The effect of the disorder induced by Cu substitution on the phonon - - PowerPoint PPT Presentation

The effect of the disorder induced by Cu substitution

• n the phonon properties of

La1-yAyMn1-xCuxMnO3 manganites

Gianluca De Marzi

Outline

• Colossal MagnetoResistance and applications
• Crystalline structure, Electronic configuration, Phase diagram
• Some hints on theoretical models
• Optical phonon in manganites
• Experimental: the role of Cu B-site doping on vibrational properties:
• Raman spectra, data analysis
• Conclusions and perspectives

Colossal Magnetoresistance Manganites

interest on A1-xBxMnO3 manganites Colossal MagnetoResistance (CMR) ρ ρ ρ ρ ρ − = ∆ =

H

MR , ρ0 = ρ=(H=0) MR < 0 and isotropic (no dependence) H J r r

• MR up to 100% (La-Ca-Mn-O at 77 K)*

*Jin et al., Science 264, 413 (1994)

Scientific and Technological interest in CMR

Scientific interest By varying x and T manganites show:

• Different typologies of spin ordering
• M-I and PM-FM transitions
• Jahn-Teller distortions and polaronic

effects

• Charge ordering
• Orbital ordering

Technological interest MagnetoResistance is “Colossal”, because:

• Permalloy (Ni/Fe),EuO MR≈2-3%
• Multilayer Cu/Co GMR≈40%
• Manganites CMR≈100%

Applications: Read/write magnetic memory devices (bobine and cassette tapes, Digital Audio Tape, etc.), and system sensible to both T and H.

Crystalline Structure Chemical formula: A1-xBxMnO3

Jahn-Teller theorem (1937):

"any non-linear molecular system in a degenerate electronic state will be unstable and will undergo distortion to form a system of lower symmetry and lower energy thereby removing the degeneracy"

Example: La-Ca-Mn-O

A = rare earth +3 (La, Pr, Y, Nd,…) B = divalent ion (Ca, Sr, Ba, …) Oxygen

Electronic Configuration

Neutral Mn electronic configuration = [Ar]3d54s2

• In AMnO3 Mn is trivalent (ionic approx.) 4 d-electrons will be responsible for its electronic

properties (AFM)

Mn3+ in

• ctahedral

co-ordination

• In BMnO3 Mn is tetravalent 3 d-electrons present (AFM)
• For a partial substitution case, A1-xBxMnO3 (0 < x < 1), Mn ions are mixed-valent.

An amazing thing is that, although AMnO3 and BMnO3 manganites are AFM insulators, at some intermediate composition A1-xBxMnO3 exhibits CMR :

Rich Phase Diagram

high T Paramagnetic Insulator (PI) x ∀

by varying x and T, manganites show:

x=0, and 1 AFM insulators x > 0.5 Charge-Ordering ≈ 0.2 ÷ 0.5 CMR (not Pr1-xCaxMnO3)

First explained by Zener (1951), Anderson & Hasegawa in the framework of the:

Double-Exchange Model

Double-Exchange

Wherever an Mn3+ and Mn4+ are in neighbouring Mn sites, there exists the possibility of eg-electron hopping from the Mn3+ to the Mn4+ via the oxygen anion. Two simultaneous electron hops are required Mn3+ onto O2- and O2- onto Mn4+

( )

∑ ∑

+ +

⋅ − + − =

L ab i ib ia ab i H L ij j i ij

c c S J c H c c t H

, ,

. . σ

σ σ σ

v v

∑

⋅ +

L j i j i AF

S S J

,

v v Whitin the solution for two classical spins* the following relationship holds: t(Θ) = t cos(Θ/2) lowering T PMFM ↔ I M raising H spin alignment I M

*Anderson and Hasegawa, De Gennes

DE explain qualitatively the experiments, but…

Millis et al.: “Double-Exchange alone

doesn’t explain the resistivity of La1-xSrxMnO3”

• Phys. Rev. Lett 74, 5144 (1995):
• DE
• verstimate

Tc

• ne
• rder
• f

magnitude(1000-3000 K )

• T dependence of ρ(T) is completely

different at T<Tc

• Experimental values for ρ are lerger than

that predicted by DE theory One has to consider the el-ph interaction, due in part to the JT splitting of the Mn eg states.

( ) ( )

∑ ∑

+ + =

+

j Q k d j Q d g H H

jb ab ja DE 2

ˆ 2 1 ˆ

σ σ

the competition between localisation and DE can be parameterised by an effective el-ph constant:

eff loc

t E kt g = = λ

results are in agreement with experimental data (by Schiffer et al.)

Optical phonons in CMR Manganites

Let us consider LaMnO3 manganite

Theory group analysis for the undistorted (cubic) perovskite gives the following irreducible representation:

Γ = 4F1u+F2u O1h (Pm3m) The JT effect distorts the octahedra, and the structure is

• rthorhombic:

D2h

16(Pnma):

60 phonons (k=0) are predicted (Iliev,98) 25 IR active 9B1u+7B2u+9B3u 24 Raman 7Ag+5B1g+7B2g+5B3g 8 Silent Au 3 Acoustic B1u+B2u+B3u

3 IR active (F1u) 3 acoustic (F1u) 1 silent (F2u) NO Raman active

Phonon Assignment for the undoped LaMnO3

IR Measurements

De Marzi et al. PRL (98)

Raman Spectra

Ilev et al., PRB 57, 2872 (1998)

Raman measurements on doped manganites

Common features in Raman spectra: maxima are mainly located at three intervals:

• 180-300 cm-1 M1
• 400-520 M2
• 580-680 M3

but…phonon assignment is still controversial

increasing doping JT reduction “more cubic” structure extremely small Raman scattering efficiency difficult measurements

• M1 corresponds to an Ag out-of-

phase x-rotation of the oxygen cage

• M2 is A2g (mainly bending)
• M3 is B2g (mainly stretching)

AE Pantoja, HJ Trodahl, J. Phys.: Cond. Matt. 13 (2001) 3741

Our samples: polycrystalline La1-ySryMn1-xCuxO3

Cu substitution at the B sites of the perovskite structure strongly influence:

• Transition temperature Tc (Sapiña et al.)
• the Mn-O-Mn angles of the MnO6 octahedra
• structural disorder

Therefore: phononic properties of manganites are modified by B-site doping aim: to study the evolution of phonons as function of B-site doping

The idea is that a MI transition can occur when the octahedral is forced to be undistorted and the Mn-O-Mn angles tends towards 180° , and this can be obtained by changing the average dimension of the atom at the A and/or B site.

Our samples: polycrystalline La1-ySryMn1-xCuxO3

• samples with Cu doping 0 < x < 0.10
• Tc is reduced by B-site substitution:
• Structure rhombohedral R-3c (x-ray

analysis by Sapiña et al.)

• Single phase compounds
• the ratio Mn4+/(Mn4++Mn3+) = 0.3 is fixed

the effects are not due to DE mechanism. La1-ySryMn1-xCuxO3 x y %Mn4+ TC 0.00

0.02 0.04 0.06 0.08 0.10

0.300

0.274 0.248 0.222 0.196 0.170

32

32 32 32 35 32

372

358 331 308 274 236

Table 1: nominal compositions for x, y doping, % of tetravalent Mn ions, and observed Tc [13]

Raman spectra of La1-ySryMn1-xCuxO3

polarized configuration

200 400 600 800 1000

Raman Intensity Raman Shift (cm

• 1)

La1

• y SryMn1-xCu xO

x y TC 0.00 0.300 372 0.02 0.274 358 0.04 0.248 331 0.06 0.222 308 0.08 0.196 274 0.10 0.170 236

• three peaks are well evident at

about 200, 400, and 600 cm-1

• shoulder at around 600 cm-1 disappear for x=0.00
• a strong background signal is present

cross-polarized configuration

200 400 600 800

Raman Intensity Raman Shift (cm

• 1)

La

x y T

0.00 0.300 372 0.02 0.274 358 0.04 0.248 331 0.06 0.222 308 0.08 0.196 274 0.10 0.170 236

• the first peak at about 200 cm-1 completely

disappears

Data analysis

200 400 600 800

Raman Intensity (a.u.) Raman Shift (cm

• 1)

LSM6 julio

• Spectra were fitted in the 130-900 cm-1

range with six lorentzian oscillators

• An example of the best fit curve and the

different components is shown for the x=0.10 sample;

• four peaks were found at 180-215, 430,

498, and 670 cm-1, and a broad background with a maximum at 450 cm-1.

Data analysis

0.02 0.04 0.06 0.08 0.10 180 190 200 210 220 300 400 500 600 700 800 900 1000 1100 1200

La1-ySryMn1-xCuxO3 Frequency (cm

• 1)

X doping

A

1g

Eg A

g

B

2g

Γ(D3d

6)= A1g+ 2A1u+ 3A2u+4Eg+5Eu+3A2g

8 IR active 3A2u+5Eu 5 Raman 1A1g+4Eg

Granado et al., PRB 58, 11435 (98):

ω1 A1g N.B. disappears for

u i

ε ε ⊥ ω2 Eg ω3 Eg (?) ω4 Eg (?)

Why didn’t Granado et

• al. see ω3 and ω4 ?

polishing effect

Analysis of the A1g mode

Raman shift of A1g is unusual. infact Sr87 is lighter than La139 shift toward lower ω! ω1 is sensitive to JT distortions (that are modified by Sr doping)

* Irwin et al. PRB 59, 9362 (1999)

Tolerance factor

( )

O B O A O B O A

r r r d d t + = =

− −

2 2 r +

• t<0.925 orthorhombic; 0.925<t<0.1 rhombohedric; t=1 cubic
• since r(Sr2+) > r(La3+)

when x increases (and y decreases) <rA> decreases

• Moreover, <rB> is increased by Cu substitution

0,968 0,970 0,972 0,974 0,976 0,978 0,980 185 190 195 200 205 210 215

x=0.00 x=0.10 Raman Shift of the A1g mode (cm

• 1)

Tolerance Factor

But we observe just the

• pposite

A1g is not an external mode Infact, A1g involves the motion of the oxygen cage *

• band shift is a

linear function

• f t
• increasing

x causes the system to be more distorted

More Analysis…

Data analysis has still not been concluded

assignment of higher frequencies peaks is unresolved:

• is the ω4 mode one of the frozen phonon related to the JT distortions ? (Dediu et al.)
• ω3 and ω4: Second order Raman scattering?
• ω3 and ω4 related to the bending and stretching modes of the MnO6/CuO6 octahedra?

Best fit are now being performed with:

( ) ( ) [ ]

( )

        + − + Γ + Γ + =

∑

= n i i i i

A A n S

1 2 2 2 2 2 2 2

1 γ ω ω ω ωγ ω ω ω ω Bose- Einstein factor Odd lorentzian

“collision-dominated” low- frequency response associated with diffusive hopping of the carriers

Conclusions and Perspectives

• Raman spectra of La1-ySryMn1-xCuxO3

polycrystalline manganites, at T=300 K and in the 100-1000 cm-1 range, both and

u i

ε ε ⊥

u i ε

ε

have been measured;

• three peaks are well evident at about 200, 400, and 600 cm-1
• shoulder at around 600 cm-1 disappear for x=0.00

ω1 A1g out-of-phase rotation of (Mn/Cu)O6

ε ε ⊥

N.B. disappears for

u i

ω2 Eg ω3 Eg (?) ω4 Eg (?)

• A1g shift shows linear dependence on tolerance

factor

• increasing x cause the system to be more

distorted

In order to remove such ambiguities on the ω3 and ω4 assignment, it is necessary to :

• take measurements as a function of T
• extend the measurements to other system

La1-yAyMn1-xMxO3 A=Sr, M= Cr, Zn; A=Ba, M= Zn, Sc further studies are also highly desirable

• measure IR reflectivity spectra (Belgrad)

ellipsometry in the NIR and vis. (Barcelona)