1. New magnetic materials for Spintronics
Juana Moreno
Department of Physics and Astronomy
Center for Computation and Technology
Louisiana State University 1

2. Computational Materials Science
Petascale computing & the development of new formalism, algorithms and codes will allow the accurate modeling of materials.
Huge opportunities for young computational materials scientists to make breakthroughs in the study of complex materials for future high-tech devices.

3. Exponential growth in computing power:

4. Analyzer -www.britannica.com
Bush Differential
Analyzer -
ENIAC - U.S. Army Photo
(D. P. Landau, UGA)
CRAY 1 -
CRAY X1 – ORNL/CCS

5. A Truly interdisciplinary approach
Biological
Sciences
Physical
Sciences
Medical Science
Engineering
Social Sciences
Computational
methods
Environmental
Sciences
Arts & Humanities
Economics
Engineering Animation, Inc.

6. What is spintronics? And why?
Spin-unpolarized current:
Electrons move with random spin
orientation
Spin-polarized current:
Electrons move with same spin
orientation

7. We need faster, smaller, more efficient chips
By 2020 a transistor in a chip may reach the size of a few atoms.
Electronics based on a new paradigm is needed!

8. Devices based on “static” spins
Giant magneto-resistance hard-disks
GMR effect (1988)
IBM hard disk (1997)
[Prinz, Science 1998]

9. The advantage of using spin….
Silicon technology
(no equivalent change)
0.1 penny/MB!
GMR introduction
(advantage of using spin in devices)

10. Devices based on spin-polarized currents
Spin Field Effect Transistor
Datta-Das (1990)
Spin precession due to spin-orbit interaction with spin-orbit splitting controlled by gate potential p p - -
type
type n n - -
type
type
Spin LED
H split the spin levels circularly polarized light. + -
spin
injector hν
Very small spin injections!

11. Metal-semiconductor interface
Difficulties to build spintronic devices?
Metal devices don’t amplify the current.
Mixed devices have problems with spin injection.
Large scattering at the interface between ferromagnetic metal-semiconductor due to conductivity mismatch.
Metal-semiconductor interface
Solution: a magnetic semiconductor as spin-injector
Easily integrated with current semiconductor technology
Multiple functionality
Amplification
Spin transistor (Datta-Das, 1990)
Problem at interface with metal-semiconductor junction!
Non-magnetic Semiconductor
Spin coherence may be preserved
Spin coherence destroyed due to
conductivity mismatch
Ferromagnetic Metal
Metal-semiconductor junction
Semiconductor-semiconductor junction
Spin transistor (Datta-Das, 1990)
Problem at interface with metal-semiconductor junction!
Non-magnetic Semiconductor
Spin coherence may be preserved
Spin coherence destroyed due to
conductivity mismatch
Ferromagnetic Metal
Metal-semiconductor junction
Semiconductor-semiconductor junction
Spin transistor (Datta-Das, 1990)
Problem at interface with metal-semiconductor junction!
Non-magnetic Semiconductor
Spin coherence may be preserved
Spin coherence destroyed due to
conductivity mismatch
Ferromagnetic Metal
Metal-semiconductor junction
Semiconductor-semiconductor junction
Spin transistor (Datta-Das, 1990)
Problem at interface with metal-semiconductor junction!
Non-magnetic Semiconductor
Spin coherence may be preserved
Spin coherence destroyed due to
conductivity mismatch
Ferromagnetic Metal
Metal-semiconductor junction
Semiconductor-semiconductor junction
Spin transistor (Datta-Das, 1990)
Problem at interface with metal-semiconductor junction!
Non-magnetic Semiconductor
Spin coherence may be preserved
Spin coherence destroyed due to
conductivity mismatch
Ferromagnetic Metal
Metal-semiconductor junction
Semiconductor-semiconductor junction
Spin transistor (Datta-Das, 1990)
Problem at interface with metal-semiconductor junction!
Non-magnetic Semiconductor
Spin coherence may be preserved
Spin coherence destroyed due to
conductivity mismatch
Ferromagnetic Metal
Metal-semiconductor junction
Semiconductor-semiconductor junction
Spin transistor (Datta-Das, 1990)
Problem at interface with metal-semiconductor junction!
Non-magnetic Semiconductor
Spin coherence may be preserved
Spin coherence destroyed due to
conductivity mismatch
Ferromagnetic Metal
Metal-semiconductor junction
Semiconductor-semiconductor junction
Spin transistor (Datta-Das, 1990)
Problem at interface with metal-semiconductor junction!
Non-magnetic Semiconductor
Spin coherence may be preserved
Spin coherence destroyed due to
conductivity mismatch
Ferromagnetic Metal
Metal-semiconductor junction
Semiconductor-semiconductor junction
Ferromagnetic Semiconductor

12. What is spintronics? And why? Ferromagnetic semiconductors
Outline
What is spintronics? And why?
Ferromagnetic semiconductors Mn

13. Search for a ferromagnetic semiconductor
1960: Cr spinels
60’s-70’s: Eu-based chalcogenides: EuO, EuS, Eu1-xGdxO
Tc up to 137 K (difficult to integrate)
80’s: II-VI alloys: CdMSe, CdMTe, ZnMTe, ZnMSe, …
M= V,Cr,Mn,Fe,Co…
(low Tc, difficult to dope)
90’s: III-V grown with Molecular Beam Epitaxy (MBE):
InMnAs (1992)
GaMnAs (1996)
Tc > 170 K
00’s: GaMnP,GaMnN, GeMn, GaMnSb
Oxides: ZnMnO
Organics

14. Ga1–xMnxAs Tc Room Temperature ( ~ 300K) Magnetization [a.u.]
Saturation magnetization (T = 0)
Mn2+
(S=5/2)
Ga0.92Mn0.08As
K. Y. Wang et al., (2005)
S. J. Potashnik et al., Phys. Rev. B (2002)

15. antiferromagnetic correlations
In the absence of free carriers: short-range super-exchange
Holes mediate ferromagnetic order:
antiferromagnetic correlations
carrier dispersion Mn spin carrier
spin Mn Mn
Zener or double-exchange Hamiltonian 15

16. Optimizing Tc for a ferromagnetic semiconductor
Crucial questions:
How to increase Tc above room temperature ?
Why the saturation magnetization is reduced ?
How to understand magnetic anisotropy ?
Is there an impurity band? 16

17. Weak-coupling mean-field theory
                      
                                       
Weak-coupling mean-field theory
[Dietl et al., Science (2000)]
Dietl’s theory is too simplistic to analyze material properties.
Room temperature

18. Theoretical difficulties in (GaMn)As
Strongly correlated system
Strongly disordered system
Non-local effects
Large spin-orbit coupling
How to model:
How multiple carrier bands modify the double-exchange mechanism?
How to include short range correlations in a dilute magnetic system?
How to incorporate realistic modeling of the host material?
 Solution: Dynamical mean field & dynamical cluster approximation

19. Dynamical Mean-Field Approximation (DMFA)
Replace the system by a single impurity in a bath of holes
2. Include local fluctuations of the embedded impurity using a frequency dependent self-energy
3. The self-energy connects the impurity Green function with the mean-field bath
mean-field function impurity self-energy
Green function 19

20. Dynamical Cluster Approximation (DCA)
DCA provides systematic non-local corrections to the DMFA
Finite size cluster self-consistently embedded in a mean-field.
Non-local correlations are short-ranged
DMFA
B. J. van Zeghbroeck (1997)
DCA
Maier, Jarrell, Pruschke & Hettler, Rev. Mod. Phys. (2005)

21. Tc for carriers with angular momentum j
Tc (j) reaches a maximum for j = 3/2!
max
Model for GaAs must include two hole bands
(j = 3/2) with light and heavy masses. Tc is suppressed due to magnetic frustration.
J. M, Fishman & Jarrell,
Phys. Rev. Lett. 96,

22. Magnetic anisotropy in (GaMn)As
Tensile strain  Out-of-plane anisotropy
 Compressive strain  In-plane anisotropy
(GaMn)As
substrate
(spin reorientation)
(GaMn)As
substrate
[100] z θ M x Φ y
[110]
[110]

23. Model for (GaAs)Mn Ga1-xMnxAs non-magnetic ion (Ga) magnetic ion (Mn)
Carrier spin
Localized spin (Mn)
Mn site
Spin-orbit coupling
& strain effects
Ga1-xMnxAs

24. exact solution at Nc = ∞
Nc = 1
(DMFA)
Nc = 14
Nc = 36

25. Results: Magnetization and Tc
Mn concentration : x = 5 %
hole concentration : p = 2.5%
Nc : Cluster size
Mn concentration : x = 5 %
Experiments

26. Results: Magnetic anisotropyz
Compressive strain
(─0.2%)
substrate
(GaMn)As θ M x Φ y
[110]
Tensile strain
(0.2%)
substrate
(GaMn)As

27. Results: Magnetic anisotropy
[100]
[110]
Compressive strain
substrate
(GaMn)As
Calculation (DCA) z θ M x Φ y
Experiments
[110]

28. Outline
What is spintronics? And why?
Ferromagnetic semiconductors
Organic magnets

29. Organic-based Materials in Spintronics
Advantages:
smoother interfaces
longer spin coherence lengths
more flexible metallicity range
scaffolding to hold metal centers
cheaper bottom-up fabrication
low-weight
mechanically flexible, ....
Disadvantages:
low conductivity
difficult to grow organic + metal
low-weight, ….

30. Magnetic Organic Semiconductors: Magnetic Organic Semiconductors:
Porphyrins & Metalloporphyrins
Q. Huo, Langmuir 2004,
J. Porphyrins and Phthalocyanines 2005
Porphyrins & Metalloporphyrins
Magnetic Organic Semiconductors:
Q. Huo, Langmuir 2004,
J. Porphyrins and Phthalocyanines 2005
Porphyrins & Metalloporphyrins
Magnetic Organic Semiconductors:
Q. Huo, Langmuir 2004,
J. Porphyrins and Phthalocyanines 2005
Porphyrins & Metalloporphyrins
Magnetic Organic Semiconductors:
Q. Huo, Langmuir 2004,
J. Porphyrins and Phthalocyanines 2005
Porphyrins & Metalloporphyrins
Magnetic Organic Semiconductors:
Q. Huo, Langmuir 2004,
J. Porphyrins and Phthalocyanines 2005
Goldberg, Crystal growth & Design, 2004, 2006
Porphyrins & Metalloporphyrins
Magnetic Organic Semiconductors:
Q. Huo, Langmuir 2004,
J. Porphyrins and Phthalocyanines 2005
Porphyrins & Metalloporphyrins
Magnetic Organic Semiconductors:
Q. Huo, Langmuir 2004,
J. Porphyrins and Phthalocyanines 2005

31. Majidi, Moreno, Schwalm, Fishman (2007)
Metalloporphyrin modeling t1 t1 Co Ni t2
Majidi, Moreno, Schwalm, Fishman (2007)

32. Conclusions: Huge opportunities in computational materials science

33. Thanks to:
Peter Reis, Ryky Nelson, Karlis Mikelsons (Georgetown), Majid Nili (PennState), Karan Aryanpour (Univ. Arizona)
Unjong Yu (GIST, South Korea), Aziz Majidi (Jakarta, Indonesia), Brian Moritz (Standford/SLAC)
Alex Brandt (Rackspace), Mason Swanson (Ohio State), Jitu Das (Carnegie Mellon), Jonathan Gluck (Swarthmore), Brittany Shannon (Monmouth)
Randy Fishman (ORNL), Mark Jarrell, Dana Browne (LSU)
Gonzalo Alvarez, Paul Kent, Thomas Maier, Fernando Reboredo (ORNL)