EXAFS STUDIES OF METAL-SUPPORT INTERACTIONS VS CLUSTER FORMATION IN IRON PARTICLES DISPERSED IN AMORPHOUS ALUMINA

(1)

HAL Id: jpa-00226113

https://hal.archives-ouvertes.fr/jpa-00226113

Submitted on 1 Jan 1986

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

EXAFS STUDIES OF METAL-SUPPORT

INTERACTIONS VS CLUSTER FORMATION IN IRON PARTICLES DISPERSED IN AMORPHOUS

ALUMINA

L. Alagna, D. Fiorani, T. Prosperi, J. Dormann

To cite this version:

L. Alagna, D. Fiorani, T. Prosperi, J. Dormann. EXAFS STUDIES OF METAL-SUPPORT IN- TERACTIONS VS CLUSTER FORMATION IN IRON PARTICLES DISPERSED IN AMOR- PHOUS ALUMINA. Journal de Physique Colloques, 1986, 47 (C8), pp.C8-1057-C8-1060.

�10.1051/jphyscol:19868205�. �jpa-00226113�

(2)

EXAFS STUDIES OF METAL-SUPPORT INTERACTIONS VS CLUSTER FORMATION IN IRON PARTICLES DISPERSED IN AMORPHOUS ALUMINA

L. ALAGNA, D. FIORANI, T. PROSPER1 and J.L. DORMANN'

I.T.S.E., Area della Ricerca d i Roma del C.N.R., P.O. Box 10.

Monterotondo Staz., I-00026 Rome, Italy

* ~ a b o r a t o i r e d e Magnetisme, C.N.R.S., F-92195 Meudon Principal Cedex, France

Abstract

Les spectres EXAFS ont Qt6 measures sur une serie de materiaux obtenus par co- pulverisation de fer et alumine sur kapton. Trois differents types de fer exi- stent selon le pourcentage de fer poulverise. Pour des pourcentages eleves (75%, 90%) les particules sont costituees sourtout de a-Fe. Pour des percentages in- termediaires (55%,65%) les particules sont amorphes et les distances observees sont une combination de distances Fe-@ et Fe-Fe, en majorit&.

The EXAFS spectra have been measured of a series of materials obtained by co- sputtering iron with alumina onto kapton. Three different types of particlesare clearly present depending on the relative percentage of iron co-sputtered. For low weight at.% Fe (25%) the particles are mainly oxidic in nature. At high weight at.% Fe (75%,90%) the particles are composed mainly of b.c.c. a-Fe. At

intermediate values (55%,65% at. weight Fe) the particles have an amorphous structure and the only detectable distance is a combination of mainly Fe-0 distances and Fe-Fe distances*

Introduction

Because of their importance in technological applications as magnetic fluids, magnetic tape and in catalysis, granular thin films are of continuing interest. The EXAFS spectra of a series of thin films formed by co-sputtering iron with alumina onto kapton tape have been measured in order to probe the degree of dispersion of the iron particles in the insulating matrix and the nature of the particle-particle and particle-matrix interactions. The samples have previously been investigated by other spectroscopic and magnetic measurements (1-4). According to Mossbauer spectroscopy (J.L. Dormann et al. to be published), between 4.2 and 300 O K the amount of oxidized iron decreases as the particle size increases whereas the relative amount of metallic iron increases, reaching 90% at a particle size of 140 A . Via a specific test for oxy- gen and iron, XPS measurements indicated that oxidised iron phases (Fe 0 and iron aluminate) are present in addition to metallic iron in the external region of the 2.3 samples, i.e. down to 600 1 (1). Finally, a thorough magnetic study provided evidence for the co-existence of both ferromagnetic and antiferromagnetic contribution to the magnetization the former being ascribed to the presence of oxidised iron species (3,4).

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:19868205

(3)

JOURNAL DE PHYSIQUE

The materials under study are listed in the Table, together with their iron content and the particle size (2) : the variation of co-sputtered iron and alumina leads to samples having different particle size in the range 40-140

.

A thorough characte risation cf the samples using electron microscopy and SAXS (G. Cocco, personal comrnu- nication), ensured that the samples measured have a homogeneous particle size, which is of crucial importance for EXAFS study.

The spectra were registered at Fe K-edge at 2eV intervals on the Puls X-ray li- ne, Adone Synchrotron, Frascati (ring energyl.5 GeV ; average current 60 mA).

Measurements were made at room temperature on 6 superimposed tapes each 1 1.I thick. The spectra of Fe203, FeA1204 and bulk Fe were also registered for use as mo- dels for extracting experimental phases and amplitudes of Fe-0, Fe-Fe and Fe-.A1 pairs.

TABLE

Sample C~el(wt.2) Fe oxidized X Fe particle R Backscatterer a from Mossbauer size (i) ²⁵

Fez03 -a I1 shell FeA1204 I1 shell

~ 1 1 ~ ~

S13 S18 S14 IV shell 515 IV shell Fe foil IV shell

a) Only the main backscatterer reported when a complex phase is present; b) The A suffix indicates that the preparation was performed with a rotating target; c) Pre- lifiinary results very similar to S13. The SAXS analysis doesn't indicate an ordered distribution of particle size; d) XRD radius; e) W A F S radius (error 0.02 A ) .

Results and Discussion

Figure 1 shows the experimental EXAFS spectra extracted using standard methods (5) and the Table lists the EXAFS results together with information derived on them from other tecniques. The suitability of EXAFS in bringing to light differences in the nature of the clusters even when amorphous materials are involved is evident from Figure 1. Co-sputtering of iron and alumina is performed in non-equilibrium conditi- ons so that the kind of ordered particle grcwth occurring is impossible to predict a priori, but a visual inspection of the EXAFS spectra indicates that t~lree different types of materials are present.

The radial distribution curves of Figure 2 (phase shifts not applied) show these types may be classed as : short range contributions alonsup to the sample with 75%

of iron content-named S14 in the Table-; 'bulk iron type' for the samples with high iron content, namely for the sample 75% in iron and, more defined, for that (S15) vith 90% iron content ; and 'amorphous type' for sample with intermediate iron con-

tent of 65% (S18 in the Table).

(4)

514, 515, Fe foil.

Analysis of the phases and amplitudes of the first shell in the sample with the lowest iron content (SllA), shows it is due to a Fe-0 contriSution alone, the second shell containing contributions from both Fe.90 and Fe..Fe pairs.

The first point to note is that in this sample the first shell contribution is due to Fe-0 bonds alone. This shows that the particles (as observed by electron microscopy (2)) are not simply composed of metallic iron surrounded by A1203 and con- firms the Mossbauer data at 4.Z°K which are explicable assuming the presence of oxidized iron in this sample.

By way of contrast, the highest iron content samples (75%-90%) clearly show long range ordering, with characteristic a,-Fe b.c.c. peaks, even though X-ray diffra- ction gave no evidence of crystalline order.

Nevertheless, there is a just significant contraction in Fe-Fe distance, as com- pared with a-Fe itself, as expected for metal cluster formation (7-9). There is no change in the bulk iron structure in the 90% iron containirg sample, as judged by the amplitude ratio of the first and fourth peak of the radial distribution function, a&

thought this ratio is doubled in the 75% iron containing sample.

Given that the Fe-Fe bond distance contraction for both first and fourth shell leads to beats at different points in the k-range, and hence only a narrow windowing range is left available, no attempt has been made to extract coordination numbers.

The samples having intermediate iron content, between 55% and 65%, (S13 and S18), are particularly interesting because they are amorphous. They both have a similar first shell distance, in agreement with information from other tecniques (1-4) whick shows that these samples are similar in nature.

The Raverage= 2.26 A may arise from i) a combination of Fe-0 and bulk Fe-Fe distances ; ii) a combination of bulk Fe-Fe and small Fen (n < 5) multimers (believed to give rise to Fe-Fe distances down to 2.02-2.03 for 0.05% iron content in Neon at 4.2"K (7).

The first possibility is preferred given that hyperfine splitting is observed in the Mossbauer spectrum at 4.2OK and Fe-A1203 is detected in XPS (1-2).

A complete fit of the experimental data is underway to confirm this suggestion.

(5)

JOURNAL DE PHYSIQUE

Figure 2. Fourier trans-

forms of the _EXAFSreported in Figure 1, in the same order. Range 2.5 i13.8 k

X-l

gaussian window.

References

1) Paparazzo E., Dormann J.L., Fiorani D., J. Electr. Spectr. and Rel. Phen. 36,

77, (1985).

2) Paparazzo E., Dormann J.L., Fiorani D., Phys. Rev.

m,

1154, (1983).

3) Dormann J.L., Fiorani D., Tholence J.L., J.Magn. Magn. Mat., 35, 117, (1983).

4) Fiorani D., Dormann J.L., Tholence J.L., Bessais L., Villers D., J. Magn. Magn.

Mat., 54-57, 173, (1986).

5) Alagna L., Tomlinson A.A.G., J. Chem. Soc. Faraday Trans. I, 3, 3009, (1982).

6) Montano P.A., Shulze W;, Teshe B., Shenoy G.K., Morrison T.I., Phys. Rev. E,

672, (1985).

7) Purdum H., Montano P.A., Shenoy G.K., Morrison T.I., Phys. Rev., E , ^4412,

(1982).

8) Balerna A., Bernieri E., Picozzi P.,Reale A., Santucci S., B~rattini E.

Mobilio S., Phys. Rev., E, 5058, (1985).

9) Apai G., Hamilton J.F., Stohr J., Thompson A., Phys. Rev. Lett. 43, 165, (1979)