Double Complex Salt [Co(NH₃)₆][Fe(CN)₆] Plasma Treatment

Abstract

The method of obtaining functional materials almost always influences the physicochemical properties of the resulting substances. The plasma treatment of solid materials is considered to be a more energy efficient method when compared with thermal destruction. Our work is the first to treat double complex salt (DCS) [Co(NH₃)₆][Fe(CN)₆] with different plasma discharge modes. We have demonstrated the possibility of obtaining a single-phase spinel with a CoFe₂O₄ structure as a result of the calcination in air of the plasma destruction product. The crystallite sizes of the obtained spinel are 40 nm, with a lattice constant 8.38 Å.

1. Introduction

Double complex salt (DCS) [Co(NH₃)₆][Fe(CN)₆] is formed by two complexing metals—cobalt (Co) and iron (Fe), with the complex cation and anion consisting of separate island complexes inside the DCS molecule [1,2]. Both the complex precursors and the DCS have simple and reproducible synthesis methods with quantitative yields. This makes DCS promising as a starting compound for various methods of obtaining functional materials.

DCS that contain Fe and/or Co are highly relevant in modern materials science. For example, DCS containing Fe and Co are discussed in [3,4]. The decomposition of DCS [Pt(NH₃)₅Cl][Me(C₂O₄)₃]…nH₂O, where Me includes Fe and Co, is discussed in [3]. Mixed ligand complexes with metal ions Co(II) and Fe(III)) of the type [ML¹L²] (where L¹ = indoline-2,3-dione and L² = naphthalene-1,2-dione) were synthesized and characterized using a number of methods in [4].

Environmental applications of DCS for pollutant gas removal are proposed in [5], where the general formula is [Co¹⁻³(LH)₂], in which (LH₂) is the vic-dioxime ligand. The authors of [6] report the synthesis, structure and characteristics of novel supramolecular salts with 1,10-phenanthroline 3D metal cations and similar malatostannate/germanate (IV) anions and 1,10-phenanthroline cations including Fe(II) and Co(II). Fe- and Co-containing DCS are observed in [7].

The compound [Co(NH₃)₆][Fe(CN)₆] is well known [8]. The works [9,10,11,12,13] describe the physical and chemical properties and promising application areas of this DCS. Obtaining functional materials by DCS thermal decomposition has been proposed, whereby the ligands are removed completely and in a way that is almost completely harmless.

Catalysts containing Co and Fe are also highly required in different application areas. For example, in [14], a Fe-Co dual single-atom catalyst was prepared by one-step molten-salt-assisted high-temperature pyrolysis. Similar Fe-Co dual single-atom catalysts were prepared in [15] according to the following strategy: a porous carbon matrix with Fe-Co diatomic sites was synthesized by a co-adsorption-pyrolysis. Ref. [16] reports a Fe-Co catalyst that is active at very low temperatures (−73 °C) and that is suitable for CO removal.

In recent years, a transition from micro-sized to nano-sized catalysts has emerged in catalysis. Currently, there are many methods for synthesizing nanoparticles (NPs) consisting of atoms of only one type of metal [17,18]. The transition to the synthesis of two- and multi-component particles is associated with a number of difficulties associated with the need to maintain the required ratio of components and their distribution throughout the volume of particles. DCS of the [M′L′][M′′L′′] type, where M′ and M′′ are complexing metals and L′ and L′′ are ligands, are promising precursors in the synthesis of bimetallic NPs.

The metal atoms in the initial complex are mixed at the molecular level, which facilitates the formation of solid solutions of bimetals during the thermal decomposition of the precursor. Thermolysis temperatures are significantly lower than the melting temperatures of individual metals, which makes this process attractive. It is important to note that the ratio of metals in the resulting NPs is constant and is strictly determined by the stoichiometry of the initial complex salt.

The structure of the obtained NPs is of great importance in terms of their catalytic properties. In particular, many authors note the importance of the spinel-like structures of the active phase for the hydrogenation processes of carbon oxides [19,20]. When obtaining spinel structures by traditional methods—coprecipitation or impregnation with subsequent heat treatment—massive spinel-like structures of various compositions are formed, and some metals pass into a catalytically inactive state. In this regard, the formation of catalytically active spinel-like structures by the method of thermolysis of DCS is of great practical interest.

According to [21,22], thermal decomposition of binary complex compounds is a promising method for obtaining spinels, due to the strict stoichiometry of the DCS. The thermolysis conditions significantly influence the result; in particular, time and temperature determine the phase composition of the resulting material. Achieving a stable phase composition of the DCS thermolysis product is an urgent task when obtaining catalytically active phases. Plasma–chemical decomposition of DCS is a novel alternative method for this. We could not find any studies on the treatment of DCS using plasma discharge so this type of DCS processing is reported here for the first time.

The use of additional methods for DCS precursor processing (low-temperature plasma processing) can also contribute to a better yield of the target product with improved characteristics. Plasma is a partially ionized gas containing electrons, ions, molecules, radicals and other excited particles. All of the reactive plasma species can actively participate in the synthesis and modification of inorganic materials [23]. Plasma can be used to synthesize some compounds that cannot be easily obtained by traditional methods [24] and enables reactions that are thermodynamically difficult under normal conditions. For example, such compounds as sulfides, nitrides and phosphides are usually obtained at high temperatures, but they can be synthesized under mild conditions with the participation of plasma [25]. Plasma synthesis increases the stability, dispersion, and purity of metal-containing NPs and significantly reduces the synthesis time [26]. A successful plasma synthesis of CoFe₂O₄ spinels has been reported [27]. Furthermore, plasma is a versatile and convenient tool for reduction, oxidation, doping, etching, coating, processing, and surface cleaning [28].

Spinels with the composition CoFe₂O₄ are widely applied [29,30]. They can serve as semiconductor photocatalysts due to ferromagnetic behavior [31], as supercapacitors [32] or as high temperature gas sensor [33].

The CoFe₂O₄ structure can be obtained by a variety of methods, such as the following: ceramic [34], hydrothermal [35,36] or solvothermal [37] synthesis; sol-gel [38,39,40]; hydroxide coprecipitation [29,39,41,42]; the microemulsion method [39]; microwave treatment [40,43,44]; the thermal decomposition of complexes [39,45]; and gel burning [42,46]. The methods listed above either require long-term high-temperature processing, high energy costs, or expensive equipment. Furthermore, almost all of these mentioned methods pollute the environment with salt decomposition products [47].

The production of a solid product with the spinel structure CoFe₂O₄ by thermal decomposition of [Co(NH₃)₆][Fe(CN)₆] is free from these disadvantages, as the target product is obtained at relatively low thermolysis temperatures [48,49,50]. Variations in the coordinated ligands change the surface structure of the thermolysis products and the size of the resulting particles [49,51]. For example, the use of urea as one of the reagents in the synthesis of spinels is very convenient [49,52,53,54] as it decomposes at low temperatures. In an air atmosphere, urea decomposes from 152 °C [55], while in a nitrogen environment it decomposes from 158 °C [56]. A large volume of gaseous non-toxic products release at this range, meaning that there is no contamination of the products.

Obtaining a single-phase target product with the CoFe₂O₄ structure is an important technological task. This work applies a complex of physical and chemical processes to solve the problem. The complex includes plasma treatment of DCS [Co(NH₃)₆][Fe(CN)₆] followed by thermal treatment (thermolysis) under various conditions. Thermal treatment after plasma is due to the need to obtain a single-phase target product with a CoFe spinel structure.

The main objective of our work is to study the effect of plasma treatment of DCS [Co(NH₃)₆][Fe(CN)₆] on the composition and properties of solid thermolysis products.

2. Materials and Methods

2.1. Obtaining DCS

The salt [Co(NH₃)₆]Cl₃ was purchased from Vekton, Russia, Saint Petersburg and the salt K₃[Fe(CN)₆] was obtained from NevaReaktiv, Russia, Saint Petersburg. DCS was obtained by mixing stoichiometric amounts of solutions [Co(NH₃)₆]Cl₃ and K₃[Fe(CN)₆. This article will only consider the DCS with the composition [Co(NH₃)₆][Fe(CN)₆], thus we will later refer to it as ‘DCS.’

2.2. Treatment of DCS with Plasma

The DCS plasma processing was carried out in a dielectric barrier discharge (DBD) installation with planar electrodes (Figure 1). The configuration of the reactor with planar electrodes was chosen as it is more suitable for the treatment of the solid materials. The DCS powder can be distributed and treated more uniformly on the surface of the plate and thus be more efficiently decomposed. In each experiment, 300 to 500 mg of DCS were placed in the reactor.

The reactor consisted of two parallel electrodes (steel plates) made of stainless steel with a diameter of 50 mm. Two quartz plates (100 × 100 × 2 mm) placed between the electrodes served as a dielectric barrier. The distance between the plates was 3.5 mm and the discharge gap was 7.5 mm. Cylindrical aluminum radiators 50 mm in diameter and 55 mm high dissipated the generated heat. A 500 mg DCS was placed between quartz plates. Argon from a cylinder (99.9999%, ITC Promekservis, Russia, Moscow) was supplied between the quartz plates using an RRG-20 regulator (Eltochpribor, Russia, Moscow, 2023). The Ar flow rate was 70 mL/min. The whole set-up was placed under a glass dome (Figure 1b) to ensure the Ar flow did not mix with the air. The high voltage power supply generated a sine wave signal with a frequency of 23 kHz. The registration of current and voltage signals, as well as Lissajous curves, was carried out using a TDS 2012B oscilloscope (Tektronix, USA, Beaverton, OR, 2010). Based on the area of the Lissajous curves, the absorbed plasma power was calculated using the following equation:

$$P=fW=f{C}_{n}A$$

(1)

where C_n is the capacitance of the capacitor connected in series with the reactor, f is the frequency of the applied voltage, and A is the area of the Lissajous curve.

The DCS was processed from 10 to 30 min in a periodic mode at different discharge powers. The sample was mixed between the processing steps. Each experiment was conducted three times. The paper contains averaged values.

The temperature of the plasma catalytic reaction zone was measured in a tubular DBD reactor using MLG 225 Laborant optical pyrometer (Metrologica LLC, Russia, Moscow, 2015). The glass dome of the DBD reactor did not allow one to estimate the temperature near the reaction zone so a different reactor configuration was chosen to accomplish the temperature measurement. Figure 2 presents the DBD reactor. The reactor comprised a quartz tube (25 mm outer diameter, 2 mm wall thickness, 160 mm length) serving as the dielectric barrier. A steel rod (8 mm diameter) with the threading functioning as the ground electrode inside the reactor, while a steel mesh (0.5 mm mesh size, 80 mm length) wrapped around the outside wall acted as the high-voltage electrode. The discharge gap was 8.5 mm. A DCS sample was placed within the reactor and fixed with ceramic wool (Al₂O₃ + SiO₂ > 97%, BERD LLC, Khimki, Russia).

2.3. Research Methods

IR spectra were recorded on a Nicolet 6700 FT-IR spectrometer (Thermo Fisher Scientific Inc., Hillsboro, OR, USA, 2010) in the 4000–400 cm⁻¹ range (KBr tablets).

X-ray phase analysis was performed for the products of DCS thermal decomposition. For each sample the analysis was undertaken twice in the range of 5–90° on a Shimadzu XRD 6000 powder diffractometer (Kyoto, Japan, 2008) equipped with a Cu-K_a source (λ = 1.5418 Å) and graphite monochromator for the diffracted beam. Recording mode was undertaken with a step of 0.1° and an exposure of 3 s. Qualitative analysis was performed using the powder diffraction database PDF-4 2021 (International Center for Diffraction Data (ICDD), Philadelphia, PA, USA).

Synchronous thermal analysis (STA) was made using HQT-4 Henven (Beijing Henven Experimental Equipment, China, Beijing, 2022). Experiments were performed in an air atmosphere. The heating rate was 10 °C/min and the temperature range was 30–1000 °C. This analysis was carried out in an Al₂O₃ crucible without a lid. STA included thermogravimetry (TG) and differential scanning calorimetry (DSC) analyses.

Static experiments on thermal decomposition were carried out in a muffle furnace SNOL 6.7/1300 (Snol, Narkūnai, Lithuania, 2004) in the temperature range of 300–1000 °C for 1 h in an Al₂O₃ crucible. The heating rate was 10 °C/min.

The carbon content of the samples was determined using an ELTRA CS-2000 analyzer (Alpha Resources, LLC, Stevensville, MI, USA, 2004). Plasma treated samples were studied by this method.

CHN elemental analysis was performed on a EuroVector EA3000 (Italy, Pavia, 2008) elemental analyzer at a temperature of up to 1000 °C. Plasma + temperature-treated samples were studied using this method.

Scanning electron microscopy (SEM) images were taken using a Tescan Amber GMH (Brno, Czech Republic, 2021) scanning electron microscope. Images were obtained using an Everhart–Thornley SE detector at ×1000–300,000 magnification and at an accelerating voltage of 1 kV.

The surface of the catalyst samples was studied by X-ray photoelectron spectroscopy (XPS) using a spectrometer EA15 (Prevac, Rogow, Poland 2021). An AlKα X-ray tube (1486.6 eV) served as the source of ionizing radiation. Before loading into the spectrometer, the samples were ground in an agate mortar and applied to a conductive carbon tape. An electron-ion charge compensation system was used to neutralize the sample charge during the experiments. All peaks were calibrated against the C 1s peak at 284.8 eV. The background type was determined according to Shirley, and the deconvolution was performed assuming that the overall peak was a sum of Gaussian curves.

3. Results and Discussion

The processing (treatment) of DCS samples in plasma was carried out under various conditions to study the degree of DCS decomposition. Two modes were applied to understand the discharge power influence on the DCS decomposition. Figure 3a shows the Lissajous curves recorded at different discharge power levels. As can be seen, the curve is a parallelogram. This is typical for a barrier discharge [57]. Comparison of Figure 3b,c, shows that the voltage values along the abscissa axis are almost similar in two different modes. The main difference is observed in the magnitude of the charge (stretching of the phenomena observed in Figure 3b,c along the ordinate axis). Increasing the current increases the power and, therefore, increases the number of microdischarges between the electrodes (Figure 3b,c).

The original orange-colored DCS powder underwent changes after plasma processing (treatment) at a power of 25 W. However, the presence of the original DCS powder on the surface of the lower quartz plate and around the treatment area was visually observed after the first treatment for 10 min, along with the reacting black DCS particles (Figure 4). The obtained mixture was designated as sample A-1 (see

Table 1. The plasma treatment conditions of the DCS samples.

Sample	Input Power, W	Plasma Treatment Duration	Remaining Mass After Treatment (%)
A-1	25	1 cycle of 10 min	67.4 ± 0.2%
A-2	25	2 cycles of 10 min each	63.2 ± 0.4%
A-3	25	3 cycles of 10 min each	61.6 ± 0.4%
A-4	25	3 cycles of 15 min each	62.1 ± 0.5%
A-5	100	1 cycle of 15 min	57.0 ± 0.4%
A-6	100	2 cycles of 10 min each	47 ± 0.2%

for the sample designation details). Later, we found that an additional treatment is required to accomplish the DCS decomposition. The amount of the original powder decreased after mixing and treating for a second time (A-2 sample). However, particles of the original DCS powder were present in the product even after the third stage of treatment (A-3 sample). The relative mass of the sample was 61.6 ± 0.4% of the original after the third treatment. The same results were obtained even after increasing the duration of the treatment cycle up to 15 min (A-4 sample). It was then concluded that the input power of 25 W was insufficient to reach the complete decomposition of the parent DCS.

When the power was increased to 100 W, the original orange-colored DCS powder was not observed, and the color of the processed powder became dark gray at the edges with a color transition to dark green in the center of the disk (Figure 5)—sample A-5. The relative mass of the sample was 57.0 ± 0.4% of the initial weight. This indicated incomplete conversion of the DCS in the plasma and the need to conduct the second cycle of the plasma treatment.

The sample mass decreased to 47 ± 0.2% of the original after mixing the sample and reprocessing (re-treating) for 10 min (A-6 sample). The original sample was processed twice to obtain sample A-6, as in Figure 6.

The mass of the sample did not change during the subsequent iteration of mixing and treatment (the third stage of treatment). This indicated the completeness of the treatment after the second cycle at a given power. The samples, which were further analyzed, and their treatment conditions are summarized in Table 1.

When measuring the temperature, the optical pyrometer was aimed at the outer electrode in the area in which the DCS was placed. Figure 7 demonstrates the temperature dependence plot within the reaction time. At the beginning of the plasma treatment (Figure 7a) the DCS sample is stable until the temperature grows rapidly to the value of ~125 °C, and then after ~5 min starts to grow to the value of 230 °C. At this stage, the DCS sample starts to decompose with a combustion-like behavior (Figure 7b). Apparently, the observed effect is related with exothermic stage of the DCS decomposition at ~230 °C.

The paper [58] demonstrates similar behavior at a [Co(NH₃)₆][Fe(CN)₆] thermolysis in an nitrogen flow. Thermal decomposition at ~230 °C is accompanied by intensive release of NH₃ and HCN. Gases formed during DCS decomposition are not ionized under plasma treatment conditions.

In Figure 7c, no plasma is generated on the right side of the discharge zone. After these gases were eliminated from the discharge zone, only Ar was present in the reactor and the plasma continued to be generated (Figure 7d). Meanwhile, the temperature decreased to the value of ~125 °C and did not increase further. After the reactor was cooled to room temperature, the second treatment was conducted without any additional manipulations with the DCS sample. Figure 7 demonstrates that, just as in the first treatment, the temperature started to increase rapidly to ~125 °C, but that no exothermic effect was observed. Thus, the steady-state reactor temperature during the plasma treatment was approximately 125 °C, with an increase to 230 °C at the DCS exothermic decomposition stage.

Figure 8 demonstrates XRD patterns of DCS samples after plasma treatment, but before additional thermal decomposition. The basis of sample A-1 is Co(NH₃)₆·Fe(CN)₆ of the rhombohedral modification. For samples A-2–A-4 the basis is a mixture of crystalline hydrate forms of iron and/or cobalt cyanides of cubic modification Fe₄[Fe(CN)₆]₃·14H₂O (PDF Card—01-073-0689), Co₃[Co(CN)₆]₂·xH₂O (PDF Card—00-022-0215), and Fe_0.667Co(CN)₄·4H₂O (PDF Card—04-013-6956). In samples A-5 and A-6 the cubic structure of magnetite is reliably identified, in which iron atoms are known to be replaced by cobalt (Co_0.22Fe_0.78)(Co_0.98Fe_1.02)O_3.965 (PDF Card—01-083-6174). The highly broadened reflection at 17.51° (5.061 Å) is probably due to the presence of precursor residues in the form of iron and/or cobalt cyanides.

Figure 9 demonstrates IR spectra of DCS samples under different treatment modes.

There is a band at 2114 cm⁻¹ on the IR spectra, which is characteristic of Fe(III)-CN. In samples before the thermal treatment, Fe(III) content is insignificant (inflection at 2138 cm⁻¹) and the main part of the iron is in the form of Fe(II)-CN (2078 cm⁻¹) and bridge Fe (2170 cm⁻¹) [59].

XRD (Figure 8) and IR (Figure 9) data revealed that the A-6 sample is the most optimal one due to it having the lowest ferrocyanide content and the greatest (Co_0.22Fe_0.78)(Co_0.98Fe_1.02)O_3.965 content (the sharpest and narrowest peaks on the XRD pattern, respectively). Thus, we selected A-6 for further study of the processes occurring during thermolysis of DCS after plasma treatment. The optimal mode for plasma treatment of the DCS is as follows: a power of 100 W and a two-fold treatment for 10 min.

Elemental analysis showed the carbon content in A-6 at the level of 8.2 wt%, while the content in the untreated initial DCS was 19.3 wt%. Thus, during the plasma treatment, at least 2/3 of the coordinated carbon was removed.

The graph (Figure 10) demonstrates curves of STA in air for DCS sample A-6 after plasma treatment. A sharp peak of weight loss on the TG is observed in the range of 250–300 °C, accompanied by a sharp exoeffect on the on the DSC curve.

After plasma treatment, sample A-6 was annealed at temperatures from 300 to 1000 °C. IR analysis of the sample calcined at 300 °C (curve A-6-300) shows the absence of vibrations in the region 2000–2200 cm⁻¹ (Figure 9), thus the sample does not contain ferrocyanide ions. Consequently, we determine that the oxidation of the remains of organic ligands (burnout of cyanoligands) in air occurs in the region of 250–300 °C. Thus we did not measure the IR spectra of plasma + temperature-treated A-6 sample for T = 400–900 °C because it is clear that ferrocyanide ions almost certainly burned out. The maximum temperature of 1000 °C was chosen based on XRD data (Figure 11) because a single-phase decomposition product with the structure of spinel CoFe₂O₄ was obtained.

XRD data (Figure 11) reveal the A-6-300 sample cubic spinel structure CoFe₂O₄ (PDF Card—04-016-3954), with a lattice constant of ~8.38 Å and a yield ~30%. Partially overlapping low intensity reflections (splitting) in the region of 31.18° (2.866 Å), 36.52° (2.459 Å), 44.52° (2.034 Å) are probably caused by the presence of a similar structure. The structure is cobalt oxide Co₃O₄ (PDF Card—04-008-2376) with a smaller lattice constant (~8.11 Å). Increasing the temperature to 800–900 °C results in a strictly single-phase CoFe₂O₄ structure. Figure 12 demonstrates the change in the parameters of the coherent scattering region (CSR) depending on the temperature of the heat treatment of the DCS. CSR is calculated using the Scherrer equation, D = 0.941λ/βcos θ, where D is the CSR or average crystallite size, λ is the X-ray wavelength (CuKα = 1.5405929 Å), and θ and β are the diffraction angle and the full width at half maximum (FWHM) of the observed peaks, respectively. As with any calculated values, this one has its particularities.

The CSR curves in Figure 12 indicate that the particle sizes of the solid product with the structure of spinel CoFe₂O₄ vary from 15 to 40 nm in the temperature range of 300–1000 °C and that the size of the lattice constant is ~8.38 Å. An increase in particle size with increasing heat treatment temperature of DCS was previously observed in [60]. When using other methods (for example, combustion, precipitation and coprecipitation) of synthesizing single-phase spinel CoFe₂O₄ in the same temperature range, the crystallite sizes are 34–70 nm, with a lattice constant ~8.34–8.37 Å [42]. Spinel obtained by DCS treated by plasma + thermolysis at 1000 °C has a crystallite size two times smaller than that of earlier studies.

The nonlinear regression that fits the experimental data (red curve on Figure 12) was calculated according to the data in

Table 2. Fitting parameters for the approximation of the theoretical data on CSR; y = CSR, x = T.

Model	Poly 4
Equation	y = A0 + A1 × x + A2 × x2 + A3 × x3 + A4 × x4
A0	−43.73 ± 16.54
A1	0.49 ± 0.12
A2	−0.0015 ± 0.0003
A3	0.00019 ± 0.00003
A4	−0.000000082 ± 0.000000001
Reduced Chi-Sqr	0.24
R2	0.999
Adjusted R2	0.996

. The approximation fits the experimental data very well, as the R² at these parameters are almost equal to 1. We see that significant members stand for members up to x².

As a single-phase spinel is formed at a temperature of 1000 °C, the sample obtained at this temperature was examined in more detail.

Elemental analysis showed the absence of nitrogen and hydrogen in the composition of sample A-6-1000; therefore, coordinated ammonia was completely removed as a result of plasma + thermolysis treatment. The carbon content is 0.9 ± 0.1 wt%, which indicates that 4.7% of the original carbon content, which is part of the cyano groups, remained in the calcination residue in the DCS. The IR spectrum of the A6-1000 sample (Figure 7) showed the absence of absorption bands characteristic of ferrocyanides.

The composition of the sample surface was studied using the XPS method (Figure 13). The surface of the sample is predominantly composed of carbon, oxygen, iron and cobalt (

Table 3. Sample surface composition according to XPS data.

Element	C	O	Si *	Fe	Co
At %	70.9	23.4	4.3	0.7	0.7
Wt %	59.7	26.3	8.4	2.7	2.9

* Silicon was added when grinding the sample in an agate mortar.

).

XPS allows one to obtain information on the composition of the surface 5–10 nm [61]. Thus, the surface itself of the sample is enriched with carbon and silicon, while iron and cobalt are predominantly concentrated deeper in the material. Figure 13b demonstrates that carbon is represented by four types of bonds: C-C in sp3 hybridization (BE = 284.9 eV, FWHM = 1.77, 58.4%), C-O-C/C-OH bonds (BE = 286.2 eV, FWHM = 1.83, 26.6%), -C = O bonds (BE = 288.1 eV, FWHM = 2.27, 11.6%) and -C(O)-O- bonds (BE = 289.6 eV, FWHM = 1.62, 3.4%) [62,63]. Deconvolution of the O1s spectrum (Figure 13e) suggests that oxygen on the surface is represented by three types of atoms: adsorbed oxygen (BE = 532.4 eV, FWHM = 2.848, 75.0%), crystal lattice oxygen (Me-O) (BE = 529.7 eV, FWHM = 1.369, 21.9%) and oxygen bound to carbon -C(O)-O- (BE = 533.9 eV, FWHM = 1.511, 3.1%) [64,65,66]. Cobalt on the surface is represented by Co²⁺ in an octahedral environment (BE = 779.8 eV, FWHM = 2.742, 64.9%) and Co²⁺ in a tetrahedral environment (BE = 782.2 eV, FWHM = 1.949, 9.3%). The peak at 785.4 eV can be correlated with the satellite peaks of cobalt (II) and the Auger line of FeLMM. Iron is a mixture of Fe(II) (BE = 532.4 eV, FWHM = 2.848, 75.0%) and Fe(III) (BE = 532.4 eV, FWHM = 2.848, 75.0%), the peak at 718.5 eV is a satellite peak of Fe and of the Auger line of CoLMM [65,66,67,68,69].

It should be noted that the CSR size is usually lower than the SEM grain size determination, as the CSR corresponds to the internal structurally ordered region of the grain and does not include strongly structurally distorted grain boundaries. The discrepancy between the CSR and the real crystallite size was observed in [70] with the example of (Gd_0.96Eu_0.01Sm_0.01Tb_0.01Er_0.01)NbO₄ ceramic solid solutions. Indeed, SEM showed that the crystal size was about 200 nm, Figure 14.

We observe a rather unusual morphology of the substance: at different scales the shape has a different structure. This seems to be a manifestation of self-organization: at the scale level of 100 μm this manifests as long rods of 20–100 μm in size (Figure 14a), with the rods consisting of individual sintered crystallites of a faceted, regular shape, but of very different sizes and with the crystallites having sizes from 200 nm to 1 μm (Figure 14b,c). At the same time, the surface of the crystallites also has a stepped structure (Figure 13d). Thus, three levels of organization of the substance are clearly traced, which is most likely a consequence of energy dissipation under highly nonequilibrium conditions. It is worth assuming that the rods are the result of plasma sintering, and that the precipitation of crystallites inside them is the result of the next stage of processing, thermal decomposition. However, such effective dissipation of different types of energy leads to the fact that the grains are highly heterogeneous inside and the internal structurally ordered regions do not exceed 40 nm (see Figure 11).

4. Conclusions

For the first time, the [Co(NH₃)₆][Fe(CN)₆] DCS was processed by a plasma discharge. A higher applied power makes decomposition more effective, as evidenced by the XRD and IR spectrometric analyses. The optimal processing mode is as follows: double processing for 10 min at 100 W with intermediate mixing. A single-phase sample of spinel with the CoFe₂O₄ structure was obtained by step-by-step processing of the [Co(NH₃)₆][Fe(CN)₆] DCS. The production of the main phase of CoFe₂O₄ begins at 300 °C. The highest yield (100%) is achieved with heat treatment at 900–1000 °C, and at 1000 °C CSRs have a size of 40 nm, while the crystallites have sizes from 200 nm to 1 μm. This is explained by self-organization at three scale levels during two types of DCS treatment. Coordinated ammonia was completely removed as a result of plasma + thermolysis treatment. The surface of the sample is predominantly composed of carbon, oxygen, iron and cobalt.