therefore, the voltage of the second harmonic can be calculated as
Eq (4.17) can be written in a symmetrical form as
where
1st harmonic |
![]() |
Perpendicular-to-plane component of magnetization as a function of applied in-plane magnetic field. |
Click on image to enlarge it |
When current is small, the ΔHoff and ΔHanis can be ignored. Than, the Hall voltage VHall,ω of 1st harmonic can be calculated from Eq.(4.9) as
substitution of Eq(4.1) into Eq.(4.20) gives the Hall voltage VHall,ω of 1st harmonic as
The ratio of voltage of 1st harmonic to the voltage of 1st harmonic can be calculated as
Measurement of anisotropy field Hanis |
||
|
||
The arrow shows the direction and magnitude of the applied in-plane magnetic field. The ball shows the magnetization direction. Without magnetic field the magnetization is perpendicularly-to-plane. Under magnetic field, the magnetization turns toward magnetic field. The field, at which the magnetization turns completely in-plane, is called the anisotropy field. The dots of the right graph shows experimental data. Measurement date: May 2018. | ||
Click on image to enlarge it |
Electrical current can induce spin torque or reduce the exchange interaction between localized electrons. This can change the direction of magnetization of a material.
Current-induced magnetization reversal in FeBTb film
|
||||
|
Two current-induced effects, which can lead to the current-induced magnetization reversal:
Both effects occur because of transfer of delocalized (conduction) spin-polarized electrons from a point to point, which alters an equilibrium spin polarization in a material.
The spin torque occurs when the delocalized spin-polarized electrons are transferred from one material to another by a drift or a diffusion current. When spin-polarized delocalized electrons are injected, it is not only change magnitude of spin accumulation, but also it changes spin direction of spin accumulated electrons. As result, the spin direction of localized and delocalized electrons becomes different. This induces the torque, which may turn or reverse the spin direction of the localized electrons.
Note: At one place an electron gas may have only one spin direction of its spin accumulation. In the case when the electrons with a different spin direction is injected, the spins quickly relax and the spin accumulation of only one spin direction remains. The final spin direction is different from initial spin direction and from the injected spin direction. Details see here and here
The spin torque may change magnetization direction in a material because of the exchange interaction between localized and delocalized electrons.
There are several effects which can cause the current-induced spin torque:
1)The spin-transfer torque.
It occurs because of transfer of spin-polarized electron from material to material by a drift or diffusive spin current. Example: the spin transfer between electrodes in a MTJ or GMR junction. The polarity of the spin-transfer torque depends on mutual magnetization directions of the electrons.
2) The spin-orbit (SO) torque.
It occurs in magnetic or non-magnetic metals in which there are substantial spin-dependent scatterings. Due to spin-dependent scatterings a spin-polarized current flows perpendicularly to the flow of spin- unpolarized drift current. The spin is accumulated at one side of a metallic wire and the spin is depleted at another side. The spin accumulation(depletion) may cause the spin torque at sides of the wire, which magnitude and direction is proportional to the drift current. The accumulated spin may have different spin direction than the spin direction of the equilibrium spin polarization.
Current-induced magnetization reversal in FeBTb film
|
||||
|
Current-induced magnetization reversal in FeBTb film
|
||||
|
(about systematic errors of 2nd harmonic measurements)
Regarding 2nd harmonic method, I have to disagree with you. The technic is reliable if you manipulate it correctly, and like any experiment there is always the risk of an “artifact” effect not taken into account. I believe that we have reached today a conclusion on how to perform an analysis using 2nd harmonic and to take into account spurious effects..
The problem of the 2nd harmonic measurement is that it has too many independent contributions, such as
1. magnetization precession due to spin injection
2. magnetic field Hoff, which is induced by the spin accumulation
3. Current dependency of anisotropy field
4. PHE/AMR effect.
Three of them can be used for magnetization reversal by an electrical current.
The fact is that the 2nd harmonic measurement does not have enough data to describe its own measured data, because of a large number of different independent contributions. The new method, which I have developed, measures each contribution individually and independently of other contributions. Each contribution has a rich and interesting Physics, which can be individually optimized for an efficient magnetization reversal.
--------------------------------
The 2nd harmonic measurement has the similar tendency as the current dependency of the magnetic field Hoff, which is induced by the spin accumulation. Therefore, it is OK to use data of the 2nd harmonic measurement in a publication, in which different tendencies are studied and discussed, and in which some systematic error is not a big issue. However, for a technology optimization, the use of a direct and more reliable measurement is better.
(about field- like torque) (from Sreyas Satheesh) I had some serious doubts regarding the field like torque terms. You had mentioned the field-like field to be independent of the magnetization direction and to be directed along the direction of the current. However, in some of the works, I had found it to be orthogonal to the direction of the current. Ref: 1)Garello, K. et al. Symmetry and magnitude of spin-orbit torques in ferromagnetic heterostructures. Nature Nanotech. 8, 587–593 (2013). 2)Miron, I. M. et al. Perpendicular switching of a single ferromagnetic layer induced by in-plane current injection. Nature 476, 189–193 (2011).
(about torque & spin dynamic & Quantum mechanic)
There is only one torque, which is damping (or anti damping torque) of the Landau-Lifshitz equation. The introduction of any possible torque of different types or a different direction violates the rules of the Quantum Mechanics.
The spin is a pure quantum- mechanical object and the torque is the object of classical physics. Therefore, strictly-speaking it is incorrect to use the torque for a description of the spin dynamics. However, it is still possible to use the torque for the spin dynamics, when the torque closely mimics and well- approximates all features of the quantum-mechanical dynamics of the spin. The reason for the use of the torque is to simplify the description and understanding of the spin dynamics. However, in contrast to the classic mechanic, in which the torque may have any direction and magnitude, the quantum- mechanical rules limit the torque to only one possible direction and make the torque strength dependent on the spin precession angle.
The spin dynamics, as any quantum mechanical process, is described by a transition between quantum levels. In the case of the spin, the lower-energy level corresponds to spin direction along the magnetic field (spin-up) and the lower-energy level corresponds to spin direction opposite to the magnetic field (spin-down) . Only possible other quantum states of the spin are the states, whose energy is between the spin-up and spin-down levels and which corresponds to the spin precession at a different spin precession angle.
For example, in an equilibrium the spin is in the spun-up state and there is no spin precession. When there is an injection of spin-down electrons, both the spin-up and spin-down quantum states are partially filled, which corresponds to the spin precession. The spin precession is larger when there are more spin-down electrons. This quantum mechanical process can be described rather well and reasonably correctly by the damping torque (or anti-damping torque) of the Landau-Lifshitz (LL) equation.
Except for the transition between the spin-up and spin-down quantum levels, which is described by the damping torque of LL Eqs, I do not see any other options for a possible quantum spin dynamic and, therefore, any possibility for introduction of the other torque. For example, another possible mechanism of the spin reversal, the parametric magnetization reversal, when the magnetization direction is modulated in the resonance with spin precession, is also described by the transition between the spin-up and spin-down quantum levels and, therefore, the same anti-damping torque of LL Eqs. You can find more explanations about this in this video (click here)
There is no such thing as the field-like torque. However, there is a magnetic field, which is induced by spin-accumulated electrons. Since the spin accumulation is created by the current, this magnetic field can be modulated by current and can be used for the parametric magnetization reversal. One of the in-plane components of this magnetic field is incorrectly associated with the damp-like torque and another in-plane component is incorrectly associated with the field-like torque. The reason for that is the symmetry of the 2nd harmonic measurement with respect to the magnetization reversal. More details about this magnetic field and its measurement you can find in this video (click here).
Two papers, which you have mentioned, are two important papers, in which the field-like torque was introduced based on the 2nd harmonic measurement. However, the problem of the 2nd harmonic measurement is that it is influenced by too many parameters and the data of measurement of the 2nd harmonic alone is not sufficient to explain the changes of all those parameters. Now the direct measurements have clarified the situation. You can find the explanation about it in the mentioned video.
In the two mentioned papers, the puzzling data of the 2nd harmonic measurement was explained by the introduction of the field-like torque. It is a great, but incorrect idea. It is a very natural way of the development of science. Some great, but incorrect models may exist until new data clarifies the situation.
Parametric magnetization reversal. Spin- orbit torque Conference presentation. Intermag 2021 |
---|
|
|
|||||||||||||||||||||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
![]() |
|||||||||||||||||||||||||||||||||||
click on image to play it |
I am strongly against a fake and "highlight" research
I will try to answer your questions as soon as possible