Hydrothermal Deposition of ZnO Layer on Fe-Based Amorphous Fibres Used for the Preparation of Cold Sintered Fibre-Based Soft Magnetic Composites

Abstract

This paper presents findings on the influence of hydrothermal deposition processing parameters (precursor concentration and deposition duration) on the characteristics of ZnO layers deposited on the surface of amorphous Fe-based fibres. The characteristics of the coated fibres, especially the deposited layer, were investigated from structural, morphological, chemical, thermal and magnetic points of view. It was found that the use of a zinc acetate-based solution of 0.05 M concentration and a deposition duration of 24 h leads to a continuous, thin and adherent layer of ZnO on the surface of the fibres. The thickness of the ZnO layer is 450–500 nm, as shown by TEM-EDX investigations. The ZnO coated fibres were used to prepare fibre-based soft magnetic composites via a cold sintering process. The composites obtained are the first reported soft magnetic composites based on amorphous fibres. The coercive field of the cold sintered composites is 3.5 to 8.6 times lower, and the maximum relative permeability is 2.4 times larger, than the best coercive field and the maximum relative permeability reported until now for a crystalline fibre-based soft magnetic composite. The initial relative permeability of cold sintered composites is constant up to the frequency of 1000 Hz.

1. Introduction

Soft magnetic composites (SMCs) are a highly important class of material, having great potential for applications in the development of more efficient electrical machines. Generally speaking, these materials consist of ferromagnetic particles coated with an insulating layer [1,2,3]. The coated particles are pressed into a mould to obtain the composite core and, if possible, annealed to release the stresses induced during the compaction process. The main advantages of SMCs are: isotropy of magnetic properties, low eddy currents at medium to high frequencies, stability of magnetic properties even at high frequencies and the ease of obtaining complex 3D shapes [4]. The main drawbacks of SMCs are: relatively low permeability and low saturation induction (induced by the presence of the insulating layer), high hysteresis losses (especially in the absence of proper annealing) and relatively low mechanical strength. A key aspect of the preparation of a good SMC is the deposition of the insulating layer on the surface of the ferromagnetic phase [5]. Both the insulation type and the deposition technique have been intensively investigated. As insulator materials, different types of thermosetting polymers such as epoxy resins, polyesters and acrylic resins have been used [1,2]. In addition, inorganic coatings have been intensively used, with the most common of these being Al₂O₃, SiO₂, ZrO₂, MgO, FePO₄ or other soft magnetic ferrites [6,7,8,9,10,11,12]. A wide variety of methods have been implemented to properly cover the ferromagnetic phase with the insulating layers, such as surface oxidation, sol-gel, hydrothermal, mixing, mechanical milling and many others [5]. The hydrothermal method has received increasing attention recently due to its versatility and ability to provide good ceramic coatings. The hydrothermal method can be defined as a synthesis route, beginning from an aqueous solution, that takes place in a closed autoclave at high temperature and high pressure. This method can be used to prepare powders, or can be applied in surface-coating technology. For example, Zheng et al. prepared Fe-based soft magnetic composites hydrothermally coated with a Li–Al–O insulating layer with superior magnetic properties and lower core losses compared to Fe-based soft magnetic composites with an Al₂O₃ insulating layer [13]. Sun et al. demonstrated that Fe particles can be coated with a SiO₂–Al₂O₃ insulating layer via the hydrothermal method, and a content of 25 wt% of SiO₂–Al₂O₃ led to optimum magnetic properties for the SMC with a low Ps of 106.9 W/kg and a highly approximate B_s of 1.28 T [14]. Li et al. prepared FeZrO₂–Fe₃O₄ SMCs characterised by high electrical resistivity and saturation induction with possible applications in high-speed electric motors, transformers, inductance and power supply switching [15]. Neamtu et al. investigated the properties of soft magnetic composites based on Fe fibres coated with SiO₂ by the hydrothermal method [16]. The results indicated that using SiO₂ as the insulating layer leads to compacts having lower hysteresis losses. Nevertheless, they concluded that during the compaction process, the SiO₂ layer is damaged, leading to the development of large eddy currents in the samples at high frequencies.

Cold sintering (CS) is a recent sintering technique, developed for the consolidation of ceramic powders [17]. The ceramic powders are initially mixed with an amount of a solvent that will dissolve the superficial layer of the powder, and then, simultaneously, pressed (at pressures of several hundreds of MPa) and heated at temperatures generally not exceeding 350 °C [18,19,20]. This technique is considered an ultra-low-energy sintering technique, and thus environmentally friendly, due to the low temperatures needed for the powders’ consolidation. A relatively large number of ceramic materials have been consolidated by cold sintering and their relative density was found to be within the range of 85%–99% [18,19,20]. For example, cold sintering of ZnO powders was reported with a temperature range of 120 °C–300 °C, at compaction pressures ranging from 140 MPa up to 500 MPa, using a mixture of water and acetic acid as solvent. ZnO powders were sintered by CS and the relative densities reported in the literature were in the range of 90% to 99% [18,19,20,21,22].

Recently, a new type of SMCs was reported: fibre-based soft magnetic composites (FSMCs) [16,23,24,25,26]. The preparation process of FSMCs is similar to that of SMCs, with the exception that the long ferromagnetic fibres are arranged in a mould in such a way as to achieve their ordered orientation along the direction of the magnetising field. Due to the absence of air-gaps in FSMCs, their magnetic permeability is usually several times higher than the permeability of SMCs [23,26].

In this paper, we investigate the influence of various parameters in the hydrothermal synthesis process applied to the insulating layer deposited onto the surface of Fe-based amorphous fibres. The coated fibres were used to prepare, via cold sintering, fibre-based soft magnetic composites. To the best of our knowledge, this may be the first paper presenting the preparation and characterisation of amorphous fibre-based soft magnetic composites via the cold sintering process (CS-FSMCs).

2. Materials and Methods

The Fe_77.5Si_7.5B₁₅ at.% amorphous fibres are commercially available and were purchased from the National Institute of Research and Development for Technical Physics (NIRDTP), Iaşi, Romania. The amorphous fibres used in this study were prepared using the method of rapid cooling from the melt in a rotating water layer (in-rotating water quenching technique). The master alloy used to produce the fibres was prepared by induction melting in an inert atmosphere using high-purity components (Fe, Si, B). Several grams of the master alloy were remelted by induction in a rotating water-quenched wire equipment crucible (quartz tube). The molten alloy was ejected from the crucible by applying an argon overpressure of 1–8 bar to the rotating water layer. The peripheral speed of the rotating water layer was between 8–12 m/s. Being a commercial product, the exact parameters of the in-rotating water-quenching process are not disclosed.

Before the deposition process via the hydrothermal process, the fibres were cleaned with a NaOH (CAS Number: 1310-73-2) alkaline solution with pH = 12, and then with hydrochloric acid, to eliminate any organic parts from the surface of the fibres. The fibres were subjected to two hydrothermal processes (surface silanisation and ZnO deposition) using analytical reagents and a methanol/ethanol solvent bath. The functionalisation of the fibres with organosilanes is an important strategy to yield better adhesion of the ZnO oxide.

For the first hydrothermal process (surface functionalisation, silanisation) the precursors were tetraethoxysilane TEOS (99.9%, CAS Number: 78-10-4) with ammonium hydroxide (32%) as a catalyst. The polymerisation reaction of silanes (TEOS) starts with the hydrolysis of the silanes, followed by condensation to form a sol, a gel or silsesquioxanes. The molar ratio used was TEOS:EtOH (1:4). After homogenisation of both solutions, 20 mL of the ammonium hydroxide solution (NH₄OH- CAS Number: 1336-21-6) was added dropwise as the activator, followed by magnetic stirring for 1 h. After transferring the precursor solution to a Teflon-lined stainless-steel autoclave, the several fibres were inserted and the clave was sealed before being further heated to 200 °C for 1 h and then cooled to room temperature.

After silanisation, the fibres were subjected to a second hydrothermal process for ZnO deposition using Zn acetate and sodium hydroxide (NaOH). A zinc acetate solution with a concentration of 0.5 M was prepared in 100 mL methanol under magnetic stirring until the solution became clear. Meanwhile, 5 M NaOH solutions were prepared by mixing NaOH in 100 mL methanol under stirring. The solution pH was raised to 12 by adding a solution of NaOH under stirring. This solution mixture was transferred into the Teflon-lined sealed stainless-steel autoclave with the fibres and kept in a laboratory oven at a temperature of 200 °C for 6 h. Then, the autoclave was taken outside and allowed to cool naturally to room temperature. The experimental procedure is repeated for a Zn acetate concentration of 0.05, 0.1 and 0.25 M by keeping the 6 h deposition duration constant. Samples were also prepared by varying the deposition time between 6–24 h and keeping the concentration of the precursor constant at 0.05 M. For a proper cold sintering experiment, the superficial layer of ZnO powder or deposited layer must be initially dissolved in the water and acetic acid solution. During heating, the dissolved ions of ZnO precipitate on the surface of the deposited layer and form the sintering necks. After hydrothermal treatment, the fibres were annealed at 350 °C for 1 h in a high-purity Ar atmosphere to remove any residual organic parts and to form the ZnO coating.

The structural characterisation of the fibres was performed via the X-ray diffraction (XRD) technique using an INEL Equinox 3000 diffractometer (INEL, Artenay, France). The angular range of 2θ = 20–110° was investigated and the radiation used was Co Kα (λ = 1.7903 Å).

The morphology of the coated fibres was analysed via scanning electron microscopy (SEM) using a JEOL JSM 5600 LV microscope (JEOL USA, Inc., Pleasanton, CA, USA), equipped with an energy dispersive X-ray spectrometer (EDX)-type Oxford Instruments UltimMax65 Oxford Instruments (UltimMax65, Bognor Regis, UK). The thickness of the SiO₂ and ZnO coatings were evaluated via transmission electron microscopy (TEM). A HELIOS NanoLab600 (Tokyo, Japan) focussed ion beam (FIB) was used to cut a thick lamella from the prepared sample. A JEOL JEM-2100F (Tokyo, Japan) at 200 kV field emission transmission electron microscope (TEM) with high spatial resolution and analytical performance was used to acquire images of areas of interest as well as for EDX investigations. The mapping was performed in scanning transmission electron microscopy (STEM) mode, which has a lattice resolution of 0.2 nm.

Fourier transform infrared (FTIR) spectroscopy was performed for the coated fibres using a Bruker Tensor 27 FTIR Spectrometer (Karlsruhe, Germany) in ATR (Attenuated Total Reflection) mode. All spectra were normalised for the highest band in the 400–2000 cm⁻¹ range.

The magnetic characteristics of the fibres were investigated with a Lake Shore VSM 7410 vibrating sample magnetometer (VSM) (Westerville, ON, USA) with a maximum applied field of 2700 Oe for the uncoated fibres and 1000 Oe for the coated fibres. The magnetic field was applied in the longitudinal and transversal direction of a single fibre approximately 5 mm in length which was cut from the original coated (or uncoated) fibre.

The thermal stability of the uncoated fibres was investigated via in-situ high-temperature X-ray diffraction (HT-XRD). The HT-XRD patterns were recorded up to 800 °C using an Anton Paar device (INEL, Artenay, France)—HTK 1200 N advanced high-temperature chamber. The heating rate was 10 °C/min and the atmosphere was a preliminary dynamic vacuum (10⁻² Torr). The frequency of the XRD spectra acquisition was one diffraction per minute.

The thermal stability of the coated fibres was studied by differential thermal analysis (DTA) and thermogravimetry (TG) on apparatus using high purity alumina powder as a reference. The measuring range was 20–400 °C, with a heating/cooling rate of 10 °C/min under a high-purity argon atmosphere. The DTA-TG equipment was coupled with a QMS 200 atmospheric sampling quadrupole spectrometer with an ionisation potential of 70 eV in the m/z = 10–90 range (m/z represents the ratio between the mass of the ion (m) to the number of elementary charges (e)).

For the preparation of the cold sintered fibre-based soft magnetic composites (CS-FSMCs), the coated fibres were first immersed in a solution of deionised water and acetic acid with a concentration of 1 M. Additionally, a quantity of 10 wt.% of fibre mass of ZnO powders was added to this solution to bridge the layer of ZnO that covers the fibres during the cold sintering experiments. The ZnO powder used in this study (CAS Number 1314-13-2) has a particle size < 5 μm and a purity of 99.9%. The immersed fibres were introduced in a toroidal mould (by wrapping the fibres on the core rod of the die) and pressed at a compaction pressure of 500 MPa. The mould was heated to 200 °C with a heating speed of 10 °C/min using a heater jacket. The pressure and the temperature were held constant for 60 min. After this interval, the pressure was released and the mould was allowed to cool naturally to room temperature. After cooling the mould, the toroidal samples were extracted from the mould and subjected to magnetic characterisation.

AC magnetic properties of the toroidal compacts (inner diameter of 12 mm, outer diameter of 18 mm and heights of approximately 4 mm) were determined at a maximum flux density (B_max) of 0.1 T in the frequency range of 50–1 kHz, using a computer-controlled Remagraph (Remacomp C—705 hysteresisgraph produced by Magnet Physik Dr. Steingroever GmbH). The same equipment was used for DC magnetic characterisation. The maximum relative permeability (μ_rmax), the coercivity (H_c) and the saturation induction (B_s) were determined at the frequency of 0.5 Hz and a magnetic field excitation of 10 kA/m.

3. Results and Discussions

3.1. Characterisation of the Amorphous Fibres

It is known that after hydrothermal deposition, annealing is needed to remove any traces of the precursors that may remain on the surface of the fibres, and to complete the reaction leading to the formation of ZnO [27]. To properly choose the precursors, and especially the correct annealing temperature, the amorphous fibres were investigated using several analytic techniques. By ‘correct’ annealing temperature, we mean a temperature that will not lead to the crystallisation of the amorphous fibres. To demonstrate the amorphous structure of Fe_77.5Si_7.5B₁₅ at.% fibres, a structural characterisation was performed via X-ray diffraction (XRD). The XRD pattern of the Fe_77.5Si_7.5B₁₅ at.% fibres is presented in Figure 1. The XRD pattern of these fibres consists of a single peak that is both broad/diffuse and low-intensity. Such an XRD pattern results from a lack of long-range atomic order, which is a characteristic of amorphous materials. More than 10 XRD patterns were recorded to validate that the fibres are amorphous along their entire length. All the recorded XRD patterns showed identical characteristics: a single broad/diffuse low-intensity peak.

The SEM images of the amorphous fibres are presented in Figure 1b,c. Figure 1b was acquired to observe the morphology of the surface of the fibres and Figure 1c was recorded to observe the cross-section of the fibre and to measure their diameter. As can be observed in Figure 1b, the surface of the fibre is smooth (having low roughness). This type of surface offers a high-quality insulation layer with a relatively constant thickness. To obtain access to the cross-section of these fibres, the amorphous fibres were cut with scissors. The fibres’ cross-sections present several shear bands, as is expected and generally observed in the case of amorphous alloys subjected to mechanical stress up to the fracture point. The diameter of the fibres was measured in the cross-section images and in the images that present the morphology of the surface of the fibres. As can be observed, the diameter of the fibres is relatively constant, ranging between 105 µm and 107 µm.

The normalised hysteresis loops of the amorphous Fe–Si–B fibres are presented in Figure 2. Two types of measurements were made, one in which the magnetic field is applied parallel to the long axis of fibre (longitudinal) and the other in which the magnetic field is applied perpendicular to the long axis of the fibre (either perpendicular or transversal). It can be seen that there is a significant difference between the shape and the principal parameters (saturation magnetisation and coercive field) of the hysteresis loops recorded in the transversal or longitudinal direction. The different shapes and parameters of the measured hysteresis loops (in the transversal or longitudinal directions) are related to the shape anisotropy of the fibres [24]. It can be seen that it is easier to magnetise the fibres along their long axis (longitudinal) rather than in the transversal direction. The reason for this is due to the demagnetising field, which is stronger when the fibres are magnetised transversally than when the fibres are magnetised longitudinally [28]. The demagnetising field acts in the opposite direction to the magnetic field creating it (the magnetising field). Consequently, the applied field in the transversal direction must be stronger to produce the same magnetic field inside the fibre (the same effect). It can be seen that the magnetic field required to reach saturation magnetisation is approximately 900 Oe in the longitudinal direction, while in the transversal direction, a magnetic field of approximately 2700 Oe is necessary. In addition, by analysing the tilt of the hysteresis loops, it is straightforward to observe that the magnetic permeability of the fibre becomes higher when the fibre is magnetised longitudinally. Similar results were also reported when the influence of the alignment of the fibres on the magnetic properties of fibre-based soft magnetic composites was investigated [24]. As expected, these fibres present strong shape anisotropy, and thus their orientation along the applied magnetic field inside the composite compacts can be beneficial to the magnetic proprieties of the composite compacts.

The thermal stability of the amorphous fibres was investigated by in-situ HT-XRD measurements. The in-situ HT-XRD analysis of the fibres is presented in Figure 3a,b. Figure 3a presents the isometric view of a 3D plot (in-situ HT-XRD) that was constructed by assembling all the diffraction patterns recorded in the temperature range 25–800 °C (the data acquisition was continuous and the acquisition time for each pattern was 60 s). The isometric view provides information about structural changes in the samples, and especially about the evolution of the peak’s intensity versus temperature. Figure 3b presents the top-down view of the in-situ HT-XRD analysis. Although this representation offers little information about the peak’s intensity, the crystallisation kinetics of the amorphous fibres can be more accurately investigated using this representation (in which the determination of the crystallisation temperatures of different phases can be more accurately observed).

After examining the data resulting from the in-situ HT-XRD analysis, the following conclusions were drawn:

• The amorphous structure of the FeSiB fibres is stable up to the temperature of 500 °C. This is visible in the in-situ HT-XRD graph as a broad peak, from approximately 45° to 60° in 2 theta, which also has low intensity.
• The crystallisation of the amorphous fibres begins at 510 °C. As a result of fibres’ crystallisation, two new phases are formed almost simultaneously. The newly formed phases are (i) α-Fe(Si), a solid solution with a cubic crystal structure belonging to the space group Im-3m, space group number 229; and (ii) Fe₂B, which is an iron boride compound with a tetragonal crystal structure belonging to the space group I4/mcm, space group number 140. The three principal peaks (the most intense) of α-Fe(Si) are situated at 52.9°, 78.1° and 100.7° and correspond, according to the JCPDS file no. 03-065-6323, to the following families of crystallographic planes: (110), (200) and (211). The peaks attributable to the Fe₂B phase are situated at 41.3°, 50.3°, 53°, 58.9°, 66.8°, 67.9°, 88.8°, 96.6° and 98.2°. These correspond, according to the JCPDS file no. 00-036-1332, to the following families of crystallographic planes: (200), (002), (211), (112), (202), (310), (312), (213) and (411). As the temperature rises, peak intensities increase (Figure 3a), indicating the growth of the crystallinity of the phases. As a concluding remark, we can determine that the crystallisation of the fibres takes place at around 510 °C and the type of crystallisation is quasi-eutectic, consisting of the almost simultaneous formation (at a difference of only 10 °C) of the α-Fe(Si) and Fe₂B phases. The first phase resulting from the crystallisation of the amorphous phase is α-Fe(Si). Also, the crystallisation of Fe₂B at almost the same temperature as α-Fe(Si) excludes the application of any annealing to the fibres at temperatures above 510 °C, since the presence of Fe borides in any soft magnetic material induces a dramatic decrease in its soft magnetic properties (especially an increase in the coercive field) [29].

3.2. Characterisation of the ZnO-Coated Amorphous Fibres

This investigation aimed to determine the annealing temperature needed for the decomposition of the zinc acetate into zinc oxide. Additionally, the absence of any reaction between the deposited layer of ZnO and the fibres during annealing was investigated. After hydrothermal deposition, the surface of the fibres is coated with zinc acetate and not with ZnO. The thermal stability of the coated fibres was investigated via differential thermal analysis (DTA) coupled with thermogravimetry (TG) and mass spectrometry (MS); the results are presented in Figure 4a,b. The fibres subjected to DTA-TG-MS analysis were coated via the hydrothermal method for 24 h using a precursor with a concentration of 0.05 M. The DTA curve (Figure 4a) presents two endothermic reactions at temperatures of 220 and 350 °C, respectively. These temperatures were interpreted as the melting point of zinc acetate (C₁₂H₁₈Zn₄O₁₃) and the temperature at which decomposition of zinc acetate to ZnO begins according to the following equation [27]:

$${\mathrm{C}}_{12}{\mathrm{H}}_{18}{\mathrm{Zn}}_4{\mathrm{O}}_{13}\stackrel{\mathrm{heat}}{\to }4\mathrm{ZnO}+3{\mathrm{CH}}_3{\mathrm{COCH}}_3\uparrow +3{\mathrm{CO}}_2\uparrow$$

(1)

According to the TG curve (Figure 4a), the sample begins to constantly lose mass in the temperature range of 100–250 °C. In the same temperature range on the MS analysis (Figure 4b), an increase in the signal from the m/z = 18 corresponding to water can be observed. In the temperature range 260–350 °C, an important mass loss is noticeable on the TG curve with an increase in the CO (m/z = 12) and CO₂ (m/z = 44) and acetone (m/z = 58) fragments. The intensity of the m/z = 44 fragment reveals that a large quantity of CO₂ is released in this temperature range due to the burn-off of organic moieties. Based on the corroboration of data obtained from DTA-TG-MS analysis, we concluded that the decomposition of the Zn acetate begins at about 260 °C and ends at about 350 °C. Similar results have been reported in the literature [27,30].

To validate that the result of the thermal decomposition of the zinc acetate on the surface of the amorphous fibres is indeed zinc oxide, XRD investigations were performed. The XRD pattern of the fibres coated via the hydrothermal method for 24 h using a precursor with a concentration of 0.05 M and annealed at 350 °C/1 h is presented in Figure 5. The XRD pattern contains the signature of two phases very different from the crystallographic point of view: an amorphous phase (the signal given by the amorphous fibres) and a crystalline phase (the ceramic coating). Following the phase-identification process, using the PDF2 database, it was concluded that all sharp XRD peaks observed correspond to the ZnO phase according to the JCPDS file no. 01-075-0576. No other phases, such as zinc acetate, silicon dioxide or iron oxides were identified, proving that—within the experimental limits of the XRD technique—the annealing temperature and duration were correctly chosen, and no reaction occurs between the fibres and the deposited layer during annealing.

Another validation of the ZnO layer formation via the hydrothermal method and annealing was given by the FTIR analysis presented in Figure 6a,b. The comparison between samples prepared with precursors of different molar concentrations (Figure 6a) reveals that increasing the precursor concentration leads to a larger amount of ZnO formed on the surface of the fibres (see the integral intensity of the ZnO absorbance band). Additionally, when comparing the samples prepared with the 0.05 M concentration precursor but with different deposition duration (Figure 6b), it was observed that increasing the deposition duration increased the amount of ZnO deposited on the surface of the fibre.

In the FTIR spectra of all samples, alongside the Zn–O absorption band (around 500 cm⁻¹), another absorption band was observed around 1000 cm⁻¹ which was attributed to the Si–O–Si absorption band. The siloxane (Si–O–Si) was created on the surface of the fibres via silanisation reaction prior to the deposition of the ZnO, to create a bridge between the metallic fibre and the coating. Instead, a band at 937 cm⁻¹ could be observed which can be assigned to the vibration silanol (Si–OH), this being characteristic of the silica network as well [31,32].

The main conclusions of this part of our study were that the fibres coated via the hydrothermal method need to be annealed at 350 °C to thermally decompose the zinc acetate to zinc oxide, and that no reactions occur between the fibre’s surface and the deposited layers (silicon dioxide and zinc acetate).

3.3. Electron Microscopy Investigations (SEM, TEM, EDX) of the Deposited Layer

All the SEM images and EDX investigations presented in this section correspond to the fibres that were hydrothermally coated and annealed at 350 °C for 1 h. According to the literature, changing the precursor concentration and the deposition duration are the most effective parameters to control the synthesis process [33,34]. In our case, the concentration of the precursor solution was modified as follows: 0.5 M, 0.25 M, 0.1 M and 0.05 M. The deposition duration was 6 h and 24 h. For all the samples, we present a set of two SEM images and the corresponding EDX analysis. The first set presents the SEM-EDX analysis at low magnification to obtain an overview of the surface of the fibre. The second set consists of SEM-EDX analysis performed at higher magnifications to determine, in more detail, the features of the deposited layer.

Typical SEM-EDX analysis of the amorphous fibres coated via the hydrothermal technique using a precursor solution of 0.5 M concentration and a deposition duration of 6 h are presented in Figure 7. The first set of SEM-EDX analyses (Figure 7a–d) revealed that some fibres are relatively well-coated (the right-side fibre) with ZnO while others are completely uncoated (the left-side fibre). SEM-EDX analysis performed at higher magnification (Figure 7e–h) revealed that large surfaces of the fibres are only partially coated. We assume that, due to the high concentration of the precursor solution, the ceramic coating formed is too thick, with large crystals, and its adherence to the fibre is unsatisfactory since it breaks and falls off under its own weight. Few such crystals are visible in Figure 7e.

The next step consisted of decreasing the precursor concentration from 0.5 M to 0.25 M. SEM-EDX analysis of the amorphous fibres coated via the hydrothermal technique using a precursor solution of 0.25 M concentration, a deposition duration of 6 h and annealed at 350 °C for 1 h is presented in Figure 8. It can be noticed that a certain improvement in the coating quality (the degree of coverage) was obtained. A larger surface of the fibres is coated with the ZnO ceramic layer. However, as can be seen in Figure 8e–h, the ceramic layer is too thick and brittle; any manipulation of the coated fibres can induce cracks and exfoliation of the ceramic layer. The exfoliation of the ceramic layer as a result of the simple placement of the fibres on the carbon tape of the SEM sample holder can be easily observed in Figure 8a–d.

To further reduce the thickness of the deposited layer of ZnO, the concentration of the precursor solution was further decreased to 0.1 M with the deposition duration maintained at 6 h. The SEM-EDX analysis of the coated and annealed fibres, using the above-mentioned parameters for the coating process, is presented in Figure 9.

Analysing the distribution maps corresponding to Zn and O (Figure 9c,d), we noticed that the degree of coverage of the fibres was substantially improved, with almost the entire surface of the fibres now being covered by the deposited ceramic layer. Also, the exfoliation phenomenon was significantly reduced. We interpret this result as being caused by the thickness of the layer being greatly reduced, itself the result of using a precursor with a lower concentration (0.1 M). Analysis performed at higher magnification (Figure 9e–h) revealed the existence of small areas where the fibres are uncoated, or the layer of ZnO was exfoliated. Such areas may lead to electrical contact between neighbouring fibres during the compaction process, leading to large magnetic losses via eddy currents. Also, the deposited layer appears to be very porous as can be observed in the SEM image presented in Figure 9e. We recall the fact that we aim to obtain a thin, continuous, adherent and compact layer of ZnO on the surface of the amorphous fibres.

Once again, the concentration of the precursor solution was reduced in an attempt to prepare the ZnO layer with the above-mentioned characteristics. The new concentration of the precursor used was 0.05 M while the deposition duration was maintained at 6 h. The SEM-EDX investigations of the coated fibres are presented in Figure 10. It can be observed that only a partial coating of the fibres was obtained when a precursor of 0.05 M concentration and a deposition duration of 6 h were used. At higher magnifications (Figure 10e–h), we can notice that the surface of the fibres is coated with small islands of ZnO that nucleated but did not have enough time for growth and coalescence to form a continuous layer.

To allow sufficient time for the nucleated islands of ZnO to grow and coalesce to form a continuous layer, the deposition duration was increased to 24 h while the concentration of the precursor was maintained at 0.05 M. The SEM-EDX analysis of the coated fibres is presented in Figure 11. A reasonably good degree of coverage of the fibre surface was obtained in this way, indicating that coalescence of the ZnO islands occurred. These results were validated at higher magnification (Figure 11e,f). In addition, the deposited layer did not appear to be porous, as occurred when the concentration of the precursor was 0.1 M and the deposition duration was 6 h. An important feature of the deposited layer is its adherence to the surface of the fibre. While in the case of the other deposition experiments mentioned previously, in which the exfoliating of the deposited layer was noticed, no exfoliating was noticed in this specific experimental setup (precursor concentration of 0.05 M and deposition duration of 24 h).

To quantify the thickness of the ZnO and SiO₂ layers, TEM and EDX analyses were performed for the fibres coated using a precursor concentration of 0.05 M concentration and a deposition duration of 24 h. The results of these TEM-EDX investigations are presented in Figure 12. In Figure 12a we present the TEM image of a cross-section through the surface of the Fe–Si–B fibre coated with SiO₂ and ZnO.

The cross-section was achieved by FIB milling and depositing a layer of Pt onto the surface of the coated fibre to protect the ZnO and SiO₂ layers from getting damaged during FIB milling. Figure 12b–f presents the elemental distribution maps of the Fe, Zn, O, Si and Pt in the cross-section. According to TEM images and the EDX analysis, the thickness of the SiO₂ layer is 150–200 nm. The ZnO layer has a thickness of 450–500 nm as can be seen in Figure 12. Analysing the TEM image, we observe that indeed both deposited layers have a dense structure with no porosity. Bearing in mind that the deposition process takes place in a sealed autoclave and the volume of the autoclave is relatively small—250 mL—the deposition conditions are uniform in the entire volume of the autoclave leading to a uniform thickness of the deposited layer.

At the end of this section of our study, we concluded that the most promising candidates for the preparation of the cold sintered fibre-based soft magnetic composites are those fibres coated under the following conditions: (i) precursor concentration of 0.1 M and deposition duration of 6 h; and (ii) precursor concentration of 0.05 M and deposition duration of 24 h.

3.4. Magnetic Properties of the Coated Fibres

The magnetic properties of the coated fibres were determined via VSM measurements by applying the magnetic field parallel to the fibres’ long axis. As can be seen in Figure 13, the saturation magnetisation (M_s) of these coated fibres decreases as the concentration of the precursor increases. Also, the increase in the deposition duration leads to a decrease in the saturation magnetisation (M_s) of the coated fibres. As validated by SEM-EDX and FTIR analysis, an increase in the precursor concentration or the deposition duration leads to an increase in the ZnO layer thickness. As a result, the sample subjected to VSM measurement contains increasingly more nonmagnetic atoms (Zn and O) and increasingly fewer magnetic atoms (Fe) leading to a decrease in mass magnetisation of the samples [35].

It is worth noticing that the sample prepared with the precursor of the highest concentration—0.5 M—does not follow the same behaviour, and its saturation magnetisation is superior to those samples prepared with the precursors with concentrations of 0.1 M and 0.25 M. This can be explained by the fact that even if its ceramic layer is the thickest, the adherence of the layer to the amorphous fibre is the weakest, and during the sample preparation for VSM measurement (cutting a sample of approximately 5 mm length), a part of ZnO coating detached from the surface of the fibres, leading to a larger mass magnetisation for this sample.

3.5. DC and AC Characteristics of the Cold Sintered Fibre-Based Soft Magnetic Composites

For the preparation of the cold sintered fibre-based soft magnetic composites (CS-FSMCs), the fibres coated via the hydrothermal method using 0.05 M (hereinafter referred to as the 0.05 M sample) and 0.1 M (hereinafter referred to as the 0.1 M sample) precursor solution concentrations were selected. We recall that the deposition duration was 24 h when a 0.05 M solution was used and 6 h when a 0.1 M solution was used. The density of the cold sintered samples was 4.5 g/cm³ for the 0.05 M sample and 4.1 g/cm³ for the 0.1 M sample. The low density of the samples can be explained by the relatively low pressure that was applied during the cold sintering process. Usually, the compaction pressure applied during the cold sintering processes does not exceed 500 MPa because there is a risk that the liquid (acetic acid with concentration of 1 M) and small particles of ZnO will leak out of the mould. In addition, it is well known that if the amorphous materials are subjected to a compression test, the stress–strain curve consists mainly of elasticity followed by a small amount of plastic strain (up to 1%) [36]. As is known, the saturation magnetisation of a material is defined as the sum of all individual atomic magnetic moments per unit volume [37]. As a result of the elastic behaviour of the amorphous materials, the density of the compacts is low and, as a consequence, the saturation induction of the samples is low. The saturation induction of the 0.1 M sample was 0.41 T and the saturation induction corresponding to the 0.05 M sample was 0.45 T. However, several positive aspects resulted from the evaluation of the DC hysteresis loops of the compacts. The values for coercivity (H_c) and maximum relative permeability (µ_rmax) are outstanding. In

Table 1. Comparison between the DC magnetic properties reported up to now for FSMCs prepared via classic routes and cold sintering.

Type of FSMCs	DC Magnetic Properties			Ref.
	Bs	Hc	µrmax
	(T)	(A/m)	-
Fe fibres with polymer (1 wt% Araldite)	1.52	353	893	[23]
Fe fibres coated with SiO2 (200 nm)	1.39	367	797	[16]
Fe fibres coated with SiO2 (200 nm) + 1 wt% Araldite	1.26	353	733	[16]
Short Fe fibres + polymer (1 wt% Araldite)	1.28	421	506	[24]
Fe fibres coated with Fe3O4 (0–514 nm)	1.49	366	886	[25]
Fe fibres coated with Fe3O4 (210–920 nm)	1.43	389	798	[25]
Fe fibres coated with Fe3O4 (309 nm–1.70 µm)	1.35	398	719	[25]
Fe77.5Si7.5B15 coated with ZnO (0.05 M sample)	0.45	40.9	2180	This work
Fe77.5Si7.5B15 coated with ZnO (0.1 M sample)	0.41	101	820	This work

is presented a comparison between the DC magnetic properties reported up to now for all FSMCs prepared via classic routes.

As can be noticed from the data presented in Table 1, the lowest coercive field reported until now was 353 A/m while the coercive field corresponding to the samples presented in this study is 3.5 to 8.6 times lower. In addition, the maximum relative permeability of the 0.05 M sample is 2.4 times larger than the largest maximum relative permeability reported up to now for an FSMC. To explain the values of maximum relative permeability reported in this study, two factors must be considered: (i) The magnetic properties of amorphous alloys (especially coercive field and permeability) are generally several times better when compared to their polycrystalline counterparts [34]. This explains the difference between the maximum relative permeability of the 0.05 M sample and other fibre-based soft magnetic composites. (ii) It was previously reported that by increasing the thickness of the insulating layer, it became increasingly difficult to move the magnetic domain walls due to the successively enhanced pinning effect [26,35]. Considering that permeability is a property that indicates how much magnetic induction is generated by the material in a given magnetic field, and the fact that the thickness of the ZnO layer is larger in the case of 0.1 M sample as compared to 0.05 M sample, this can explain the large value of maximum relative permeability for the 0.05 M sample. The maximum relative permeability of the 0.1 M sample is 8% lower than the largest maximum relative permeability reported up to now for an FSMC. The low values of the coercive field and high values of maximum relative permeability for the CS-FSMCs as compared to classically prepared FSMCs can also be explained based on the difference between the deformation behaviour of amorphous fibres versus Fe fibres. When the amorphous fibres are subjected to compression, during the compaction process, they undergo elastic deformation, whereas polycrystalline Fe fibres undergo elastic and plastic deformation [34]. The plastic deformation induces stresses and crystallographic defects into the Fe fibres that act as pinning centres for the domain walls’ movement. It is known that the presence of stressed regions and crystallographic defects in a soft magnetic material hinder the displacement of the domain walls leading to an increase in its coercive field and a decrease in the magnetic permeability [23].

The evolution of the initial relative permeability and the total core losses versus frequency for the CS-FSMCs is presented in Figure 14 By way of comparison, the curves determined on a toroidal core (having the same size as the CS-FSMCs) based on 0.5 mm thick laminates of non-oriented electrical steel with 3% Si are presented. This type of electrical steel was selected since is one of the most commonly used grades in the construction of electric motors.

It can be noticed that the initial relative permeability (µ_ri) of the 0.05 M sample is 2.8 times higher than the initial relative permeability of the 0.1 M sample (Figure 14a). The reason for this could be the fact that the insulation layer is thinner for the 0.05 M sample and the density of this sample is higher. The initial relative permeability of both compacts is constant up to the frequency of 1000 Hz. If we analyse the initial relative permeability of the magnetic core based on Fe–Si laminates, we notice that the permeability is constant only up to 100 Hz. Above this frequency, a steeper decrease is observed. The decrease in the permeability of a magnetic core upon increasing the excitation frequency is generally associated with the excessive development of eddy currents in the samples. The frequency at which the permeability of a magnetic core starts to decrease is known as the cut-off frequency; beyond this frequency, it is no longer recommended to use the magnetic core. As a concluding remark, the magnetic core based on Fe–Si laminates can be used up to 100 Hz while the CS-FSMCs can be used up to 1000 Hz.

The evolution of the total core losses versus frequency is presented in Figure 14b. It can be noticed that, regardless of the frequency, the lowest total losses correspond to the magnetic core based on electrical steel. The analysis of total losses of the CS-FSMC samples revealed that the lowest losses correspond to the 0.05 M sample. This can be explained by the characteristics of the insulating layer that cover the fibres. The ZnO layer of the 0.05 M sample is thinner and more continuous than the ZnO layer of the 0.1 M sample. It was previously reported that by increasing the thickness of the insulating layer, the displacement of the domain walls becomes increasingly difficult [25,35]. This leads to an increase in the hysteresis losses of the compact. The fact that the insulating layer of the 0.05 M sample is more continuous will limit the development of eddy currents. Considering that the total losses of a magnetic core are mainly composed of hysteresis losses and eddy current losses, it is understandable why the 0.05 M sample has the lowest losses.

4. Conclusions

This paper investigates the influence of the processing parameters of hydrothermal deposition (precursor concentration and deposition duration) on the characteristics of the ZnO layer deposited on the surface of amorphous Fe-based fibres. Analysis of Fe-based fibres revealed that the fibres have an amorphous structure as demonstrated by XRD analysis, with the diameter of the fibres measured by SEM analysis ranging between 105 µm and 107 µm. Our in-situ HT-XRD analysis proved that the crystallisation of the amorphous fibres began at a temperature of 510 °C and led to the formation of α-Fe(Si) and Fe₂B phases. The amorphous fibres were subjected to a double hydrothermal deposition process (surface silanisation and ZnO deposition). The investigation of the coated fibres highlighted the following findings:

(i)

The decomposition of the Zn acetate starts at about 260 °C and ends at about 350 °C as proved by DTA-TG-MS analysis. As the result of the thermally decomposed precursor solution, the fibres are covered with a layer of ZnO as proved by XRD investigation.

(ii)

Hydrothermal deposition for 6 h using a precursor of 0.5 M and 0.25 M leads to a thick and brittle ZnO layer with low adherence to the fibres. The most promising coatings were obtained when the concentration of the precursor solution was reduced to 0.1 M (deposition duration of 6 h) and 0.05 M (deposition duration of 24 h).

(iii)

According to the TEM image and the EDX analysis, the thickness of the SiO₂ layer is 150–200 nm and the ZnO layer has a thickness of 450–500 nm for the 0.05 M sample.

(iv)

The saturation magnetisation of the coated fibres decreases as the concentration of the precursor and the deposition duration increase.

Cold sintered fibre-based soft magnetic composites (CS-FSMCs) were successfully prepared using hydrothermally coated fibres. Although the saturation induction of the CS-FSMCs is low (0.41 T and 0.45 T), their coercive field is 3.5 to 8.6 times lower and the maximum relative permeability is 2.4 times larger in the case of the 0.05 M sample than the best coercive field and the maximum relative permeability reported up to now for an FSMC. The initial relative permeability of both CS-FSMCs is constant up to the frequency of 1000 Hz. Because the insulating layer of the 0.05 M sample is thinner and more continuous than the layer of the 0.1 M sample, the lowest total core losses correspond to the 0.05 M sample.

In our opinion, the use of hydrothermal deposition can open new perspectives concerning the deposition of insulation layers that need to cover the ferromagnetic phase (fibres or particles) to limit the development of eddy currents. Moreover, this cold sintering process has good potential and may lead to the development of new types of soft magnetic composites and an increased number of applications.