Inﬂuence of Hydrogen Reduction Stage Conditions on the Microwave Properties of Fine Iron Powders Obtained via a Spray-Pyrolysis Technique

: The relationship between the chemical purity of one-size particles and microwave properties in ferromagnetic materials is not clearly studied. Ferromagnetic nanostructured iron powders were synthesized from iron nitrate solution using ultrasonic spray-pyrolysis and then reduced in H 2 ﬂow at 350, 400, 450, and 500 ◦ C. A rise in the concentration of solutions of a precursor from 10 to 20 wt. % led to an increase in mean particle size. The interrelationship was studied between chemical composition and the microwave dispersion of the powders obtained. An increase in the temperature of reduction changes the chemical composition and increases the amplitude of complex microwave permeability, which was studied using solid-state physics methods (XRD, STA, SEM, and VNA). It was found that annealing at 400 ◦ C is the optimal treatment that allows the production of iron powders, consisting of about 90% of α -Fe phase, possessing a particle surface with low roughness and porosity, and demonstrating intense microwave absorption. Annealing at a higher temperature (500 ◦ C) causes an even higher increase in permeability but leads to the destruction of nanostructured spheres into smaller particles due to grain growth. This destruction causes an abrupt increase in permittivity and therefore signiﬁcantly reduces potential applications of the product. The insight into chemical–magnetic relationships of these materials enhances the data for design applications in magnetic ﬁeld sensing.


Introduction
The features of the complex microwave permeability µ + i·µ" and permittivity ε + i·ε" of composites filled with magnetic powders depend on structure-property relationships between structure elements and on the chemical state of the components. In the first case, the filler particle's shape (spheres, discs, flakes, etc.), the matrix-inclusion interface, agglomeration, and dispersion of inclusions in the matrix govern these properties. These phenomena underlie a classical understanding of the propagation of electromagnetic waves in a medium [1,2]. According to effective medium theory (e.g., the Maxwell-Garnett equation [3]), an increase in the inclusion content in the composites leads to the growth of the absorption properties. This was demonstrated elsewhere, e.g., by using carbonyl iron powder [4] and ferrite [5]. In the second case-the chemical composition and the presence of impurities in materials-the crystalline and magnetic structure of inclusions influences both static and dynamic magnetic properties [2,6].
A few studies have reported a controlled change in the microwave properties of materials, using the principle of changing the chemical composition. For example, in soft nickel ferrites, the non-magnetic α-Fe 2 O 3 phase simultaneously reduces saturation magnetization and improves frequency stability (decreases frequency dispersion) permeability [7]. from prepared precursor aqueous solutions with concentrations 10 (0.5 M) and 20 wt. % (1 M) with a diameter from 0.5 to 5 µm. Aerosol particles passed through the 1000 • C heated zone of tube furnace LF-50/500-1200 (LOIP, Russia) in a quartz tube reactor (φ 45 mm, 100 cm in length). The flow rate of the carrier gas (air, 5 l·min -1 ) was achieved by an oil vacuum pump through the reactor, including a cancellated filter (metal mesh, 10 µm aperture) and a system of three gas bubblers filled with water, which were used to decrease the flow and collect waste that overpasses the filter. Therefore, starting from iron (III) nitrate precursor solution (hereafter identified as 10 and 20 sol. %) two powders of iron oxide Fe 2 O 3 were collected on the filter. The product yield was approximately 1 g per hour.

Iron Powder Preparation by the Hematite Reduction in Hydrogen
The iron powders were obtained from the hematite powder. The hematite red powders (10, 20 sol. %) spread into a quartz crucible were annealed at 350, 400, 450, and 500 • C for 75 min in hydrogen flow. This gave eight black iron powders in total. Hydrogen was synthesized using hydrogen gas generator GV-12 (Meta-Chrom, Russia) with a productiveness of 12 l·min -1 . Before starting heating and after annealing, the reactor was blown off by high-purity inert argon gas for cleaning from atmospheric air and passivation of the surface of the powder during cooling to room temperature forming a protective oxide film [21].

Characterization
The structure and chemical phase composition of synthesized iron powders was studied by dint of a Difray 401 diffractometer (Scientific Instruments, Russia) at room temperature. The X-ray powder diffraction method (XRD) was applied for the measurements from 14 • to 140 • 2θ degrees using Bragg-Brentano geometry and chromium radiation (2.2901 Å). XRD results were compared with the powder diffraction file (PDF) database. The quantitative phase analysis and the average size of the crystallites were evaluated by the reference intensity ratio (RIR) method and the Scherrer formula, respectively.
Scanning electron microscopy (SEM) was performed using a LEO EVO-50 XVP (Carl Zeiss, Jena, Germany) microscope. The particle size distribution of the eight studied samples was examined manually from the SEM images, measuring approximately 500 individual particles for each powder.
The simultaneous thermal analysis (STA) instrument Netzsch STA 449 F3 Jupiter (Netzsch, Germany) was used to carry out the oxidation experiments [22]. The weight of iron powders taken for analysis was controlled at 10-20 mg to decrease the influence of the temperature gradient in the sample. Lower weightings give lower-quality STA curves, and higher weightings show problems with nonuniform heating. Simultaneous thermogravimetry (TG) and differential scanning calorimetry (DSC) measurements were performed in a corundum crucible under atmospheric pressure conditions in the temperature range from 30 to 1000 • C with a heating rate of 10 • C·min -1 . Dry air was used as an oxidant.
The composite samples for the microwave measurements were prepared by mixing molten paraffin wax matrix with prepared iron powders [23,24]. S-parameters of toroidal composite samples with filler fraction 75 wt. % in all cases were measured with a standard 7/3 mm coaxial line using a vector network analyzer (VNA) N5224B (Keysight, Santa Rosa, CA, USA). The measurement ports were calibrated using the TRL calibration method. The frequency dependencies of complex microwave permeability and permittivity were determined by the Nicolson-Ross-Weir [3,25] technique in the frequency range of 0.1 to 30 GHz.
The measured permeability was fitted by two Lorentzian curves [26], from which the amplitudes and resonance frequencies were found. To calculate Acher's parameter, a mass fraction of 75 % was recalculated into 25 vol. % (using the tabular density of pure iron, 7.88 g/cm 3 ), which was the effective volume fraction of the metal, but not the volume fraction of inclusions. The fitted frequency dispersion of microwave permeability was applied for the calculation of the Acher's parameter, i.e., a quantitative magnetic characteristic of the composition of the material [27,28], following the formula given in [20]. Sharp peaks observed in the high-frequency region of the complex microwave permittivity are attributed to higher-order modes and not related to intrinsic material parameters of the composites [29].

X-ray Diffraction Analysis
X-ray diffraction analysis of the iron powders indicated the presence of high-intensity peaks of body-centered cubic (bcc) α-Fe and hexagonal α-Fe 2 O 3 . Figure 1 shows XRD peaks of Fe 10 sol. % and 20 sol. % samples reduced from iron oxide to iron in hydrogen at 350, 400, 450, and 500 • C. According to the PDF-60696 card, the tabular value of the lattice constant of the pure iron phase is 2.866 Å and the position of peaks are 68.78 • (110) and 106.03 • (200) 2θ. Characteristics of the iron phase coincide well with described tabular values (vide Table 1). For the appraisal of the lattice parameters of the hexagonal crystal system, a minimum of two peaks is required. Due to the lower intensity of peaks, located at 54.04 • and 101.35 • 2θ, the lattice parameters of α-Fe 2 O 3 were estimated only for a few samples, close in value to a = 5.04 Å and c = 13.77 Å, which corresponds to the tabular value [30].
Magnetism 2023, 2, FOR PEER REVIEW 4 CA, USA). The measurement ports were calibrated using the TRL calibration method. The frequency dependencies of complex microwave permeability and permittivity were determined by the Nicolson-Ross-Weir [3,25] technique in the frequency range of 0.1 to 30 GHz. The measured permeability was fitted by two Lorentzian curves [26], from which the amplitudes and resonance frequencies were found. To calculate Acher's parameter, a mass fraction of 75 % was recalculated into 25 vol. % (using the tabular density of pure iron, 7.88 g/cm 3 ), which was the effective volume fraction of the metal, but not the volume fraction of inclusions. The fitted frequency dispersion of microwave permeability was applied for the calculation of the Acher's parameter, i.e., a quantitative magnetic characteristic of the composition of the material [27,28], following the formula given in [20]. Sharp peaks observed in the high-frequency region of the complex microwave permittivity are attributed to higher-order modes and not related to intrinsic material parameters of the composites [29].

X-ray Diffraction Analysis
X-ray diffraction analysis of the iron powders indicated the presence of highintensity peaks of body-centered cubic (bcc) α-Fe and hexagonal α-Fe2O3. Figure 1 shows XRD peaks of Fe 10 sol. % and 20 sol. % samples reduced from iron oxide to iron in hydrogen at 350, 400, 450, and 500 °C. According to the PDF-60696 card, the tabular value of the lattice constant of the pure iron phase is 2.866 Å and the position of peaks are 68.78° (110) and 106.03° (200) 2θ. Characteristics of the iron phase coincide well with described tabular values (vide Table 1). For the appraisal of the lattice parameters of the hexagonal crystal system, a minimum of two peaks is required. Due to the lower intensity of peaks, located at 54.04° and 101.35° 2θ, the lattice parameters of α-Fe2O3 were estimated only for a few samples, close in value to a = 5.04 Å and c = 13.77 Å, which corresponds to the tabular value [30].  Table 1. XRD data and the size of iron powders produced with different temperatures of reduction T red , where a-lattice constant, CS-crystallite size, ω-an evaluated mass fraction of α-Fe and α-Fe 2 O 3 in each sample, d av -average diameter of powder particles. The X-ray diffraction technique provides average structure information (degree of crystallinity, phase identification, lattice parameter) [31], but no account is taken of the particle's morphology, i.e., grains of particles or porosity. The crystallite sizes were estimated from the full width at half maximum (FWHM) of the diffraction profile. For Fe 10 sol. % and 20 sol. %, the sample crystallite size of the α-Fe phase increased from 27 to 49 nm with the increase in reduction temperature.
Semi-quantitative phase analysis was estimated using the reference intensity method. The method is based on scale factor differences between reference and measured peaks. The reference standard of corundum values for phases α-Fe (I/I c = 10.77) and α-Fe 2 O 3 (I/I c = 4) were used. The reduction degree increased from 85 to 95 wt. % when the temperature of reduction increased from 350 to 500 • C.
The estimation of containing phase ω and crystallite size is provided in Table 1. Crystallite size and purity of iron powders increased proportionally to the reduction temperature for both samples (Fe 10 and 20 sol. %). Hydrogen content in the gas mixture, the rate and temperature of reduction, and a smaller size of powder, i.e., a larger specific area, influenced the reduction degree [32]. The Fe/Fe 2 O 3 ratio was more efficiently reduced at the high-temperature reduction (500 • C), than at 350 • C.

Scanning Electron Microscopy
The SEM microphotographs showed that the increase in temperature of reduction did not change the overall particle size but resulted in the increase of porosity (see Figure 2b-d,f-h). The iron powders demonstrated a spherical nanostructured shape, consisting of agglomerated grains of iron, which is typical for particles obtained using the spray-pyrolysis technique [14]. The reduction at 350 °C provided iron particles with a slick, low-roughness surface (Figure 2a,e). At higher temperatures, pores became coarser and larger. Distinct grain structure and porosity were observed at 400-500 °C temperature of reduction. Between these samples (400-500 °C), minimal grain size and porosity as well as crystallite size (37 nm) and a high mass fraction (~90%) were demonstrated in both samples after reducing at 400 °C (Figure 2b,f). This notable characteristic indicates that reducing at 400 °C provided powder with good properties, including morphology and chemical composition.
The concentration of precursor solution influenced the mean size of the particle size distribution of both iron and hematite powders. The average size of Fe2O3 was estimated at 0.5 and 0.7 μm, respectively (for 10 and 20 sol. %, see Figure A1). Small growth of mean particle size proportionally to the temperature of reduction was observed ( Figure 3 and Table 1). Reduction processes insignificantly increased the mean size of the particles at temperatures below 450 °C. The mass fraction of α-Fe increased when the reduction temperature increased, but high temperature led to the destruction of nanostructured spherical particles to smaller particles of which the former was composed. For iron powder (10 sol. %), spheres broke down at 500 °C (Figure 2d). For another sample (20 sol. %), similar destruction presumably can be obtained at higher temperatures than 500 °C. Enlargement of the agglomerated grains as well as pore size with the temperature of reduction were observed. The reduction process caused grain sintering, enlargement of pores between grains [33], and rise of crystallite size (see Table 1). Pore formation can probably be explained as follows. Reduction of iron (III) oxide to iron occurs with losses in mass and volume of the sample. Straight calculation following the 2Fe 2 O 3 = 4Fe + 3O 2 reaction gives a mass loss of 42.97%. According to [34], the loss in volume during the process is~30%. A simultaneous increase in crystallite and grain size of iron with an increase in temperature results in the formation of the observed "mesh" structure. Smooth particles can be obtained with the direct addition of hydrogen during synthesis [15]. A similar structure has previously been reported for Co [35] but has never been seen before in iron powders.
The reduction at 350 • C provided iron particles with a slick, low-roughness surface (Figure 2a,e). At higher temperatures, pores became coarser and larger. Distinct grain structure and porosity were observed at 400-500 • C temperature of reduction. Between these samples (400-500 • C), minimal grain size and porosity as well as crystallite size (37 nm) and a high mass fraction (~90%) were demonstrated in both samples after reducing at 400 • C (Figure 2b,f). This notable characteristic indicates that reducing at 400 • C provided powder with good properties, including morphology and chemical composition.
The concentration of precursor solution influenced the mean size of the particle size distribution of both iron and hematite powders. The average size of Fe 2 O 3 was estimated at 0.5 and 0.7 µm, respectively (for 10 and 20 sol. %, see Figure A1). Small growth of mean particle size proportionally to the temperature of reduction was observed ( Figure 3 and Table 1). Reduction processes insignificantly increased the mean size of the particles at temperatures below 450 • C. The mass fraction of α-Fe increased when the reduction temperature increased, but high temperature led to the destruction of nanostructured spherical particles to smaller particles of which the former was composed. For iron powder (10 sol. %), spheres broke down at 500 • C (Figure 2d). For another sample (20 sol. %), similar destruction presumably can be obtained at higher temperatures than 500 • C. The reduction at 350 °C provided iron particles with a slick, low-roughness surface (Figure 2a,e). At higher temperatures, pores became coarser and larger. Distinct grain structure and porosity were observed at 400-500 °C temperature of reduction. Between these samples (400-500 °C), minimal grain size and porosity as well as crystallite size (37 nm) and a high mass fraction (~90%) were demonstrated in both samples after reducing at 400 °C (Figure 2b,f). This notable characteristic indicates that reducing at 400 °C provided powder with good properties, including morphology and chemical composition.
The concentration of precursor solution influenced the mean size of the particle size distribution of both iron and hematite powders. The average size of Fe2O3 was estimated at 0.5 and 0.7 μm, respectively (for 10 and 20 sol. %, see Figure A1). Small growth of mean particle size proportionally to the temperature of reduction was observed ( Figure 3 and Table 1). Reduction processes insignificantly increased the mean size of the particles at temperatures below 450 °C. The mass fraction of α-Fe increased when the reduction temperature increased, but high temperature led to the destruction of nanostructured spherical particles to smaller particles of which the former was composed. For iron powder (10 sol. %), spheres broke down at 500 °C (Figure 2d). For another sample (20 sol. %), similar destruction presumably can be obtained at higher temperatures than 500 °C.

Thermal Analysis
Simultaneous (TG-DSC) thermal analysis was performed for all obtained iron powders ( Table 2). In the TG-DSC analysis of the samples "500 • C" the presence of exothermic reaction between iron powder and oxidizer in the range from 100 to 570 • C was observed ( Figure 4). The value of the oxidized iron powder mass gain was the closest to the theoretical value, calculated from the chemical reaction, which was 42.97%. That indicated iron powder purity and can be compared with quantitative X-ray diffraction results (see Table 1).  The thermal effects started above 100 °C and finished at 600 °C. The onset temperature shifted to higher temperatures with an increase in the reduction temperature. In Figure 4, both samples demonstrate two exothermic reaction peaks, which is typical for other powders. The total heat released, derived from the area under the curve, was 7.285 and 6.570 kJ·g -1 for samples 10 and 20 sol. %, respectively, reduced at 500 °C. Weak deviations in size [36], surface area, and porosity of particles as well as in grains and crystallinity [37] caused a variability of the values of onset temperature, the total released energy, and the mass gain. The thermal effects started above 100 • C and finished at 600 • C. The onset temperature shifted to higher temperatures with an increase in the reduction temperature. In Figure 4, both samples demonstrate two exothermic reaction peaks, which is typical for other powders. The total heat released, derived from the area under the curve, was 7.285 and 6.570 kJ·g -1 for samples 10 and 20 sol. %, respectively, reduced at 500 • C. Weak deviations in size [36], surface area, and porosity of particles as well as in grains and crystallinity [37] caused a variability of the values of onset temperature, the total released energy, and the mass gain.
As can be observed in Figure 4, the mass gain occurred in two stages. The deconvolution of the DSC-peaks is related to iron oxidation steps in the following order of reactions: [38]. The oxidation of iron at the first stage can be described with the parallel production of magnetite Fe 3 O 4 and hematite Fe 2 O 3 . Then, in the second stage, residual Fe 3 O 4 is oxidized to Fe 2 O 3 [36]. Three of the four deconvoluted peaks may be described with these reactions. The fourth peak, observed at temperatures above 570 • C [39] (vide Figure 4a), may be related to the wüstite formation. The correlation of the fitting curves obtained by deconvolution was greater than 0.99.
Considering STA, XRD, and SEM results, the 400 • C temperature is the optimum for annealing hematite powder in a hydrogen atmosphere to produce the chemically pure (~90%) iron powder with a sleek surface of particles, low porosity and large crystallite size (36-37 nm). The chemical purity of a metal is essential for its magnetic properties.

Microwave Measurement
The measured frequency dependencies of microwave permittivity and permeability are illustrated in Figures 5 and 6. The composites demonstrated intense microwave absorption. The imaginary permeability of the composite filled with the most chemically pure powder (T red = 500 • C) exhibited almost as high maximum amplitude as that of carbonyl iron [4], which is usually considered a "standard" commercially available form of pure iron powder (see Figure 6, curve "66 and 81 wt. %"). The ferromagnetic resonance frequency (FMR), which is the frequency where the real permeability equals unity, was approximately 20 GHz for all studied composites. The imaginary permeability exhibited two magnetic loss peaks: the low-frequency (LF) peak, located at 0.3-2 GHz, depending on sample properties, and the high-frequency (HF) peak, at around 8-11 GHz. Precise peak positions as well as Acher's parameters are given in Table 3. Both LF and HF peak positions were found to depend on the chemical purity of iron but not on the size of the particles (see Table 3: peak positions for each temperature were almost the same between samples). The LF peak was also found to be of higher relative amplitude for particles of Fe 10 sol. % sample than that of particles of Fe 20 sol. %. The low-frequency maximum in the imaginary permeability curve is usually deemed to be attributed to the effect of eddy currents [2], and the motion of domain walls [26,40]. In this case, the LF peak position is too high to be dealt with a domain wall motion. The physical origins of this absorption peak require further study. The HF maximum was related to the natural ferromagnetic resonance [4].      An increase in the temperature of hydrogen reduction led to an increase in the amplitude and a decrease in the resonance frequency (damping maxima) for both LF and HF peaks in all cases. The correlation observed was: (a) the higher the chemical purity of α-Fe had been; and (b) the lower had been the measured damping maxima. Smaller (10 sol %) particles of pure α-Fe showed an LF peak of a higher quality factor than larger particles (20 sol %).
In addition to these changes, there is a qualitative distinction between "350 • C" and "400 • C" vs. "450 • C" and "500 • C" samples. An increase in temperature higher than 400 • C by treatment with hydrogen decreased the Q-factor of the ferromagnetic resonance peak and decreased magnetic losses above~15 GHz. Therefore, 400 • C is the threshold temperature that qualitatively changes the microwave permeability of these powders. Two phenomena contribute to these changes: an increase in the chemical purity of the iron powder and a decrease in effective particle size caused by the partial destruction of the spheres, which was observed in Figure 2. Comparing properties of spherical powders of the same size but different chemical purity of α-Fe, it was observed that chemical composition significantly affects microwave permeability. This result adds substantial information to the knowledge that allows dynamic magnetic properties to fine-tune the desired composite materials.
The calculated integral Acher's parameter was within the range of 0.170 to 0.215, which was typical for spherical particles of carbonyl iron. A monotonic increase in K A value with the increase in temperature of reduction (except the "400 • C" samples of both samples) correlates well with the increase in a fraction of α-Fe determined using both XRD and STA techniques. This may be interpreted as an increase in the effective volume of pure iron in the sample. The "400 • C" samples showed high Acher's parameter which indicates high efficiency in microwave absorption among other samples.
The high permittivity for the composite filled with the most chemically pure iron is attributed to the destruction of initial spherical particles which is especially noticeable in Figure 2. The presence of not only spherical particles of~1-micron size but rather 0.1-1 size distribution noticeably affected the permittivity value.

Conclusions
Ferromagnetic nanostructured iron powders were obtained using a two-stage process involving spray-pyrolysis and reduction in hydrogen. The powders were obtained from iron nitrate aqueous solution (10 and 20 wt. %). In the first stage, the hematite powders with a mean diameter of 0.5 and 0.7 µm were synthesized, respectively, to the concentration of initial solutions. In the second stage, iron powders were obtained from hematite in hydrogen flow. The reduction process was carried out at 350, 400, 450, and 500 • C. The reduction of two iron oxide powders, 0.5 and 0.7 µm in diameter, at given temperatures, resulted in eight sample types to study in total.
It was found that the temperature reduction influenced grain size, crystallite size, porosity, and chemical purity of the product. As the reduction developed, iron oxide loses weight and volume, and then the newly formed a-Fe crystalizes. These processes resulted in particles essentially of the same size as initial iron oxide, but of porous, mesh structure, consisting of iron grains. The threshold temperature of reduction that gave distinct porous iron particles was found to be 400 • C. Reduction at 500 • C caused even higher grain growth, which led to partial destruction of the initially spherical particles. XRD showed the presence of α-Fe and α-Fe 2 O 3 phases in all samples. The a-Fe phase purity was estimated to be 85 mass. % when reduced at 350 • C, and 95 mass. % when reduced at 500 • C. Thermogravimetric data coincided well with the XRD measurements.
Variations in the reduction temperature significantly affected microwave permeability. It is important to note that the reduction temperature did not change the shape and size of the particles, which are well known to govern microwave permeability and permittivity. An increase in the chemical purity of iron particles increased the value of Acher's parameter, which is the quantitative characteristic of the dynamic magnetic performance of a material. Additionally, an increase in the chemical purity of iron increased the amplitude of both the real and imaginary parts of the permeability and shifted magnetic loss peaks to lower frequencies. Among other samples, powders obtained at 400 • C demonstrated distinct characteristics. Those possessed simultaneously high chemical purity, low porosity, high Acher's parameter, and high magnetic losses at around 15 GHz frequency range.
The results obtained enhance our understanding of the relationship between the chemical conditions of iron and the dynamic magnetic properties of iron-based composite materials.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

Appendix A
Magnetism 2023, 2, FOR PEER REVIEW 11 performance of a material. Additionally, an increase in the chemical purity of iron increased the amplitude of both the real and imaginary parts of the permeability and shifted magnetic loss peaks to lower frequencies. Among other samples, powders obtained at 400 °C demonstrated distinct characteristics. Those possessed simultaneously high chemical purity, low porosity, high Acher's parameter, and high magnetic losses at around 15 GHz frequency range. The results obtained enhance our understanding of the relationship between the chemical conditions of iron and the dynamic magnetic properties of iron-based composite materials.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.