Phase separation effects in diluted magnetic semiconductors

Phase separation effects in diluted magnetic semiconductors

collaborators:

T. Andrearczyk, P. Kossacki, J. Jaroszyński, M. Sawicki – Warsaw

F. Matsukura, H. Ohno – Sendai

K. Edmonds, C.T. Foxon, B.L. Gallagher, K.Y. Wang – Nottingham

J. Cibert, D. Ferrand – Grenoble

G. Bauer, A. Bonanni, W. Jantsch – Linz

D. Kechrakos, N. Papanikolaou, K. N. Trohidou -- Athens

support: Ohno Semiconductor Spintronics ERATO Project of JST

NANOSPIN -- EC projects

Humboldt Foundation

Tomasz DIETLInstitute of Physics, Polish Academy of Sciences

Institute of Theoretical Physics, Warsaw University

Introduction

Ga1-xMnxAs: resistance vs. temperature and Curie temperature vs. x

Ga1-xMnxAs: resistance vs. temperature and Curie temperature vs. x

• ferromagnetism on both sides of metal-insulator transitions• ferromagnetism disappears in the absence of holes

Matsukura et al. (Tohoku) PRB’98

III-V DMS

Effect of acceptor doping on magnetic susceptibility in Zn1-xMnxTe:P

Sawicki et al. (Warsaw) pss’02

-1 vs. T

• ferromagnetism driven by hole doping • competition between intrinsic short-range AFM and hole-induced long-range FM

II-VI DMS

Ferromagnetic temperature in p-(Zn,Mn)TeFerromagnetic temperature in p-(Zn,Mn)Te

Ferrand et al. (Grenoble, Warsaw) PRB’01Sawicki et al. (Warsaw) pss’02

1

10

1

10

3030

Fer

rom

agn

etic

Tem

p.

T F /

x eff

(K) 10

1710

1810

1910

205x1020

Hole concentration (cm-3)

(Zn,Mn)Te:P

(Zn,Mn)Te:N

InsulatingMetallic

• ferromagnetism on both sides of metal-insulator transition

1/

Where are we?

-1 0 1 2 3

-0.05

0.00

0.05

0.10

T = 175 K

T = 172 K8% (Ga,Mn)As

M[1

10](T

) / M

Sat

(5K

) [

r.u

. ]

Magnetic Field [ Oe ]

Wang/ Sawicki (Nottingham, Warsaw)ICPS’04

remanent magnetisation and 1/ vs. Thysteresis loops

MREM

TC = 173 K

TC CW

Semiconductor materials showing hysteresis and spontaneous magnetisation at 300 K

wz-c-(Ga,Mn)N, (In,Mn)N, (Al,Mn)N, (Ga,Cr)N, (Al,Cr)N (Ga,Fe)N (Ga,Gd)N, (Ga,Eu)N (Ga,Mn)As, (In,Mn)As, (Ga,Mn)Sb, (Ga,Mn)P:C

(Zn,Mn)O, (Zn,Ni)O, (Zn,Co)O, (Zn,V)O, (Zn,Fe,Cu)O, (Zn,Cu)O

(Zn,Cr)Te

(Ti,Co)O2, (Ti,V)O2, (Ti,Cr)O2, (Sn,Co)O2, (Sn,Fe)O2, (Hf,Co)O2

(Cd,Ge,Mn)P2, (Zn,Ge,Mn)P2, (Cd,Ge,Mn)As2, (Zn,Sn,Mn)As2

(Ge,Mn), (Ge,Cr), (Ge,Mn,Fe) (La,Ca)B6, C, C60, HfO2, (Ga,Gd)N – materials in which magnetic

moment is claimed to do not come from 3d or 4f shell will not be discussed cf. G. Bouzerar

SQUID studies of DMS in Warsaw

M. Sawicki et al.:

wz-c-(Ga,Mn)N, (Ga,Fe)N

(Ga,Mn)As

(Zn,Mn)Te:N, P

(Cd,Mn)Te, (Cd,Mn)Se

(Cd,Cr)Te, (Zn,Cr)Se

(Zn,Mn)O, (Zn,Co)O, (Zn,Cr)O

Today’s talk

• „low” TC ferro DMS

-- metallic side -- insulator side – electronic phase separation

• „high” TC ferro DMS

– chemical phase separation

cf. A. Moreno

Metallic side of metal-to-insulator transition

p-d Zener/RKKY model of hole-controlled ferromagnetism in DMS

Driving force: lowering of the hole energy due to redistribution between hole spin subbands split by p-d exchange interaction

T.D. et al.,’97-Jungwirth et al. (Austin/Prague) ’99-

k

EF

p-d Zener/RKKY model of hole-controlled ferromagnetism in DMS

Driving force: lowering of the hole energy due to redistribution between hole spin subbands split by p-d exchange interaction, ~ M

T.D. et al.,’97-MacDonald et al. (Austin/Prague) ’99-

No adjustable parameters

TC ~ 2(s)DOS

Essential ingredient: Complexity of the valence band structurehas to be taken into account

M

k

EF

Mn-based p-type DMS to which p-d Zener model has been found to apply

Expl.: Tohoku, Tokyo, Grenoble, Wuerzburg, PSU, Notre Dame, UCSB, Nottingham, …

xMn = 5%p = 3.5x1020 cm-3

• TC CW

• TC (p,x) consistent with

p-d Zener model• not double exchange

Insulator side of metal-to-insulator transitionAnderson-Mott localization

Small hole concentration rs > 2.4 because of either:

-- small acceptor concentration

-- large compensation

-- depletion by gates

-- depletion at surfaces and interfaces

e.g. TAMR devices of (Ga,Mn)AS Ruster et al. (Wuerzburg) PRL’05

Giddings et al. (Hitachi, Nottingham) PRL’05

Insulator side of metal-to-insulator transition

Suggested model:

percolation of bound magnetic polarons

Bhatt et al. (Princeton) PRL’02; Das Sarma et al., PRL’02,’04, ....

p-type(II,Mn)VI

(III,Mn)V

Resistivity and magnetisation in (Ga,Mn)As

4 K

Co-existence of ferromagnetic and paramagnetic components in non-metallic samples

F. Matsukura et al..(Tohoku) PRB ’98, SSC’97

104

102

100

10-2

1.5 2 5 10Temperature (K)

Res

istiv

ity (

Ohm

cm

)

B = 0

B = 11 T

(Zn,Mn)Te:Nx = 3.8%

p = 3x1019 cm-3

Collosal negative magnetoresistance on insulator side of MIT

Ferrand et al. (Grenoble, Warsaw) PRB’02

104

102

100

10-2

1.5 2 5 10 Temperature (K)

Res

istiv

ity (

Ohm

cm

)

B = 0

B = 11 T

(Zn,Mn)Te:Nx = 3.8%

p = 3x1019 cm-3

Collosal negative magnetoresistance on insulator side of MIT

Ferrand et al. (Grenoble, Warsaw) PRB’02

Katsumoto et al. (Tokyo) pss’98Reminiscent to CMR oxides

Ferromagnetism on insulator side of MIT-- competing models

• Percolation of bound magnetic polarons

• Ferromagnetic metallic-like regions embeded in insulating paramagnetic matrix

electronic nanoscale phase separation

To tell the model:

• inelastic neutron scattering Kepa et al. (Warsaw, NIST) PRL’03

• search for collosal MR in modulation-doped quantum wells, where no BMP are expected Jaroszynski et al. (Warsaw, NHMFL) cond-mat/0509

• Monte Carlo + Schroedinger eq. with magnetic disorder

Dechrakos et al. (Athenes, Warsaw) PRL’05

cf. E.L. Nagaev, E. Dagotto et al.

Probing competing AF and FM interactions by inelastic neutron scattering in p-(Zn,Mn)Te

Probing competing AF and FM interactions by inelastic neutron scattering in p-(Zn,Mn)Te

Kępa et al. (Warsaw, NIST) PRL’03

inelastic neutron scatteringof n.n. Mn pairs

large single crystals of Zn0.95Mn0.05Te:P

p = 5x1018 cm-3, TCW = 2 K

Insulator side of the MIT

Zn0.95Mn0.05Te

Hint = -2(JAF + Jh)SiSj

JAF < 0 super-exchange Jh > 0 hole-induced

Hole induced contributionHole induced contribution

empty dots - no holes, full dots – with holes

E = 2Jh = 0.03 0.006 meV

E RKKY = 0.020 meV

E BMP = 0.12 meV

Resistivity vs. carrier density at various Tin (Cd,Mn)Te/(Cd,Mg)Te:I quantum well

Jaroszynski et al. (Warsaw, NHMFL) cond-mat/0509 submitted to PRL

Electron density (cm-2)

Resistivity vs. carrier density at various Tin (Cd,Mn)Te/(Cd,Mg)Te:I quantum well

Jaroszynski et al. (Warsaw, NHMFL) cond-mat/0509 submitted to PRL

Electron density (cm-2)

Resistivity vs. carrier density at various Tin (Cd,Mn)Te/(Cd,Mg)Te:I quantum well

Electron density (cm-2)

Interpretation: nanoscale electronic phase separation into metallic ferromagnetic regions embeded in isolating paramagnetic matrix

Localization length >> rs

Ferromagnetic coupling via weakly-localised holesFerromagnetic coupling via weakly-localised holes

•At the distance between Mn ions wave function can be regarded as

extended =>only part of the spins contribute to the ferromagnetic signal

Random distribution of acceptors and spins Metallic and ferromagnetic lakes embedded in insulating matrix

High TC ferro DMS

Experimental indications of room temperature ferromagnetism in (Zn,Cr)Te

K. Ando et al., PRL’03

Effect of doping

Ando et al.. (Tsukuba) PRL’03Ozaki et al. (Tsukuba) APL’05

Ferromagnetism of (Ga,Mn)N – effect of doping

Reed et al. (NCSU) APL’05

(Ga,Mn)Nx = 0.2%

TC >> 300 K

(Ga,Mn)N, x = 0.2%TC 0 for Si doping

(Ga,Mn)N:Si

GaAs + MnAs precipitatesGaAs + MnAs precipitates

depending on growth conditions precipitates or spinodal decomposition

Moreno et al. (Berlin) JAP’02

control magnetic properties De Boeck et al. (IMEC) APL’96 enhance magnetooptical effects (MCD) Akinaga et al. (Tsukuba) APL’00; Shimizu et al. (Tokyo) APL’01

affect conductance and Hall effect

not seen in HRXRDMoreno et al. (Berlin) JAP’02

Heimbrodt et al. (Marburg) PRB’04

spinodal decomposition

hexMnAs

GaAs

TC 320 K

H (Oe)

zbMnAs

GaAs

TC 350 K

Model for high TC DMS1. DMS in question undergo spinodal decomposition into TM reach

and TM poor phases that conserve the structure of host crystal [(Ga,Mn)As (Ge,Mn) — TEM; (Ga,Mn)N -- synchrotron radiation microprobe

Martinez-Criado et al.. (ESR, Schottky) APL’05]

2. TM reach phase is a high TC ferromagnetic metal or ferrimagnetic

insulator, which accounts for spontaneous magnetisation at RT

Model for high TC DMS1. DMS in question undergo spinodal decomposition into TM reach

and TM poor phases that conserve the structure of host crystal [(Ga,Mn)As (Ge,Mn) — TEM; (Ga,Mn)N -- synchrotron radiation microprobe Martinez-Criado et al.. (ESR, Schottky) APL’05

2. TM reach phase is a high TC ferromagnetic metal or ferrimagnetic

insulator, which accounts for spontaneous magnetisation at RT

3. Because of Coulomb repulsion spinodal decomposition is blocked if TM is charged – TM charge state is controlled by co-doping with shallow impurities T.D., submitted to Nature Mat.

Mn+3EF

GaN

EF

Mn+2

GaN:Si

Cr+2

EF

ZnTe

EF

Cr+3

ZnTe:N

SUMMARY

Three classes of DMS showing ferromagnetic properties:

1. Magnetically uniform hole-controlled ferromagnetic DMS p-d Zener model + real v.b. structure

2. Magnetically non-uniform ferro DMS exhibiting electronic nanoscale phase separation driven by:

-- quenched disorder: carrier density fluctuations on insulating side of MIT -- competition between FM and AFM interactions

Griffiths phase (?)

Monte Carlo simulations with random acceptor and spin distributions

3. Magnetically non-uniform ferromagnetic DMS exhibiting chemical nanoscale phase separation: -- annealed disorder (at growth temperature) -- controlled by magnetic ion charge state

new method of self-organised growth of nanostructures

(La,Ca)MnO3

DMS: interactions determine spatial distribution of both carriers and localized spins

END