Supermagnetic Nano-Bifunctional Catalyst from Rice Husk: Synthesis, Characterization and Application for Conversion of Used Cooking Oil to Biodiesel

The present work investigated the biodiesel production from used cooking oil catalyzed by nano-bifunctional supermagnetic heterogeneous catalysts (RHC/K2O/Fe) derived from rice husk doped with K2O and Fe synthesized by the wet impregnation method. The synthesized catalysts (RHC/K2O/Fe) were characterized for crystallinity by X-ray diffraction spectroscopy (XRD), total acidity and basicity using CO2/NH3-TPD, textural properties through Brunauer-Emmett-Teller (BET), thermal stability via thermogravimetric analyzer (TGA), functional group determination by Fourier-transform infrared spectroscopy (FTIR), surface morphology through field emission scanning electron microscopy (FESEM), and magnetic properties by vibrating sample magnetometer (VSM). The VSM result demonstrated that the super-paramagnetic catalyst (RHC/K2O-20%/Fe-5%) could be simply separated and regained after the reaction using an external magnetic field. The operating conditions such as catalyst loading, methanol/oil molar ratio, temperature, and reaction duration were studied. The screened RHC/K2O-20%/Fe-5% catalyst was selected for further optimization and the optimum reaction parameters found were 4 wt % of catalyst, a molar ratio of methanol/oil of 12:1, 4 h reaction duration, and 75 ◦C reaction temperature with a maximal yield of 98.6%. The reusability study and reactivation results revealed that the nano-bifunctional magnetic catalyst (RHC/K2O-20%/Fe-5%) could be preserved by high catalytic activity even after being reused five times.


Introduction
An increasing supply of petrochemical resources to fulfill the high energy demands for industrialization and metropolitan growth over the last decades has caused the depletion of non-renewable conventional fuels [1,2]. The emission of toxic gases from the burning of fossil carbon with randomly oriented carbon sheets [39]. The two additional peaks of 44.2° and 64.6° on RHC corresponded to the characteristic peaks of SiO2 with the monoclinic phase (ICSD File No. 01-083-1413) that consisted of more ordered graphite-like carbon lattice structure [40]. After loading of K2CO3 and FeCl3 with different concentrations, intensities of the main peaks of RHC decreased gradually and more new peaks exhibited in the XRD pattern. K2O peaks appeared at 2ϴ: 13.3° (ICSD File No. 00-039-1380), 28.8° (ICSD File No. 01-018-0994) and 50.6° (ICSD File No. 00-019-0927). The peak of K2CO3 appeared at 2ϴ: 33.2° (ICSD File No. 00-049-1093) [41]. Meanwhile, peaks related to Fe groups were located at 2ϴ: 31.8° of α-Fe2O3 (ICSD File No. 01-088-2359), 41.0°, and 45.2° of Fe2O3 (ICSD File No.01-075-0449). Both crystal systems of α-Fe2O3 and Fe2O3 were rhombohedral. The intensities of potassium and iron were increased gradually as the dopants' concentrations varied from 20% to 30% and 5% to 10%, respectively [29]. The XRD pattern of RHC/K2O-20/Fe-10% clearly showed that the small diffraction peak of K2CO3 compared to other catalysts due to the surface of K2CO3 particles were dispersed and hindered by Fe particles due to which the X-ray beam unable to diffract on catalysts' surface [42]. Small peaks of Potassium Iron Silicate (K2Fe2SiO29(O4)58) and Potassium Iron Oxide (K3FeO2) were observed at 2ϴ:66.8° (ICSD File No. 00-043-0264) and 74.6° (ICSD File No. 00-048-0957) with cubic and tetragonal phase respectively. Alternatively, the formation of both of the peaks was due to the interaction of K2O, Fe2O3, and SiO2 substances on the catalyst during the calcination process [43]. Both acidic and basic properties of RH, RHC, and RHC derived catalysts were determined using CO2 and NH3 -TPD, respectively, as tabulated in Table 1. Based on the results, the derived catalysts from RHC were exhibited as bifunctional properties that consisted of both basic and acid sites. The patterns showed major CO2 and NH3 desorption at various degrees of temperature that indicated the level of the basic and acidic strength of the catalyst. The peaks that appeared below 360 °C represented a weak desorption site [44] while peaks of moderate and strong desorption site arose at between 360 °C and 700 °C and above 700 °C, respectively [45].

Temperature Programmed Desorption (TPD)
Both acidic and basic properties of RH, RHC, and RHC derived catalysts were determined using CO 2 and NH 3 -TPD, respectively, as tabulated in Table 1. Based on the results, the derived catalysts from RHC were exhibited as bifunctional properties that consisted of both basic and acid sites. The patterns showed major CO 2 and NH 3 desorption at various degrees of temperature that indicated the level of the basic and acidic strength of the catalyst. The peaks that appeared below 360 • C represented a weak desorption site [44] while peaks of moderate and strong desorption site arose at between 360 • C and 700 • C and above 700 • C, respectively [45]. Generally, RH and RHC exhibited weak and moderate strengths of base and acid properties with total desorption, with basic density of 2.07 mmol g −1 , 2.72 mmol g −1 and acidic density 15.57 mmol g −1 , 15.01 mmol g −1 , respectively. As the rice husk char was doped with K 2 CO 3 (from 15% to 30%) and FeCl 3 .6H 2 O (from 5% to 10%), the strength and total basicity and acidity increased significantly. TPD-CO 2 analysis showed the desorption of CO 2 gas at a temperature range of 550 • C to 850 • C, which was maximized at around 707 • C to 788 • C, could be considered as the strong strength of basic sites from K 2 O. The basicity value of the synthesized catalyst loaded with 20% of potassium was significantly higher than lower K 2 O loading. Meanwhile, the further increment of K 2 O loading caused the total absorption-desorption of CO 2 to reduce due to less interaction between the probe gas and fully covered active sites [44,46]. The same goes for the acidity properties of the catalysts; as the loading percentage of K 2 O increased, the total desorption of NH 3 reduced. In addition, the basicity of RHC/K 2 O-20%/Fe-10% reduced after being doped with 10% of Fe due its acidity strength that suppressed the basic site that might neutralize the catalyst and cause no conversion of biodiesel to happen. Besides that, random distributions of excess agglomerated Fe (10% of Fe) on the catalyst might cover the active K-active site on porous RHC support which imbalanced the basicity characteristic, also reduced the catalyst pore size that will prevent the penetration of methanol and triglyceride; hence, it discourage the transesterification to occur [47]. The excess phase of K 2 O and Fe 2 O 3 on the surface of RHC support led to reduction of its porosity, hence decreasing the BET surface area [48]. The biodiesel yield depended on to the basicity of the catalyst where the higher the basicity, the greater the FAME formation, as stated in Table 1, in which RHC/K 2 O-20%/Fe-5% showed better biodiesel yield among others. The further increment of dopant compounds will slightly reduce of catalytic activity of converting triglycerides into fatty acid methyl ester (FAME) which related to the mass transfer and poor mixing between reactants and catalyst performance [49][50][51]. Thus, magnetic RHC/K 2 O-20%/Fe-5% was chosen for further characterization analysis and catalyst optimization in biodiesel production.

Surface Area and Porosity Measurement
Textural properties of rice husk (RH), rice husk char (RHC) and magnetic RHC/K 2 O/Fe catalysts were evaluated by adsorption and desorption of N 2 at 77 K. Table 1 shows that the specific surface area for RH was 3.97 m 2 g −1 whereas the surface is for RHC 203.54 m 2 g −1 after being calcined was at 450 • C. Carbonisation of biomass at high temperature can increase the carbon content and produce high porosity material [52]. A study conducted by Zhao [29] stated the use of potassium hydroxide in pre-treatment as a chemical activation agent to improve the surface area and pore diameter of RHC. A significant reduction of surface area from 203.5 4 m 2 g −1 to 57.89 m 2 g −1 may be due to the surface of catalyst precursor which was covered by potassium and ferric doping, and the decrement of pore volume from 0.1644 cm 3 g −1 to 0.0588 cm 3 g −1 was due to the blockage of catalyst pores which confirmed the impregnation of K and Fe on the catalyst's surface. As shown in Figure S1a, RH, RHC and RHC nano-magnetic catalysts exhibited a Type-IV adsorption isotherms at a lower p/p 0 values with narrow hysteresis loop of desorption, according to the International Union of Pure and Applied Chemistry (IUPAC) classification which proved the presence of mesoporous structure and may be characterized by the formation of slit-shaped pores [53].
The perfect transesterification reaction depends on the pore structure of the heterogeneous catalyst. The size of the distribution of the catalyst is shown in Figure S1b RHC, and RHC nano-magnetic catalysts demonstrated comparable pore size distributions in the mesoporous pore diameter distribution range of 2 nm to 50 nm [50]. Boz et al. [20] reported that the diameter of triglycerides and glycerol molecules was 2.5 nm, which was small enough to pass through into the porous catalysts and would be employed throughout the transesterification reaction.

Thermal Gravimetric Analysis
The thermal stability of RH and RHC/K 2 O-20%/Fe-5% catalyst was ascertained by using thermogravimetric analysis (TGA)/differential thermal analysis (DTA) as presented in Figure 2. Three main stages of thermal decomposition consisted of removal of the adsorbed water molecule at a temperature below 100 • C, followed by the secondary stage of primary pyrolysis temperature range between 200 • C to 650 • C and the third phase of decomposition at high temperature over 700 • C [20], as shown in Figure 3.
The first stage of thermal decomposition of RH at the range of 26 • C to136 • C with a weight loss of 6.59% was associated with the elimination of adsorbed water molecule on the RH surface. The second phase with a temperature range of 152 • C to 303 • C corresponded to the decomposition of hemicellulose and cellulose with a weight loss of 15.18%. The degradation of lignin occurred at a higher temperature range from 303 • C to 633 • C with a major weight loss of 44.40%. The leftover after 632 • C to 991 • C was annotated as solid residue [54][55][56]. The TGA-DTA curves of RHC did not exhibit apparent mass loss event, which demonstrated that the synthesized RHC has good thermal stability. For the supermagnetic nano-bifunctional RHC/K 2 O-20%/Fe-5% catalyst, the first weight loss of 17.11% at about 26 • C to 176 • C may be due to the elimination of adsorbed or chemisorbed H 2 O molecule [57]. Another weight loss of 3.65% between 523 • C to 648 • C was attributed to the dehydroxylation of OH ions of supermagnetic RHC nano-bifunctional catalysts as well as disintegrate of FeCl 3 into its oxide (Fe 2 O 3 ). The decomposition of K 2 CO 3 to K 2 O exhibited a major weight loss of 18.43% at a temperature range between 705 • C to 1000 • C.
Catalysts 2020, 10, x FOR PEER REVIEW 6 of 21 Chemistry (IUPAC) classification which proved the presence of mesoporous structure and may be characterized by the formation of slit-shaped pores [53]. The perfect transesterification reaction depends on the pore structure of the heterogeneous catalyst. The size of the distribution of the catalyst is shown in Figure S1b RHC, and RHC nanomagnetic catalysts demonstrated comparable pore size distributions in the mesoporous pore diameter distribution range of 2 nm to 50 nm [50]. Boz et al. [20] reported that the diameter of triglycerides and glycerol molecules was 2.5 nm, which was small enough to pass through into the porous catalysts and would be employed throughout the transesterification reaction.

Thermal Gravimetric Analysis
The thermal stability of RH and RHC/K2O-20%/Fe-5% catalyst was ascertained by using thermogravimetric analysis (TGA)/differential thermal analysis (DTA) as presented in Figure 2. Three main stages of thermal decomposition consisted of removal of the adsorbed water molecule at a temperature below 100 °C, followed by the secondary stage of primary pyrolysis temperature range between 200 °C to 650 °C and the third phase of decomposition at high temperature over 700 °C [20], as shown in Figure 3.
The first stage of thermal decomposition of RH at the range of 26 °C to136 °C with a weight loss of 6.59% was associated with the elimination of adsorbed water molecule on the RH surface. The second phase with a temperature range of 152 °C to 303 °C corresponded to the decomposition of hemicellulose and cellulose with a weight loss of 15.18%. The degradation of lignin occurred at a higher temperature range from 303 °C to 633 °C with a major weight loss of 44.40%. The leftover after 632 °C to 991 °C was annotated as solid residue [54][55][56]. The TGA-DTA curves of RHC did not exhibit apparent mass loss event, which demonstrated that the synthesized RHC has good thermal stability. For the supermagnetic nano-bifunctional RHC/K2O-20%/Fe-5% catalyst, the first weight loss of 17.11% at about 26 °C to 176 °C may be due to the elimination of adsorbed or chemisorbed H2O molecule [57]. Another weight loss of 3.65% between 523 °C to 648 °C was attributed to the dehydroxylation of OH ions of supermagnetic RHC nano-bifunctional catalysts as well as disintegrate of FeCl3 into its oxide (Fe2O3). The decomposition of K2CO3 to K2O exhibited a major weight loss of 18.43% at a temperature range between 705 °C to 1000 °C.

Functional Group Analysis
The Fourier Transform Infrared (FTIR) spectra of rice husk (RH), rice husk char (RHC), and supermagnetic RHC nano-bifunctional catalyst are illustrated in Figure 3. Broad absorption bands at range 3488.67 cm −1 to 3359.29 cm −1 and 1672.20 cm −1 to 1637.26 cm −1 corresponded to the hydroxyl group -OH and silanol, Si-OH stretching and bending, respectively that indicated physisorption of a water molecule on the surface of RH, RHC and magnetic RHC/K2O-20%/Fe-5% catalyst [58]. The band around 2917.66 cm −1 was assigned to the asymmetric and symmetric methylene stretching vibration of the saturated aliphatic compound present in the cellulose and hemicellulose of RH [59]. A spectra band around 1732.30 cm −1 was associated with the acetyl and ester group of the C=O bond in hemicellulose while lignin infrared spectra appeared at 1509.22 cm −1 , which represented C=C aromatic vibration. The lignin absorption spectra characteristic was found at 1463.53 cm −1 , and cellulose plane bending vibrations, CH2 of RH were clearly visible at 1420.90 cm −1 . The spectra of 1312.94 cm −1 showed the C-H deformation in RH's cellulose and hemicellulose. The high-intensity band spectrum around 1031.81 cm −1 was detected due to the asymmetric and symmetric band stretching of the Si-O-C and Si-O-Si bonds corresponding to the silica and the cellulose of RH. The peaks related to lignin and cellulose disappeared in the RHC spectrum band after the RH calcination due to biomass degradation at high temperatures. New spectra band which corresponded to the asymmetrical stretching vibration of the siloxane, Si-O-Si bond was formed and slightly shifted to 1057.80 cm −1 of RHC after calcination of RH. A sharp peak of symmetric stretching vibration of Si-O and vibration-bending modes of the Si-O-Si bond appeared at 800.29 cm −1 and 466.67 cm −1 in both RH and RHC spectra respectively. The silica functional group spectra remained before and after the calcination process and provided a mesoporous structure for RHC as support [60]. A sharp and intense spectrum peak at 1418.36 cm −1 of RHC/K2O-20%/Fe-5% was attributed to the carbonate species from K2CO3 doped on the RHC support [40]. The peaks at 1331.76 cm −1 and 837.37 cm −1 were assigned to the stretching vibration C-O and the deformation vibration of C-O-C, respectively [61]. The weak vibration of M+-O stretching adsorption bands at 544.80 cm −1 , and 463.04 cm −1 belonged to Fe-O and K-O bonding, respectively [62].The deformation bands of M + -OH converted to Si-O-M + located at 964.71 cm −1 and 912.26 cm −1 were assigned to Fe and K, respectively [63]. Upon the formation of Si-O-M + , the intensity of the siloxane absorption band was reduced and shifted to lower frequency due to the reduction of electron density [64].

Functional Group Analysis
The Fourier Transform Infrared (FTIR) spectra of rice husk (RH), rice husk char (RHC), and supermagnetic RHC nano-bifunctional catalyst are illustrated in Figure 3. Broad absorption bands at range 3488.67 cm −1 to 3359.29 cm −1 and 1672.20 cm −1 to 1637.26 cm −1 corresponded to the hydroxyl group -OH and silanol, Si-OH stretching and bending, respectively that indicated physisorption of a water molecule on the surface of RH, RHC and magnetic RHC/K 2 O-20%/Fe-5% catalyst [58]. The band around 2917.66 cm −1 was assigned to the asymmetric and symmetric methylene stretching vibration of the saturated aliphatic compound present in the cellulose and hemicellulose of RH [59]. A spectra band around 1732.30 cm −1 was associated with the acetyl and ester group of the C=O bond in hemicellulose while lignin infrared spectra appeared at 1509.22 cm −1 , which represented C=C aromatic vibration. The lignin absorption spectra characteristic was found at 1463.53 cm −1 , and cellulose plane bending vibrations, CH 2 of RH were clearly visible at 1420.90 cm −1 . The spectra of 1312.94 cm −1 showed the C-H deformation in RH's cellulose and hemicellulose. The high-intensity band spectrum around 1031.81 cm −1 was detected due to the asymmetric and symmetric band stretching of the Si-O-C and Si-O-Si bonds corresponding to the silica and the cellulose of RH. The peaks related to lignin and cellulose disappeared in the RHC spectrum band after the RH calcination due to biomass degradation at high temperatures. New spectra band which corresponded to the asymmetrical stretching vibration of the siloxane, Si-O-Si bond was formed and slightly shifted to 1057.80 cm −1 of RHC after calcination of RH. A sharp peak of symmetric stretching vibration of Si-O and vibration-bending modes of the Si-O-Si bond appeared at 800.29 cm −1 and 466.67 cm −1 in both RH and RHC spectra respectively. The silica functional group spectra remained before and after the calcination process and provided a mesoporous structure for RHC as support [60]. A sharp and intense spectrum peak at 1418.36 cm −1 of RHC/K 2 O-20%/Fe-5% was attributed to the carbonate species from K 2 CO 3 doped on the RHC support [40]. The peaks at 1331.76 cm −1 and 837.37 cm −1 were assigned to the stretching vibration C-O and the deformation vibration of C-O-C, respectively [61]. The weak vibration of M+-O stretching adsorption bands at 544.80 cm −1 , and 463.04 cm −1 belonged to Fe-O and K-O bonding, respectively [62].The deformation bands of M + -OH converted to Si-O-M + located at 964.71 cm −1 and 912.26 cm −1 were assigned to Fe and K, respectively [63]. Upon the formation of Si-O-M + , the intensity of the siloxane absorption band was reduced and shifted to lower frequency due to the reduction of electron density [64].

Morphological and Energy Dispersive X-ray Spectroscopy Analysis
The surface morphologies of RH, RHC, and supermagnetic RHC/K 2 O-20%/Fe-5% nanobifunctional catalysts were studied using FESEM, as shown in Figure 4. Figure 4a shows that the RH sample consists of an external surface with a well-organized corrugated structure [54]. The asymmetrical porous layer of RH can be observed through the cross-section view with a magnification of 10,000×, as shown in Figure 4b. Highly porous structure with a large surface area of RHC, as shown in Figure 4c, was formed after carbonization due to thermal decomposition of organic material from RH. Figure 4d shows that nano-magnetic RHC/K 2 O-20%/Fe-5% catalyst was made up of the agglomeration of irregular small and large particles, mainly potassium and iron attached to the porous surface of the RHC precursor. The elemental compositions of RH, RHC, and nano-magnetic RHC/K 2 O-20%/Fe-5% were verified by energy-dispersive X-ray spectroscopy analysis (EDX) with a detector mounted on a microscope. The RH was made up of 39.83 wt % C, 53.29 wt % O, 6.56 wt % Si and a trace amount of 0.33 wt % K. Meanwhile, the elemental compositions of RHC were; 61.07 wt % C, 27.50 wt % O, 9.76 wt % Si and 1.67 wt % K after carbonization and pre-treatment process. The nano-magnetic catalyst consisted 41.02 wt % C, 27.25 wt % O, 3.60 wt % Si, 21.34 wt % K, and 2.01 wt % Fe. The interaction between K and Fe of the prepared catalyst resulted in a well-developed porous structure believed to have contributed to its high catalytic activity. It was observed that the presence of K indicated a high basicity of prepared catalyst, which can contribute to the transesterification of used cooking oil. Also, the high oxygen atomic percentage suggested that the metallic and synergetic compound formed consisted of oxygen atom mainly bonded to the KFeSi composite rather than being bonded to individual metal oxides. The experimental weight percentage of potassium in nano-magnetic catalyst exceeded its theoretical weight due to the amount of potassium from the pre-treatment process that was not completely washed off from RHC. The surface morphologies of RH, RHC, and supermagnetic RHC/K2O-20%/Fe-5% nanobifunctional catalysts were studied using FESEM, as shown in Figure 4. Figure 4a shows that the RH sample consists of an external surface with a well-organized corrugated structure [54]. The asymmetrical porous layer of RH can be observed through the cross-section view with a magnification of 10,000×, as shown in Figure 4b. Highly porous structure with a large surface area of RHC, as shown in Figure 4c, was formed after carbonization due to thermal decomposition of organic material from RH. Figure 4d shows that nano-magnetic RHC/K2O-20%/Fe-5% catalyst was made up of the agglomeration of irregular small and large particles, mainly potassium and iron attached to the porous surface of the RHC precursor. The elemental compositions of RH, RHC, and nano-magnetic RHC/K2O-20%/Fe-5% were verified by energy-dispersive X-ray spectroscopy analysis (EDX) with a detector mounted on a microscope. The RH was made up of 39.83 wt % C, 53.29 wt % O, 6.56 wt % Si and a trace amount of 0.33 wt % K. Meanwhile, the elemental compositions of RHC were; 61.07 wt % C, 27.50 wt % O, 9.76 wt % Si and 1.67 wt % K after carbonization and pre-treatment process. The nano-magnetic catalyst consisted 41.02 wt % C, 27.25 wt % O, 3.60 wt % Si, 21.34 wt % K, and 2.01 wt % Fe. The interaction between K and Fe of the prepared catalyst resulted in a well-developed porous structure believed to have contributed to its high catalytic activity. It was observed that the presence of K indicated a high basicity of prepared catalyst, which can contribute to the transesterification of used cooking oil. Also, the high oxygen atomic percentage suggested that the metallic and synergetic compound formed consisted of oxygen atom mainly bonded to the KFeSi composite rather than being bonded to individual metal oxides. The experimental weight percentage of potassium in nanomagnetic catalyst exceeded its theoretical weight due to the amount of potassium from the pretreatment process that was not completely washed off from RHC.

Vibrating Sampling Magnetometer Analysis
The magnetic property of RHC/K 2 O-20%/Fe-5% was studied using a vibrating sampling magnetometer (VSM), as shown in Figure 5. The M-H loops of fresh magnetic catalyst showed strong magnetization of 7.88 emug −1 , and there was a closed hysteresis loop of the catalysts which can be concluded as a superparamagnetic material at room temperature that can be easily separated from the biodiesel reaction system by introducing an external magnet [61,65]. The magnetic property of RHC/K2O-20%/Fe-5% was studied using a vibrating sampling magnetometer (VSM), as shown in Figure 5. The M-H loops of fresh magnetic catalyst showed strong magnetization of 7.88 emug −1 , and there was a closed hysteresis loop of the catalysts which can be concluded as a superparamagnetic material at room temperature that can be easily separated from the biodiesel reaction system by introducing an external magnet [61,65].

Effects of Catalyst Amount for Catalyst Screening and Catalytic Performance
The catalytic activity of transesterification of used cooking oil (UCO) was carried out in the presence of synthesized catalysts at the following reaction conditions; methanol to oil molar ratio of 10:1 at a temperature of 65 °C, for 4 h with various catalyst loadings of 1 to 8 wt % ( Figure 6). Among the four catalysts tested, the RHC/K2O-20%/Fe-5% presented the best reaction activity and produced the highest yield of 91.8%; thus, all the optimizations were carried out with this catalyst material. The presence of optimum strength for both acidic and basic properties on the surface of catalysts could affect the formation of highly reactive methoxide species in transesterification. The catalyst of RHC/K2O-15%/Fe-5%, RHC/K2O-25%/Fe-5% and RHC/K2O-30%/Fe-5% likely showed the same result of biodiesel production. Generally, a higher concentration of K-loading is inessential as the best catalyst because the very high basicity characteristic tends to cause difficulty for fatty acid methyl ester (FAME) molecule to desorb after the reaction and vice versa for lower loading concentration in which the reaction itself will not initiate due to low basicity for adsorption of reactant molecules [66]. According to the attained results, the biodiesel yields increased as the catalyst loading was increased from 1 to 4 wt % respectively. Meanwhile, further increase of catalyst loading up to 8 wt % showed the drop of biodiesel yield due to agglomeration of catalyst particles, causing the reactant mixture to become, viscous which led to poor blending between the catalyst and reactants [23]. The result exhibited that 4 wt % of RHC/K2O-20%/Fe-5% was the optimal amount of catalyst loading for the reaction.

Effects of Catalyst Amount for Catalyst Screening and Catalytic Performance
The catalytic activity of transesterification of used cooking oil (UCO) was carried out in the presence of synthesized catalysts at the following reaction conditions; methanol to oil molar ratio of 10:1 at a temperature of 65 • C, for 4 h with various catalyst loadings of 1 to 8 wt % ( Figure 6). Among the four catalysts tested, the RHC/K 2 O-20%/Fe-5% presented the best reaction activity and produced the highest yield of 91.8%; thus, all the optimizations were carried out with this catalyst material. The presence of optimum strength for both acidic and basic properties on the surface of catalysts could affect the formation of highly reactive methoxide species in transesterification. The catalyst of RHC/K 2 O-15%/Fe-5%, RHC/K 2 O-25%/Fe-5% and RHC/K 2 O-30%/Fe-5% likely showed the same result of biodiesel production. Generally, a higher concentration of K-loading is inessential as the best catalyst because the very high basicity characteristic tends to cause difficulty for fatty acid methyl ester (FAME) molecule to desorb after the reaction and vice versa for lower loading concentration in which the reaction itself will not initiate due to low basicity for adsorption of reactant molecules [66]. According to the attained results, the biodiesel yields increased as the catalyst loading was increased from 1 to 4 wt % respectively. Meanwhile, further increase of catalyst loading up to 8 wt % showed the drop of biodiesel yield due to agglomeration of catalyst particles, causing the reactant mixture to become, viscous which led to poor blending between the catalyst and reactants [23]. The result exhibited that 4 wt % of RHC/K 2 O-20%/Fe-5% was the optimal amount of catalyst loading for the reaction.

Effects of Methanol to Oil Molar Ratio
Transesterification reaction requires methanol as the basic reactant to react with UCO, and the molar ratio of methanol to oil is one of the significant parameters which influence the conversion of triglycerides to fatty acid methyl ester (FAME). Basically, the transesterification reaction is an equilibrium reaction where the conversion of biodiesel could be raised by applying the excessive methanol to force the equilibrium to the product formation [67,68]. The relationship between methanol to oil molar ratio is shown in Figure 7a. Based on the experimental data, the conversion rate increased from 71.4% to 92.5% as the methanol to oil molar ratio increased from 6:1 to 12:1, and the best result of the molar ratio was 12:1. Meanwhile, biodiesel production decreased to 70.44% with an additional increase of the molar ratio of methanol to oil to 27:1. The excesses amount of oil to methanol ratio enables the contact among reagents and catalyst to fasten the reaction, whereas, the surplus methanol give negative impact on transesterification because it could diminish the catalyst concentration and UCO that hinders the triglyceride molecules to the active sites, which leads to the reduction of biodiesel yield [69]. Also, the transesterification of excess methanol to oil ratio will favor glycerolysis reaction, which interferes with the phase separation of glycerol due to an increase in glycerol solubility in methanol, which drives the equilibrium to the reactants and reduces the ester yield [29,67]. As per the above results, the 12:1 methanol to oil molar ratio was selected as the optimum ratio.

Effects of Methanol to Oil Molar Ratio
Transesterification reaction requires methanol as the basic reactant to react with UCO, and the molar ratio of methanol to oil is one of the significant parameters which influence the conversion of triglycerides to fatty acid methyl ester (FAME). Basically, the transesterification reaction is an equilibrium reaction where the conversion of biodiesel could be raised by applying the excessive methanol to force the equilibrium to the product formation [67,68]. The relationship between methanol to oil molar ratio is shown in Figure 7a. Based on the experimental data, the conversion rate increased from 71.4% to 92.5% as the methanol to oil molar ratio increased from 6:1 to 12:1, and the best result of the molar ratio was 12:1. Meanwhile, biodiesel production decreased to 70.44% with an additional increase of the molar ratio of methanol to oil to 27:1. The excesses amount of oil to methanol ratio enables the contact among reagents and catalyst to fasten the reaction, whereas, the surplus methanol give negative impact on transesterification because it could diminish the catalyst concentration and UCO that hinders the triglyceride molecules to the active sites, which leads to the reduction of biodiesel yield [69]. Also, the transesterification of excess methanol to oil ratio will favor glycerolysis reaction, which interferes with the phase separation of glycerol due to an increase in glycerol solubility in methanol, which drives the equilibrium to the reactants and reduces the ester yield [29,67]. As per the above results, the 12:1 methanol to oil molar ratio was selected as the optimum ratio.

Effects of Reaction Temperature
Transesterification was performed to determine the optimum reaction temperature by changing temperature range from 35 °C to 105 °C through the increment of 10 °C. The reaction rate can be improved exponentially by the increase of reaction temperature. Also, the reactant was more mixable which permits higher reaction rate [70]. Figure 7b displays the effect of reaction temperature on the biodiesel yield. At temperature between 35 °C to 55 °C, the conversion yield of methyl ester was below 60% due to the limitations of mass transfer and reactant solubility while at higher temperature, the reactants gained enough kinetic energy, which speeds up the mass transfer rate among the oil, methanol and catalyst phases, which results in the maximum conversion, 98.6% at 75 °C. However, when the reaction temperature proceeded above 75 °C, a slight loss in activity was observed because of high vaporization of the methanol that remained in the vapor phase. In addition, at a higher temperature reaction, the polarity of methanol decreased and resulted in a low concentration of methoxide species in reaction mixture [29,66,71].

Effects of Reaction Temperature
Transesterification was performed to determine the optimum reaction temperature by changing temperature range from 35 • C to 105 • C through the increment of 10 • C. The reaction rate can be improved exponentially by the increase of reaction temperature. Also, the reactant was more mixable which permits higher reaction rate [70]. Figure 7b displays the effect of reaction temperature on the biodiesel yield. At temperature between 35 • C to 55 • C, the conversion yield of methyl ester was below 60% due to the limitations of mass transfer and reactant solubility while at higher temperature, the reactants gained enough kinetic energy, which speeds up the mass transfer rate among the oil, methanol and catalyst phases, which results in the maximum conversion, 98.6% at 75 • C. However, when the reaction temperature proceeded above 75 • C, a slight loss in activity was observed because of high vaporization of the methanol that remained in the vapor phase. In addition, at a higher temperature reaction, the polarity of methanol decreased and resulted in a low concentration of methoxide species in reaction mixture [29,66,71].

Effects of Reaction Duration
The optimized reaction time can play a vital role in producing economical biodiesel. The result shown in Figure 7c indicated that the conversion of FAME improved significantly as the reaction time was amplified, and a maximum conversion was obtained at an optimum reaction time of 4 h. The low biodiesel yield at lower conversion time intervals was reported due to insufficient conversion of FAME while the transesterification decreased as the reaction extended more than 4 h due to the formation of soap [23,72]. Besides, the decreasing of biodiesel yield beyond the optimized duration can be deduced to the reversible nature of the conversion reaction of oil to biodiesel [73].

Reusability and Study of Deactivation of Supramagnetic RHC/K 2 O-20%/Fe-5% Nano-Bifunctional Catalyst
The recovery and reusability of heterogeneous catalysts are one of the significant factors that influence the economics of biodiesel productions. The regenerated supramagnetic RHC/K 2 O-20%/Fe-5% nano-bifunctional catalyst was evaluated at the optimal reaction conditions for reusability study, as presented in Figure 8a. It was noticed that after a second cycle, the biodiesel yield dropped from 98.4% to 91.3%. Whereas, in the third cycle, the biodiesel yield was reduced to 88.7%, and it was observed that the catalytic activity considerably reduced to 44.1% after the sixth cycle. Thus, the produced catalyst can be used for at least five cycles with significant yield. The reduction in the biodiesel yield was probably due to the deposition of remaining organic impurities or product intermediate (such as oil, biodiesel, glycerol, and FFAs) on the active site of the nano-magnetic RHC/K 2 O-20%/Fe-5% catalyst which hindered the reagents' access [74]. Contamination by moisture and carbon dioxide in the air during the separation process by the external magnetic field also contributed to the deactivation of the nano-magnetic catalyst [18].
Further studies of catalyst deactivation were conducted using N 2 adsorption-desorption, TPD-CO 2 /NH 3 , XRD, and VSM characterization for fresh RHC/K 2 O-20%/Fe-5% and spent RHC/K 2 O-20%/Fe-5% (6th cycle). The specific surface area and pore volume of spent catalyst dropped from 57.89 m 2 g −1 to 2.78 m 2 g −1 and 0.0588 cm 3 g −1 to 0.0102 cm 3 g −1 respectively, probably due to the attachment of large molecules from transesterification products such as FAME, glycerol and unreacted oil which caused less interaction for adsorption-desorption between nitrogen gas and catalyst surface. The basicity and acidity concentration of the catalyst were related to its surface area, where low adsorption of probe gases (CO 2 /NH 3 ) showed lower basicity and acidity strength compared to the fresh catalyst, as stated in Table 2. The XRD pattern (Figure 8b) of the spent nano-magnetic catalyst predicted no noticeable structural variation as compared to the fresh catalyst, whereas the intensity of few diffraction peaks of the spent supermagnetic nano-bifunctional catalyst was lower than the fresh nano-magnetic catalyst after six consecutive cycles. Also, the crystal structure of the reused catalyst was altered. The masking of the impurities on the active site of supermagnetic nano-bifunctional spent catalyst did affect the magnetization strength, which reduced to 2.76 emug −1 as shown in Figure 8c. Instead of deactivation of nano-magnetic catalyst by a poisoning mechanism through the attachment of large molecules [18], the repeating process of catalyst reactivation by calcining at high temperature of 550 • C or known as thermal degradation might cause the mesoporous catalyst active site to collapse and result in the decrease in surface area and pore volume which would thus reduce the catalytic activity [75]. Table 2. Textural properties and total CO 2 , NH 3 desorption of fresh nano-magnetic RHC/K 2 O-20%/Fe-5% and spent nano-magnetic RHC/K 2 O-20%/Fe-5%.

Comparative Study of Catalytic Activity with Literature Reported on Biochar Supported Catalysts
The comparative study of catalytic activity in the transesterification reaction between nanomagnetic RHC/K 2 O-20%/Fe-5% catalyst and other reported biochar supported catalysts is listed in Table 3. The compiled data suggested that the as-synthesized RHC/K 2 O-20%/Fe-5% catalyst depicted that catalyst produced in the present study was comparable in terms of catalytic activity to previously reported biochar catalysts in the literature. It was noted that the RHC/K2O-20%/Fe-5% catalyst was an effective and efficient supermagnetic nano-bifunctional catalyst for the transesterification of UCO. The confirmation of biodiesel production was characterized using 1 H-NMR and FTIR spectroscopy and was compared with used cooking oil. The 1 H-NMR spectrum of biodiesel and UCO is presented in Figure S2a,b. The typical peak of deuterated chloroform (CDCl 3 ) was detected at 7.29 ppm [72]. The spectra of biodiesel had a characteristic peak of singlet methoxy proton at 3.61 ppm, and a triplet of an α-methylene group of esters approximately found at 2.24 ppm [76]. These two peaks were not observed in UCO, therefore validating the conversion of triglyceride into respective fatty acid methyl ester. The other signals of multiplet formed at 5.26 ppm-5.34 ppm were identified as olefinic hydrogen [77]. A multiplet peak signal at 0.83 ppm-0.86 ppm was recognized as terminal methyl protons. A multiplet signal of methylene protons of carbon chain was located at 1.22 ppm-1.34 pm, and the signal of β-carbonyl methylene protons was observed at 1.56 ppm-1.60 ppm [63,78,79].
The functional groups and vibration bands of the synthesized biodiesel and UCO were identified in which both FTIR spectra of biodiesel and UCO samples were nearly alike but significantly different in the methyl ester functional peak observed as shown in Figure S3.

Synthesis of Rice Husk Char
The rice husk (RH) was washed with distilled water (H 2 O) to remove rice residue, clay particles, and sieved it. After multiple washing with distilled water, the RH was filtered and oven-dried (105 • C) until it achieved constant weight. The rice husk char (RHC) was prepared by carbonization at 450 • C for 3 h under nitrogen (N 2 ) flow. The RHC was treated with 1 M potassium hydroxide to eliminate traces of minerals and enhance the pore size of the char. The treated RHC was washed with hot distilled water until pH 7 and dried in the oven at 105 • C [31,82].

Preparation of Nano-Magnetic RHC/K 2 O/Fe Catalyst
A series of RHC/K 2 O/Fe catalysts with varying potassium oxide loadings (K 2 O: 15, 20 25 and 30 wt %) with 5% and 10% iron (III) oxide were prepared by wet impregnation method of RHC supported with aqueous solution of potassium carbonate (K 2 CO 3 ) and iron (III) chloride hexahydrate (FeCl 3 .6H 2 O). Both solutions of K 2 CO 3 and FeCl 3 .6H 2 O were dropped simultaneously into wet RHC and mixed vigorously for 6 h. The mixture was filtered and dried in the oven overnight at 70 • C. The dried solid base catalyst was carbonized in the furnace at 700 • C for 3 h with a heating rate of 3 • C min −1 [83].
Temperature programmed desorption method of carbon dioxide and ammonia gas (CO 2 -TPD/NH 3 -TPD) as probe gas were used to analyze the basicity and acidity strength of supermagnetic nano-bifunctional catalysts derived from rice husk using TPD/R/O/1100 series (Thermo Finnigan, Kyoto, JPN) apparatus coupled with thermal conductivity detector (TCD). The RHC/K 2 O/Fe catalysts samples were pretreated under N 2 gas flow for 1 h at 150 • C, during the acid site distribution analysis. The absorption of NH 3 was carried out under the flow of NH 3 stream for 1 h, and then the RH, RHC, and RHC/K 2 O/Fe catalysts were again flowed with N 2 gas for 45 min to remove excess NH 3 gas prior to analysis. Also, the basicity of RH, RHC, and RHC/K 2 O/Fe catalysts sample were subjected to only a single pretreatment under the flow of CO 2 gas. The chemisorption of CO 2 and NH 3 was carried out by CO 2 and NH 3 gas at a flow rate of 20 mL min −1 for 1 h, respectively. The additional CO 2 or NH 3 was then flushed with Helium (He) gas for 30 min at a flow rate of 30 mL min −1 . The desorption of CO 2 and NH 3 proceeded with the flow of He gas as carrier gas with a ramp of 10 • C min −1 from 50 • C to 950 • C and held for 30 min. The amounts of desorbed CO 2 and NH 3 were measured and noted [84].
The textural properties such as surface area, pore volume and pore size distribution of RH, RHC, and RHC/K 2 O/Fe catalysts samples were performed by using Brunauer-Emmer-Teller (BET) and Barrett-Joyner-Halenda (BJH) techniques from the desorption loop method with a Micromeritics ASAP-2020 (Micromeritics Instrument, Norcross, GA, USA). The catalyst sample was degassed at 120 • C to remove adsorbed contaminants and water vapor on the surface of the catalyst [85].
The morphologies of RH, RHC, and nano-magnetic RHC catalysts were determined with field-emission scanning electron microscopy (FESEM, Novananosem-230). The sample was sputtercoated with 7 nm gold (Au) to enhance the conductivity for better image resolution [86].
The magnetization of supermagnetic nano-bifunctional RHC/K 2 O-20%/Fe-5% was measured using a vibrating sample magnetometer (VSM, Lake Shore 7404 series, Lake Shore Cryotronics, Westerville, OH, USA). The VSM was calibrated with a pure nickel sphere with a standard value of 6.95 emu g −1 .
A pinch of the pre-weighed magnetic catalyst was sandwiched with clear duck tapes and attached on the surface of the VSM sample tail. The sample tail was inserted into a sample chamber, and the VSM was operated with 1.4 Tesla at room temperature.

Catalytical Activity Reactions Via Transesterification
Transesterification of biodiesel was carried out in 100 mL of 3-necked round-bottomed flask equipped with a water-cooled condenser and mechanical stirring under atmospheric pressure. Approximately 10 g of used cooking oil (UCO), the desired amount of methanol, and RHC/K 2 O/Fe catalyst was weighed and placed in the round-bottomed flask. The reaction sample was heated at 65 • C for 4 h and stirred by using the magnetic stirrer. After transesterification, the magnetic base catalyst was removed by using an external magnetic bar, and two phases of products were formed, which consisted of glycerol and biodiesel at the bottom and top layers, respectively. The excess methanol was removed by using a rotary evaporator. The unreacted oil from the synthesized biodiesel was removed by washing with hot distilled water, and the excess water was dried by adding ammonium sulfate ((NH 4 ) 2 SO 4 , Merck, Darmstadt, DE, 99.0%) based on the weight of oil. Next, the purified biodiesel was analyzed by gas chromatography, and methyl heptadecanoate was used as the internal standard with hexane (Merck, Darmstadt, DE, 98.5 %) as a solvent. The diverse parametric conditions for the transesterification of UCO allowed to ensure the achievement of an optimized reaction [87].

Biodiesel Analysis
The qualitative properties of purified methyl ester were examined using gas chromatography (GC) with flame ionization detector (FID) detector and a polar BPX-70 capillary column (60 m length; 0.25 mm ID; 0.25 µm capillary diameter) (Trajan Scientific, Ringwood, VIC, AUS). The n-hexane was used as GC solvent; whereas, methyl heptadecanoate was applied as an internal standard. The reference standards were prepared as; 500 ppm of each standard methyl palmitate, methyl oleate, methyl myristate, methyl stearate, and methyl linoleate. Briefly, 1 µm of prepared biodiesel sample was injected into the injection port; the GC oven temperature was fixed to increase from 100 • C-250 • C with 10 • C min −1 of the temperature ramp rate. The FAME yield (%) was measured using the formula stated below [11]: where; A (mm 2 ) = Total area of peaks; A meh (mm 2 ) = Peak area of methyl heptadecanoate; C meh (ppm) = Concentration of methyl heptadecanoate; V meh (mL) = Volume of methyl heptadecanoate; Wt (g) = Weight of the product. In addition, the confirmation of biodiesel production was characterized by using 1 H-NMR and FTIR spectroscopy and was compared with used cooking oil. 1 H-NMR spectra were recorded on Bruker AVANCE-500 MHz instrument (Bruker Corporation, Billerica, MA, USA) and the deuterated chloroform (CDCl 3 , Merck, Darmstadt, DE, 99.96 %and tetramethylsilane (TMS) were used as a solvent and internal standard, respectively. The FTIR spectra was recorded using Thermo Nicolet-7000 Spectrometer equipped with attenuated total reflectance (ATR) sampling accessory. A Diamond-ATR accessory with a removable ZnSe crystal was mounted on the sample compartment of the spectrometers. The liquid sample of UCO and biodiesel were filled onto the trough plate, respectively, and the cover was placed over the trough plate. The infrared spectra were scanned from 400-4000 cm −1 with a spectral resolution of 4 cm −1 .

Catalyst Regeneration and Spent Catalyst Characterisation
The spent nano-magnetic catalyst from the synthesized biodiesel was collected by using the external magnetic field and washed with acetone and n-hexane to remove the impurities trapped on the catalyst surface. Then, the washed catalyst was dried in the oven for 6 h at 80 • C and was reactivated at 550 • C for 2 h in a furnace with nitrogen flow. The recovered nano-magnetic catalyst was reused at optimum conditions to study the reusability of the catalyst and characterization of spent nano-bifunctional supermagnetic catalyst using BET, TPD, XRD, and VSM were performed.

Conclusions
The super-magnetic nano-bifunctional RHC/K 2 O/Fe catalyst was successfully synthesized for the transesterification of UCO. The optimum super-magnetic nano-bifunctional RHC/K 2 O-20%/Fe-5% catalyst, possessed a specific surface area of 57.89 m 2 g −1 and bifunctional properties for basic and acid densities of 4.43 mmolg −1 and 24.55 mmolg −1 , respectively. The supermagnetic nano-bifunctional RHC/K 2 O-20%/Fe-5% catalyst, showed excellent catalytic activity and rigidity that can be reused until the fifth cycle. Its recoverability by using an external magnetic field has presented the most effective separation process due to its high magnetic property (7.88 emug −1 ). The optimized conditions (amount of catalyst, 4 wt; methanol/oil molar ratio, 12:1; reaction time, 4 h; reaction temperature, 75 • C) gave the highest biodiesel yield (98.6%) through the conventional reflux system. A slight deactivation of the catalyst was found mainly due to non-selective poisoning in which the reactants deposited on the surface of active sites. Therefore, the supermagnetic nano-bifunctional RHC/K 2 O-20%/Fe-5% catalyst depicts to be an ideal magnetized catalyst as a replacement for conventional homogeneous and heterogeneous catalysts due to its potential for conversion of UCO to biodiesel, its effectiveness, and its practicability.