This paper introduces a potentially scalable electro-infiltration process to produce exchange-coupled hard magnetic nanocomposite thin films. Fe-Co/CoPt nanocomposite films are fabricated by deposition of CoFe2O4 nanoparticles onto Si substrate, followed by electroplating of CoPt. Samples are subsequently annealed under H2 to reduce the CoFe2O4 to magnetically soft Fe-Co and also induce L10 ordering in the CoPt. Resultant films exhibit 0.97 T saturation magnetization, 0.70 T remanent magnetization, 127 kA/m coercivity and 21.8 kJ/m3 maximum energy density. First order reversal curve (FORC) analysis and δM plot are used to prove the exchange coupling between soft and hard magnetic phases.

Since the concept of exchange-spring magnet was proposed by Kneller and Hawig in 1991,1 nanocomposite materials comprising soft and hard magnetic phases have been extensively studied through theoretical modeling (e.g. Refs. 2 and 3) and experiments in the pursuit of strong permanent magnets that outperform any single-phase materials. Two-dimensional bilayer thin film structures have experimentally confirmed the exchange-coupling-based enhancement of energy density.4,5 To realize thicker, and three-dimensional exchange-coupled materials, sputtering systems have been modified for gas-phase nanoparticle deposition (cluster deposition) to create an inclusion-matrix structure for nanocomposites.6–8 In comparison to that approach, this study utilizes magnetic nanoparticle building blocks and an electro-infiltration process9,10 to fabricate three-dimensional exchange-coupled Fe-Co/CoPt nanocomposite films with relatively simple experimental infrastructure.

The target structure of the nanocomposite is closely packed, nanoscale, soft magnetic Fe-Co inclusions embedded in a hard magnetic L10 CoPt matrix. As shown in Fig. 1, CoFe2O4 nanoparticles are first consolidated on a Si substrate to form a porous particle layer. Next, equiatomic CoPt is electroplated through the particles, forming a two-phase three-dimensional composite. The nanocomposite is then annealed to convert CoPt from the low-coercivity A1 phase to the high-coercivity L10 phase and simultaneously to reduce the CoFe2O4 into high magnetization Fe-Co. The details of the fabrication process are described further below.

The process starts with a (100) Si substrate that is sputtered with a 25 nm TiN insulation layer to prevent diffusion during annealing, a 10 nm Ti adhesion layer, a 30 nm Pt seed layer, and another 10 nm Ti adhesion layer. A 10.5 μm thick AZ 9260 (MicroChemicals GmbH) photoresist is then spin-coated on top of the substrate and patterned via standard photolithography to form pattern-defining molds in the shapes of either 3.6 mm × 3.6 mm square or 3.6 mm diameter circle.

CoFe2O4 nanoparticles are synthesized using aqueous co-precipitation.11 CoFe2O4 particles are used instead of metal Fe-Co particles, since the latter are prone to oxidation and are difficult to store and process. Co-precipitation is selected because there is no organic ligand involved in the process, which can create impurities and inhibit the exchange interaction at the soft-hard interface in the final nanocomposite. To maximize the nanocomposite energy density, various models suggest the optimal diameter of Fe-Co inclusions in a range of 6 nm to 15 nm.1–3 The synthesized CoFe2O4 nanoparticles exhibit a hydrodynamic diameter of 10.7±1.4 nm, as measured by dynamic light scattering. Considering mass loss and density change, the average diameter of Fe-Co particles after annealing is expected to be 8.3 nm (ignoring diffusion effects). To briefly introduce the synthesis process, 100 mL 0.1 M FeCl3 and 100 mL 0.05 M Co(NO3)2 precursor solutions are mixed and stirred. Then 100 mL 1.8 M NaOH solution is added, and the entire solution is heated at 95°C for 12 h. After synthesis, the particles are collected by an external magnet and washed with 500 mL deionized water 3 times, stirred with 60 mL 2 M HNO3 for 5 min, and washed another 4 times, followed by 3 steps of 24 h dialysis in 4 L deionized water in order to remove the excess salt in solution. After vortex mixing and sonication, the solution is let stand to allow agglomerated nanoparticles to precipitate, and then the supernatant containing the stably suspended particles is collected for further processing. A 2 g/L concentration is utilized for the subsequent deposition of particle films.

Before deposition of the nanoparticles, the Si substrate is dipped in diluted HF to remove Ti and expose the Pt seed layer. 30 μL CoFe2O4 particle solution is dispensed onto the substrate and dried in a freeze box to form a consolidated particle layer. Next, the substrate is dipped into deionized water to wet the particles, and CoPt is electroplated under a constant current density of 100 mA/cm2 for 5 min using the recipe described in Ref. 12. The photoresist is then stripped using acetone followed by isopropanol. The substrate is placed in a sonication bath for 5 min to remove any excess unbound CoFe2O4 particles on the film surface. The substrate is then annealed in forming gas (4% H2, 96% N2) using a tube furnace (Lindberg/Blue M HTF55322C) at 675°C for 30 min with a 22°C/min ramp up rate followed by natural cool down. Additional films of just CoFe2O4 nanoparticles and just electroplated CoPt are also fabricated and annealed as control samples.

A scanning electron microscope (FEI Nova NanoSEM 430) equipped with an energy-dispersive x-ray spectroscope (EDS) detector (EDAX) is used for imaging, film thickness measurement and elemental composition analysis. The in-plane magnetic properties are characterized by hysteresis loop, first order reversal curve (FORC) diagram, and δM plot using a vibrating sample magnetometer (VSM) (ADE Technologies Model EV9) with a saturating field of 2.26 T.

Figure 2 shows cross-section SEM images of consolidated CoFe2O4 nanoparticles, an annealed nanoparticle layer (without CoPt), an annealed single-phase CoPt layer (without particles), and an annealed Fe-Co/CoPt nanocomposite film with a representative thickness of 350 nm. CoFe2O4 nanoparticles (Fig. 2a) are distinguishable in the consolidated particle film and Fe-Co crystals (Fig. 2b) are formed after annealing. Both CoPt film (Fig. 2c) and Fe-Co/CoPt nanocomposite film (Fig. 2d) exhibit grain structures, but Fe-Co phase and CoPt phase inside the nanocomposite film cannot be recognized from the secondary electron images. The elemental compositions for the single-phase materials and the nanocomposite, based on the EDS data, are shown in Table I. The synthesized CoFe2O4 nanoparticles exhibit a 1.8:1 atomic ratio between Fe and Co, close to the 2:1 ratio in the ferrite composition, and the ratio remains after the particles are annealed. The oxygen observed in the annealed particle layer may be attributed to the oxidization of Fe-Co alloy after annealing. The composition ratio of Co to Pt is measured to be around 39:61, which deviates from the ideal ratio of 50:50 required for the L10 phase. However, the Ti seed layer contributes additional signal in the measured composition ratio.

As shown in the Fig. 3a, the resultant Fe-Co/CoPt nanocomposite exhibits a relatively smooth hysteresis curve indicative of exchange coupled medium with a small kink (as opposed to a mixture of two phases, which may exhibit a pronounced kink). The nanocomposite exhibits a high saturation magnetization of 0.97 T and a remanent magnetization of 0.70 T, much higher than the 0.45 T and 0.42 T, respectively, for the CoPt film. The coercivity of 127 kA/m is 20x larger than the coercivity of the reduced/annealed Fe-Co particle film (6.6 kA/m), but about one-sixth that of the CoPt film (784 kA/m). Consequently, the nanocomposite maximum energy density of 21.8 kJ/m3 is slightly lower than the CoPt film (26.8 kJ/m3). In the Fig. 4a, the FORC diagram13 of the nanocomposite shows two magnetic phases: the dominant exchange-coupled phase exhibits switching fields ranging from 50 to 120 kA/m, and a smaller phase appears with ∼5 kA/m switching field, which is presumably uncoupled Fe-Co, consistent with the small kink observed in the hysteresis curve. The δM plot14 of the Fe-Co/CoPt nanocomposite shown in Fig. 4b also confirms the existence of exchange-coupling between the soft and hard magnetic phases. The positive δM values at 20–90 kA/m indicate exchange coupling and match with the dominant phase in the FORC diagram. The maximum energy density of the nanocomposite can be further improved by avoiding the uncoupled soft phase.

Fe-Co/CoPt nanocomposite films are fabricated using an electro-infiltration process, which utilizes magnetic nanoparticles and electroplating to form three-dimensional composite structures. The nanocomposite film exhibits a saturation magnetization of 0.97 T, a remanent magnetization of 0.70 T and a coercivity of 127 kA/m, resulting in a maximum energy density of 21.8 kJ/m3. The exchange coupling between soft magnetic Fe-Co phase and hard magnetic CoPt phase was confirmed in FORC diagram and δM plot. The maximum energy density is expected to improve in a fully coupled nanocomposite. Additionally these results motivate follow-on efforts to more thoroughly study the Fe-Co/CoPt nanocomposite structure (e.g. grain size, grain boundaries, volume fractions, diffusion effects) through further x-ray diffraction (XRD) and transmission electron microscopy (TEM) measurements and analysis.

The authors thank the staff of the UF NRF and MAIC user facilities for assistance in the microfabrication and material characterization. This research was supported in part by the US National Science Foundation (CMMI-1451993).

1.
E. F.
Kneller
and
R.
Hawig
,
IEEE Trans. Magn.
27
(
4
) (
1991
).
2.
R.
Skomski
and
J. M. D.
Coey
,
Phys. Rev. B
48
(
21
) (
1993
).
3.
M.
Ghidini
,
G.
Asti
,
R.
Pellicelli
,
C.
Pernechele
, and
M.
Solzi
,
J. Magn. Magn. Mater.
316
,
159
165
(
2007
).
4.
E. E.
Fullerton
,
J. S.
Jiang
,
M.
Grimsditch
,
C. H.
Sowers
, and
S. D.
Bader
,
Phys. Rev. B
58
(
18
) (
1998
).
5.
G.
Giannopoulos
,
L.
Reichel
,
A.
Markou
,
I.
Panagiotopoulos
,
V.
Psycharis
,
C.
Damm
,
S.
Fähler
,
I.
Khan
,
J.
Hong
, and
D.
Niarchos
,
J. Appl. Phys.
117
,
223909
(
2015
).
6.
Y.
Liu
,
T. A.
George
,
R.
Skomski
, and
D. J.
Sellmyer
,
Appl. Phys. Lett.
99
,
172504
(
2011
).
7.
B.
Balamurugan
,
D. J.
Sellmyer
,
G. C.
Hadjipanayis
, and
R.
Skomski
,
Scr. Mater.
67
,
542
547
(
2012
).
8.
X.
Liu
,
S.
He
,
J.-M.
Qiu
, and
J.-P.
Wang
,
Appl. Phys. Lett.
98
,
222507
(
2011
).
9.
X.
Wen
,
S. J.
Kelly
,
J. S.
Andrew
, and
D. P.
Arnold
,
AIP Adv.
6
,
056111
(
2016
).
10.
Y.
Hayashi
,
S.
Hashi
,
H.
Kura
,
T.
Yanai
,
T.
Ogawa
,
K.
Ishiyama
,
M.
Nakano
, and
H.
Fukunaga
,
Jpn. J. Appl. Phys.
54
,
075201
(
2015
).
11.
F.
Tourinho
,
R.
Franck
, and
R.
Massart
,
J. Mater. Sci.
25
,
3249
3254
(
1990
).
12.
O. D.
Oniku
,
B.
Qiand
, and
D. P.
Arnold
,
J. Magn. Magn. Mater.
416
,
417
428
(
2016
).
13.
C. R.
Pike
,
A. P.
Roberts
, and
K. L.
Verosub
,
J. Appl. Phys.
85
,
6660
(
1999
).
14.
P. E.
Kelly
,
K.
O’Grady
,
P. I.
Mayo
, and
R. W.
Chantrell
,
IEEE Trans. Magn.
25
(
5
) (
1989
).