DSpace at VNU: Nanostructure and magnetic properties of Fe0.56Cu0.44 films

A contribution of the positive surface magnetic anisotropy KS¼ þ0:1 mJ/m2 to the coercivity is deduced.. When the size of the magnetic particles is reduced to a few tens of nanometers, t

Nanostructure and magnetic properties of Fe 0.56 Cu 0.44 films

N.H Duca,*, D.T Huong Gianga, A Fnidikib, J Teilletb

a Cryogenic Laboratory, Faculty of Physics, Vietnam National University, Nguyen Trai, Thanh Xuan, Hanoi 334, Viet Nam

b GPM-UMR 6634, Site Universitaire du Madrillet, B.P 12, 76801 Saint-Etienne-Du-Rouvray Cedex, France

Abstract

X-ray diffraction, high-resolution transmission electron microscopy, M.ossbauer effect and magnetisation investigations have been performed on sputtered Fe0.56Cu0.44 thin films, in which the Fe concentration is near the percolation threshold Segregation into FCC-Cu rich and BCC-Fe rich phases takes place in the as-deposited film This state is described in terms of the interfacial transient concentration and ferromagnetism, which originate from the ferromagnetic BCC-Fe in the centre of the individual grains, the iron-rich crystalline Fe(Cu) alloy lying near the interface and the paramagnetic copper-rich FCC-Cu(Fe) matrix Annealing effects cause not only the evolution of the grain size of the BCC-nanocrystalites, but also the enrichment of Fe in this phase, leading to an increase of the interfacial sharpness In addition, the magnetic coercive field is found to be enhanced The coercivity shows a value as large as 24.2 mT for the sample annealed at 400
C A contribution of the positive surface magnetic anisotropy (KS¼ þ0:1 mJ/m2) to the coercivity is deduced

r2003 Elsevier Science B.V All rights reserved

PACS: 75.50.Bb; 75.50.Tt; 75.50.Vv; 75.70.Rf

Keywords: Granular film; M ossbauer spectra; Magnetic coercivity; Surface magnetic anisotropy

1 Introduction

Granular solids, composed of nanometer-sized

magnetic metal particles in a non-magnetic matrix,

currently provide a wealth of scientific interest as

well as of potential applications for magnetic

recording, optical devices and sensors[1–2] When

the size of the magnetic particles is reduced to a

few tens of nanometers, they exhibit a number of

outstanding physical properties Its origins may be

related, on the one hand, to the particle size, and

on the other hand to the high density of

topological defects arising at the grain surfaces The surface atoms create a new phase and any characteristic property of this phase may become a dominant one for the whole system Indeed, surface effects on the giant magnetoresistance

Enhancement of the magnetic coercivity (up to a value as large as about 60 mT at room temperature

as well as about 250 mT at T ¼ 2 K) has been reported for Fe- and Co-based granular films by several authors[1,4–11] In these cases, Chen et al

dominates the magnetic properties Their so-called relaxation model is rather suitable to explain peculiarities in magnetic properties of granular

*Corresponding author.

E-mail address: duc@netnam.org.vn (N.H Duc).

0304-8853/03/$ - see front matter r 2003 Elsevier Science B.V All rights reserved.

doi:10.1016/S0304-8853(03)00061-1

systems However, the approach seems to be

imperfect in, for instance, discussing the sign of

the surface anisotropy Recently, a contribution of

the surface anisotropy to the magnetic coercivity

was reported for granular Cu0.8Fe0.2 films below

the percolation threshold [12] In these films,

isolated BCC-Fe nanoparticles (size from 1 to

35 nm) were considered to be well embedded in a

non-magnetic FCC-Cu(Fe) matrix In Ref [12],

both the strength and the sign of the surface

magnetic anisotropy was deduced and discussed

Magnetic properties of granular systems, however,

depend not only on the particle size, but also on

their distribution and density

In this paper, we focus attention on the

microstructure and magnetic properties of

sput-tered Fe0.56Cu0.44 films, in which the Fe

concen-tration is near the percolation threshold

2 Experimental

Fe0.56Cu0.44thin films were deposited on a glass

substrate at room temperature by using a triode

rf-sputtering system To avoid corrosion and

oxidation, the film stacks were covered with a

10 nm-thick Nb layer on top The film thickness

was 380 nm The composition was analysed using

energy dispersive X-ray (EDX) spectroscopy

After depositing, the samples were annealed in a

vacuum of 5  105Torr, in the temperature range

from 100
C to 400
C

The structure of the samples was investigated by

X-ray diffraction using a cobalt anticathode

(lCo-Ka¼ 0:1790 nm) The grain size was

(FWHM) of the principal diffraction peaks

using the Scherrer relation and confirmed by

high-resolution transition electron microscopy

(HRTEM)

The magnetisation was measured with a

vibrat-ing sample magnetometer (VSM) in magnetic

fields up to 1.4 T applied in the film-plane and

film-normal directions

(CEMS) at room temperature were recorded using

a conventional spectrometer equipped with a

home-made helium–methane proportional

coun-ter The source was a57Co in rhodium matrix The films were set perpendicular to the incident g-beam The spectra were fitted with a least-squares technique using a histogram method relative to discrete distributions, constraining the line width

of each elementary spectrum to be the same Isomer shifts are given relative to a-Fe at 300 K The average ‘‘cone-angle’’ b between the incident g-ray direction (i.e the film-normal direction) and that of the hyperfine field Bhf (or the Fe-magnetic moment direction) is estimated from the line-intensity ratios 3: x : 1 : 1 : x : 3 of the six

M.ossbauer lines, where x is related to b by sin2b ¼ 2x=ð4 þ xÞ:

3 Experimental results and discussion 3.1 Microstructure

The X-ray diffraction patterns are presented in

annealed at 100
C, 200
C, 300
C, 350
C and

400
C, respectively For the as-deposited film, one observes a broad Bragg peak, which seems to be formed by an overlap of (1 1 0)-BCC and (1 1 1)-FCC reflections (see the principal peak around

y ¼ 26
) This indicates a coexistence of fine Fe-rich BCC and Cu-Fe-rich FCC nanograins No appreciable change is observed after annealing between TA¼ 100
C and 200
C After annealing

at TAX300
C, however, a clear splitting of the principal peak is observed The lower-angle peak is located at yFCC-Cu¼ 25:5
(i.e (1 1 1) reflections of FCC-Cu) and it almost remains at the same position with increasing TA; while the higher diffraction angle yBCC-Fe (i.e (1 1 0) reflections of BCC-Fe) slightly increases This reflects not only the formation of FCC-Cu and BCC-Fe grains, but also the Fe enrichment in the BCC-phase The separation of the BCC-Fe (1 1 0) and BCC-Cu (1 1 1) peaks is more visible with further increasing

TA: Finally, at TA¼ 400
C, the XRD result exhibits the presence of eight rather sharp peaks corresponding to five theoretical peaks of FCC-Cu and three theoretical peaks of BCC-Fe This indicates a total decomposition of Fe and Cu in the sample This is consistent with results reported

previously by Childress et al.[11] The average size

of the BCC-Fe particles (dFe) is estimated from

FWHM using the Scherrer relation The obtained

results are listed in Table 1 Note that, for

Fe0.56Cu0.44films, dFeincreases from 9 nm for the

as-deposited film to 50 nm for the film annealed at

400
C For Fe0.20Cu0.80 films however, dFeis less

than 30 nm [9] The cross-section TEM 2-D color

Figs 2(a) and (b), respectively The corresponding

electron diffraction patterns are shown in the

insets These results confirm not only the particle

size of the BCC-Fe crystallites, but also their

interconnections In particular, it can also be seen

that, while the segregation was almost completed

in the 400
C-annealed Fe0.56Cu0.44 film, a clear

transient concentration is still observed in the 100

C-film This result is rather useful to discuss the

information on the hyperfine parameters deduced

from M.ossbauer spectrometry studies (see below)

3.2 M.ossbauer spectra

corre-sponding hyperfine field distributions PðBhfÞ are

shown for the investigated Fe0.56Cu0.44 granular

films For the as-deposited film, the CEM

spec-trum consists of both magnetic and paramagnetic

contributions (see Fig 3(a)) This sample is al-ready magnetic with the relative magnetic and paramagnetic M.ossbauer fractions of 85% and 15%, respectively An average hyperfine field of

average ‘‘cone-angle’’ /bS ¼ 68
can be reported However, it is worthwhile to mention that the hyperfine field is distributed in a rather broad range (from 20 to 34 T) These findings imply that the Cu-rich FCC matrix is paramagnetic The BCC Fe-rich nanograins are strongly coupled, but they exhibit an interfacial transient concentration and ferromagnetism, which originates from the ferromagnetic BCC-Fe in the centre of the individual grains, the iron-rich crystalline Fe(Cu) alloy lying near the interface and the paramagnetic

15 20 25 30 35 40 45 50 55 60

Theta (degrees)

(200)

(a) (b) (c) (d) (e) (f)

Fig 1 X-ray diffraction patterns of the Fe 0.56 Cu 0.44 thin films: (a) the as-deposited film; after annealing at (b) 100
C; (c) 200
C; (d) 300
C; (e) 350
C and (f) 400
C.

Table 1 The grain size (d Fe ), magnetic coercivity (m0H C ), saturation magnetisation (m0M S ) and effective anisotropy constant (K eff ) for Fe 0.56 Cu 0.44 granular films

Sample dFe(nm) m0HC(mT) m0MS(T) Keff (kJ/m 3 ) As-deposited 9 3.0 0.705 5.12

T A ¼ 100
C 12 3.4 0.695 5.63

T A ¼ 200
C 13 3.4 0.710 5.66

T A ¼ 300
C 18 7.4 0.717 12.65

T A ¼ 350
C 26 17.4 0.712 29.74

T A ¼ 400
C 50 24.2 0.718 41.40

copper-rich FCC-Cu(Fe) matrix This is in good

agreement with the TEM color mapping results

After annealing at TA¼ 100
C and 200
C, the

CEM spectra are almost identical (see Figs 3(b)–

(c)) A weak increase in /BhfS (up to 32.0 T) and a

decrease in /bS (down to 56
) started to occur in

the sample annealed at 300
C (see Table 2) In

accordance with the XRD results, this increase of

/BhfS is associated to the evolution of the grain

size as well as the enrichment of Fe in the

BCC-phase The decrease of the /bS value reflects the

enhancement of an out of plane magnetisation

TA¼ 350
C, the hyperfine field distribution is

narrowing, /BhfS ¼ 32:4 T and the paramagnetic

fraction remains 6% only Finally, after annealing

sextuplet of BCC-Fe (isomer shift d ¼ 0 and

/BhfS ¼ 33 T, see Table 2) This confirms the

complete decomposition of Fe and Cu metals In

this case, a sharp BCC-Fe/FCC-Cu interface is

thought to be formed

3.3 Magnetisation and magnetic coercivity

Figs 4(a)–(c) illustrates the magnetic hysteresis

loops measured in magnetic fields applied in the

film-plane and along the film-normal directions for

the as-deposited film and the films annealed at

TA¼ 200
C and 400
C, respectively The mag-netic data are summarised in Table 1 Note that the as-deposited sample is a soft magnetic material with saturation magnetisation MS¼ 0:705 T and coercivity m0HC¼ 3 mT: With increasing anneal-ing temperature, the saturation magnetisation slightly increases The coercivity m0HC; however, remains almost constant for TAp200
C, but strongly increases for TAX300
C A m0HC value

as large as about 24.2 mT is achieved at room temperature for film annealed at TA¼ 400
C This coercivity is quite large as compared with that of pure BCC-Fe (m0HCB1 mT) In addition, we also observed an indication of the existence of a (spontaneous) perpendicular magnetisation

measured in magnetic fields applied along the film-normal direction in Figs 4(a) and (b)) This was already mentioned in the discussion of the average M.ossbauer ‘cone-angle’ /bS in Section 3.2 Below, we will connect it to a contribution of the surface magnetisation

According to the conventional theory, the

can be estimated from the coercivity m0HC and the saturation magnetisation MS by using the relation:

Keff¼1

Fe

Cu

Fig 2 TEM 2-D color mapping images of the Fe 0.56 Cu 0.44 thin films annealed at 100
C (a) and 400
C (b) TEM diffraction patterns

of the corresponding phases are shown in the insets.

where a is a proportionality constant, which depends on the magnetisation reversal mechanism According to the Stoner–Wohlfarth model, i.e for randomly distributed single domain particles, one has aE1 ([7]and references therein) The obtained

Keff values are listed inTable 1 Assuming an assembly of fine, spherical Fe particles with average diameter dFe; Keff can be expressed in terms of the volume (KV) and surface (KS) anisotropy constants, as[12]

0

1.00 1.02

1.00 1.01

1.00 1.01

1.00 1.02

1.00 1.02

1.00 1.03

Velocity (mm/s)

0 10 20

30

(a)

0 10 20 30

0 10 20 30

0 10 20 30

0 20 40 60 80

0 50 100

0 10 20 30

(f )

(b)

(c)

(d)

(e)

B hf (T)

Fig 3 M ossbauer spectra and hyperfine-field distributions of the Fe 0.56 Cu 0.44 thin films: (a) the as-deposited film; after annealing at (b) 100
C; (c) 200
C; (d) 300
C; (e) 350
C and (f) 400
C.

Table 2

Hyperfine parameters for Fe 0.56 Cu 0.44 granular films: isomer

shift (d), hyperfine field (/B hf S), M ossbauer cone-angle (/bS)

and paramagnetic fraction (A para )

Sample d (mm/s) /B hf S (T) /bS (deg.) A para (%)

As-deposited 0.052 29.4 68 15

T A ¼ 100
C 0.040 29.4 68 15

T A ¼ 200
C 0.040 30.5 68 15

T A ¼ 300
C 0.010 32.0 56 14

T A ¼ 350
C 0.006 32.4 55 6

InFig 5, dFeKeffis plotted as a function of dFe As

can be seen from this figure, a large deviation from

linearity is observed for small grain sizes A similar

result was observed in Fe0.2Cu0.8films[12] In Ref

[12], this behaviour was attributed to the variation

of the surface anisotropy in the series due to the varying degree of the segregation, and then to a different modification of the band-structure in the surface phase At present, as already indicated by the M.ossbauer data, this can also be ascribed to modifications of the interfacial transient concen-tration and ferromagnetism Approximately, we can deduce the values of KVand KSfrom the large grain size region, where a sharp interface is expected to exist It turns out from this approach that KV¼ 51 kJ/m3 and KS¼ þ0:1 mJ/m2 The observed negative sign of KV means that the Fe magnetic moments in the grain cores tend to be oriented in the film plane The positive sign of KS; however, indicates that Fe-magnetic moments in the surface phase tend to be oriented along the surface normal [12] Finally, it is interesting to mention that the obtained KVvalue is rather close

to the values of 50 and 40 kJ/m3, reported for the cubic anisotropy constant of the pure BCC-Fe [4] and the granular Fe0.2Cu0.8 films [12], respec-tively The obtained value of the surface aniso-tropy KS, however, is larger than that (+0.04 mJ/

m2) of the Fe0.2Cu0.8 granular films Thus, the grain interconnections, do not reduce, but enhance

KS: Consequently, this can be associated with the intergrain magnetic coupling effects However, one should also take into account effects of the Fe-concentration dependence of the surface magnetic anisotropy

-24

-18

-12

-6

0

6

12

18

24

⊥

//

(a)

TA = 30°C

-3 kA

µ° H(T)

-20

-16

-12

-8

-4

0

4

8

12

16

20

⊥

//

(c)

-3 kA

-16

-12

-8

-4

0

4

8

12

16

⊥

//

(b)

TA = 200 °C

-3 kA

Fig 4 Magnetic hysteresis loops for the Fe 0.56 Cu 0.44 thin films:

(a) the as-deposited films; after annealing at (b) 200
C and (c)

400
C.

-3.0 -2.0 -1.0 0.0 1.0

d Fe (nm)

2 )

Fig 5 The plot of d Fe K eff as a function of d Fe for the

Fe 0.56 Cu 0.44 thin films.

4 Concluding remarks

Magnetic properties of Fe0.56Cu0.44films can be

described in terms of the initial formation of

BCC-Fe nanograins with interfacial transient

concen-tration and ferromagnetism With increasing

annealing temperature, a complete separation is

reached and the interfaces become sharper The

enhancement of the magnetic coercivity observed

in the granular films was associated with the

surface anisotropy contribution, which is thought

to be strengthened by the intergrain magnetic

coupling as well as by the Fe-concentration

enrichment

Acknowledgements

This work was partly supported by the State

Program for Natural Scientific Researches of

Vietnam, within project 420.301

References

[1] B Abeles, in: R Wolfe (Ed.), Applied Solid State Science: Advances in Materials and Device Research, Academic Press, New York, 1976, p 1.

[2] C.I Chen, J Appl Phys 69 (1991) 5267.

[3] C Alof, B Stahl, M Ghafari, R Hahn, J Appl Phys 88 (2000) 4212.

[4] C.L Chien, J Appl Phys 69 (1991) 5276.

[5] J.Q Xiao, J.S Jiang, C.L Chien, Phys Rev Lett 68 (1992) 3749.

[6] G Xiao, CL Chien, Appl Phys Lett 51 (1987) 1280 [7] C Chen, O Kitakami, Y Shimada, J Appl Phys 84 (1998) 2184.

[8] T Hayashi, S Hirono, M Tomita, S Umemura, Nature (London) 381 (1996) 72.

[9] T Murayama, M Miyamura, S Kondoh, J Appl Phys.

76 (1994) 5361.

[10] C Chen, O Kitakami, Y Shimada, J Appl Phys 86 (1999) 2161.

[11] J.R Childress, C.L Chien, M Nathan, Appl Phys Lett.

50 (1990) 95.

[12] N.H Duc, N.A Tuan, A Fnidiki, C Dorien, J Teillet, J Ben Youssef, H Le Gall, J Phys.: Condens Matter 14 (2002) 6657.