Disentangling bulk and surface contribution in correlated systems and buried interfaces

Giancarlo Panaccione Lab. TASC - INFM - CNR, Trieste, Italy

Outline

-- VOLPE project: Characteristics and Performances

- Results:- Transition Metal Oxides (Ruthenates)- Enhanced s-contribution @ Fermi level (In2O3, PbO2) - Buried interfaces (spin valve interface, diluted magnetic semiconductors)

- Conclusions and perspectives

Long Term Proposal till 2008 + Open to users

“Bulk sensitive Photoemission on highly correlated materials”C. S. Fadley, Univ. Davis, USM. Grioni, EPFL, SwitzerlandJ. Joyce, Los Alamos, US

Core levels + Valence band

VOLPE’s history(VOLume PhotoEmission from solids with Synchrotron Radiation)

Operational since 2005 at ESRF (ID16 beamline)EU project funded in 2001 – Trieste, Univ. Rome III, ESRF, EPFL, LURE

J. Joyce, Los Alamos, USO. Tjernberg, RIT, SwedenL. H. Tjeng, Univ. Köln, Germany

Core levels + Valence bandKinetic energy up to 6-10 keV

Statistics comparable to LE-PESEnergy resolution down to < 50 meV

Needs

1) High Res/High flux beamline (low cross sections)2) Adapted Spectrometer performances (high stability/low noise)

(no commercial equivalent in 2001)

High Resolution /High Flux beamline ID16@ ESRF, inelastic scattering

A – 3 x U35 undulators

B – Si(1,1,1) double crystal mono

C – Si(n,n,n) post-monochromator

D – toroidal mirror

E – sample

F – electron analyzerE

B

CD

F

A

Channel-cut in backscattering (87o)

(3,3,3) @ 6 keV:

∆E=50 meV & 6.2•1011 ph/sec/100 mA

(4,4,4) @ 8 keV:

∆E=37 meV & 7.6•1011 ph/sec/100 mA

(5,5,5) @ 10 keV:

∆E=14 meV & 2.6•1011 ph/sec/100 mA

Combination of higher reflection order Channel-cut

e.g. (2,2,0) @ 6 keV:

∆E=340 meV & 3•1012 ph/sec/100 mA

Tunable photon energy

B

Dispersive element

+

VOLPE’s spectrometer

(Univ. Rome III and INFM)Input lensHigh retarding factor Constant linear mag.Small field of view

71 meV @ 5933 eV50 meV analyser

250

P. Torelli et al. Rev. Sci. Instr. 76, 023909 (2005)F. Offi et al. NIM A, 550, 454 (2005)G. Panaccione et al. NIM A 547, 56 (2005)

(ELETTRA)2D delay line detector

High stability power supplies

+

0.3 0.2 0.1 0.0 -0.1 -0.2 -0.30

50

100

150

200

250

Exp. FIT

Inte

nsity

(C

ount

s)

Binding Energy (eV)

Au fermi edgeT=20Khν=5933eVE

P=20 eV, slit 0.4mm

∆Eexp = 71meV

Dark counts< 0.3 events / (sec * cm2)

150

100

Pix

el

150100500Pixel

(a)(a)

2D cross delay anode (ELETTRA)G. Cautero et al. NIM A 595, 447 (2008)P. Torelli et al. Rev. Sci. Instr. 76, 023909 (2005)

Linearity guaranteed

Decoupling up to 12 kV50

0

Pix

el

150

100

50

0

Inte

nsity

(arb

. uni

ts)

369.0368.5368.0367.5Binding energy (eV)

(b) Technical parametersTime resolution: 27psSpatial resolution: <70µmPulse pair resolution: <10ns (multihit capability)Real time coincidence check @ 2MHz Configuration (1D, 2D, single channel)Full time resolved information available

500 Å

200

0

Wedge samples

Si substrate

Si 1s intensity attenuation vs. overlayer thickness

3000

4000

5000

6000

In

tens

ity (

cps)

Co wedgeSi - 1s photoemission

Sacchi et al. submitted to PRB

26 Å

39 Å

52 Å

65 Å

78 Å

Quantifying bulk sensitivity

I (x) = I0 e-x/λ

M. Sacchi et al. PRB 71, 155117 (2005)

Mind the rare-earths!!!!

4078 4079 4080 4081 4082 4083

0

1000

2000

Electron kinetic energy (eV)

91 Å

104 Å 117 Å 130 Å

Jablonski and C.J. Powell, J.El. Spectr. Relat. Phen100, 137 (1999).S. Tanuma et al. Surf. Interface Anal. 35, 268 (2003)M. P. Seah and W. A. Dench, Surf. Interface Anal. 1, 2 (1979)

2 3 4 5 6 7 8 9 100123456789

10

elements inor. comp. Co Cu Ge Gd2O3

λ (n

m)

Electron energy (keV)

Mind the rare-earths!!!!

80

90

100

8

10

12

14

16

18

Con

trib

utio

n (%

)

λ = 5 Å

Quantifying bulk sensitivity: contribution of the surface Information depth (ID)

(layer thickness from which 95% of the total signal is produced):

λ = 5 Å → ID 15 Åλ = 25 Å → ID 75 Åλ = 60 Å → ID 180 Å

0 20 40 60 80 100 120 140 160 180 2000

10

20

30

40

50

60

70

80

λ = 5 Å ~ E = 20 eV λ = 25 Å ~ E = 1.5 keV λ = 65 Å ~ E = 5 keV

Con

trib

utio

n (%

)

Thickness (Å)

λ

x

exI−

=)(

0 5 10 15 20 25 30 35 400

2

4

6

Con

trib

utio

n (%

)

Depth (Å)

M. Sacchi et al., Phys. Rev. B 71, 155177 (2005)F. Offi et al., Phys. Rev. B 77, 201101 R (2008)

hν = 1253 eV, small transfer of spectral weightdespite MIT turns samples into powder

1. MIT in V2O3

Why bulk sensitivity with HAXPES?

D.S. Toledano et al., Surf. Science 449, 19 (2000)

V 2pO 1s

MIT in V2O3 with HAXPES:V 2p core level

V2O

3 metallic phase

hν= 5934 eV

Phot

oem

issi

on I

nten

sity

(A

rb. U

nits

)

V 2p

Extra peaks disappear when crossing the MIT - Huge transfer of spectral weight

hν= 5934 eV V 2p core level

Phot

oem

issi

on I

nten

sity

(A

rb. U

nits

)

300 K (PM phase) 140 K (AFI phase)

G. Panaccione et al, PRL 97, 116401 (2006)

5400 5405 5410 5415

Phot

oem

issi

on I

nten

sity

(A

rb. U

nits

)

Kinetic Energy (eV)

V 2p Room Temp

528 526 524 522 520 518 516 514 512 510

Difference

Binding Energy (eV)Ph

otoe

mis

sion

Int

ensi

ty (

Arb

. Uni

ts)

Presence of ‘extra-peaks’ (well screened features)

only in bulk sensitive PES., only in metallic phaseV2O3, agreement M. Taguchi et al. PRB 71, 155102 (2005)

M. Taguchi et al. PRL 95, 177002 (2005) G. Panaccione et al. PRB 77, 125133 (2008) NCCOK. Horiba et al. PRL 93, 236401 (2004) F. Offi et al. PRB 77, 174422 (2008) ManganitesSuga’s group VO2, CrO2 and many other systems

M. Uruma et al. Suga’s grouparXiv:0711.2160v1 [cond-mat.str-el]

Why bulk sensitivity with HAXPES?

2. Ruthenates: Well screened features visible already in the soft X-ray range

BUT

Do Surface and Bulk behave identically?

arXiv:0711.2160v1 [cond-mat.str-el]

Contribution to PES- Surface

- Subsurface - Bulk

Phot

oem

issi

on I

nten

sity

(A

rb. U

nits

) Sr3(Mn

xRu

1-x)

2O

7

x = 5% Sr 3d

T= 20Khν = 5999 eV

VOLPE @ normal emission Almost single line – better spectral purity Sekiyama et al. PRB 70, 060506(R) 2004

Hard X-ray PES:Surface contribution strongly suppressed

5856 5858 5860 5862 5864

Phot

oem

issi

on I

nten

sity

(A

rb. U

nits

)

Kinetic Energy (eV)

Mn substitutes RuInterplay Ru 4d - Mn 3d

Mn 3+ acceptorExtra eg occupies dx2-y2

3d-4d interplay responsible of MIT with AFI order (possible OO)

Extra peaks due to competitionbetween local and non local screening

Sr3 (MnxRu2-x)O7: Evolution vs. doping? Surface vs bulk?Direct access to Ru with core level PES

Chainani and Taguchi talks

M. Taguchi et al. PRL 95, 177002 (2005)

M.A. Van Veenendal, Phys. Rev. B 74, 085118 (2006)M.A. Van Veenendaal and G. Sawatzky PRL 70, 2459 (1993)

A. Kotani and K. Okada, J. Phys. Soc. Jpn. 74, 653 (2005)

NORMAL EMISSION 45 DEGREE INCIDENCE

well screened featuresSr3(Mn

xRu

2-x)O

7

Ru 3d core level80 Khν = 460 eV

Pho

toem

issi

on In

tens

ity (

Arb

. uni

ts)

Satellite peaks in TMO core levels: soft X-raysClear doping dependence in the screening channels

APE beamline, ELETTRAE. Annese, GP et al.

165 170 175 180

Sr 3p1/2

Pho

toem

issi

on In

tens

ity (

Arb

. uni

ts)

Kinetic energy (eV)

parent x = 0.05 Mn x = 0.2 Mn

In collaboration with A. Damascelli, G. Sawatzky, UBC, Vancouver, Canada

Ru 3d

Sr 3p1/2

P

hoto

emis

sion

Inte

nsity

(ar

b. U

nits

)

Sr3(Ru

1-xMn

x)

2O

7

20 Khν = 7595 eV

Evolution vs. doping confirmed BUT

Different ratio well screened/main peaks

Satellite peaks in TMO core levels: Hard X-rays

7302 7305 7308 7311

7300 7305 7310 7315

Ru 3d

Kinetic Energy (eV)

Kinetic energy (eV)

Pho

toem

issi

on In

tens

ity (

arb.

Uni

ts)

Parent (Mn 0%) Doped (Mn 20%)

Ru 3d5/2

Enhancing s-contribution with HaXPES:Fermi level in Lead Dioxide (β-PbO2)

III

III

x 15

hν = 1486.6 eV

(a)

(b)

x 15

I

II

III

hν = 56 eV

5

PbO2 Total DOS

I

II

III i ii

D.J. Payne, R.G. Egdell, et al. PRB 75, 153102 (2007)

D.J. Payne, et al J. Phys. C, in press

Binding energy / eV

-2024681012

x 15

IIIIII

hν = 7700 eV

hν = 6000 eV

x 15

IIIIII

(c)

(d)

0

Binding energy (eV)-50510

Density of states (states per eV

per cell)

0

2

O 2p PDOSPb 6s PDOSPb 6p PDOS

Calculation

A. Walsh et al, PRL 100, 167402 (2008)

Evolution of the s-like intensityat Fermi level vs. Sn doping

Combination of XES and HaXPES

Failure of band bending model

Upper limit 2.9 eV instead of 3.75 eV

Revision needed for interface band offset,e.g. In2O3/Si will be < 1 eV

A lA l

Alq3

Buried interfaces with HaXPES:Organic spin valve - (LSMO-Alq3)TEM, J. Chapman, Univ. Glasgow, UK

F. Borgatti, I. Bergenti, A. Dediu et al. CNR-Bologna @ VOLPE

Epitaxial LSMO (20 nm) film Alq3 spin/charge layer (50-100 nm),

Thin tunnel barrier (AlOx 1-2 nm)Co top electrode (10-20 nm)

Alq3(50 nm)/Co(15 nm) Alq3(50 nm)/AlOx(2 nm)/Co(15 nm).

no A lOx

with AlOx

O1s

Yie

ld (

arb.

un.)

no AlOx

with AlOx

C1s

Core-level of Alq3/Co interfacewithout (black) and with (red) presence of AlOx film (1-2 nm)

Buried interfaces with HaXPES:Organic spin valve - (LSMO-Alq3)

As-grown samples

7060 7070 7300 7310 7320

7190 7200 6030 6040

Yie

ld (

arb.

un.)

no AlOx

with AlOx

N1s

Yie

ld (

arb.

un.)

Kinetic Energy (eV)

no AlOx

with AlOx

Al1s

Kinetic Energy (eV)F. Borgatti et al. Submitted APL

15 nm capping

Doping control: Ferromagnetic SemiconductorsModel system: Ga1-xMnxAs

Low T growth = Mn interstitial + As antisitesPost annealing Tc ↑ conductivity ↑

BUT

Tc ↓ if thickness ↑Surface vs. bulk magnetic properties

Tc < 200 KA.H. Macdonald, Nature Mat. 4, 195 (2005)

λFe

θanti dx

Fe

GaMnAs

dxedI e

x

λ−

∝hν

Mr

F. Maccherozzi et al.

Phys. Rev. Lett. 101, 267201 (2008).

Au cap

( )xρ ( )xδ

Collaboration with:

Prof. W. Wegscheider (Univ. of Regensburg)

Prof. C. H. Back (Univ. of Regensburg)

Prof H. Ebert, (LMU, Muenchen)

Fe

(GaMn)As

XMCD Fe L2,3 edges

I(+)I(-)

XMCD

H=100 Oe

0.6

0.5

0.4

0.3

0.2

0.1

0.0

TE

Y (ar

b. u

nits

)

-0.20

-0.15

-0.10

-0.05

0.00

0.05

xmcd

(%

)

sample 2 - P0(tFe = 0.71 nm)

sample 2 - P1(tFe = 0.71 nm)

sample 1(tFe = 4 nm)(tFe=1.42 nm)

5 ML 10 ML 28 ML

50

-20

-10

Mn layers magnetized anti-parallel to Fe; Mn Ferromagnetic at RT

Magnetic coupling: XMCD at Fe and Mn L2,3 (APE beamline)

730720710700

-0.20

730720710700

energy (eV)730720710700

0.5

0.4

0.3

0.2

0.1

0.0

TE

Y

650645640635

0.10

0.05

0.00

-0.05

xmcd

(%

)

650645640635

energy (eV)650645640635

Sample 2 - P0(tFe= 0.71 nm)

Sample 2 - P1(tFe= 1.42 nm)

Sample 1(tFe= 4 nm)

5 ML 10 ML 28 ML

XMCD Mn L2,3 edges

I(+)I(-)

XMCD

H=100 Oe

Evolution of Mn 2p core lines vs. Fe Thickness and doping

In as-grown and capped samples (NO SPUTTERING)

inte

nsity

(ar

b. u

nits

)

O 1s(BE=532)

Mn 2pC1s(BE=285eV)

As3d (BE=42)O3s

Ga3d(BE=20)As oxide

inte

nsity

(ar

b. u

nits

)

Mn 2p

As-cap

2 min

4 min

OAs

Fe/(GaMn)As interfaces: next step with HAXPES

inte

nsity

(ar

b. u

nits

)

600 400 200 0binding energy (eV)

60 40 20 0

as-grown

1 min

2 min

3 min

4 min

680 660 640 620binding energy (eV)

air-exposed

See also B. Schmid, R. Claessen, W. Drube et al. PRB 78, 075319 (2008)

F. Maccherozzi et al., APE beamline, in preparation

Examples of ‘useful’ HAXPES experiments

-Surface properties different from the bulk (and you search for bulk properties , e.g. some TMOs)(correlation, reconstruction, high sensitivity to contamination, no easy cleavage plan)

-Surface or interfaces properties different from the bulk(and you would like to tailor new interface properties , e.g. Fe/(GaMn)As)

- Buried systems- Buried systems(as grown, capped)

- sp contribution at Fermi level and cross section change may be important

Competitive with XAS in TEY

Valence band + core level neededComparison with soft X-ray needed

AcknowledgementsVOLPE group and collaborators

ESRF (France), G. Monaco, S. Huotari, A. Fondacaro, C. Henriquet, L. Simonelli

ELETTRA , G.Cautero, P. Lacovig, P. Pittana, M. CauteroTASC, E. Annese, U. Manju, I. VobornikUniv. Rome III (Italy), F. Offi, G. StefaniSOLEIL and LCP (France), M. Sacchi Univ. DAVIS, (CA, US) C.S. FadleyUniv. British Columbia, Canada, A. Damascelli, G. Sawatzky

THEORYA. Poteryaev, A. Georges,

M. Gatti, L. Reining(CHPT- Ecole Polytechnique,

Palaiseau, France)

SAMPLESP. Metcalf (Purdue Univ, Indiana, US)H. Takagi (Univ. Tokyo, Japan)A. Guarino, A. Vecchione (Lab. SuperMat CNR, Italy)R. Delaunay (LCP, Paris, France)R.J. Cava (Princeton)

Univ. British Columbia, Canada, A. Damascelli, G. SawatzkyUniv. Modena (Italy), S3-CNR, P. Torelli, G. Paolicelli