2001
DOI: 10.1103/physrevb.63.064407
|View full text |Cite
|
Sign up to set email alerts
|

Interface resistance of disordered magnetic multilayers

Abstract: We study the effect of interface disorder on the spin-dependent interface resistances of Co/Cu, Fe/Cr, and Au/Ag multilayers using a newly developed method for calculating transmission matrices from first-principles. The efficient implementation using tight-binding linear-muffin-tin orbitals allows us to model interface disorder using large lateral supercells whereby specular and diffuse scattering are treated on an equal footing. Without introducing any free parameters, quantitative agreement with experiment … Show more

Help me understand this report
View preprint versions

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1

Citation Types

15
138
1

Year Published

2002
2002
2008
2008

Publication Types

Select...
7
1
1

Relationship

2
7

Authors

Journals

citations
Cited by 117 publications
(154 citation statements)
references
References 23 publications
15
138
1
Order By: Relevance
“…The theory is valid under two conditions: ͑i͒ the system should be diffuse, i.e., the elastic mean free path ᐉ ͑including scattering at interfaces͒ should be smaller than typical sample scales and ͑ii͒ the ferromagnetic elements should have an exchange splitting ⌬, which is large enough that the magnetic coherence length ᐉ c ϭប/ͱ2m⌬ Ͻmin(ᐉd F ), where d F is the thickness of the ferromagnetic layer. These conditions are usually fulfilled in transitionmetal systems: Deviations from diffuse behavior, such as quantum-size effects and breakdown of the series resistor model, are small or controversial, 20,24 whereas the magnetic coherence length is of the same order as the lattice constant in high-T c transition-metal ferromagnets. 10,25 We obtain identical results by two methods: The first one is a combination of the Boltzmann-like method of Schep et al 18 for collinear systems and the random-matrix theory of Waintal et al 14 The second one is an extension of magnetoelectronic circuit theory 6 to arbitrary resistors.…”
Section: Fig 1 Different Realizations Of Perpendicular Spin Valvesmentioning
confidence: 99%
“…The theory is valid under two conditions: ͑i͒ the system should be diffuse, i.e., the elastic mean free path ᐉ ͑including scattering at interfaces͒ should be smaller than typical sample scales and ͑ii͒ the ferromagnetic elements should have an exchange splitting ⌬, which is large enough that the magnetic coherence length ᐉ c ϭប/ͱ2m⌬ Ͻmin(ᐉd F ), where d F is the thickness of the ferromagnetic layer. These conditions are usually fulfilled in transitionmetal systems: Deviations from diffuse behavior, such as quantum-size effects and breakdown of the series resistor model, are small or controversial, 20,24 whereas the magnetic coherence length is of the same order as the lattice constant in high-T c transition-metal ferromagnets. 10,25 We obtain identical results by two methods: The first one is a combination of the Boltzmann-like method of Schep et al 18 for collinear systems and the random-matrix theory of Waintal et al 14 The second one is an extension of magnetoelectronic circuit theory 6 to arbitrary resistors.…”
Section: Fig 1 Different Realizations Of Perpendicular Spin Valvesmentioning
confidence: 99%
“…Moreover, according to Xia et al 37 g ↑↓ / A Ӎ 5·10 14 ⍀ −1 m −2 . For a quantitative assessment of the relations ͑5͒ and ͑9͒ we assume Ϸ 5°͑sin 2 = 0.01͒ and =10 11 s −1 ͑ប =65 eV͒.…”
Section: Spin Pumping Effectmentioning
confidence: 99%
“…The spin polarization of electrons emitted by Py has been measured to be p ഠ 0.4 in point contacts [19], the magnetization per atom is f ഠ 1.2, and the Landé factor is g L ഠ 2.1 [12]. The interface conductance of metallic interfaces with Fe or Co is of the order of 10 15 V 21 m 22 , with significant but not drastic dependences on interface morphology or material combination [20]. This corresponds to roughly one conducting channel per interface atom.…”
mentioning
confidence: 99%
“…His expression for the damping coefficient [Eq. (20) in Ref. [9] ] scales like ours as a function of layer thickness, but differs as a function of material parameters.…”
mentioning
confidence: 99%