From Disposal to Reuse: Production of Sustainable Fatty Acid Alkyl Esters Derived from Residual Oil Using a Biphasic Magnetic Catalyst

The development of technologies to promote residual oil reuse has been encouraged, aiming to reduce the environmental impact and promote sustainability. In this study, a biphasic magnetic catalyst with composition equal to ZnO-Ni0.5Zn0.5Fe2O4 was synthesized and applied to the fatty acid alkyl ester (FAAE) production from residual oil. The ZnO-Ni0.5Zn0.5Fe2O4 catalyst was synthesized by combustion reaction and characterized by X-ray diffraction (XRD), textural analysis, Fourier transform infrared spectroscopy (FTIR), Raman spectroscopy, thermogravimetry, particle size distribution, scanning electron microscopy, magnetic measurements, quantification of acidic sites by TPD-NH3, and catalytic tests. The efficiency of catalyst synthesis was evaluated by XRD, FTIR, and Raman spectroscopy experiments. Granulometric analysis and SEM images confirmed the presence of the agglomerates and particles with a wide size range. The catalyst presented soft magnetic behavior, with high saturation magnetization. Additionally, the catalytic activity of the ZnO-Ni0.5Zn0.5Fe2O4 system showed an average conversion of 73% for the methyl route. The results indicate that the reuse of residual oil is feasible for FAAE production, contributing to sustainable fuel development. Moreover, it allows the reintroduction of waste oil into the biodiesel production chain, reducing cost after process optimization.


Introduction
There has been growing concern regarding environmental issues in recent years due to the accumulation and disposal of postconsumer residues [1,2]. In this scenario, residual cooking oil has attracted attention since when it is improperly disposed of, it can negatively impact the environment [3]. Residual cooking oil is highly polluting and causes water and soil contamination since

Materials
In this research, the Ni 0.5 Zn 0.5 Fe 2 O 4 magnetic catalyst was synthesized via a combustion reaction from the following chemical reagents: nickel nitrate hexahydrate (Ni(NO 3 ) 2 6H 2 O), hexahydrate zinc nitrate (Zn(NO 3 ) 2 ·6H 2 O), iron (III) nitrate nonahydrate (Fe(NO 3 ) 3 ·9H 2 O), and urea. All chemical reagents used were purchased from the Dinâmica (Brazil), with purities between 98% and 99% [30]. The catalyst's performance was evaluated on its ability to convert frying oil residual into FAAE via simultaneous transesterification and esterification reactions (TES reactions). The residual soybean oil used, originating from fried pasta, with temperatures close to 180 • C and an average usage time of 3 h, was collected from pastry shops in Campina Grande, located in Paraíba state, Brazil. The oil was filtered using white strip filter paper. The physicochemical parameters of the residual oil were accomplished in agreement with the AOCS Cd 3d-63 standard, and the result showed a value of 14.8 ± 0.005 mg of KOH/g of the sample, methyl alcohol (CH 3 OH)-purity 99.8% (Dinâmica), and ethyl alcohol (CH 3 CH 2 OH)-purity 99.5% (Dinâmica) [30].

Catalyst Synthesis
The combustion reactions were accomplished in a pilot plant built in agreement with the patent BR 10 2012 002181-3 [30,31]. The pilot plant comprised a stainless-steel container connected to a conical reactor with a 200 g/batch capacity. The system reaches a maximum temperature equal to 350 • C after 60 min.
Before the synthesis of the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 magnetic catalyst, the initial composition of the precursor solution was calculated based on the total valence of oxidizing agents and reducing reagents using propellant and explosive chemistry concepts. The autoignition procedure (combustion) was performed in the stoichiometric mixture of metallic nitrates and urea allocated in a stainless-steel container in a conical reactor, with a production capacity of 200 g/batch [32].

Catalyst Heat Treatment
The catalyst obtained from the combustion reaction in polyphasic form was subjected to a thermal treatment at different temperatures (650, 700, and 800 • C), with a heating rate of 30 • C/min and 5 min of a burning level using an EDG 3000 oven, to evaluate the evolution of crystalline phases of the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 -Fe 2 O 3 system [30].

Catalytic Test
The performance of the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 -Fe 2 O 3 magnetic catalyst was evaluated in the synthesis of FAAEs from residual oil via the TES reaction. Before synthesis, the residual oil was filtered (filter paper 15.00 ± 0.15 cm) to remove the suspended particulate matter. The catalytic tests were conducted in duplicates using a pressurized stainless-steel reactor equipped with a pressure gauge, a thermocouple inlet duct, and a borosilicate glass (80 mL). The experiment conditions were 30 g oil mass, time 1 h, and an alcohol/oil ratio of 15:1. The heating and agitation of the system were carried out with the aid of an IKA C-MAG HS 7 plate model, external electrical resistance, and a magnetic bar of approximately 2.5 cm. After the reactions, the catalytic tests' products were centrifuged to separate the catalyst, purified, and dried in an oven at 110 • C for 20 min, with manual stirring at 5-min intervals [30].

Characterization of the Catalyst
The ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 magnetic catalyst synthesized was characterized by X-ray diffraction (XRD) using a BRUKER X-ray diffractometer (model D2 PHASER, Cu-Kα radiation), operating at Sustainability 2020, 12, 10159 4 of 18 30 kV and 10 mA. The angular step and counting time used were 0.016 • and 44 min, respectively [33]. The crystallite size was calculated with the aid of the Scherrer equation [30,34] and from the peak of the most intense basal reflection, see Equation (1).
where θ is the wavelength of the X-ray beam, β is full width at half maximum (FWHM), h is the Bragg scattering angle, and K (=0.89) is the shape factor. The identification of the main crystalline phases was made with the DiffracPlus Suite Eva software and Joint Committee on Powder Diffraction Standards (JCPDS). The quantification of each main crystalline phase was carried out by Rietveld refinement [30,35] with the aid of DIFFRACplus TOPAS Version 3.0 (Bruker AXS). The residual error of the Rietveld refinement was calculated from Equation (2). On Equation (2), I Obs and I Calc are the observed and calculated intensities for each step, respectively. The W i parameter is calculated from the equation W i = 1/I Obs .
The catalyst surface was characterized using the nitrogen gas adsorption and desorption technique. All experiments were carried out in Quantachorme model NOVA 3200 equipment. The surface area and pore diameter were calculated using the Brunauer, Emmett, and Teller (BET) and Brunauer, Joyner, and Halenda (BJH) methods, respectively [30].
The morphological aspects of the catalyst sample were acquired by scanning electron microscopy (SEM; brand Tescan, model Vega3). The laser diffraction technique was used to measure particle size distribution using an SZ-100 series (HORIBA Scientific) nanoparticle analyzer [30].
Hysteresis plots were measured at room temperature using a vibrating sample magnetometer (VSM, Lake Shore model 7404), with a maximum applied magnetic field of 13700 G. Saturation magnetization (M s ), remaining magnetization (M r ), and coercive field (H c ) were the properties obtained from this experiment [30].
The acidity of the catalyst was determined through desorption analysis at the programmed ammonia temperature (TPD-NH 3 ) in the SAMP3 multipurpose analysis system. Approximately 100 mg of sample was pretreated at 400 • C under a helium atmosphere (30 mL·min −1 ). Then, the temperature was reduced to 100 • C, and the sample was subjected to an ammonia current for chemical adsorption for 45 min. In the final step of the adsorption process, NH 3 molecules were removed at 100 • C for 1 h, with the helium flow rate at 30 mL·min −1 . The thermograms were obtained on heating (from 100 to 800 • C) at 10 • C·min −1 under a helium flow rate of 30 mL·min −1 [30].
FTIR spectra between 4000 and 200 cm −1 were obtained using a BRUKER Vertex 70 spectrometer, with 64 scans. The FTIR spectra were recorded in transmittance units in the wavenumber range.
Raman spectra were recorded on a RENISHAW spectrophotometer (model InVia Raman microscope) using an Ar + laser, with a power of 100 mW and a wavelength of 514 nm. The 50x objective was used.
The percentages of methyl or ethyl esters were determined via gas chromatography, using a chromatograph (VARIAN 450c) instrument with a flame ionization detector and a capillary column as the stationary phase (Varian Ultimetal "Select Biodiesel Glycerides RG"; dimensions: 15 m × 0.32 mm × 0.45 mm). The initial injection temperature was 100 • C, the oven temperature was 180 • C, and the detector operated at a temperature of 380 • C [30].
The acidity index (official AOCS method, Cd 3d-63) was used to characterize both the residual oil and the products resulting from the catalytic tests. It was possible to quantify the mass yield of synthesized FAAEs, considering the initial mass of the residual oil in the TES reaction, assuming that the complete reaction of a specified amount (x) of residual oil leads to the achievement of 100% yield mass (X) of FAAEs. Therefore, the percentages of mass yields were defined and calculated as the values that express the reactions' final products' masses after the purification processes [30].

Results and Discussion
3.1. Characterization of the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 Catalyst 3.1.1. X-ray Diffraction Figure 1 shows the X-ray diffractograms of the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 catalyst heat-treated at different temperatures (650, 700, and 800 • C). In summary, the synthesized catalyst showed a heterogeneous composition, confirming its polyphasic character (Figure 1a). It was also observed that the reverse spinel phase of Ni-Zn ferrite was a major phase, and there an evolution in this phase with an increase in the firing temperature (Figure 1b,c). When the firing temperature increased from 650 to 700 • C, there was a decrease in the secondary phases of ZnO (Z) and Fe 2 O 3 (H), as well as a significant crystallization of the Ni-Zn phase. However, the catalysts treated at these temperatures (650 and 700 • C) were still polyphasic, with discrete peaks in the secondary phases ( Figure 1 shows the X-ray diffractograms of the ZnO-Ni0.5Zn0.5Fe2O4 catalyst heat-treated at different temperatures (650, 700, and 800 °C). In summary, the synthesized catalyst showed a heterogeneous composition, confirming its polyphasic character (Figure 1a). It was also observed that the reverse spinel phase of Ni-Zn ferrite was a major phase, and there an evolution in this phase with an increase in the firing temperature (Figure 1b,c). When the firing temperature increased from 650 to 700 °C, there was a decrease in the secondary phases of ZnO (Z) and Fe2O3 (H), as well as a significant crystallization of the Ni-Zn phase. However, the catalysts treated at these temperatures (650 and 700 °C) were still polyphasic, with discrete peaks in the secondary phases (ZnO and Fe2O3). The Fe2O3 phase was identified using the standard JCPDS 40-1139 card, while the ZnO phase was detected according to the standard card JCPDS 36-1451.  Figure 1 shows that the phase of interest was synthesized at relatively low temperatures since Olhero et al. [39] synthesized Ni-Zn ferrite by coprecipitation and evaluated the evolution of the   Figure 1 shows that the phase of interest was synthesized at relatively low temperatures since Olhero et al. [39] synthesized Ni-Zn ferrite by coprecipitation and evaluated the evolution of the crystalline phases of the system. The authors found that the phase of interest only formed at 1200 • C. On the other hand, Hajalilou et al. [40] developed a Ni-Zn system by high-energy grinding; however, the experimental procedure has a relatively high cost and takes a long time since that the samples wait for an estimated time of 30 h to be subjected to a high-energy grind in a controlled atmosphere of oxygen for 2 h, with the phase of interest being obtained at 500 • C.
Still, in Figure 1, it was observed that a firing temperature equal to 800 • C was enough to obtain a catalyst with a composition formed mostly of the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 phase (Figure 1d). It is worth mentioning that the firing time used was only 5 min. The synthesis was simple, fast, and does not add high energy expenditure nor high costs, as described in the literature [39,40]. The heat treatment performed at 800 • C was sufficient to obtain the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 biphasic catalyst, with a high degree of crystallinity. Thus, it is possible to conclude that the experimental procedure (combustion reaction in pilot-scale) used to obtain the biphasic catalyst has technological potential. The crystallite size and crystallinity of the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 biphasic catalyst, heat-treated at different temperatures (650, 700, and 800 • C), are listed in Table 1. The ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 catalyst, heat-treated at 800 • C, presented an average crystallinity of 54%, close to the experimental values found in the literature [41]. The crystallite size, determined from the XRD data, was 32 nm, indicating its nanometric character (≤100 nm) [12]. The crystallite size of the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 catalyst is within the crystallite size range of 7.23-46 nm, as reported by Li et al. [42], who studied a similar system (ZnFe 2 O 4 ) synthesized by the combustion method. It is also possible to notice that the heat treatment applied slightly increased the structural parameters of the catalyst, evolving from 43.0% to 54.7% in crystalline percentage and from 25.0 to 32 nm in terms of crystallite size. Figure 2 shows the X-ray diffractogram and Rietveld analysis of the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 catalyst synthesized at 800 • C. The experimental points (blue and black) were adjusted to the intensity (red line) with the aid of the TOPAS program, according to ICDD card 184163 for Ni-Zn ferrite and ICSD card 26170 for ZnO. It was observed that the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 catalyst (800 • C) presented peaks that are characteristic of the major phase of spinel (Ni-Zn ferrite). However, the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 system presented a higher percentage of the Ni-Zn phase and, at the same time, a nonsignificant amount of zinc oxide (0.46%).
The network parameters and percentage quantities of the phases in the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 catalyst (800 • C) are shown in Table 2. It was observed that the calculated value of the lattice parameter was 8.38 Å, which is close to the theoretical (8.3827 Å; JCPDS card 52-0278) and experimental values (8.35 Å; [43]) already reported. Additionally, it was possible to observe the maintenance of the unit cell with cubic symmetry, suggesting that there was no structural change due to the high proportion of Ni-Zn ferrite. Regarding phase quantification, the presence of the Ni-Zn ferrite phase in high quantity (99.54%) and segregated ZnO phases (0.46%) were identified in the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 catalyst that was heat-treated at 800 • C, which is in agreement with the XRD results (see Figure 1). Sustainability 2020, 12, x FOR PEER REVIEW 7 of 19 Figure 2. Refined X-ray diffractogram for the ZnO-Ni0.5Zn0.5Fe2O4 catalyst (800 °C).
The network parameters and percentage quantities of the phases in the ZnO-Ni0.5Zn0.5Fe2O4 catalyst (800 °C) are shown in Table 2. It was observed that the calculated value of the lattice parameter was 8.38 Å, which is close to the theoretical (8.3827 Å; JCPDS card 52-0278) and experimental values (8.35 Å; [43]) already reported. Additionally, it was possible to observe the maintenance of the unit cell with cubic symmetry, suggesting that there was no structural change due to the high proportion of Ni-Zn ferrite. Regarding phase quantification, the presence of the Ni-Zn ferrite phase in high quantity (99.54%) and segregated ZnO phases (0.46%) were identified in the ZnO-Ni0.5Zn0.5Fe2O4 catalyst that was heat-treated at 800 °C, which is in agreement with the XRD results (see Figure 1). The goodness-of-it (GOF) and the weighted (Rwp) and expected (Rexp) indices are the parameters that indicate quality for Rietveld refinement. When Rwp and Rexp values are close, they have the same trend, so it is possible to consider the refinement as satisfactory. Finally, refinement is considered satisfactory if the goodness-of-fit (GOF) is less than 1.7. In practice, values below 5 reflect an optimized refinement, as reported in the literature [44,45]. Therefore, based on the refinement results showed in Table 2, it can be considered that the results obtained were satisfactory for the ZnO-Ni0.5Zn0.5Fe2O4 catalyst.  The goodness-of-it (GOF) and the weighted (Rwp) and expected (Rexp) indices are the parameters that indicate quality for Rietveld refinement. When R wp and R exp values are close, they have the same trend, so it is possible to consider the refinement as satisfactory. Finally, refinement is considered satisfactory if the goodness-of-fit (GOF) is less than 1.7. In practice, values below 5 reflect an optimized refinement, as reported in the literature [44,45]. Therefore, based on the refinement results showed in Table 2, it can be considered that the results obtained were satisfactory for the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 catalyst. Figure 3 shows the adsorption/desorption isotherms of N 2 due to the textural characterization of the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 catalyst (800 • C). It is observed that the catalyst presented a subtle hysteresis, which is characteristic of mesoporous materials. Besides the existence of hysteresis, mesoporous materials are also characterized by the absence of nitrogen adsorption limitation, i.e., P/P0 value is high [41,46]. According to the classification recommended by IUPAC, this isotherm behavior is of type III, which characterizes an adsorption process that occurs in nonporous or macroporous materials [30].

Texture Analysis
Concerning the hysteresis cycle (Figure 3), it is possible to observe a practically constant level for P/P0 values up to 0.2 cm 3 /g. After that, a branch of hysteresis in the region of high relative pressure, with a very steep slope, is observed. Such behavior is characteristic of an H 3 type loop profile, generally associated with the nonrigid aggregates of particles, as seen later in the SEM imaging, originating pores in the slits. The literature [46] indicates that when the isotherm presents an inflection of approximately 0.2 cm 3 /g (relative pressure (P/P 0 )), it indicates the presence of micropores in the morphological structure.
Sustainability 2020, 12, x FOR PEER REVIEW 8 of 19 Figure 3 shows the adsorption/desorption isotherms of N2 due to the textural characterization of the ZnO-Ni0.5Zn0.5Fe2O4 catalyst (800 °C). It is observed that the catalyst presented a subtle hysteresis, which is characteristic of mesoporous materials. Besides the existence of hysteresis, mesoporous materials are also characterized by the absence of nitrogen adsorption limitation, i.e., P/P0 value is high [41,46]. According to the classification recommended by IUPAC, this isotherm behavior is of type III, which characterizes an adsorption process that occurs in nonporous or macroporous materials [30]. Concerning the hysteresis cycle (Figure 3), it is possible to observe a practically constant level for P/P0 values up to 0.2 cm 3 /g. After that, a branch of hysteresis in the region of high relative pressure, with a very steep slope, is observed. Such behavior is characteristic of an H3 type loop profile, generally associated with the nonrigid aggregates of particles, as seen later in the SEM imaging, originating pores in the slits. The literature [46] indicates that when the isotherm presents an inflection of approximately 0.2 cm 3 /g (relative pressure (P/P0)), it indicates the presence of micropores in the morphological structure.
The average value of the specific surface area of the ZnO-Ni0.5Zn0.5Fe2O4 catalyst (800 °C) was around 14.8 m 2 g −1 , which is lower than the experimental values reported in the literature for NiFe2O4 [47]. However, for the ZnO-Ni0.5Zn0.5Fe2O4 catalyst, when subjected to heat treatment, large particles are formed. This occurs due to the diffusion of atoms and the formation of the neck between the small and large particles, generating the growth of the particles [48,49]. Moreover, the pore diameter (3.37 nm) and the isotherms' profiles corroborate to indicate that the synthesized catalyst has a disorderly surface characteristic, with nonporous regions and other regions that present mesopores or macropores of different shapes and sizes [50]. Figure 4 shows the SEM images acquired from the ZnO-Ni0.5Zn0.5Fe2O4 catalyst that was heattreated at 800 °C. It was observed that the ZnO-Ni0.5Zn0.5Fe2O4 catalyst presented irregular agglomerates (Figure 4a) with pores in its surface. Clusters (Figure 4a) and particles (Figure 4b) were observed in the SEM images, with sizes close to 100 and 10 μm, respectively. The diffusion of small particles occurred with the consequent agglomeration of them on the large particles' surface. This indicates that the heat treatment caused the diffusion of small particles and, consequently, caused the  [47]. However, for the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 catalyst, when subjected to heat treatment, large particles are formed. This occurs due to the diffusion of atoms and the formation of the neck between the small and large particles, generating the growth of the particles [48,49]. Moreover, the pore diameter (3.37 nm) and the isotherms' profiles corroborate to indicate that the synthesized catalyst has a disorderly surface characteristic, with nonporous regions and other regions that present mesopores or macropores of different shapes and sizes [50]. Figure 4 shows the SEM images acquired from the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 catalyst that was heat-treated at 800 • C. It was observed that the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 catalyst presented irregular agglomerates (Figure 4a) with pores in its surface. Clusters (Figure 4a) and particles (Figure 4b) were observed in the SEM images, with sizes close to 100 and 10 µm, respectively. The diffusion of small particles occurred with the consequent agglomeration of them on the large particles' surface. This indicates that the heat treatment caused the diffusion of small particles and, consequently, caused the growth of the agglomerates, corroborating with the results of the specific surface area presented in the textural analysis. Figure 5 shows the distribution and accumulated distribution curves measured from the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 catalyst synthesized at 800 • C. It was observed that the accumulative distribution curve of the clusters (blue curve) and the frequency histogram of the distribution of the populations of clusters present the same particle diameter (histogram in black). In summary, the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 catalyst, heat-treated at 800 • C, showed a relatively wide particle size distribution, corroborating with the SEM images ( Figure 4). In addition, symmetrical and Sustainability 2020, 12, 10159 9 of 18 monomodal behavior was identified, with an average diameter (50% in the accumulated curve) of 39.58 nm. When this value is compared with the crystallite size (32 nm), it appears that the value obtained in the particle size distribution for D50 was higher. This indicates that the synthesized material has irregular nanometer-sized clusters.   Figure 5 shows the distribution and accumulated distribution curves measured from the ZnO-Ni0.5Zn0.5Fe2O4 catalyst synthesized at 800 °C. It was observed that the accumulative distribution curve of the clusters (blue curve) and the frequency histogram of the distribution of the populations of clusters present the same particle diameter (histogram in black). In summary, the ZnO-Ni0.5Zn0.5Fe2O4 catalyst, heat-treated at 800 °C, showed a relatively wide particle size distribution, corroborating with the SEM images ( Figure 4). In addition, symmetrical and monomodal behavior was identified, with an average diameter (50% in the accumulated curve) of 39.58 nm. When this value is compared with the crystallite size (32 nm), it appears that the value obtained in the particle size distribution for D50 was higher. This indicates that the synthesized material has irregular nanometer-sized clusters.

Fourier Transform Infrared Spectroscopy (FTIR) and
Raman Figure 6a,b shows the FTIR and Raman spectrums measured from the ZnO-Ni0.5Zn0.5Fe2O4 catalyst, heat-treated at 800 °C. Figure 6a shows bands below 1000 cm −1 , characteristic of ferrites with a spinel-like crystalline structure [51]. Bands at 537 and 449 cm −1 were also observed and are attributed to the tetrahedral and octahedral forms of ZnO-Ni0.5Zn0.5Fe2O4, which is in agreement with the XRD experiments (see Figure 1). As shown in the literature [52], the vibrational frequency around 600-500 cm − Figure 6a shows bands below 1000 cm −1 , characteristic of ferrites with a spinel-like crystalline structure [51]. Bands at 537 and 449 cm −1 were also observed and are attributed to the tetrahedral and octahedral forms of ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 , which is in agreement with the XRD experiments (see Figure 1). As shown in the literature [52], the vibrational frequency around 600-500 cm −1 is related to the Fe 3+ O 2− complex at the tetrahedral site. On the other hand, the vibrational frequency around 450-350 cm −1 is related to the Fe 3+ O 2− and M 2+ O 2− complexes (where M 2+ = Zn 2+ , Mn 2+ , Fe 2+ ) at the octahedral site. 3.1.5. Fourier Transform Infrared Spectroscopy (FTIR) and Raman Figure 6a,b shows the FTIR and Raman spectrums measured from the ZnO-Ni0.5Zn0.5Fe2O4 catalyst, heat-treated at 800 °C. Figure 6a shows bands below 1000 cm −1 , characteristic of ferrites with a spinel-like crystalline structure [51]. Bands at 537 and 449 cm −1 were also observed and are attributed to the tetrahedral and octahedral forms of ZnO-Ni0.5Zn0.5Fe2O4, which is in agreement with the XRD experiments (see Figure 1). As shown in the literature [52], the vibrational frequency around 600-500 cm −   In Figure 6b, the bands near 200 (F 2g1 ), 310 (E g ), 467 (F 2g2 ), and 672 (A 1g ) cm −1 are characteristic of the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 catalyst [53]. The A 1g mode is related to the symmetrical stretching of the oxygen atoms in the Fe-O bonds (tetrahedral bonds). The E g and F 2g1 bands are related to the symmetrical flexion of oxygen concerning Fe and the entire tetrahedron's translational movement. Lastly, the F 2g2 band occurs due to the asymmetric elongation of the Fe-O bond. Raman spectrum agrees with the XRD results since it was also possible to identify the major presence of ferrite. However, unlike the Rietveld analysis, bands related to the ZnO phase were not identified. This is probably related to the low concentration of that phase in the synthesized catalyst at 800 • C.
3.1.6. Thermogravimetry (TG) Figure 7 shows the TG and DTG curves measured from the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 catalyst, heat-treatment at 800 • C. The thermal profile showed that it was possible to identify three decomposition stages, with a total loss of mass of 2.1%. The first stage occurred in the range of 36-100 • C and is related to spontaneous evaporation of the free water absorbed on the catalyst surface. The second stage occurred at 211-314 • C and corresponded to the loss of mass due to the continuous oxidation of organic matter and inorganic salt decomposition. The third stage occurred between 546 and 596 • C, corroborating the mass loss seen in the literature [54].

Magnetic Evaluation
The magnetic behavior of the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 catalyst, heat-treated at 800 • C, was evaluated from the magnetization (M) versus magnetic field (H) curve (see Figure 8). Due to the low values of remaining magnetization (6.0 my/g) and coercivity (0.06 (G)), it is possible to see that the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 catalyst presents a soft magnetic behavior (i.e., easily magnetizes and demagnetizes. The hysteresis curve presents a well-defined S shape and a very narrow internal area, suggesting the catalyst studied has ferrimagnetic behavior [55]. The maximum saturation magnetization was 59.4 emu/g, higher than the values already reported in the literature [12]. The heat treatment applied to synthesize the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 catalyst favored a considerable increase in saturation magnetization. 3.1.6. Thermogravimetry (TG) Figure 7 shows the TG and DTG curves measured from the ZnO-Ni0.5Zn0.5Fe2O4 catalyst, heattreatment at 800 °C. The thermal profile showed that it was possible to identify three decomposition stages, with a total loss of mass of 2.1%. The first stage occurred in the range of 36-100 °C and is related to spontaneous evaporation of the free water absorbed on the catalyst surface. The second stage occurred at 211-314 ° C and corresponded to the loss of mass due to the continuous oxidation of organic matter and inorganic salt decomposition. The third stage occurred between 546 and 596 °C, corroborating the mass loss seen in the literature [54].

Magnetic Evaluation
The magnetic behavior of the ZnO-Ni0.5Zn0.5Fe2O4 catalyst, heat-treated at 800 °C, was evaluated from the magnetization (M) versus magnetic field (H) curve (see Figure 8). Due to the low values of remaining magnetization (6.0 my/g) and coercivity (0.06 (G)), it is possible to see that the ZnO-Ni0.5Zn0.5Fe2O4 catalyst presents a soft magnetic behavior (i.e., easily magnetizes and demagnetizes. The hysteresis curve presents a well-defined S shape and a very narrow internal area, suggesting the catalyst studied has ferrimagnetic behavior [55]. The maximum saturation magnetization was 59.4 emu/g, higher than the values already reported in the literature [12]. The heat treatment applied to synthesize the ZnO-Ni0.5Zn0.5Fe2O4 catalyst favored a considerable increase in saturation magnetization.  In general, Ni-Zn-based ferrites are ferrimagnetic; that is, they exhibit magnetic properties in the presence of an external magnetic field. However, when the external magnetic field is removed, it is impossible to observe coercivity or the remaining magnetization. In this way, Ni-Zn-based ferrites are easily removed from the suspension by applying an external magnetic field during the product separation process. Magnetic separation is a relevant alternative to filtration and/or centrifugation since it prevents the catalyst's loss and increases the capacity for reuse. As a consequence, there is a cost-benefit of the catalyst for industrial-scale production. 3.1.8. NH3 Temperature Programmed Desorption Analysis (TPD) In general, Ni-Zn-based ferrites are ferrimagnetic; that is, they exhibit magnetic properties in the presence of an external magnetic field. However, when the external magnetic field is removed, it is impossible to observe coercivity or the remaining magnetization. In this way, Ni-Zn-based ferrites are easily removed from the suspension by applying an external magnetic field during the product separation process. Magnetic separation is a relevant alternative to filtration and/or centrifugation since it prevents the catalyst's loss and increases the capacity for reuse. As a consequence, there is a cost-benefit of the catalyst for industrial-scale production.

NH 3 Temperature Programmed Desorption Analysis (TPD)
The active acid sites of the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 catalyst (800 • C) were determined using TPD-NH 3 analysis (see Figure 9). It is possible to see two NH 3 desorption peaks in the TPD-NH 3 analysis, where the first peak (~218 • C, corresponding to weak and moderate acidic sites) and the second peak (~525 • C) are attributed to acidic sites of a strong nature. These results are in agreement with the literature [56,57]. The acidity of the ZnO-Ni0.5Zn0.5Fe2O4 catalyst, heat-treated at 800 °C, was calculated by integrating the deconvolution area from the TPD-NH3 curve. The result indicated two types of NH3 desorption sites: the first peak refers to weak and moderate acid sites, with a temperature range between 100 °C and 350 °C; the other peak refers to the strong acidity sites that were identified between 450 °C and 650 °C, corroborating with the literature [58]. Finally, the results obtained from the TPD-NH3 curve indicate a concentration of weak-moderate and strong acid sites, with 8.59 μmol/g and 10.13 μmol/g of NH3, respectively. Therefore, the studied Ni0.5Zn0.5Fe2O4 catalyst showed a total acidity of 18.72 μmol/g of NH3. Thus, it is possible to conclude that the Ni0.5Zn0.5Fe2O4 magnetic catalyst has an acidic character. Figure 10 shows the catalytic performance, acidity reduction, and mass yield of the ZnO-Ni0.5Zn0.5Fe2O4 catalyst in TES reactions of the residual oil, performed in the presence of ethanol and methanol. At this moment, the dependent variable is the alcoholic route, with the following reaction conditions fixed-alcohol/oil ratio of 1/15, residual oil mass of 30 g, the reaction time of 1 h, temperature 200 °C, and 3% catalyst percentage (0.9 g). The ester conversion was 73%, and the mass yield was around 80% for the methyl alcoholic route. The ethyl route showed lower conversions, with 59.1% for the ester conversion and mass yield in approximately 70%. The acidity of the FAAEs produced reduced significantly when using the methyl alcoholic route (reaching ~90%), while the ethyl route reached −41%, suggesting that a large part of the available raw material reacted to the test conditions. The acidity of the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 catalyst, heat-treated at 800 • C, was calculated by integrating the deconvolution area from the TPD-NH 3 curve. The result indicated two types of NH 3 desorption sites: the first peak refers to weak and moderate acid sites, with a temperature range between 100 • C and 350 • C; the other peak refers to the strong acidity sites that were identified between 450 • C and 650 • C, corroborating with the literature [58]. Finally, the results obtained from the TPD-NH 3 curve indicate a concentration of weak-moderate and strong acid sites, with 8.59 µmol/g and 10.13 µmol/g of NH 3 , respectively. Therefore, the studied Ni 0.5 Zn 0.5 Fe 2 O 4 catalyst showed a total acidity of 18.72 µmol/g of NH 3 . Thus, it is possible to conclude that the Ni 0.5 Zn 0.5 Fe 2 O 4 magnetic catalyst has an acidic character. Figure 10 shows the catalytic performance, acidity reduction, and mass yield of the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 catalyst in TES reactions of the residual oil, performed in the presence of ethanol and methanol. At this moment, the dependent variable is the alcoholic route, with the following reaction conditions fixed-alcohol/oil ratio of 1/15, residual oil mass of 30 g, the reaction time of 1 h, temperature 200 • C, and 3% catalyst percentage (0.9 g). The ester conversion was 73%, and the mass yield was around 80% for the methyl alcoholic route. The ethyl route showed lower conversions, with 59.1% for the ester conversion and mass yield in approximately 70%. The acidity of the FAAEs produced reduced significantly when using the methyl alcoholic route (reaching~90%), while the ethyl route reached −41%, suggesting that a large part of the available raw material reacted to the test conditions. Sustainability 2020, 12, x FOR PEER REVIEW 14 of 19 Figure 10. Results of conversion to ester, yield in mass, and reduction of acidity in TES reactions in the presence of the ZnO-Ni0.5Zn0.5Fe2O4 catalysts using residual oil.

Catalytic Performance
In the TES reactions tested, there was a difference of approximately 14.29% in the conversion values in favor of methanol. Such behavior is probably due to the physical and chemical nature (short-chain and polarity) of methanol, leading to higher conversions than alcohol. Therefore, the type of alcohol used can be one factor for the disparity in the conversions resulting from the transesterification and esterification reactions since the other parameters involved in the processes, such as temperature and time, were kept unchanged [59]. Methanol is frequently used to obtain shortchains and polar, and the fact that it is free of water makes it easier to separate esters and glycerin, culminating in greater conversions. Thus, it is possible that the water present in ethanol interfered, for example, in the TES reaction by ethyl route, reducing the final reaction product (fatty acid ethyl esters). Consequently, it provided the formation of an emulsion in the mixture of esters due to the formation of insoluble fatty acids during the reaction process.
The results obtained in this work corroborate with those reported by Dantas et al. (2017) [60], in which 85% ester conversions were verified for transesterification of soybean oil. Mapossa et al. (2017) [12] found a low conversion to fatty acid ethyl esters when they tested the alcoholic ethylic route by esterifying the mixture of soybean oil oleic acid, showing conversions in the ester of 44.49, 43.09, and 43.95%, respectively.
It is important to note that the quality standards of Brazilian biodiesel are established according to ANP 07/2008 (Brazil) and EN 14214 (European Union). These two standards determine the minimum and maximum concentration values for the different species present in the biodiesel produced. The ester content determined by following Brazilian and European standards is at least 96.5% of the FAAEs. Therefore, this study has shown that the maximum conversion of the ester content obtained was 73%, which is close to the value established by regulatory standards, leading to the belief that if the reaction conditions used are changed, such as the molar ratio of alcohol:oil, amount of catalyst, and time and temperature, they can be potential candidates to optimize the biodiesel production process. Through this work, it was possible to develop the final product with a wide range of applications, contributing to the discovery and expansion of heterogeneous magnetic catalysts on a nanoscale, where the cost is relatively low, especially when using technological products aimed at society itself. Figure 11 presents profiles of the percentage values of glyceride composition that was not converted in the reaction. The values were identified and quantified by gas chromatography analysis. In the TES reactions tested, there was a difference of approximately 14.29% in the conversion values in favor of methanol. Such behavior is probably due to the physical and chemical nature (short-chain and polarity) of methanol, leading to higher conversions than alcohol. Therefore, the type of alcohol used can be one factor for the disparity in the conversions resulting from the transesterification and esterification reactions since the other parameters involved in the processes, such as temperature and time, were kept unchanged [59]. Methanol is frequently used to obtain short-chains and polar, and the fact that it is free of water makes it easier to separate esters and glycerin, culminating in greater conversions. Thus, it is possible that the water present in ethanol interfered, for example, in the TES reaction by ethyl route, reducing the final reaction product (fatty acid ethyl esters). Consequently, it provided the formation of an emulsion in the mixture of esters due to the formation of insoluble fatty acids during the reaction process.
The results obtained in this work corroborate with those reported by Dantas et al. (2017) [60], in which 85% ester conversions were verified for transesterification of soybean oil. Mapossa et al. (2017) [12] found a low conversion to fatty acid ethyl esters when they tested the alcoholic ethylic route by esterifying the mixture of soybean oil oleic acid, showing conversions in the ester of 44.49, 43.09, and 43.95%, respectively.
It is important to note that the quality standards of Brazilian biodiesel are established according to ANP 07/2008 (Brazil) and EN 14214 (European Union). These two standards determine the minimum and maximum concentration values for the different species present in the biodiesel produced. The ester content determined by following Brazilian and European standards is at least 96.5% of the FAAEs. Therefore, this study has shown that the maximum conversion of the ester content obtained was 73%, which is close to the value established by regulatory standards, leading to the belief that if the reaction conditions used are changed, such as the molar ratio of alcohol:oil, amount of catalyst, and time and temperature, they can be potential candidates to optimize the biodiesel production process. Through this work, it was possible to develop the final product with a wide range of applications, contributing to the discovery and expansion of heterogeneous magnetic catalysts on a nanoscale, where the cost is relatively low, especially when using technological products aimed at society itself. Figure 11 presents profiles of the percentage values of glyceride composition that was not converted in the reaction. The values were identified and quantified by gas chromatography analysis. Sustainability 2020, 12, x FOR PEER REVIEW 15 of 19 Figure 11. Result of the percentage of glycerides not converted to esters during the catalytic tests of the ZnO-Ni0.5Zn0.5Fe2O4 catalyst by ethyl (Et) and methyl (Met) routes.
The comparison of Figures 10 and 11 shows that the synthesized catalyst shows catalytic activity, indicating that the esterification and transesterification reactions occur simultaneously. However, the methyl route was more effective in esterification due to the high percentages of free fatty acids consumed in the reaction. The similarity of the mass yield profiles (Figure 10) with the profiles of triglyceride residues not consumed in the reaction ( Figure 11) suggest that the ethyl route has more significant reaction interference in transesterification, directing the reaction to the formation stages of mono-and diacylglycerides in higher percentages than the methyl route. Similar behavior was reported by Dantas et al. 2016 [12], who studied similar heterogeneous magnetic catalytic systems in a TES reaction performed by a mixture of refined soybean oil with oleic acid, which simulates the high acidity characteristic of residual oils due to large quantities of free fatty acids, tested by methyl and ethyl routes, which showed 91.4% and 77.8% conversion to methyl and ethyl esters, respectively.

Conclusions
The combustion reaction was sufficient to obtain the ZnO-Ni0.5Zn0.5Fe2O4 magnetic catalyst and, subsequently, the heat-treatment at 800 °C favored a biphasic transformation. X-ray diffraction confirmed the inverse spinel majority phase's presence, with an average crystallite size of 32 nm. The morphology was investigated by SEM images, which showed the clusters formed by irregular particles, confirming the particle size distribution. The catalyst revealed the typical ferrimagnetic behavior of soft materials, with a high saturation magnetization of 59.4 emu/g. The ZnO-Ni0.5Zn0.5Fe2O4 system was more catalytically active for the methyl route, with an average conversion of 73%, while the ethyl route achieved a conversion of 59.1% into biodiesel. The use of the ZnO-Ni0.5Zn0.5Fe2O4 catalyst for biodiesel production derived from residual oil was effective, with the viability of industrial production. In general, the transformation of an oil residue into biodiesel is important since it adds value to a discarded material. Additionally, the reuse of residual oil minimizes the environmental and social impacts, contributing to sustainable development. The comparison of Figures 10 and 11 shows that the synthesized catalyst shows catalytic activity, indicating that the esterification and transesterification reactions occur simultaneously. However, the methyl route was more effective in esterification due to the high percentages of free fatty acids consumed in the reaction. The similarity of the mass yield profiles (Figure 10) with the profiles of triglyceride residues not consumed in the reaction ( Figure 11) suggest that the ethyl route has more significant reaction interference in transesterification, directing the reaction to the formation stages of mono-and diacylglycerides in higher percentages than the methyl route. Similar behavior was reported by Dantas et al. 2016 [12], who studied similar heterogeneous magnetic catalytic systems in a TES reaction performed by a mixture of refined soybean oil with oleic acid, which simulates the high acidity characteristic of residual oils due to large quantities of free fatty acids, tested by methyl and ethyl routes, which showed 91.4% and 77.8% conversion to methyl and ethyl esters, respectively.

Conclusions
The combustion reaction was sufficient to obtain the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 magnetic catalyst and, subsequently, the heat-treatment at 800 • C favored a biphasic transformation. X-ray diffraction confirmed the inverse spinel majority phase's presence, with an average crystallite size of 32 nm. The morphology was investigated by SEM images, which showed the clusters formed by irregular particles, confirming the particle size distribution. The catalyst revealed the typical ferrimagnetic behavior of soft materials, with a high saturation magnetization of 59.4 emu/g. The ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 system was more catalytically active for the methyl route, with an average conversion of 73%, while the ethyl route achieved a conversion of 59.1% into biodiesel. The use of the ZnO-Ni 0.5 Zn 0.5 Fe 2 O 4 catalyst for biodiesel production derived from residual oil was effective, with the viability of industrial production. In general, the transformation of an oil residue into biodiesel is important since it adds value to a discarded material. Additionally, the reuse of residual oil minimizes the environmental and social impacts, contributing to sustainable development.