Synthesis of Fine Iron–Cobalt Alloy Particles by the Co-Reduction Of

Synthesis of Fine IronCobalt Alloy Particles by the Co-Reduction of Precursors with Solvated Electrons in SodiumAmmonia Solution

Osamu Terakado+1, Yuichiro Uno+2 and Masahiro Hirasawa

Department of Molecular Design and Engineering, Graduate School of Engineering, Nagoya University, Nagoya 464-8603, Japan

Fine ironcobalt alloy particles were synthesized by the co-reduction of their precursors with sodium in liquid ammonia. The reaction took place instantaneously because of the highly reductive solvated electrons. Very reactive, ﬁne particles of the order of ³10 nm size were obtained. Annealing process is crucial for the magnetic properties of the products, the saturation magnetization being ranging from 29 to 238 A·m2·kg¹1. One spot reduction with solvated electrons and the consequent annealing treatment in the synthesis vessel has potential ability for the synthesis of a variety of metal and alloy particles. [doi:10.2320/matertrans.M2013375]

(Received October 3, 2013; Accepted December 19, 2013; Published January 31, 2014) Keywords: ironcobalt alloy, reduction, solvated electrons, liquid ammonia

1. Introduction fine FeCo particles by the co-reduction of their chloride precursors with sodiumammonia solutions. On the basis of The synthesis of nano-scale metal and alloy materials is of the magnetic properties of the bulk alloy,12) the composition great interest in the past decades. Because of the excellent soft of Fe2Co has been focused, at which the saturation magnet- magnetic properties of the ironcobalt alloy, considerable ization shows a maximum value of 245 A·m2·kg¹1. The attention has been also paid for synthetic processes of nano- overall reaction is described as: scale FeCo alloy particles. Various processes have been CoCl2 þ 2FeCl2 þ 6Na ¼ CoFe2 þ 6NaCl: ð1Þ studied including thermal decomposition of carbonyl precursors,1) sonochemical method,2) and other chemical proc- As the synthesized particles are highly reactive and unstable esses such as microemulsion method3) and polyol process.4) in ambient atmosphere, we focus the annealing treatment of In chemical synthetic routes, a general pathway is the the materials. reduction of the precursor compounds, so that the exploring of reducing reagents is one of the key points in the research 2. Experimental Procedure of the wet processes. Alkali metal is dissolved in liquid ammonia to form 2.1 Materials metastable solutions. One of the interesting properties of the Because metalammonia solutions are less stable in solution is that its physical properties are drastically changed ambient atmosphere, all the syntheses were carried out by the metal concentration, resulting from the excess in pyrex glass vessels connected to a vacuum line 5) ¹3 13) electrons released from the metal. At low metal concen- (³10 mbar). All the vessels were cleaned with H2SO4 tration, the solution has characteristic blue colour arising HNO3 acid mixture in order to avoid the decomposition of from the electrons solvated by ammonia molecules. With the solution, i.e., amide formation, which is promoted by increasing the metal concentration, it changes to shiny impurities on glass walls. metallic colour. Ammonia, purchased from Taiyo Nissan, was liquefied Metalammonia solutions are strongly reductive in nature, and reacted with sodium metal (98%, Kanto Chem. Co., Ltd.) and are used as a reducing reagent particularly in organic at ¹60°C until blue-coloured solution was obtained. Water syntheses.6) With regard to its application to inorganic impurities in ammonia can be rigorously removed by this syntheses, only a few works have been reported, as far as process. The purified ammonia was then vacuum distilled the authors are aware. Zhu and Sadoway reported the into a pyrex glass storage tank containing lithium nitrate. The preparation of some elemental (Ta, Cu) and an alloy (Nb3Al) saturation of the salt enables the store of the liquid ammonia nanoparticles synthesized by sodiumammonia solution.7) at ambient temperature. Alkalides are relevant to the solvated electrons, where the alkali metal anions are produced when dissolved in non- 2.2 Synthesis of fine FeCo particles reducible solvents. Dye proposed the alkalide route for the The mixture of FeCl2 (99.9%, Kojundo Chem. Lab. Co., 8) nanoparticle synthesis, and series of the works, including Ltd.) and CoCl2 (97%, Wako Pure Chem. Co., Ltd.) powders the synthesis of ironcobalt nanoparticles, have been reported at stoichiometric ratio, i.e., eq. (1), was introduced in a by Wagner’s group.911) bottom of a pyrex glass synthesis vessel. Pieces of sodium In the present paper, we address the applicability of metal were put in a small V-shaped vessel which was solvated electron route, describing the synthesis process of connected via glass joint to the top of the synthesis vessel. By turning the V-shaped vessel, sodium metals are dropped into +1Corresponding author, E-mail: [email protected] the bottom of the synthesis cell. Typical composition of the +2Graduate Student, Nagoya University. Present address: Isuzu Motors solution was FeCl2 of 382 mg, CoCl2 of 195 mg, Na of Ltd., Fujisawa 252-0881, Japan 160 mg, and the solution volume of 10 mL. As discussed in 518 O. Terakado, Y. Uno and M. Hirasawa

3.1, the sodium composition was lower than the stoichio- CoFe O metric ratio given in eq. (1), typically 80% of the 2 4 Co stoichiometry. Therefore, we focus on the characterization Fe of reaction products rather than the product yield in the CoFe Co Fe present paper. 3 7 Prior to the reduction of the salts, they were dried under vacuum at 200°C. Ammonia was introduced to the synthesis vessel through vacuum distillation, and the salts were mixed in liquid ammonia at ¹60°C. Sodium was then introduced from the V-shaped compartment. The colour of the solution unit) Intensity (arb. turned to blue, and the reaction took place instantaneously. After the solvent was evaporated, the synthesized powders were taken from the cell, and washed with distilled water. 10° 20° 30° 40° 50° 60° 70° 80° 90° “ ” The product is hereafter called as as-made sample. In most 2θ of the cases, however, the products were taken from the cell in an argon-ﬁlled glove bag, in which black-coloured Fig. 1 XRD pattern of the as-made sample synthesized by reduction of the products were washed with ethanol. precursors with sodiumammonia solution. Diffraction peak positions of ﬁ Some annealing treatments were carried out for the some related compounds are also shown in the gure as reference. reaction products with an electric furnace. In some cases, the annealing was conducted in the pyrex cell prior to taking the sample in a glove bag. Detailed conditions are described in the section of results and discussion.

2.3 Characterization of the reaction products The products were characterized by an X-ray diffractom- eter (Shimadzu, XRD-6100), a scanning electron microscope with an EDS analyser (JEOL, JSM-6330F and JED-2140), and a transmission electron microscope (Hitachi, H-8). The 1μm 50nm magnetic properties were examined by a vibrating sample Fig. 2 SEM (left) and TEM (right) images of as-made powder. magnetometer (Toei Kogyo, VSM-5) at room temperature at the sweep speed of 0.8 MA·m¹1 per 3 min. of amorphous phase as well as crystalline, the latter of which 3. Results and Discussions is probably due to the XRD peaks of cobalt ferrite. As shown in Fig. 3, an elemental mapping of the products 3.1 Characterization of as-made product with a scanning transmission electron microscope (STEM) Figure 1 shows an X-ray diffraction pattern of an as-made suggests the co-reduction of the precursors, as both elements product sample obtained by the reduction reaction. Weak co-exist in the product particles without signiﬁcant segrega- peaks of the cobalt ferrite, CoFe2O4, are observed in the tion. Thus, the result supports the comparable kinetics of the diffractogram. A typical atomic composition of products, reduction of the elements, as anticipated above. determined by a SEM-EDS analysis, is cobalt 9%, iron 21%, In regard to the formation mechanism of the nanoparticles, oxygen 69%, sodium 1% and chlorine of 0.2%. As chlorine its full understanding is one of the most challenging tasks in and sodium are essentially removed from the product, the the synthesis of nanoparticles via chemical route. Here, we reduction of the precursor chlorides takes place in the refer to the work by Zubris et al., where the reaction kinetics treatment with sodiumammonia solution. The as-made of FeCo nanoalloy formation by thermal decomposition of sample, when exposed in ambient atmosphere, results in cabonyl precursors was studied.1) The formation mechanism the exothermic reaction. Thus, it is considered that the of nanoalloy systems is supposed to be co-transformation, oxidation of the synthesized alloy took place during the co-aggregation of different metallic clusters. The reduction handling of the sample, resulting in the substantial oxygen kinetics of each component is one of the crucial factors for content in the product. It should be here also noted that the the phase separation. In the present case, the reduction existence of un-reacted sodium in the synthesis cell leads to kinetics is presumably very fast because of the strong the formation of alkaline compounds during the washing reductive nature of the solvated electrons, rather different treatment of the reaction product, giving rise to the formation from the case of carbonyl decomposition of the reaction time of iron or cobalt hydroxides. Therefore, the sodium content scale of ³103 s. Moreover, the phase diagram of binary should be less than the stoichiometric ratio. Even for the FeCo system exhibits co-solubility over the most of the less sodium content, the atomic ratio of the product is compositional range. Thus, similar reaction rates and Fe : Co ³ 2 : 1, implying the comparable kinetics of the homogeneous formation of the alloy particles are expected reduction reaction of both the chlorides. in the reduction of divalent iron and cobalt salts. The present The morphology of the reaction products was ﬁne powder results support this consideration, though detailed studies, with the particle size of ³10 nm, as shown in the SEM and such as kinetics and high resolution TEM measurement under TEM images (Fig. 2). The TEM image shows the existence careful sample treatment, are desired. Synthesis of Fine IronCobalt Alloy Particles by the Co-Reduction of Precursors with Solvated Electrons in SodiumAmmonia Solution 519

500nm

Fig. 3 STEM image (left) and the corresponding cobalt (middle) and iron (right) mapping images of as-made powder prepared by sodiumammonia solution.

Table 1 Typical atomic composition of the annealed samples determined with SEM-EDS analysis. (a) Co Fe O Na Cl post-anneal 30 53 16 1.3 0.1 p low O2 11 30 57 0.8 0.1 pre-anneal 13 28 56 2.0 0.2

3.2 Influence of the annealing treatment As reported in literature, fine FeCo alloy powders are easily oxidized.2,11) We have, therefore, explored the 1µm following annealing treatments for the preparation of air- stable FeCo alloy particles: (1) post-anneal, where the as-made sample was annealed for 4 h under reducing (b) atmosphere of a gas (95%He5%H2) flow at 600°C, (2) p low O2 treatment, where the sample was exposed under the atmosphere with low oxygen partial pressure. This was achieved by the heating of Cu and Cu2O powder mixture at 150°C held in a glass container attached next to the synthesis cell, and (3) pre-anneal treatment, where sample was annealed at 400°C for 4 h in the synthesis cell after the solvent evaporation. Note that, in the last two types of treatments, the annealing was conducted in the synthesis cell and the products were, then, washed with ethanol in an 1µm argon-filled grove bag, while the sample of post-annealing treatment was exposed in ambient atmosphere before the annealing treatment. (c) Representative atomic composition of the annealed sample is summarized in Table 1. Obviously, the oxygen content is reduced in comparison with the as-made powder (see 3.1). Thus, the annealing treatment is effective for the reduction of the oxidized species. Figure 4 shows the SEM images of the samples after the annealing treatments. The growth in particle size (³100 nm) of the post-annealed sample with distinct particles is obvious, while the samples treated in synthesis cell are not well- fi de ned as that of as-made sample, and do not exhibit the 1µm significant change in particle size. As reported in literature, the annealing at high temperature results in the agglomeration Fig. 4 SEM images of the sample after different annealing treatments: and the particle growth.2,10,11) (a) annealed under HeH2 atmosphere at 600°C for 4 h (post-anneal),

It is anticipated that the annealing under reducing (b) exposed under low oxygen atmosphere for 6 h (low pO2 ), (c) annealed atmosphere (post-anneal) results in the reduction of the under vacuum at 400°C for 4 h in the synthesis cell (pre-anneal). 520 O. Terakado, Y. Uno and M. Hirasawa

CoFe O 2 4 Co -1 200 post-anneal ·kg

Fe 2 CoFe pre-anneal low p Co Fe 100 O2 3 7 / A·m / M 0 as-made Post-anneal -100

-200 Magnetization, Low pO2 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8 Magnetic Field, H / MA·m-1 Intensity (arb. unit) Intensity (arb. Pre-anneal Fig. 6 Room temperature magnetization loop of the synthesized powder.

10° 20° 30° 40° 50° 60° 70° 80° 90° Table 2 Magnetic property of synthesized particles. Saturation Remanence Remanence 2θ Coercivity, Sample magnetization, magnetization, ratio, / ¹1 / 2 ¹1 / 2 ¹1 / / Hc kA·m Fig. 5 XRD pattern of the synthesized powders prepared by different Ms A·m ·kg Mr A·m ·kg S Mr Ms annealing treatments. Diffraction peak positions of some related as-made 29 7 0.24 28 fi compounds are also shown in the gure as reference. post-anneal 238 23 0.097 11 p low O2 32 8 0.25 30 oxides. On the other hand, the annealing under inert pre-anneal 76 18 0.24 33 atmosphere or the treatment under low oxygen partial pressure in the synthesis cell can lead to the formation of protective oxide layer or the agglomeration and the particle growth, preventing the rigorous oxidation when improve the magnetic properties, as the hysteresis loop is exposed in ambient atmosphere. The XRD patterns of the comparable with that of as-made sample. On the other hand, annealed samples are shown in Fig. 5. The post-annealing the annealing in the synthesis cell enhanced the saturation results in the essentially full reduction of the as-made magnetization up to 76 A·m2·kg¹1. product, and the sharp diffraction peaks of the fcc structure Therefore, materials with different magnetic properties can of the metal are observed. As for the influence of the be obtained by an appropriate annealing treatment. Further exposure under low oxygen partial pressure, the existence optimization in the annealing process can improve the of the oxide is observed. However, in comparison to the as- magnetic properties of the FeCo fine particles. made sample (Fig. 1), the complete oxidation is obviously avoided, as intense diffraction peaks of the alloy are 4. Conclusion observed. The pre-anneal treatment results in the further intensification of the peaks of the alloy. Moreover, the Synthesis of fine FeCo particles by co-reduction of signal-to-noise ratio becomes higher than that of the low precursors with sodiumammonia solution is studied in the p O2 treatment, which presumably arises from the annealing present work. The reduction reaction takes place instanta- at high temperature. These results clearly show that both neously, and the chlorine of the precursors are fully removed p the low O2 treatment and pre-annealing are effective for by the treatment with sodium ammonia solutions. Annealing the protection against the rigorous oxidation of the fine treatment is a key issue for the preparation of air-stable particles. particles. Among the examined processes, the pre-anneal process, conducted in the synthesis cell after the reduction, 3.3 Magnetic property of the products avoids the rigorous oxidation, as observed in the XRD The results of the VSM measurements of the synthesized pattern of the product. powder are shown in Fig. 6 and Table 2. Taking the values of In summary, synthesis with solvated electrons is a the remanence ratio into account, all the samples have similar powerful method for the synthesis of fine particles. Further characteristics with an assembly of randomly oriented improvement in the annealing treatment or the process of the uniaxial particles. The as-made sample has the saturation protective layer can achieve the synthesis of fine particles magnetization of 29 A·m2·kg¹1, while the room temperature with variety of magnetic properties. magnetization loop of the post-annealed sample exhibits highest saturation magnetization of 238 A·m2·kg¹1 with Acknowledgments smallest coercivity of 11 kA·m¹1, comparable to the values 12) of the bulk Fe2Co alloy. The latter compound is, thus, The authors acknowledge Prof. Hidefumi Asano and Dr. representative soft-magnetic material. As for the annealing Tetsuya Miyawaki, Nagoya University, for the support of p treatment in the synthetic cell, low O2 treatment doe not VSM measurements. Synthesis of Fine IronCobalt Alloy Particles by the Co-Reduction of Precursors with Solvated Electrons in SodiumAmmonia Solution 521

5) J. C. Thompson: Electrons in liquid ammonia, (Clarendon, Oxford, REFERENCES 1976). 6) G. W. Watt: Chem. Rev. 46 (1950) 317379. 1) M. Zubris, R. B. King, H. Garmestani and R. Tannenbaum: J. Mater. 7) H. Zhu and D. R. Sadoway: J. Mater. Res. 16 (2001) 25442549. Chem. 15 (2005) 12771285. 8) J. L. Dye and K. L. Tsai: Faraday Discuss. 92 (1991) 4555. 2) Q. Li, H. Li, V. G. Pol, I. Bruckental, Y. Koltypin, J. Calderon-Moreno, 9) J. A. Nelson, L. H. Bennett and M. J. Wagner: J. Mater. Chem. 13 I. Nowik and A. Gedanken: New J. Chem. 27 (2003) 11941199. (2003) 857860. 3) J. Ahmed, B. Kumar, A. M. Mugweru, P. Trinh, K. V. Ramanujachary, 10) K. E. Mooney, J. A. Nelson and M. J. Wagner: Chem. Mater. 16 (2004) S. E. Loﬂand, Govind and A. K. Ganguli: J. Phys. C 114 (2010) 31553161. 1877918784. 11) K. E. Mooney and M. J. Wagner: J. Mater. Chem. 19 (2009) 611616. 4) D. Kodama, K. Shinoda, K. Sato, Y. Konno, R. J. Joseyphus, K. 12) P. Weiss and R. Forrer: Ann. Phys. 12 (1929) 279374. Motomiya, H. Takahashi, T. Matsumoto, Y. Sato, K. Tohji and B. 13) O. Terakado, T. Kamiyama and Y. Nakamura: J. Chem. Soc. Faraday Jayadevan: Adv. Mater. 18 (2006) 31543159. Trans. 94 (1998) 867869.