Thermal and Magnetic Properties of Mechanically Alloyed fcc Cu-Fe

Supersaturated Solid Solutions

Jean-Claude Crivello1, Tohru Nobuki1 and Toshiro Kuji2;*

1School of High Technology for Human Welfare, Tokai University, Numazu 410-0395, Japan2Courses of Materials Science and Chemistry Unified Graduate School, Tokai University, Numazu 410-0395, Japan

Copper and iron are immiscible elements according to the equilibrium phase diagram, but they can form metastable phases by mechanicalalloying process.

In this present work, mixtures of Cu-Fe powders in the range 0, 12, 25 and 40 atomic% of Fe have been prepared by ball milling. Theanalysis of alloyed samples shows a single phase described in the fcc-Cu structure, except for the 40%-Fe compound, which presents theadditional bcc-Fe phase. The study of microstructure and magnetic properties under thermal treatments suggests a decomposition of themetastable phase with increasing temperature.

In the Cu-richer domain, the fcc cell parameter increases with increasing Fe content. This effect is explained from the fact that theferromagnetic Fe phase is dispersed in the form of nanosized particles in the paramagnetic Cu matrix, in agreement with previous reports.[doi:10.2320/matertrans.MRA2007189]

(Received August 2, 2007; Accepted December 25, 2007; Published February 20, 2008)

Keywords: copper iron alloys, mechanical alloying, non-equilibrium compound, microstructure, thermal and magnetic properties

1. Introduction

Cu-Fe alloys are expected to be high strength and highelectric conductive materials.1) It is also indicated that thesealloys exhibit giant magneto-resistance and other outstandingphysical properties.2) But the applications associated to theirgreat potentials are limited since Cu-Fe alloys cannot formstable phases: they are described in a peritectic system with amiscibility gap.3) According to the Miedema model,4) theenthalpy of mixing are largely positive (�Hmix ’ þ13

kJmol�1 at 50–50 composition), in agreement with thevalues calculated recently with an empirical potential de-scription.5)

Mechanical alloying (MA), as for example the high energyball milling (BM), was proved to be useful to synthesize var-ious phases even with positive mixing enthalpy.6–8) It hasbeen suggested that the energy stored in the large grain-boun-dary and interfacial surface areas serves as a driving force foralloy formation. In the early 1990’s, several groups reportedthe formation of metastable solid solutions in the Cu-Fe sys-tem by these techniques.9–12) It has been shown that the me-chanical process reduces the crystalline size as the millingtime increases. More recently, the Cu-Fe alloys propertieswere reviewed by several groups.13–15) Particularly, the mag-netic properties of Cu-Fe have been studied extensively. Itshould be noticed that fcc-Cu supersaturated solid solutionswith the intrusion of Fe atoms have a large magnetic momentcomparable to bcc �-Fe.9,16,17) Ino et al.18,19) concluded thatthe ferromagnetic state of the fcc Cu-Fe alloys is realizedby the expansion of the lattice in accordance with theBethe-Slater curve. In addition, Chien et al.20) have shownthat the Curie temperature is very sensitive to the initial com-position. Furthermore, the annealing of the MA specimenseffects the decomposition of the solid solution into fcc-Cuand Fe.21) However, all the thermal and magnetic experi-ments were performed at temperatures lower than 400�C.

The purpose of the present paper is to clarify thedependence of the alloy stability towards temperature andparticularly according to its microstructure and magneticproperties. Measurements of Curie temperature for twosamples Cu75Fe25 and Cu60Fe40 are presented with associatedmagnetization measurements above 400�C.

2. Experimental Details

The alloys were prepared from mixtures of Cu and Fepowders (Kojundo Chemical Lab., size < 150 mm, 99.9%purity). The milling was conducted with 20 h of BM alloyingusing a high energy Nisshin-Giken Super-Misuni NEV #8,under argon atmosphere and a rotation speed of 720 rpm. Thevial and the balls were in zirconia. 14 balls (10mm diameter)were used with the ball to powder weight ratio (BPR) of10 : 1.After preparation of samples on the focused ions beanapparatus (Hitachi FB-2000A), transmission electron mi-croscopy (TEM) analysis was determined using high reso-lution TEM (Hitachi FE-TEM HF 2200TV), operating at200 kV. The scanning electron microscopy (SEM) used is theHitachi S-4000 model (20 kV), equipped with the HoribaEMAX-2770 EDX spectrometer.The differential scanning calorimetry (DSC) measurementswere performed with a Rigaku DSC8230 instrument with aheating rate of 10K/min under flow of purified argon gas.The powders were characterized by X-ray diffraction (XRDtype MAC science MXP3 and Shimazu XRD-6100) withCu K� radiation at 40 kV–30mA settings, 2� from 40� to140� at room temperature, and 2� from 40� to 55� withheating at 50, 150 and 300�C under vacuum. The resultingprofiles were refined with the Rietveld program Fullprof.22)

Magnetic hysteresis and Curie point measurements werecarried out using vibrating sample magnetometer (VSM)apparatuses with a heating rate of 10K/min.

*Corresponding author, E-mail: tkuji@urchin.fc.u-tokai.ac.jp

Materials Transactions, Vol. 49, No. 3 (2008) pp. 527 to 531#2008 The Japan Institute of Metals

3. Results and Discussion

3.1 Structural and thermal analysisThe XRD patterns for the 4 samples after milling are

shown in Fig. 1. The single fcc-Cu phase remains for all thecompounds from 0% to 25% of Fe. On the pattern of the 40%sample, bcc-Fe peaks appear and show a separation into twophases. These results are not completely similar to previous

works,12,16,23) where it was found that the single fcc phaseextends up to 60% of Fe in fcc solution. By increasing theatomic composition of Fe, the average intensity of peaksdecreases compared to the background line, which means thatcrystallization process is poor. The second effect is thedecrease of fcc peaks position with smaller angles andassociated increase of fcc cell parameter. The fcc latticeparameter expansion has already been reported by severalexperiments.12,18) This effect cannot be explained by thesmaller atom size of Fe compared to Cu (radius 0.126 against0.128 nm), but by the magnetovolume effect: when themagnetic iron nanoparticules are alloyed into the sub-matrixof copper, the interatomic potential for Cu has a strongercore-core repulsion due to the filled d-band shell.20,24) Thiseffect leads to the extension of the fcc cell parameter.2,25)

The effect of the temperature on the lattice expansion wasalso investigated. At low angles, the in situ X-ray diffractionwas made by gradual increase of temperature from 50, 150 to300�C. As an example, XRD patterns of the 12% Fecompound are presented in Fig. 2. Considering this sample,the decomposition into two phases (fcc-Cu and bcc-Fe)appears since annealing starts at 50�C. After all the heatingprocesses, irreversible decomposition remains at roomtemperature. By the Rietveld method, fcc cell parametershave been identified. Figure 3 sums up these results, obtainedwith accuracies of �2 < 4 and refined agreement factorsRwp < 9%. They show a linear increase of fcc latticeparameter with increasing Fe content and with increasingtemperature. As a reference in Fig. 3, the value of non-alloyed Cu parameter is lower than all results found, exceptfor the ball-milled pure-Cu at room temperature.

With DSC analysis, the decomposition process was clearlyidentified. Figure 4 shows DSC traces for 20mg of the 25%and 40%-Fe samples. Each presents an exothermic peakrespectively at TX ¼ 424�C and 392�C, corresponding to thetemperature of decomposition into two fcc and bcc phases(results checked by XRD). The solution stability decreaseswith increasing Fe-content. Thermal decomposition of thesupersaturated fcc solution has already been reported atsimilar temperatures.10,12,16,21)

40 50 60 70 80 90 100 110 120 130 140

2 θ (°)

(d) 60% Cu + 40% Fe (2 phases)

(c) 75% Cu + 25% Fe

(b) 88% Cu + 12% Fe

Inte

nsity

, I /

arb.

uni

ts

(a) 100% Cu

Fig. 1 Observed (points), calculated (line) and difference (bottom line) of

XRD powder diffraction patterns of Cu100�XFeX milled for (a) X ¼ 0, (b)

X ¼ 12, (c) X ¼ 25, (d) X ¼ 40. The Bragg positions of fcc-Cu and bcc-Fe

are indicated by vertical bars.

Fig. 2 XRD powder diffraction patterns of Cu88Fe12 after alloying, during

heating at 50-150-300�C, and after treatments at room temperature.

528 J.-C. Crivello, T. Nobuki and T. Kuji

3.2 MicroscopyFigure 5 shows an overview of Cu75Fe25 powder on SEM

microscopy. The sample after milling (a) and after heattreatment at 750�C for 1 hour (b) present particles with thesame size range (about 300 mm). Modifications on themicrostructure coming from heating can be identified by asmaller scale: TEM bright-field micrographs and correspond-ing Z-contrast STEM images of the same sample presentsignificant differences (Fig. 6). Whereas the compound justafter milling (a,a0) has a homogeneous structure, the post-annealing compound (b,b0) presents different regions ofcontrast. The white frame in Fig. 6(b0) is analyzed in Fig. 7,which contains elemental distribution maps. The correspond-ing EDX results are summarized on Table 1: the elementscomposition analysis in Fig. 7 imply a composition veryclose to the initial composition Cu:Fe (74.96:25.04). TheTEM micrograph shows a distribution of brighter regionswith a small size range (from 100 nm to 400 nm). EDX spotanalysis suggests that this brighter region corresponds to ironparticles. The darker region is richer in copper.

3.3 Magnetic propertiesShown in Fig. 8 are the hysteresis loops of magnetization

measured for 0.35 g of Cu75Fe25 and Cu60Fe40 powders. Thecoercive field HC measured is respectively about 455Oe and365Oe, which are larger than pure bcc-Fe under the sameconditions.

Figure 9 shows the magnetization at 5 kOe of the same two25% and 40%-Fe samples according to temperature T , duringcontinuous heating and subsequent cooling. Magnetizationdrops significantly near to 200�C, slightly lower than 2 othermeasurements at the Cu50Fe50 composition: 230�C forYavari et al.10) and 240�C for Jiang et al.26) By increasingT higher than 400�C, a second magnetization drop shows atypical signature of ferromagnetic iron at respectivelyTC ’ 775�C and 780�C, close to the Curie temperature ofFe (770�C). We concluded that this anomalous behavior isjustified by means of spinodal decomposition at inflectionpoint TX, which corresponds to the decomposition into bcc Feand fcc Cu predicted by DSC analysis at TX. This chemicalphase-separation was already found for richer Fe composi-tions.13,21,27)

4. Conclusions

We have shown that the high-energy ball milling ofelemental powders can be used to alloy the immisciblecopper-iron system with formation of metastable solidsolutions. Regarding the solubility of Fe elements inCu100�XFeX sub-matrix, we found that single fcc phaseoccurs until X � 25% of Fe. As reported previously, theincrease of fcc cell parameter by increasing Fe-content maybe attributed to magnetovolume effect. The effect of heatingcauses iron particles to precipitate out of the copper matrix as

Fig. 3 fcc lattice parameter of Cu100�XFeX according to the % Fe-content

at room temperature and during heating at 50-150-300�C.

EXO

Fig. 4 DSC traces from 20mg Cu60Fe40 (dashed line) and Cu75Fe25 (line)

samples.

Fig. 5 SEM micrograph of the Cu75Fe25 sample, after milling (a) and post-annealing (b).

Thermal and Magnetic Properties of Mechanically Alloyed fcc Cu-Fe Supersaturated Solid Solutions 529

Fig. 6 Bright-field TEM micrograph and corresponding Z-contrast STEM image of the Cu75Fe25 sample, after milling (a,a0) and post-

annealing (b,b0).

Fig. 7 Bright-field TEM micrograph from Fig. 6(b0) with the corresponding distribution image of Cu and Fe.

Fig. 8 Hysteresis loops of magnetization measured at 5 kOe from 0.35 g Cu60Fe40 (dashed line) and Cu75Fe25 (line) samples.

530 J.-C. Crivello, T. Nobuki and T. Kuji

a decomposition process at temperature around 400�C. Thisresult was been shown in mutual agreements by severalanalyses in the present work.

Acknowledgements

This work was partially supported financially by theFrench Ministry for Foreign Affairs. The authors thankDr. Sumida for the help during the in-situ XRD analysis inTokyo University. Dr. Saito of Chiba Institute of Technologyand associate Professor Chiba of Tokai University areacknowledged for their support in the instrumental VSManalysis.

REFERENCES

1) C. Biselli and D. G. Morris: Acta Metall. Mater. 42 (1994) 163–176.

2) S. S. Saxena, J. Tang, Y. S. Lee and C. J. O’Connor: J. Appl. Phys. 76

(1994) 6820–6822.

3) L. J. Swartzendruber: Binary Alloy Phase Diagrams, second ed., ed by

T. B. Massalski, (ASM, New York, 1990) pp. 1408–1409.

4) A. R. Miedema: Philips Tech. Rev. 36 (1976) 217.

5) A. Caro, P. E. A. Turchi, M. Caro and E. M. Lopasso: J. Nucl. Mater.

336 (2005) 233–242.

6) C. Suryanarayana: Prog. Mater. Sci. 46 (2001) 1–1184.

7) T. Kuji, S. Nakayama, N. Hanzawa and Y. Tabira: J. Alloys Compd.

356 (2003) 456–460.

8) H. Hotta, M. Chiba, T. Kuji and H. Uchida: J. Japan Inst. Metals. 70

(2006) 662–665.

9) K. Uenishi, K. F. Kobayashi, S. Nasu, H. Hatano, K. N. Ishibara and

P. H. Shingu: Z. Metallkd 83 (1992) 132–135.

10) A. R. Yavari, P. J. Desre and T. Benameur: Phys. Rev. Lett. 68 (1992)

2235–2238.

11) A. R. Yavari: Phys. Rev. Lett. 70 (1993) 3521–3522.

12) J. Eckert, R. Birringer, J. C. Holzer, C. E. Krill III and W. L. Johnson:

Mat. Res. Soc. Symp. Proc. 238 (1992) 739–744.

13) O. Drbohlav and A. R. Yavari: Acta Metall. Mater. 43 (1995) 1799–

1809.

14) J. Z. Jiang, C. Gente and R. Bormann: Mater. Sci. Eng., A 242 (1998)

268–277.

15) L. B. Hong and B. Fultz: Acta Mater. 46 (1998) 2937–2946.

16) E. Ma, M. Atzmon and F. E Pinkerton: J. Appl. Phys. 74 (1993) 955–

962.

17) G. Mazzone and M. V. Antisari: Phys. Rev. B 54 (1996) 441–446.

18) H. Ino, K. Hayashi, T. Otsuka, D. Isobe, K. Tokumitsu and K. Oda:

Mater. Sci. Eng., A 304–306 (2001) 972–974.

19) H. Ino, M. Kincho, K. Oda and K. Tokumitsu: Mater. Sci. Eng., A 312

(2001) 262–266.

20) C. L. Chien, S. H. Liou, D. Kofalt, Wu Yu, T. Egami, T. J. Watson and

T. R. McGuire: Phys. Rev. B 33 (1986) 3247–3250.

21) P. Crespo, A. Hernando, R. Yavari, O. Drbohlav, A. Garcıa Escorial,

J. M. Barandiaran and I. Orue: Phys. Rev. B 48 (1993) 7134–7139.

22) Rodrıguez-Carvajal: J. Proceedings of the XV congress of international

union of crystallography, (satellite meeting on powder diffraction,

Toulouse, France, 1990) pp. 127.

23) M. Qi, M. Zhu and D. Z. Yang: J. Mater. Sci. 13 (1994) 966–968.

24) V. Vitek and Y. Minonishi: Surf. Sci. 144 (1984) 110–123.

25) Y. R. Uhm, W. W. Kim and C. K. Rhee: Phys. Stat. Sol. A 201 (2004)

1802–1805.

26) J. Z. Jiang, Q. A. Pankhurst, C. E. Johnson, C. Gente and R. Bormann:

J. Phys.: Condens. Matter 6 (1994) L227–L232.

27) P. Crespo, A. Hernando and A. Garcıa Escorial: Phys. Rev. B 49 (1994)

13227–13230.

Table 1 Chemical compositions measured in Fig. 6 and corresponding

spots EDX analysis of the Cu75Fe25 sample.

area spot 1 spot 2

% Cu-K 74.96 77.15 50.01

% Fe-K 25.04 22.85 49.99

Fig. 9 Magnetization according to the temperature at 5 kOe from 0.03 g

Cu60Fe40 (dashed line) and Cu75Fe25 (line) samples.

Thermal and Magnetic Properties of Mechanically Alloyed fcc Cu-Fe Supersaturated Solid Solutions 531