Ion beam analysis of

Fe-based Heusler alloys/Ge hetero-epitaxial

interfaces toward spin transistors

Y. MaedaDepartment of Computer Science and Electronics,

Kyushu Institute of Technology, Japan

Advanced Science Research Center Japan Atomic Energy Agency

ASCO-NANOMAT 2013

Vladivostok, Aug. 26, 2013

ver 3.0

Our Research

Silicide Science and Technology

Ion Beam Synthesis of iron silicides

Iron silicide nanostructures for

light emitter

Iron silicide photonic crystals

Ion Beam Analysis of ferromagnetic

silicide/semiconductor hybrid structure

Photonic crystals

β-FeSi2, Fe3Si

Si(111)

-FeSi2(220)3 nm

Si(111)

-FeSi2(220)3 nm

100nm

R=14nm

R=7nm

Exciton size

100nm

R=14nm

R=7nm

Exciton size

0

500

1000

1500

2000

2500

3000

3500

4000

0.7 0.75 0.8 0.85 0.9 0.95

PL

In

ten

sity

Photon Energy (eV)

FWHM=18.9meV

0.805eV

IBS -FeSi2 precipitates

intrinsic

emission

(A band)

Si(111)

-FeSi2(220)3 nm

Si(111)

-FeSi2(220)3 nm

100nm

R=14nm

R=7nm

Exciton size

100nm

R=14nm

R=7nm

Exciton size

0

500

1000

1500

2000

2500

3000

3500

4000

0.7 0.75 0.8 0.85 0.9 0.95

PL

In

ten

sity

Photon Energy (eV)

FWHM=18.9meV

0.805eV

IBS -FeSi2 precipitates

intrinsic

emission

(A band)

Collaborator

Prof. M. Miyao, Prof. K. Hamaya, Prof. T. Sadoh

Kyushu University,

on MBE of ferromagnetic alloy films on semiconductors

Dr. M. Narumi, Dr. S. Sakai, Dr. K. Takahashi

Japan Atomic Energy Agency (JAEA)

on Ion beam analysis and development of beam lines

toward low temperature ion channeling experiments

Students graduated from Maeda Laboratory,

at Kyoto University, on RBS measurements.

Prof. Y. Terai, Dr. Y. Ando, Osaka University

on characterization of films

Outline

1. Motivation

Ion Beam Analysis of materials for Spintronics.

2. Ion Channeling and its application to

quantitative evaluation of disordering of some Heusler

alloy films epitaxially grown on Ge(111).

3. Topic : Spin injection behavior related to quality of

epitaxial growth of ferromagnetic (FM) Heusler alloy

films Fe3Si on Si(111).

4. Conclusions

One of Goals of Spintronics, Spin FET

Gate

Ge Spin injection

Fe3-xMnxSi

Gate

Ge Ge Ge Spin injection

Fe3-xMnxSi

Ge(111)

Ge(111)

Fe2MnSi(111)

Ge(111)

Fe2MnSi(111)

GateFerromagnetic

(FM) source

INS-layer

FM drain

High quality

heteroepitaxial

interface realized

by low temperature

MBE technique.

spin current

Spin-FET has a same

device structure as MOSFET.

We expect it on

very high frequency operation

and very low consuming power.

FM source epitaxially grown on

Ge(111)

K. Ueda et al: Appl. Phys.

Lett. 93, (2008) 112108-1-

3.

Efficiency of Spin Injection

from FM into SC

FM

SC2

FM

FMSC

11

1

/

RR

P

PP

PFM ： Spin polarity of FM

Rsc ： Electrical resistivity of SC

RF ： Electrical resistivity of FM

In usual case, Rsc>>RF, so that <<1

that is, Resistivity gap problem,

we have to overcome it for Spintronics.

0

0.2

0.4

0.6

0.8

1

1 10 100 1000

0.60.70.80.91

Rsc/RFM

Possible solutions

• Using half metallic FM material as an

electrodes such as Fe2MnSi etc.

(if possible, the problem remains.)

• Using Magnetic SC with highly polarized

spin states. (in future, possible?)

• Using spin injection by electron tunneling

through a Schottoky barrier. (Actually

promising)

K. Takanashi, JJAP (2011)

Spin injection by electron tunneling

though Schottky barrier

n-type SCFM metal

W

f

Jtherm

Jt

Fermi level

interface statehighly spin

polarized state

spin flop

Scheme of spin injection by electron tunneling

through a Schottoky barrier

thermal emission

induces random

injectiontunneling

Conditions for

Spin injection

We need thinner barrier

width enough to cause

electron tunneling and

higher barrier hight

enough to prevent

thermal emission.

Delta dope technology and

Reduction of interface states

by high quality MBE

Fe3-xMnxSi

スピ

ン偏

極度

PF

Mn組成 x x, Mn at B site

Sp

in p

ola

rity

P

FN

(%

)

Fe3-xMnxSi (FMS) : Spin polarization

100

80

60

40

20

0

-20

A

C

B

C

D

A

L21 type Heusler alloy

with ordered lattice Fe2MnSi

A

B

C

D

Fe

Mn

Fe

Si

Stoichiometric Fe2MnSi is estimated

to be half metallic. Mn occupation at

the B site affects spin polarization as

shown in right figure.

atomic row along

<111>

M. Lezaic et al.: Phys. Rev. B 83 (2011) 094434.

K. Ueda et al: Appl. Phys. Lett. 93, (2008) 112108-1-3.

Fe2CoSi

Low temperature MBE realized High quality Epitaxy

RHEED

Low Temperature MBE

Growth temperature TG is crucial for high

quality epitaxy of FM F3Si/Ge

0

20

40

60

80

100

0 10 20 30 40 50 60 70 80

Co

nce

ntr

atio

n (

at%

)

Depth(nm)

Fe

Si

Ge

Ge (111)

Fe3Si(111)

Ge(111)

Fe3Si(111)

B//Ge[011]B//Ge[011]

DO3-Fe3Si

HAADF

0

20

40

60

80

100

0 10 20 30 40 50 60 70 80

Co

nce

ntr

ation

(at%

)

Depth (nm)

Intermixing layer

Ge (111)

Fe3Si(111)Fe

Si

Ge

DO3 ordereda

Ge(111)

Fe3Si(111)

a

B//Ge[011]B//Ge[011]

TG=130oC TG=300oC

Low TG is much better, we can prevent active

interdiffusion between FM and Ge substrates

and keep a DO3 ordered lattice during growth.

intermixing layer

A2(random)

Low temperature MBE

technique

HAADF

L21 super lattice reflection

XTEM, SAD observations of Fe2MnSi(111)/Ge(111)

Fe2MnSi(111)//Ge(111）

L21

111

311

g//Ge[011]

Fe2MnSi<111>//Ge<111>

Ge(111)

Ge(111)

Fe2MnSi(111)

Ge(111)

Fe2MnSi(111)

XTEM image SAD pattern

XTEM brings direct information of epitaxial interfaces, however, dose not teach

the in-plain information at spin injection interfaces to us.

K. Ueda et al: Appl. Phys. Lett. 93, (2008) 112108-1-3.

2.

Introduction to Ion Channeling

and its application to Quantitative analysis

of disordering of some Heusler alloy films

epitaxially grown on Ge(111).

cited from SCRIBA, University of Surrey

Kinds of IBA

Ion Beam Analysis involves the use of an energetic ion beam to probe

the surface of a material to reveal details of the elemental and structural

details of it make up.

Why do we need IBA?IBA: Ion Beam Analysis

FM

SC (substrate)

Spin injection

a given direction

XTEM observation

very local information

SEM Surface observation Ion beam (2MeV 4He+)

Backscattered

ions from

surface to

interface

Using IBA, we can obtain not only atomic information at a tentative spin injection

interface but also information from surface and the interface under nondestructive

sample conditions.

Advantages of IBA

• Nondestructive

under low dose of incident ions

• element resolution analysis

• Depth analysis

• Average information at analyzing macro-area

• Usage of channeling techniques for

analysis of depth distribution of disorder, location of impurity

Rutherford Backscattering

• Random scattering gives information of

depth distribution of elements involved in

a sample.

• Channeling gives information of

amount and depth distribution of disorder,

location of impurity in the lattice site, and

composition and thickness of disordered

surface layer.

Wei-Kan Chu et al.,

Backscattering Spectrometry, AP

Ｇｅ<111> raw

Projection of atomic raw tilted by 5 degrees

seems a random atomic arrangement.

Ion Channeling in crystal

2MeV-He+

SSBD

Goniometer

“Channeling in crystals”

L.C. Feldman, J. W. Mayer, S. T.

Picraux, “Materials analysis by ion

channeling”, Academic Press, 1982.

Crystal

incident Ions from a specific direction into crystals can travel inside of the crystal

channel, so that backscattering yield and stopping power decreases drastically.

Alignment of crystal axis, Case of Ge<111>

q =X2+Y2

4He+

Backscatter

angle 165o

sample tilt X

Tilt Y

-4 -3 -2 -1 0 1 2 3 4

Ba

ckscatt

erin

g y

ield Y = 4º

{110} {110}{211}

a

b

c

Tilt angle X (degrees)

Y (º)

X (º)

Y=2º

Y=4º

{110}{110} {211}

ab c

Ge<111>d

e f

-4 -3 -2 -1 0 1 2 3 4

Y = 2 º {110} {110}

{211}d

e

f

Tilt angle X (degrees)

On Ge(111) plain, the zone axis of

Ge<111> can be found by tilting along

two specific direction, X and Y.

This alignment of the axis controls

accuracy of ion beam channeling.

Dual Beam Analysis Line(MD2)

ion channeling set-up at low temperature

MD2SC1

（Cited from Homepage of JAEA-TIARA)

RBS measurements at JAEA-TIARA

B-1_R2.txt

Simulated

Si

Mn

Fe

Ge

Channel500490480470460450440430420410400390380370360350340330320

Co

un

ts

18,000

17,500

17,000

16,500

16,000

15,500

15,000

14,500

14,000

13,500

13,000

12,500

12,000

11,500

11,000

10,500

10,000

9,500

9,000

8,500

8,000

7,500

7,000

6,500

6,000

5,500

5,000

4,500

4,000

3,500

3,000

2,500

2,000

1,500

1,000

500

0

1050 1100 1150 1200 1250 1300 1350 1400 1450 1500 1550 1600

Energy [keV]

Fe channel

Mn channelSi channel

Fe, Mn

Si

Ge substrate

Fe3-xMnxSi x =0.84

SIMNRAⓇ fitting

Ge

Rutherford Backscattering (RBS)

Random scattering spectrum

(Random RBS)

q>4 deg.

2MeV-4He+

RBS measurements from aligned to

random spectrum

0

2000

4000

6000

8000

300 350 400 450 500

後方

散乱

収量

(cp

s)

チャンネルナンバー

q =3.67 º

0.87 º

0 º

0.37 º

0.67 º

channeling g // Ge<１１１>

Channel Number

2MeV 4He+

Tilt angle q

Ge(111)

FMS

Ge<111>

q Bg

2MeV 4He+

Tilt angle q

Ge(111)

FMS

Ge<111>

q Bg

Backsca

tteri

ng

Yie

ld (

cps)

0

2000

4000

6000

8000

300 350 400 450 500

後方

散乱

収量

(cp

s)

チャンネルナンバー

q =3.67 º

0.87 º

0 º

0.37 º

0.67 º

channeling g // Ge<１１１>

Channel Number

2MeV 4He+

Tilt angle q

Ge(111)

FMS

Ge<111>

q Bg

2MeV 4He+

Tilt angle q

Ge(111)

FMS

Ge<111>

q Bg

Backsca

tteri

ng

Yie

ld (

cps) Fe, Mn

Si

When a sample is tilted from aligned angle, RBS spectrum changes

continuously from channeling to random scattering.

Angular yield profiles, Channeling parameter

0.01

0.1

1

-4 -3 -2 -1 0 1 2 3 4

対数

後方

散乱

収量

（規格

化）

傾斜角 q (°)

cmin=0.045

y 1/2=0.85 º

Ge

<111>

0

0.2

0.4

0.6

0.8

1

1.2

1.4

-4 -3 -2 -1 0 1 2 3 4

後方

散乱

収量

（規格

化）

傾斜角 q (°)

Ge

<111>

y1/2=0.85 º

cmin

x=0.84x=0.84

Tilt angle (degrees) Tilt angle (degrees)

Lo

g N

orm

aliz

ed

Yie

ld

Norm

aliz

ed Y

ield

m

2/1

0min 2lnexp11)(

y

qqcqc

Using the upper equation, deconvolution of angular yield profile for Fe and Mn

can be performed and channeling parameter cmin and y1/2 were obtained.

minimum yield

Fe3-xMnxSi/Ge(111) hybrid structure at interfaces

0.01

0.1

1

-4 -3 -2 -1 0 1 2 3 4

0.01

0.1

1

-4 -3 -2 -1 0 1 2 3 40.01

0.1

1

-4 -3 -2 -1 0 1 2 3 4

0.01

0.1

1

-4 -3 -2 -1 0 1 2 3 40.01

0.1

1

-4 -3 -2 -1 0 1 2 3 4

0.01

0.1

1

-4 -3 -2 -1 0 1 2 3 40.01

0.1

1

-4 -3 -2 -1 0 1 2 3 4

0.01

0.1

1

-4 -3 -2 -1 0 1 2 3 4

x=0 x=0.36

x=0.72 x=0.84

cmin=0.017 cmin=0.024

cmin=0.045 cmin=0.031

y 1/2=0.98 º y 1/2=0.91 º

y 1/2=0.88 ºy 1/2=0.85 º

Tilt angle q (degrees)

Log

Yie

ld n

orm

aliz

ed

Ge

<111>

Ge

<111>

Ge

<111>

Ge

<111>

Log

Yie

ld n

orm

aliz

ed

Fe3Si/Ge

We observed systematic increase of the minimum yield and decrease of the

critical half angle as Mn content increased. All FMS with each Mn content

maintained epitaxy with Ge(111). Y. Maeda et al.: MRS Proc. 1119E (2008) 1119-L05-02.

Fe3-xMnxSi

aligned ion beam

along Ge<111>

Resolution in depth

~2nm under geometric

condition employed.

q

0.01

0.1

1

-4 -3 -2 -1 0 1 2 3 4

cmin=0.013

Tilt Angle (degree)

Norm

aliz

ed Y

ield

Fe2CoSi/Ge(111)

0.010

0.100

1.000

-4 -3 -2 -1 0 1 2 3 4

cmin

=0.037

y1/2

=0.94

Tilt Angle (degree)

No

rma

lize

d Y

ield

110K300K

110K300K

Tilt Angle (degree)

No

rma

lize

d Y

ield

0.010

0.100

1.000

-4 -3 -2 -1 0 1 2 3 4

cmin

=0.018

y1/2

=0.99

0.010

0.100

1.000

-4 -3 -2 -1 0 1 2 3 4

cmin

=0.024

y1/2

=0.95

0.010

0.100

1.000

-4 -3 -2 -1 0 1 2 3 4

cmin

=0.045

y1/2

=0.85

0.010

0.100

1.000

-4 -3 -2 -1 0 1 2 3 4

cmin

=0.037

y1/2

=0.94

Tilt Angle (degree)

No

rma

lize

d Y

ield

110K300K

110K300K

Tilt Angle (degree)

No

rma

lize

d Y

ield

0.010

0.100

1.000

-4 -3 -2 -1 0 1 2 3 4

cmin

=0.018

y1/2

=0.99

0.010

0.100

1.000

-4 -3 -2 -1 0 1 2 3 4

cmin

=0.024

y1/2

=0.95

110K300K

Tilt Angle (degree)

No

rma

lize

d Y

ield

0.010

0.100

1.000

-4 -3 -2 -1 0 1 2 3 4

cmin

=0.018

y1/2

=0.99

0.010

0.100

1.000

-4 -3 -2 -1 0 1 2 3 4

cmin

=0.024

y1/2

=0.95

0.010

0.100

1.000

-4 -3 -2 -1 0 1 2 3 4

cmin

=0.045

y1/2

=0.85

Fe2MnSi/Ge(111)

Norm

aliz

ed Y

ield

Tilt Angle (degree)

cmin=0.045

y1/2=

0.88deg.

y1/2=

0.85deg.

<111> <111>

300K

Angular yield profiles at the interfaces with Ge(111)

FMS/Ge FCS/Ge

Summary of results

0

0.01

0.02

0.03

0.04

0.05

0.8 0.9 1 1.1 1.2 1.3

21at%Mn

18at%Mn

9at%Mn

Fe3Si

Fe2CoSi

Half angle y1/2 (degree)

Min

imum

yie

ld c

min

Perfect crystal line

with Debye temperature

D=420 K

Perfect crystal line

with Debye temperature

D=420 K

300 K

0

0.01

0.02

0.03

0.04

0.05

0.8 0.9 1 1.1 1.2 1.3

21at%Mn

18at%Mn

9at%Mn

Fe3Si

Fe2CoSi

Half angle y1/2 (degree)

Min

imum

yie

ld c

min

Perfect crystal line

with Debye temperature

D=420 K

Perfect crystal line

with Debye temperature

D=420 K

300 K

FMS

FS

FCS

worse

better

better

worse

Toward quantitative analysis

using channeling parameter

J. H. Barrett, Phys. Rev. B 3 (1971) 1527.

D. S. Gemmell, Rev. Mod. Phys. 46 (1974) 129.

Barrett-Gemmell model teaches us the way to quantitative

analysis using channeling parameter (cmin, y1/2).

d

u

uNd

2/1

2

2

min

126

118.18

y

c

<u>: total atomic displacement

N: atomic density

d: interatomic spacing in axial directions

Fortunately, we can solve above simultaneous equations using

elementary mathematics.

1

How can we deduce <u> from

channeling parameter?

ABAu

NdB

dA

42

1)A(

8.18,

126

),(

2

2

min

2

2/1

2/1min

。

cy

yc

2

deduced from angular yield profile

Y. Maeda et al.: MRS Proc. 1119E (2008) 1119-L05-02.

Y. Maeda et al.: Thin Solid Films 519 (2011) 8461-8467.

Remarks: under the condition at the same energy of incident ions

3

Total atomic displacement along

atomic row

222

sth uuu

For perfect crystals or crystals with small imperfection being

not detectable22

thuu

Total displacement can be written by sum of contributions of

thermal vibration and static displacement due to some

imperfection in the atomic row.

4

Calculation of <uth>

Debye model teaches us the one dimensional thermal

vibration of a given atom, if the Debye temperature D of

material is known.

　　　

。

Tydt

e

t

yy

y

yu

Dy

t

Dith

0

2/1

,1

1)(

4

1)(1.12)A(

f

f

i : reduced mass

Only for i-th element,

5 Wei-Kan Chu et al.,

Backscattering Spectrometry, AP

First order Debye function

Ty

dte

t

yy

D

y

t

　　　

0,

1

1)(f

Case of Fe2MnSi

D=420K

T=300K

y=1.4

6

0.5

0.6

0.7

0.8

0.9

1

0 0.5 1 1.5 2

f(y

)

y

Calculation of static displacement

<us> due to imperfection

0

)(

)(

22

22

22

s

th

ths

th

u

uuii

uuu

uui

corresponding to perfect crystals

or crystals with not detectable imperfection

The static displacement can be obtained from subtraction between measured

total displacement and calculated thermal vibration.

100

657.5

657.5011.0653.5)(

xxGe(111)

Fe3-xMnxSi

x (%)

0

0.02

0.04

0.06

0.08

0.1

0.12

-0.1 -0.05 0 0.05 0.1 0.15

静的

原子

変位

<u

s>

(Å

)

格子不整合率 (%)

x =0.84

x =0.72

x =0.36

x =0

Lattice mismatch between FMS and Ge

at 300K

Lattice mismatch (%)

Sta

tic a

tom

ic d

isp

lac

em

en

t <

us>

(A

)

Fe3Si/Ge

Static atomic displacement <us> at interface

Strong correlation between

<us> and lattice mismatch

at the interface was found.

This means that imperfection

may mainly be introduced by

the lattice mismatch.

This difference may come from

B site randomness due to Mn

occupation.

Zero mismatch

Ferromagnetic resonance (FMR)

measurements

-80

-60

-40

-20

0

20

-400 -200 0 200 400

V E

MF/

w (

V

/ m

m)

H-HFMR

(Oe)

c

Si Sub.

Fe4Si

Fe3SiPd

2 nm

Si Sub.

Fe4Si

Fe3SiPd

Si Sub.

Fe4Si

Fe3SiPd

2 nm

Si Sub.

Fe2Si

Fe3Si

Pd

2 nm

Si Sub.

Fe2Si

Fe3Si

Pd

Si Sub.

Fe2Si

Fe3Si

Si Sub.

Fe2Si

Fe3Si

Pd

2 nm

Pd/Fe3Si/Siq

R.T.

VV

FM

Pd

Si sub.

qHH

2 mm

~1 mm

Pure spin current

These FMR measurements give the first confirmation that crystal quality of Fe3SI

directly controls performance of spin injection.

(Y. Andoh, APEC-SILICIDE 2013)

Measurements

Temp.：300 K

Microwave power：200 mW

Frequency : 9.6 GHz

ESR（Bruker EMX10/12）

DC component (EMF)

Conclusions

• Ion beam analysis (IBA) is helpful and powerful

for analysis of epitaxial hetero-interfaces.

• It is possible to deduce quantitative and average

information such as atomic displacements at a

give depth or interface from the ion channeling

parameter.

• Most recently, spin injection experiments are

going more active. IBA data will be helpful for

understanding its properties.

Thank you for your kind

attention!

ASCO-NANOMAT 2013

Vladivostok, Aug. 26, 2013

Ion channeling of Fe3Si, Fe4Si/Si(111)

amorphous Crystalline

Si Sub.

Fe3Si

Si Sub.

Fe3Si

Si Sub.

Fe4 Si

Fe3Si

2 nm

LTMBE

Results of ion channeling and static atomic displacement

and spin injection experiment

(%) at IF cmin y1/2(deg.) <us>(A) emf (mV)

Fe3Si/Si +4.13 0.18 0.99 0.25 ~68

Fe4Si/Si +4.15 0.23 0.95 0.52 ~18

Fe3Si/Ge -0.05 0.02 0.98 0.09

Dynamical spin injection into Pd layer using Fe3Si

① Large EMF

High quality Single crystalline

Fe3Si

② Control experiments

-60

-40

-20

0

-400 0 400H-H

FMR (Oe)

VE

MF /

w (

V

/mm

)

Only Fe3Si Al / Fe3Si

-400 0 400

-60

-40

-20

0

H-HFMR

(Oe)

VE

MF /

w (

V

/mm

)

Fe3Si

To realize highly efficient generation of pure spin current…

High quality FM layer with uniform magnetic properties.

Small Gilbert damping constant

⇒ Ferromagnetic silicide Fe3Si with homogeneous quality

is promising.

-80

-60

-40

-20

0

-400 0 400H-H

FMR (Oe)

VE

MF /

w (

V

/mm

)

Py

-400 0 400

-60

-40

-20

0

VE

MF /

w (

V

/mm

)

H-HFMR

(Oe)

20 times

(Y. Andoh, APEC-SILICIDE 2013)

-200 0 200

-60

-40

-20

0

H-HFMR

(Oe)

VE

MF /

w (

V

/mm

)

-70

-60

-50

-40

-30

-20

-10

0

10

-200 0 200

H - HFMR

(Oe)

VE

MF /

w (

V

/mm

)

Pd/PyPd/Fe3Si Pd/Co6Fe4

-70

-60

-50

-40

-30

-20

-10

0

10

-200 0 200

H - HFMR

(Oe)

VE

MF /

w (

V

/mm

)

VISHE/w

= 67.1 V/mm

VISHE/w

= 3.05 V/mmVISHE/w

= 2.92 V/mm

Comparison of EMF for various FM samples

Single crystal Single crystal

(Y. Andoh, APEC-SILICIDE 2013)

Fe2MnSi(111)/Ge(111) Td=200oC

Intermixing layer

0.015

0.02

0.025

0.03

0.035

0.04

0.045

0.05

0.84

0.86

0.88

0.9

0.92

0.94

0.96

0.98

1

0 0.2 0.4 0.6 0.8 1最

小収

量 c

min

チャ

ネリ

ング

臨界

角 y

1/2 (°

)

BサイトのMn組成 x

Axial channeling parameter (cmin, y1/2)

Fe3-xMnxSi

A

C

B

C

D

A

A

B

C

D

<111>

Fe

Mn

Fe

Si

Mn content at B site

Min

imum

yie

ld c

min

Cri

tical ang

le for

channelin

g (

degre

es)

We found that increase of Mn content at the B site brings randomness of the atomic

row along <111> direction. This behavior has been observed as increase of a

Debye-Waller factor in EXAF, XANES measurements. Randomness along <111>

may be introduced by weak chemical bonds around Mn atoms at the B site.

NMR and Moessbauer measurements

revealed that in every Mn content,

Mn atoms preferably occupy at the B site

in the lattice.

Y. Maeda et al.: Thin Solid Films 519 (2011) 8461-8467.