Impact of Ni Content on the Structure and Sonophotocatalytic Activity of Ni-Zn-Co Ferrite Nanoparticles

The structure, morphology, and sonophotocatalytic activity of Ni-Zn-Co ferrite nanoparticles, embedded in a SiO2 matrix and produced by a modified sol-gel method, followed by thermal treatment, were investigated. The thermal analysis confirmed the formation of metal succinate precursors up to 200 °C, their decomposition to metal oxides and the formation of Ni-Zn-Co ferrites up to 500 °C. The crystalline phases, crystallite size and lattice parameter were determined based on X-ray diffraction patterns. Transmission electron microscopy revealed the shape, size, and distribution pattern of the ferrite nanoparticles. The particle sizes ranged between 34 and 40 nm. All the samples showed optical responses in the visible range. The best sonophotocatalytic activity against the rhodamine B solution under visible irradiation was obtained for Ni0.3Zn0.3Co0.4Fe2O4@SiO2.


Introduction
Despite the measures taken to reduce pollution, industrial effluents containing dyes and pigments used in the textile industry often resurface in the surrounding environment. Dyes are complex organic structures with a high resistance to chemical and biological degradation, high water solubility, and have a negative impact on the environment, particularly aquatic ecosystems [1,2]. Therefore, the efficient treatment of industrial effluents and wastewaters containing dyes is crucial for environmental protection.
The photocatalytic degradation of dyes is a simple, cost-effective, and environmentally friendly approach for wastewater treatment as it allows the decomposition of complex organic structures into CO 2 and water [3,4]. In the last few years, the use of sonophotocatalysis for the degradation of a wide range of organic pollutants in aqueous systems has been the topic of several studies [5,6]. Sonophotocatalysis use the synergistic effects of ultrasonic waves, UV-Vis irradiation and photocatalyst to form highly reactive free radicals in an aqueous medium that further react with dyes and lead to their degradation [7]. By providing additional nuclei, the heterogeneous catalyst enhances the formation of cavitation bubbles, which in turn increases the formation of reactive radicals through water pyrolysis [5]. The mechanism of sonophotocatalytic degradation, as well as the main advantages of combined ultrasound and photocatalytic processes, are presented by Abdurahman et al. [5]. The high energy consumption of sonophotocatalysis makes its large-scale application difficult, however, the high costs could be compensated by the low time required for the degradation of organic compounds [7]. Due to their magnetic properties, their recovery using an external magnetic field and their reuse is possible [8,9]. emission spectrometry (ICP-OES). The formation of the crystalline phase, crystallite size, and lattice constant were monitored by X-ray diffraction (XRD). The surface (specific surface area and porosity) was investigated using the Brunauer-Emmett-Teller (BET) method. The sonophotocatalytic properties of the samples were evaluated under visible light irradiation, assisted by sonication against RhB.

Thermal Analysis
The TG-DTA curves of all of the samples show three endothermic and two exothermic processes characterized by very close, overlapping peaks ( Figure 1).
The endothermic effect at 61-70 • C, accompanied by 3-5% mass loss, is attributed to the loss of crystallization and constitution water. The endothermic effect at 136-144 • C, accompanied by 17-26% mass loss, is assigned to the formation of divalent metal precursors (Co, Ni, and Zn succinates), while the endothermic effect at 187-201 • C accompanied by 9-14% mass loss, is ascribed to the formation of a trivalent metal precursor (Fe succinate). The distinct behavior of Fe succinate, compared to the divalent metal (Co, Ni, Zn) succinates, can be attributed to the redox reaction between Fe(NO 3 ) 3 and 1,4BD, as well as to the stronger acidity of the aqua-cation [Fe(H 2 O) 6 ] 3 [43]. The overlapping exothermic effects, at 270-292 • C and 310-325 • C, accompanied by 19-25% mass loss, are attributed to the decomposition of metal succinates to metal oxides, which leads to the formation of ferrites.
The exothermic peak, characteristic of the decomposition of divalent metal succinates, decreases with the increasing Zn content and shifts toward higher temperatures, leading to the increase of the exothermic peak, attributed to the decomposition of the Fe succinates. The SiO 2 matrix suffers various transformations during the thermal process, making the demarcation of the processes attributed to the formation and decomposition of succinate precursors difficult [44]. The total mass loss increases in the following order: Ni 0. 5  In the case of the samples dried at 40 • C, the intense band at 1379-1389 cm −1 is associated with the N-O bonds stretching vibration in metal nitrates. This band disappears in the case of samples dried at 200 • C, indicating the decomposition of nitrates [44]. The bands at 2958-2963 cm −1 and 2872-2888 cm −1 are specific to the symmetric and asymmetric vibration of C-H bonds in 1,4-BD or succinate precursors. These bands also disappear in samples heattreated at 200 • C. The bands at 1578-1605 and 3200-3210 cm −1 are attributed to the stretching and bending vibrations of O-H in 1,4-BD and adsorbed molecular water [44,45]. In the samples dried at 200 • C, the band at 3200-3210 cm −1 is shifted towards higher wavenumbers (3421-3437 cm −1 ), indicating that the metal succinates are hygroscopic [44,45]. The presence of this absorption band could also be due to the O-H stretching vibration and Si-OH deformation vibration caused by the hydrolysis of -Si(OC 2 H 5 ) 4 [44,45]. For samples dried at 40 and 200 • C, the bands at 557-568 cm −1 are attributed to Ni-O and Zn-O vibrations, while the band at 433-452 cm −1 is attributed to the Fe-O vibration in the nitrates [44,45]. In samples at 40 • C, the band at 683-393 cm −1 is assigned to the Co-O bond vibration in the cobalt nitrate [44,45]. The formation of the SiO 2 matrix in the samples dried at 40 [44,45]. Figure 3a shows the FT-IR spectra of NCs thermally treated at 1000 • C. The band at 618-626 cm −

FT-IR Analysis
As vibrational modes in FT-IR spectroscopy are determined by the bond type, the symmetry of the lattice sites and the elements in the crystal lattice, the monitoring of the ferrite formation process is possible [44]. The FT-IR spectra of the gels dried at 40 and 200 • C are presented in Figure 2.
As vibrational modes in FT-IR spectroscopy are determined by the bond type, the symmetry of the lattice sites and the elements in the crystal lattice, the monitoring of the ferrite formation process is possible [44]. The FT-IR spectra of the gels dried at 40 and 200 °C are presented in Figure 2. In the case of the samples dried at 40 °C, the intense band at 1379-1389 cm −1 is associated with the N-O bonds stretching vibration in metal nitrates. This band disappears in the case of samples dried at 200 °C, indicating the decomposition of nitrates [44]. The bands at 2958-2963 cm −1 and 2872-2888 cm −1 are specific to the symmetric and asymmetric vibration of C-H bonds in 1,4-BD or succinate precursors. These bands also disappear in samples heat-treated at 200 °C. The bands at 1578-1605 and 3200-3210 cm −1 are attributed to the stretching and bending vibrations of O-H in 1,4-BD and adsorbed molecular water [44,45]. In the samples dried at 200 °C, the band at 3200-3210 cm −1 is shifted towards higher wavenumbers (3421-3437 cm −1 ), indicating that the metal succinates are hygroscopic [44,45]. The presence of this absorption band could also be due to the O-H stretching vibration and Si-OH deformation vibration caused by the hydrolysis of -Si(OC2H5)4 [44,45]. For samples dried at 40 and 200 °C, the bands at 557-568 cm −1 are attributed to Ni-O and Zn-O vibrations, while the band at 433-452 cm −1 is attributed to the Fe-O vibration in the nitrates [44,45]. In samples at 40 °C, the band at 683-393 cm −1 is assigned to the Co-O bond vibration in the cobalt nitrate [44,45].  [44][45][46].
In comparison to the samples dried at 40 °C and 200 °C, in the samples thermally treated at 1000 °C, the wavenumbers specific to Co-O bond vibration decrease, and the wavenumbers specific to M(II)-O increase. The Jan-Teller effect, determined by the presence of Fe 2+ ions in the sublattices, can lead to band splitting, small bands and/or shoulders [47]. The Fe 2+ ions may result from the hopping process, namely M 2+ + Fe 3+ ↔ M 3+ + Fe 2+ (M = Co, Ni, Zn). The Co 3+ ion may migrate to tetrahedral (A) sites, while the Fe 2+ ions remain in their sites [31]. (1) (3)

XRD Analysis
The XRD patterns (Figure 3b) confirm the presence of well-crystallized ferrites in all of the samples, while the positions and intensities of the diffraction lines support the spi- In comparison to the samples dried at 40 • C and 200 • C, in the samples thermally treated at 1000 • C, the wavenumbers specific to Co-O bond vibration decrease, and the wavenumbers specific to M(II)-O increase. The Jan-Teller effect, determined by the presence of Fe 2+ ions in the sublattices, can lead to band splitting, small bands and/or shoulders [47]. The Fe 2+ ions may result from the hopping process, namely M 2+ + Fe 3+ ↔ M 3+ + Fe 2+ (M = Co, Ni, Zn). The Co 3+ ion may migrate to tetrahedral (A) sites, while the Fe 2+ ions remain in their sites [31].
The bands at 790-795 cm −1 are characteristic for the vibration of the Si-O bond in SiO 2 matrix, while those at 1090-1095 cm −1 and 485-490 cm −1 are characteristic to the stretching and bending vibration of Si-O-Si chains and show a low degree of polycondensation of the SiO 2 network [44,45]. The difference in band position could be attributed to the difference in M-O distance in the tetrahedral and octahedral sites [39].

XRD Analysis
The XRD patterns (Figure 3b [41,48]. In all of the samples, the local crystal structure is cubic spinel-type, belonging to the Fd-3m space group [21,40,41] [39]. The excess of metal oxides in insoluble secondary phases (Fe 2 O 3 ) can contribute to the densification by generating high pore volumes and demagnetizing fields. During synthesis, the homogeneity of the metal oxide particles may result in higher defects and pore volumes in the final products [36,42].
The average crystallites size (D XRD ) was calculated using the Scherrer equation (Equation (1)).
where a is the lattice constant (Å). The average crystallite size lies in the nanocrystalline range and increases with the increasing Ni content, while the lattice parameter (a) decreases with the increasing Ni content (Table 1). The change in the lattice constant (a) generates internal stress and suppresses additional grain growth during thermal treatment [44,50,51]. The tetrahedral (A) sites have smaller radii (0.52 Å) than the octahedral (B) site (0.81 Å) [9]. The ionic radii of Ni 2+ (0.69 Å), Zn 2+ (0.74 Å) and Co 2+ (0.75 Å) ions are larger than the ionic radius of Fe 3+ (0.64 Å) [3,38,52]. The amorphous to crystalline phase transformation and the relative content of crystalline phases, after thermal treatment at 1000 • C, were assessed using the relative degree of crystallinity (DC), calculated as the ratio between the area of diffraction peaks and the total area of diffraction peaks and halos. The DC increases with the increase of the crystallite size and Ni content. The Reference Intensity Ratio (RIR) method was used for the quantitative phase analysis of NCs thermally treated at 1000 • C.

Elemental Analysis
The Ni/Zn/Co molar ratios, determined by microwave digestion and combined with inductively coupled plasma optical emission spectrometry, are in good agreement with the theoretical values (Table 1).

BET Analysis
Due to the low amount of adsorbed/desorbed nitrogen, the determination of porosity and specific surface area (SSA) for the samples thermally treated at 1000 • C was not possible. The SSA below the method detection limit (0.5 m 2 /g) suggests that all ferrites have a nonporous structure, probably due to particle agglomeration that limits nitrogen absorption.

TEM Analysis
The TEM images of the mixed Ni-Zn-Co ferrites following thermal treatment at 1000 • C ( Figure 4) reveal spherical, small (high Zn content), or large (high Ni content) nanoparticles that form large spongy aggregates.
The formation of agglomerates with irregular morphology composed of high Zn content ferrite particles and the homogenous dispersion of high Ni content ferrite particles is also remarked. The small grains have a high surface area to volume ratio and allow faster oxygen diffusion than the larger grains, leading to an increase in the stoichiometry of the sample [35]. Although the small particles are closely arranged together, a clear boundary between adjacent particles is still observed. The average particle size is 34-36 nm, the difference being attributed to the grain boundary motion that exerts a dragging force, while the pores delay the force over the grain [41]. Moreover, the driving force increases the grain boundaries over the pores, resulting in lower pore volume and higher density [41]. The average crystallite size estimated by XRD is close to the particle size determined by TEM, the slight differences being attributed to the amorphous SiO 2 matrix and large-size nanoparticles [43,44,50].

UV-VIS Analysis
The optical response of the samples was evaluated by UV-Vis spectroscopy. The UV-Vis absorption (Figure 5a) shows that all the samples have a broad response in the visible range. Based on the absorption spectra and using the Tauc's relation [50], the band gap energy of the samples was evaluated (Figure 5b).

UV-VIS Analysis
The optical response of the samples was evaluated by UV-Vis spectroscopy. The UV-Vis absorption (Figure 5a) shows that all the samples have a broad response in the visible range. Based on the absorption spectra and using the Tauc's relation [50], the band gap energy of the samples was evaluated (Figure 5b). The band gap energy values are in the range 1.21-1.49 eV, are lower than that of NiFe2O4 (2.2 eV), ZnFe2O4 (1.91 eV) and CoFe2O4 (2.31 eV), and are comparable to those reported for CoFe2O4 xerogel calcined at 500 °C (1.5 eV) [33,50]. The band gap of our samples was also lower than those of NixCo1-xFe2O4 (1.37-1.78 eV) obtained by coprecipitation [52]. The optical band gap of the samples is due to the d-d transition. The crystal field splits the d level in the eg and t2g levels and the band gap energy depends on octahedral (B) and tetrahedral (A) sites. The band gap energy, in the case of the octahedral site, is higher than that of the tetrahedral (A) site [53]. The variation of the band gap energy, by replacing Zn 2+ ions with Ni 2+ , can be explained by the redistribution of Ni 2+ ions between the octahedral (B) and tetrahedral (A) sites. In the XRD data, the peaks corresponding to the plane (220) and (422) are sensitive to the tetrahedral (A) site, whereas the peak corresponding to the (222) plane is sensitive to the octahedral (B) site [54,55].    [33,50]. The band gap of our samples was also lower than those of Ni x Co 1-x Fe 2 O 4 (1.37-1.78 eV) obtained by coprecipitation [52]. The optical band gap of the samples is due to the d-d transition. The crystal field splits the d level in the e g and t 2g levels and the band gap energy depends on octahedral (B) and tetrahedral (A) sites. The band gap energy, in the case of the octahedral site, is higher than that of the tetrahedral (A) site [53]. The variation of the band gap energy, by replacing Zn 2+ ions with Ni 2+ , can be explained by the redistribution of Ni 2+ ions between the octahedral (B) and tetrahedral (A) sites. In the XRD data, the peaks corresponding to the plane (220) and (422) are sensitive to the tetrahedral (A) site, whereas the peak corresponding to the (222) plane is sensitive to the octahedral (B) site [54,55]. The values of the I(220)/I(222) ratio, for the samples annealed at

Sonophotocatalytic Activity
The sonophotocatalytic activity of the samples was evaluated using an RhB synthetic solution under visible irradiation. Before visible irradiation, the samples were kept in the dark for 1 h to reach the adsorption-desorption equilibrium. The adsorption capacity of the sample varied between 7-28%. The adsorption properties depend on the surface sites and specific surface area. In our case, the samples had almost identical particle sizes; thus, those surface sites were responsible for the adsorption properties. The removal rate ( Figure 6) was evaluated after 7 h of visible irradiation and varied between 16 and 75%. Similar removal efficiencies (83.9%) for methylene blue were obtained using Ni-Cu-Zn ferrite@SiO 2 @TiO 2 by Chen et al. [52,55].

Sonophotocatalytic Activity
The sonophotocatalytic activity of the samples was evaluated using an RhB syn solution under visible irradiation. Before visible irradiation, the samples were kept dark for 1 h to reach the adsorption-desorption equilibrium. The adsorption capac the sample varied between 7-28%. The adsorption properties depend on the surfac and specific surface area. In our case, the samples had almost identical particle sizes those surface sites were responsible for the adsorption properties. The removal rate ure 6) was evaluated after 7 h of visible irradiation and varied between 16 and 75%. S removal efficiencies (83.9%) for methylene blue were obtained using Ni-Cu-Z rite@SiO2@TiO2 by Chen et al. [52,55].  The sample with similar Zn 2+ and Ni 2+ ions content (Ni0.3Zn0.3Co0.4Fe2O4@SiO2) s the highest removal capacity, indicating that the equilibrium between Ni-ferrite an ferrite assures the best photocatalytic performance. In addition, based on the quanti crystalline phase analysis, this sample contains a lower amount of α-Fe2O3 (2%) com with samples Zn0.6Co0.4Fe2O4@SiO2 (1), Ni0.1Zn0.5Co0.4Fe2O4@SiO2 Ni0.2Zn0.4Co0.4Fe2O4@SiO2 (3), which means that in the case of this sample, α-Fe2O3 do significantly influence photocatalytic activity.
For this sample, the photodegradation kinetic was analyzed with respect to t sorbance of RhB using the pseudo-first order kinetic model (Equation (6)).

− * = •
where At is the absorbance of RhB at time t, A 0 * is the absorbance of RhB at time t0 is the apparent kinetic constant (min −1 ). A linear relationship with the irradiation (Figure 7), with a rate constant of 2.79 × 10 −3 min was obtained. For this sample, the photodegradation kinetic was analyzed with respect to the absorbance of RhB using the pseudo-first order kinetic model (Equation (6)).
where A t is the absorbance of RhB at time t, A * 0 is the absorbance of RhB at time t 0 and k i is the apparent kinetic constant (min −1 ). A linear relationship with the irradiation time (Figure 7), with a rate constant of 2.79 × 10 −3 min was obtained.

Characterization
The formation and decomposition of the carboxylate-type precursor were inv gated by thermogravimetry (TG) and differential thermal analysis (DTA) using the Q600 thermogravimeter, in air, up to 1000 °C, at 10 °C·min −1 heating rate, using alum standards. The FT-IR spectra were recorded on KBr pellets containing 1% samples u a Perkin Elmer Spectrum BX II spectrometer, while the XRD patterns were recorde room temperature using a Bruker D8 Advance diffractometer with CuKα1 radiation 1.54060 Å ). The Ni/Zn/Co molar ratios were confirmed using Perkin Elmer ICP-OES tima 5300 DV (Norwalk, CT, USA) after closed-vessel microwave-assisted aqua regia gestion using a Speedwave Xpert system (Berghof, Germany). The specific surface (SSA) was obtained using the BET model from N2 adsorption-desorption isotherms orded at −196 °C by a Sorptomatic 1990 (Thermo Fisher Scientific) instrument. The U VIS absorption spectra were recorded using a JASCO V570 UV-VIS-NIR spectrophot eter, equipped with a JASCO ARN-475 absolute reflectivity measurement accessory.

Characterization
The formation and decomposition of the carboxylate-type precursor were investigated by thermogravimetry (TG) and differential thermal analysis (DTA) using the SDT Q600 thermogravimeter, in air, up to 1000 • C, at 10 • C·min −1 heating rate, using alumina standards. The FT-IR spectra were recorded on KBr pellets containing 1% samples using a Perkin Elmer Spectrum BX II spectrometer, while the XRD patterns were recorded at room temperature using a Bruker D8 Advance diffractometer with CuK α1 radiation (λ = 1.54060 Å). The Ni/Zn/Co molar ratios were confirmed using Perkin Elmer ICP-OES Optima 5300 DV (Norwalk, CT, USA) after closed-vessel microwave-assisted aqua regia digestion using a Speedwave Xpert system (Berghof, Germany). The specific surface area (SSA) was obtained using the BET model from N 2 adsorption-desorption isotherms recorded at −196 • C by a Sorptomatic 1990 (Thermo Fisher Scientific) instrument. The UV-VIS absorption spectra were recorded using a JASCO V570 UV-VIS-NIR spectrophotometer, equipped with a JASCO ARN-475 absolute reflectivity measurement accessory.

Sonophotocatalysis
The sonophotocatalytic activity of the samples was evaluated against RhB solution under visible light irradiation in a Laboratory-Visible-Reactor system using a 400 W halogen lamp (Osram) and an ultrasonic bath. The catalyst (10 mg) was suspended in an aqueous solution of RhB (1.0 × 10 −5 mol L −1 , 20 mL), and the mixture was stirred in the dark to achieve the adsorption equilibrium on the catalyst surface. Each degradation experiment was conducted for 240 min. Samples from a given mixture (3.5 mL) were withdrawn for analysis every 60 min. After separating the catalyst from the suspensions with a permanent magnet, the solution was analyzed using a UV-Vis spectrophotometer by recording the maximum absorbance of RhB at 554 nm. The sonophotocatalytic activity was estimated based on the calculated degradation rate. Before the sonophotodegradation experiments, the RhB adsorption on the surface of the nanoparticles was analyzed. The adsorption was verified in the dark by mixing the photocatalyst into the RhB solution for 60 min until the adsorption-desorption equilibrium was reached.

Conclusions
Ni-Zn-Co ferrites, with different Ni:Zn:Co ratios (Zn 0. 6 ), embedded in SiO 2 were obtained by sol-gel method, followed by thermal treatment at 1000 • C. The thermal analysis revealed the formation and decomposition of metal succinate precursors in two stages, with distinct formation and decomposition of divalent (Ni 2+ , Zn 2+ , Co 2+ ) and trivalent (Fe 3+ ) succinates. The shapes of the DTA curves are similar, with the exception of the divalent metal's succinate decomposition stage, where for samples with high Ni content, the intensity of the exothermic peak decreases and shifts to higher temperatures. The total mass losses vary between 54.4-58.5%. The precursor formation and their decomposition into ferrites, as well as the formation of the silica matrix, are also confirmed by the FT-IR spectra. The XRD revealed the presence of well-crystallized ferrites along two crystalline phases of the SiO 2 matrix (cristobalite and tridymite). In samples with high Zn content, traces of hematite were also identified. The agglomeration of particles and the particle size of Ni-Zn-Co ferrites increase with the increasing Ni content, from 34 nm to 40 nm. All of the samples show a good optical response in the visible range, the best sonophotocatalytic performance being found for the Ni 0.3 Zn 0.3 Co 0.4 Fe 2 O 4 @SiO 2 sample, most likely due to the equilibrium between Ni-ferrite and Zn-ferrite.