Core-Shell Formation and Juxtaposition in Fe and Si Hybrid Clusters

Prepared by Controlling the Collision Stages

Naokage Tanaka1;*1, Kenji Sumiyama1;*2, Ryoji Katoh1, Takehiko Hihara1,Kazuhisa Sato2, Toyohiko J. Konno2 and Ko Mibu3

1Department of Materials Science and Engineering, Nagoya Institute of Technology, Nagoya 466-8555, Japan2Institute for Materials Research, Tohoku University, Sendai 980-8577, Japan3Department of Engineering Physics, Electronics and Mechanics, Nagoya Institute of Technology, Nagoya 466-8555, Japan

Fe and Si hybrid clusters are co-deposited on substrates in a double-glow-discharge-source-type plasma-gas-condensation system, andobserved by transmission electron microscopy, scanning transmission electron microscopy, X-ray diffraction and Mossbauer spectroscopy.Core-shell Fe/Si clusters, in which chemically heterogeneous bcc Fe-Si cores are covered by amorphous-like Si shells, are obtained on thesubstrate when Fe and Si cluster nuclei collide with each other at an early stage (without setting the partition plate between the two sources). Onthe other hand, Fe and Si clusters are juxtaposed on the substrate when their nuclei collide with each other at a late stage (with setting of thepartition plate). These results demonstrate that we can obtain various hybridized states of Fe and Si clusters by controlling the collision stages.[doi:10.2320/matertrans.M2010177]

(Received May 19, 2010; Accepted August 11, 2010; Published October 6, 2010)

Keywords: hybrid cluster, core-shell morphology, scanning transmission electron microscopy, Mossbauer spectroscopy, X-ray diffraction

1. Introduction

Since blending and alloying are key techniques to improvematerials quality and functionality, equilibrium phase dia-grams have been established in traditional metallurgy.1)

Several non-equilibrium alloys have also been obtained byrapid quenching from vapor states.2) From a mesoscopicpoint of view, it is interesting to blend two kinds ofnanometer size clusters via vapor-phase routes and to realizevarious degrees of hybridized states, which will be useful formagnetic and catalytic applications.3,4)

Recently we have developed a double-glow-discharge-source-type plasma-gas-condensation cluster deposition(PGCCD) system whose deposition chamber was simplyevacuated by a mechanical booster pump.5) Using thePGCCD system, we can generate clusters of two kinds ofmaterials, control their collision stages by setting a remov-able partition plate in the sputtering room and growth duct,and co-depositing them on a substrate. When their collisionstages are delayed with the setting of a partition plate, Fe orCo clusters can be randomly juxtaposed with Al, Si or Niclusters, where Al and Si clusters are amorphous.5–9) On theother hand, when their collision stages are expedited withoutsetting the partition plate, core-shell clusters, in which Feor Co cores are surrounded by amorphous Si or Al shells,and alloy clusters, in which Fe and Ni are mixed with eachother, can be obtained,5–9) depending on the combination ofmaterials.

The following two effects are conceivable origins for thecore-shell cluster formations. (i) an oxidation effect: in thepoor background vacuum condition, Si and Al atoms, andtheir nuclei are easily oxidized because the affinities of Siand Al to O are much higher than those to Fe.10) If Si and

Al nuclei are oxidized before colliding with Fe nuclei, theycannot form alloys.1) (ii) a surface effect: the surface energyof Si is much lower than that of Fe.11) When Fe and Si atoms,and their nuclei collide with each other at their growth stage,Si nuclei cover Fe nuclei.

In order to minimize the oxidization effect, we haveattained a better background vacuum condition of thePGCCD system by adding a deposition room evacuated bya turbo molecular pump and setting a nozzle and a skimmerbetween the growth duct and deposition room (see Fig. 1).12)

In the improved PGCCD system, Fe and Si clusters arejuxtaposed by setting the partition plate, while Fe-rich core-Si rich shell clusters are formed without setting the partitionplate. Moreover, the chemical analyses have suggested thatthe cores are bcc Fe or heterogeneous bcc Fe-Si alloys, theirdiameters increasing with the overall cluster sizes, and theshells are 2 nm thin amorphous Si layers.13) In order toelucidate the detailed structure and morphology of Fe/Sicore-shell clusters, this paper describes experimental resultsof transmission electron microscopy (TEM), scanning trans-mission electron microscopy (STEM), X-ray diffraction andMossbauer spectroscopy on Fe/Si hybrid clusters preparedvia the late and early collision stages.

2. Experimental

The improved PGCCD system has two DC glow dischargesources (see Fig. 1).12,13) After setting Fe and Si (doped with10�2%Al) disc plates (80mm in diameters and 5mm inthicknesses) on the target holders of the sputtering rooms 1and 2, these rooms were evacuated down to 3� 10�4 Pa andthe intermediate room down to 3� 10�5 Pa by a compoundmolecular pump, CMP (not shown in Fig. 1). The depositionroom was evacuated down to 6� 10�6 Pa by a turbomolecular pump, TMP. For glow discharge sputtering,CMP was shut down, argon gas (99.9999 vol%) wasintroduced through gas-inlets with variable leak valves and

*1Graduate Student, Nagoya Institute of Technology. Present address:

Hitachi Maxel Co. Ltd., Osaka*2Corresponding author, E-mail: sumiyama@nitech.ac.jp

Materials Transactions, Vol. 51, No. 11 (2010) pp. 1990 to 1996#2010 The Japan Institute of Metals

the intermediate room was evacuated by a mechanicalbooster pump, MBP (600m3/h). The argon gas flow ratewas kept at 3:3� 10�6 m3/s, leading to the following argongas pressures: 350 Pa in the sputtering room, 2.3 Pa in theintermediate room and 0.3 Pa in the deposition room. WhenDC electric power of 200–400W was supplied to the Fe andSi targets, Fe and Si atoms were sputtered out of the targets.Fe and Si clusters were prepared by independently operatingone glow discharge source, and Fe/Si hybrid clusters byoperating the two glow discharge sources with and withoutsetting the partition plate. They were deposited on carbon-coated micro-grids and thin quartz glass substrates.

Using a transmission electron microscope (TEM, HitachiCo., HF-2000, operating at 200 kV), in which an energydispersive X-ray (EDX) analyzer was installed, morpholo-gies, sizes and average chemical compositions were observedfor lightly deposited specimens and selected area electrondiffraction (ED) patterns for heavily deposited specimens.The detailed morphological and chemical analyses werecarried out using a scanning transmission electron micro-scope (STEM, FEI Titan 80-300, operating at 300 kV) inwhich a field emission gun and an energy dispersive X-rayspectrometer (EDS) were installed. For a high-angle annulardark-field (HAADF) STEM imaging, a detector with theinner angle greater than 50mrad was used. X-ray diffractionpatterns were observed by an X-ray diffractometer (Rigaku4037) in the Bragg-Brentano geometry with Cu-K� radia-tions monochromatized by a graphite crystal. Internalconversion electron Mossbauer spectra were observed by aconstant acceleration type spectrometer at room temperatureusing a He-1%(CH3)3CH gas-flow counter with a 57Cosource in a Rh matrix. The velocity scale was calibrated by astandard Fe film.

3. Results

3.1 Transmission electron microscope studiesFirst we summarize the experimental results of TEM

images and ED patterns on Fe and Si clusters prepared byoperating the single glow discharge cluster source.13) Bcc Feand �-Fe2O3 or Fe3O4 phases coexist in the Fe clusters whoseaverage size is 9.5 nm, while diamond-like Si and SiO2

phases coexist in the Si clusters whose average size is 27 nm.The oxide phases are attributed to oxidation of Fe and Siclusters when exposed to ambient atmosphere. Fe/Si hybridclusters were prepared and studied as follows.3.1.1 Fe/Si hybrid clusters prepared without setting a

partition plateFigures 2(a) and (b) show a TEM image and a size

distribution histogram of Fe/Si hybrid clusters with theaverage composition of Fe-35 at%Si and the average clustersize of about 13 nm. The inset figure of Fig. 2(a) clearlyreveals that the cluster has a core-shell morphology. An EDpattern (Fig. 2(c)) indicates that a bcc Fe phase predominatesand a �-Fe2O3 or Fe3O4 phase slightly coexists, but that nodiamond-like Si phase exists in Fe/Si hybrid clusters eventhough the average chemical composition is Fe-64 at%Si.The lattice constant estimated from the diffraction line is0.283 nm, which is much smaller than that of Fe clusters andbulk Fe (0.2868 nm), suggesting Fe-Si alloy formation.14)

Figure 3 shows HAADF-STEM images ((a),(b)) andSTEM-EDS elemental mapping ((c) Si-K, (d) Fe-K) of Fe/Si hybrid clusters with the average chemical composition of

Fig. 2 TEM observations of Fe/Si hybrid clusters prepared without setting

the partition plate. (a) a TEM image, an expanded (the inset) image, and

(b) a size distribution histogram estimated from Fig. 2(a) for Fe/Si hybrid

clusters whose average composition is Fe-35 at%Si. (c) an electron

diffraction pattern of Fe/Si hybrid clusters whose average composition is

Fe-64 at%Si.

Fig. 1 Schemes of an improved double source plasma-gas-condensation

system. (a) a top view of the system and (b) a side view of the sputtering

room from the left hand side of Fig. 1(a).

Core-Shell Formation and Juxtaposition in Fe and Si Hybrid Clusters Prepared by Controlling the Collision Stages 1991

Fe-34 at%Si. For about a 20 nm region in Fig. 3(b), the brightcontrasted Fe-rich part (Fig. 3(d)) is wholly covered by theweak contrasted Si-rich part (Fig. 3(c)): they equicentrallyoverlap with each other. Quantitative STEM-EDS analysesreveal that the core region is Fe-rich (Fig. 3(e): Fe-6 at%Si),while a considerable amount of Si is detected at the shellregion (Fig. 3(f): Fe-38 at%Si). Although Fe peaks aredetected in both core and shell parts, we cannot distinguishwhether Fe and Si form alloy phases or not.3.1.2 Fe/Si hybrid clusters prepared with setting a

partition plateFigures 4(a) and (b) show a TEM image and a size

distribution histogram of Fe/Si hybrid clusters with theaverage composition of Fe-40 at%Si: the cluster sizes arewidely distributed from 6 to 40 nm. Referring to the results ofsingle source experiments,13) the smaller clusters are Fe andthe larger ones Si. An ED pattern (Fig. 4(c)) shows that bccFe, Fe-oxide and diamond-like Si phases coexist in Fe/Sihybrid clusters whose average composition is Fe-67 at%Si,suggesting no Fe-Si alloy formation even in such a Si-richcomposition.

Figures 5 and 6 show HAADF-STEM images and STEM-EDS elemental mapping of Fe/Si hybrid clusters with theaverage compositions of Fe-40 at%Si and Fe-34 at%Si,

respectively. For about a 20 nm region in Fig. 5(b), brightcontrast parts are Fe-rich (Fig. 5(d)) and weak contrast onesare Si-rich (Fig. 5(c)): they are juxtaposed and partiallyoverlap with each other. STEM-EDS analyses reveal thata large particle with weak contrast is composed of Si(Fig. 5(e)), while the characteristic X-rays both from Fe-Kand Si-K are detected at small clusters with bright contrast,due to the overlapping of small Fe clusters and a large Sicluster (Fig. 5(f)). It is also found from Fig. 6 that smallclusters are Fe (particles 1, 2, 6 in Fig. 6(a): Figs. 6(b), (c),(g)) and large clusters are Si (particle 3 in Fig. 6(a):Fig. 6(d)), while some particles contain both Fe and Si(particles 4, 5 in Fig. 6(a): Figs. 6(e), (f)).

In Figs. 5 and 6, moreover, dark contrast cores are detectedin some Fe and Fe-rich clusters (see particles 1 and 2 inFig. 6(a): Figs. 6(b) and (c)). They can be ascribed tooxidation of Fe clusters15,16) and will be described in detailsfor annealed specimens of Fe and Si hybrid clusters.17)

3.2 X-Ray diffraction studiesFigure 7 shows X-ray diffraction patterns of (a) Fe/Si

hybrid clusters with the chemical composition of Fe-64 at%Siprepared without setting the partition plate, (b) Fe/Si hybridclusters with the chemical composition of Fe-63 at%Siprepared when setting the partition plate and (c) Fe clusters.Here, the broad peaks at around 2� ¼ 22� are ascribed toquartz glass substrates. In Fig. 7(b) the diffraction linesallotted to a bcc phase and a diamond-like Si phase arepredominant, whereas the ones allotted to a �-Fe2O3 orFe3O4 phase are also perceived (for example at around2� ¼ 36�). The angles of the bcc diffraction lines in Fig. 7(b)are almost the same as those in Fig. 7(c), suggesting no

Fig. 3 STEM observations of Fe/Si hybrid clusters with the average

composition of Fe-34 at%Si prepared without setting the partition plate.

(a) and (b) HAADF-STEM images. (c) Si-K and (d) Fe-K maps observed

by STEM-EDS elemental mapping. EDS spectra obtained from a core

(Fig. 3(e)) and a shell (Fig. 3(f)).

Fig. 4 TEM observations of Fe/Si hybrid clusters prepared with setting the

partition plate. (a) a TEM image and (b) a size distribution histogram

estimated from Fig. 4(a) for Fe/Si hybrid clusters whose average

composition is Fe-40 at%Si. (c) an electron diffraction pattern of Fe/Si

hybrid clusters whose average composition is Fe-67 at%Si.

1992 N. Tanaka et al.

Fe-Si alloy formation. In Fig. 7(a) the diffraction linesallotted to a bcc phase are predominant, where the anglesof bcc diffraction lines are higher than those in Fig. 7(c),suggesting Fe-Si alloy formation.

Fe/Si hybrid clusters were prepared without setting thepartition plate and by supplying different electric powers tothe Fe and Si targets for changing the chemical compositions.In X-ray diffraction patterns of these Fe/Si hybrid clusters(see Fig. 8(a)), the angle of a bcc diffraction line becomeshigher and the line-width increases with the increase in theSi composition. As shown in Fig. 8(b), the lattice constantsestimated from Fig. 8(a) gradually decrease with the increasein the Si composition, where the decreasing rate is muchslower than that for bulk bcc Fe-Si alloys.14) These resultsdemonstrate that the Fe-rich bcc cores are made of hetero-geneous Fe-Si alloys. The lattice constants of 0.284 and0.283 nm for Fe/Si hybrid clusters, whose average chemicalcompositions are Fe-38 at%Si and Fe-70 at%Si, correspondto those of the bulk Fe-20 at%Si and Fe-25 at%Si alloys.14)

This figure also suggests that the Fe and Si atoms are mixedwithin the Fe-rich bcc phase field of the equilibrium phasediagram, covering the disordered bcc phase (0–10 at%Si),the B2 type ordered phase (10–11 at%Si) and the DO3 typeordered phase (11–25 at%Si).1)

3.3 Mossbauer effect studiesFigures 9(a) and (b) show conversion electron Mossbauer

spectra measured at room temperature for (a) Fe clusters and(b) Fe/Si hybrid clusters with the average composition ofFe-64 at%Si prepared without setting the partition plate.Choosing the best fitting parameters (an average magnetichyperfine field, HF, an isomer shift, IS and a quadrupolesplitting, QS) listed in Table 1,18) the observed broadspectrum can be divided into the four (bcc Fe, Fe2þ, Fe3þ

and superparamagnetic) components depicted in the samefigure. In particular, the line-width of the sextet in Fig. 9(b)is much wider than that in Fig. 9(a). The broad satellitesextet (lower hyperfine field components) is due to thepresence of Si atoms on the nearest neighbor positions ofcentral Fe atoms,19) being attributed to bcc Fe-Si alloyswhose compositions are widely distributed. Owing to thevery broad Mossbauer spectrum, we can neither evaluate thecompositional distribution among individual Fe-Si alloy

Fig. 5 STEM observations of Fe/Si hybrid clusters with the average

composition of Fe-40 at%Si prepared with setting the partition plate. (a)

and (b) HAADF-STEM images. (c) Si-K and (d) Fe-K maps observed by

STEM-EDS elemental mapping spectra obtained from a large particle

(Fig. 5(e)) and a small particle with bright contrast (Fig. 5(f)).

Fig. 6 (a) an HAADF-STEM image of Fe/Si hybrid clusters with the

average composition of Fe-34 at%Si prepared with setting the partition

plate. Quantitative EDS analyses reveal that small clusters are Fe

(Figs. 6(b),(c),(g)) and large particles are Si (Fig. 6(d)), while some

particles contain both Fe and Si (Figs. 6(e),(f)).

Core-Shell Formation and Juxtaposition in Fe and Si Hybrid Clusters Prepared by Controlling the Collision Stages 1993

cores nor confirm presence of amorphous Fe-Si alloy phaseswhose hyperfine fields are much smaller.20) As shown inTable 1, however, the intensity of the sextet is higher andthose of doublets are lower for Fe/Si hybrid clusters incomparison with these for Fe clusters, suggesting that Sishells protect oxidation of large Fe-rich cores.7) The line-width of the singlet is also narrower in Fe/Si hybrid clustersthan in Fe clusters probably because the magnetic interactionbetween small Fe clusters is weaker in core-shell type Fe/Sihybrid clusters than in Fe clusters, accentuating the super-paramagnetic character.

4. Discussion

The present experimental results indicate how we cancontrol the morphology of the Fe/Si hybrid clusters. Fe andSi nuclei are formed in the sputtering rooms of the PGCCDsystem. When the partition plate is set, the Fe and Si nucleifurther grow in the growth duct and they collide with eachother only near the substrate (at a late stage). In such largeclusters whose surface to volume ratios are small, effects oflattice softening and surface melting are not sufficient forthem to merge with each other.21,22) Therefore, Fe and Siclusters are juxtaposed on the substrate, i.e. no alloy cluster isformed even at their contact interfaces.

Fig. 7 X-ray diffraction patterns observed using Cu radiations for Fe/Si

hybrid and Fe clusters. (a) an Fe/Si hybrid clusters with the average

composition of Fe-64 at%Si prepared without setting the partition plate,

(b) Fe/Si hybrid clusters with the average composition of Fe-63 at%Si

prepared with setting the partition plate and (c) Fe clusters.

Fig. 8 (a) X-ray diffraction patterns observed using Cu radiations for Fe/Si

hybrid clusters prepared without setting the partition plate by supplying

the different electric powers to the Fe and Si targets. (b) the closed circles

are lattice constants of Fe/Si hybrid clusters estimated from Fig. 8(a) and

the open circles those of bulk bcc Fe-Si alloys.14)

Fig. 9 Internal conversion electronMossbauer spectra at room temperature

for (a) Fe clusters and (b) Fe/Si hybrid clusters with the average

composition of Fe-64 at%Si prepared without setting the partition plate.

Dots are experimental spectra and the interpolated lines along the dots are

the total sum of the four spectral components. Individual spectral

components, a sextet, two doublets and a singlet fitted to the experimental

ones, are also shown by lines. The values of an average magnetic hyperfine

field (HF), isomer shifts (IS) and quadrupole splittings (QS) are described

in Table 1.

1994 N. Tanaka et al.

When the partition plate is not set, on the other hand, Feand Si nuclei collide with each other in the sputtering roomand growth duct (at an early stage). Since the surface energyof Si is much lower than that of Fe,11) Si nuclei cover surfacesof Fe nuclei. At the contact interfaces of Fe and Si nuclei,internal diffusions of Fe and Si atoms are enhanced due tosurface melting and lattice softening effects.21,22) The limiteddiffusion at low temperature leads to heterogeneous bcc Fe-Sicores covered by Si shells because Fe matrices can contain Sisolutes up to 25 at%Si, but Si matrices hardly contain Fesolutes in the Fe-Si equilibrium phase diagram.1) A surfacesegregation effect also contributes to maintain the core-shellmorphology.23) A similar alloy formation has been observedin a TEM column where small hemispherical particles (withless than 10 nm) on substrates accept post-impinging atoms atambient temperature.24,25) Such instantaneous alloying isaccelerated for element combinations having negative mix-ing enthalpies. Moreover, amorphous Fe-Si alloy layers havebeen obtained by solid state reaction when thin Fe films areepitaxially grown on Si substrates at ambient temperature.26)

At this moment we cannot exclude presence of amorphousSi-rich Fe-Si alloys in the shell regions.

Finally it is worth to mention why several ordered andintermetallic-compound phases have not been obtained in thebcc Fe-Si core regions despite their very large cohesiveenergies.1,11) In bulk Fe-Si alloys at high temperatures, thediffusion coefficient of Fe is much larger and the activationenergy is lower in the DO3 type ordered Fe3Si alloys than inbcc Fe metals and disordered bcc Fe-Si alloys, while thediffusion coefficient of Si is smaller and the activation energyis higher in the ordered alloys than in the disordered alloys.27)

This is ascribed to the high vacancy concentrations as aconsequence of the low vacancy formation energy of Fe atomsites in ordered Fe3Si alloys. Indeed, ordered Fe3Si alloyclusters have been obtained when Fe-Si clusters grow viaa similar gas phase synthesis route using a single glowdischarge source and composite targets.28) In the presentexperiment, the ordered Fe3Si alloy phase might be formednear contact interfaces between Fe and Si clusters, but itcould not be detected, probably due to the serious chemicalheterogeneity. When the amorphous Fe-Si layers on Sisubstrates were annealed or thin Fe films were deposited onhigh-temperature Si substrates, B20-type ordered FeSi alloylayers were formed at 570K, and transformed to �(CaF2

type)-FeSi2 islands and �-FeSi2 rods at higher temper-atures.26,29) In the bulk B20-type ordered ("-) FeSi alloy,diffusivities of both Fe and Si are small, and the value of Fe is

much smaller than that of Si.30) In Fe-Si cores, therefore, thelimited diffusions of Fe and Si atoms at low temperaturehinder formation of "-FeSi and �-FeSi2 phases.

5. Conclusion

The observations of transmission electron microscopy,scanning transmission electron microscopy, X-ray diffractionand Mossbauer spectroscopy have clearly confirmed that thestructures and morphologies of Fe/Si hybrid clusters can bevaried by controlling the collision stages in the double-glow-discharge-source-type PGCCD system. Fe and Si clusters arejuxtaposed on the substrate when these nuclei collide witheach other at a late stage. On the other hand, chemicallyheterogeneous bcc Fe-Si cores are covered by amorphousSi shells on the substrate when they collide with each otherat an early stage. Degrees of homogeneity in Fe/Si hybridclusters will be controlled by changing the electric power ofthe glow discharge source and the substrate temperature.

Acknowledgments

The authors wish to thank Mr. Y. Sasaki for hisexperimental contribution. This work was supported byIntellectual Cluster Project by the Ministry of Education,Culture, Sports, Science and Technology (MEXT), Japan,Aichi Prefecture, Nagoya City and Aichi Science andTechnology Foundation, and a Grant-in-Aid for ScientificResearch given by MEXT, Japan. The measurement ofMossbauer spectra was supported by the NanotechnologyNetwork Project of MEXT, Japan. One of the authors (KS)appreciates the support from Nagoya Industrial ScienceResearch Institute. The authors were also indebted toEmeritus Professor G. M. Graham of Toronto University,Canada for correcting the paper draft.

REFERENCES

1) T. B. Massalski, H. Okamoto, P. R. Subramanian and L. Kacprzak ed.:

Binary Alloy Phase Diagrams, 2nd ed., (ASM International, Ohio,

1990) pp. 1771–1772.

2) K. Sumiyama: Phys. Stat. Sol. (a) 126 (1991) 291–312.

3) H. Gleiter: Prog. Mat. Sci. 33 (1989) 223–315.

4) A. S. Edelstein and R. C. Cammarata: Nanomaterials: Syntesis,

Properties and Applications, (Institute of Physics Publishing, Bristol,

1996).

5) R. Katoh, T. Hihara, D. L. Peng and K. Sumiyama: Appl. Phys. Lett. 82

(2003) 2688–2690.

Table 1 Fractions of the spectral components and the fitting parameters (isomer shifts, IS, quadrupole splittings, QS and an average

magnetic hyperfine field, HF) for internal conversion electron Mossbauer spectra shown in Fig. 9.

Fe clusters Fe/Si hybrid clusters

Sextet

(bcc Fe)

0.39

(IS ¼ 0:00mm/s; HF ¼ 33T)

0.54

(IS ¼ 0:04mm/s; HF ¼ 33T)

Doublet

(Fe2þ)

0.12

(IS ¼ 0:90mm/s; QS ¼ 1:10mm/s)

0.10

(IS ¼ 1:42mm/s; QS ¼ 0:94mm/s)

Doublet

(Fe3þ)

0.37

(IS ¼ 0:35mm/s; QS ¼ 0:97mm/s)

0.29

(IS ¼ 0:31mm/s; QS ¼ 1:02mm/s)

Singlet

(superparamagnet)

0.12

(IS ¼ 0:00mm/s)

0.07

(IS ¼ 0:00mm/s)

Core-Shell Formation and Juxtaposition in Fe and Si Hybrid Clusters Prepared by Controlling the Collision Stages 1995

6) R. Katoh, K. Nonaka, K. Sumiyama, D. L. Peng and T. Hihara: Mater.

Trans. 49 (2008) 1830–1835.

7) R. Katoh, T. Hihara, D. L. Peng and K. Sumiyama: Appl. Phys. Lett. 87

(2005) 252501.

8) R. Katoh, T. Hihara, D. L. Peng and K. Sumiyama: J. Appl. Phys. 100

(2006) 034308.

9) A. Monden, R. Katoh, D. L. Peng and K. Sumiyama: Mater. Trans. 47

(2006) 1949–1952.

10) A. F. Wells ed.: Structural Inorganic Chemistry, 5th ed. (Oxford

Science Publications, Oxford, 1987).

11) F. R. de Boer, R. Boom, W. C. M. Mattens, A. R. Miedema and A. K.

Niessen: Cohesion in Metals—Transition Metal Alloys, 2nd ed. (North-

Holland, Amsterdam, 1989).

12) K. Sumiyama, A. Monden, R. Katoh, N. Tanaka, D. L. Peng and T.

Hihara: Mater. Trans. 50 (2009) 516–522.

13) K. Sumiyama, R. Katoh, S. Kadowaki and T. Hihara: J. Nanopart. Res.,

12 (2010) 2589–2596.

14) W. B. Pearson: A Handbook of Lattice Spacings and Structures of

Metals and Alloys, (Pergamon Press, London, 1958) pp. 655–658.

15) Y. Yin, R. M. Rioux, C. K. Erdonmez, S. Hughes, G. A. Somorjai and

A. P. Alvisatos: Science 304 (2004) 711–714.

16) R. Nakamura, D. Tokozakura, J.-G. Lee, H. Mori and H. Nakajima:

Materia Japan 47 (2008) 368–374.

17) R. Katoh, S. Kadowaki, K. Sumiyama, N. Tanaka and T. Hihara: in

preparation.

18) N. N. Greenwood and T. C. Gibb:Mossbauer Spectroscopy, (Chapman

and Hall, London, 1971).

19) M. B. Stearns: Phys. Rev. 129 (1963) 1136–1144.

20) G. Marchal, Ph. Mangin and Chr. Janot: Solid State Commun. 18

(1976) 739–742.

21) S. Sugano and H. Koizumi: Microcluster Physics, 2nd ed., (Springer-

Verlag, Berlin, 1998).

22) P. Jensen: Rev. Mod. Phys. 71 (1999) 1695–1735.

23) Y. H. Xu and J. P. Wang: Adv. Mat. 20 (2008) 994–999.

24) H. Mori, M. Komatsu, K. Takeda and H. Fujita: Philos. Mag. Lett. 63

(1991) 173–178.

25) H. Yasuda and H. Mori: Phys. Rev. Lett. 69 (1992) 3747–3750.

26) J. H. Won, A. Kovacs, M. Naito, M. Ishimaru and Y. Hirotsu: J. Appl.

Phys. 102 (2007) 103512.

27) A. Gude and H. Mehrer: Phil. Mag. A 76 (1997) 1–29.

28) Y. Jing, Y. H. Xu and J. P. Wang: J. Appl. Phys. 105 (2009) 07B520.

29) J. H. Won, K. Sato, M. Ishimaru and Y. Hirotsu: J. Appl. Phys. 100

(2006) 014307.

30) M. Salamon and H. Mehrer: Phil. Mag. A 79 (1999) 2137–2155.

1996 N. Tanaka et al.