Synthesized heterogeneous nano‐catalyst from cow teeth for fatty acid methyl esters production through transesterification of waste vegetable frying oil and methanol: Characterization and optimization studies

Using readily available biomass, it is possible to produce fatty acid methyl esters (FAME) at a low cost. This study focused on the transesterification process of waste frying vegetable oil to generate FAME (biodiesel) using a cost‐effective heterogeneous nano‐catalyst synthesized from waste cow teeth (CT). The cow teeth nano‐catalyst (CTNC) was synthesized via a three‐step method of calcination, hydration‐dehydration, and re‐calcination. The synthesized CTNC samples were characterized. The Box–Behnken design of response surface methodology (RSM) was used to obtain the optimal operating conditions that maximize the FAME yield (Y 1 ) and the effects of the transesterification process conditions on FAME production. The results revealed that CTNC is of microporous structure with higher crystallinity. It consists of hydroxyapatite, beta‐tricalcium phosphate, and calcium oxide, having a mean particle size of 43.96 nm, a specific surface area of 124.77 m 2 /g, and a pore volume of 0.303 cm 3 /g. The process conditions that significantly influenced the EWFVOME yield were the CTNC loading, reaction time, and MTOR. A maximum of 98.70% (Y 1 ) was obtained at the optimum transesterification process variables of reaction time (2 h), MTOR (12:1), and CTNC loading (3.75 wt%). The biodiesel fuel properties were found to be within the acceptable values of ASTM D6751 and EN 14214 fuel standards.


INTRODUCTION
Biodiesel, which is a fatty acid alkyl ester (FAAE), is considered one of the best replacements for fossil fuels because it is clean, renewable, non-toxic, biodegradable, low carbon emitter, and eco-friendly. 1,2 Thus, biodiesel continues to gain more attention as a cost-effective alternative and renewable fuel due to the depletion, price instability, non-renewability, and adverse effects of fossil oil resources on the human-environment scenario. 3 As a result of the associated setbacks with fossil fuels, great efforts are being made globally toward generating and promoting sustainable and renewable energy sources for sustainable environmental development. 1 The main reason biodiesel has been considered an engine fuel is that it has similar characteristics to conventional diesel, making it applicable in compression ignition (CI) engines without modifying the engine. Generally, fatty acid alkyl ester has a higher flash point than conventional diesel fuel, gives better lubricity, and undergoes complete combustion, increasing the engine energy output. 2,3 However, the high production cost of fatty acid alkyl ester due to edible oils and convectional catalysts has significantly impeded the commercialization of biodiesel globally. Therefore, considering waste materials to decrease fatty acid alkyl ester production cost is a significant enthusiasm in fatty acid alkyl ester research. Using wastes such as animal teeth and bones as catalysts and waste frying vegetable oil (WFVO) for FAME production can solve the high production expenditure of FAME. Studies on the change of rice bran oil, 4 sludge palm oil, 5 waste frying oil, 6 moringa oil, 7 palm kernel oil, 8 and shea butter oil 9 have been carried out by some researchers for FAAE production. FAAE is generated from the transesterification reaction between triglycerides and alcohol, which is initiated and enhanced by either homogeneous or heterogeneous catalysts or enzymes. The catalysts (homogenous and heterogeneous) can be grouped into acid or alkali-based catalysts. 10 The use of homogeneous catalysts (HOC) has some limitations. The limitations of using (HOC) in fatty acid alkyl ester production include product separation difficulty, short service life, corrosiveness, expensive cost, and non-reusability. 3,11 These limitations have resulted in the high cost of the transesterification process. 3,11 Thus, heterogeneous catalysts (HC) can be used to minimize these limitations related to HOC.
The utilization of HC in a transesterification reaction provides many merits such as recyclability, long-lived, ease of separation, and cheapness. However, the major limitation of using heterogeneous catalysts in fatty acid alkyl ester production is that it results in high energy utilization and low diffusion rate. The high energy utilization occurs because of the agitation problem of the oil, alcohol, and catalyst. 12 The low diffusion rate results from the low basicity of the heterogeneous catalysts and reduced surface area, which possibly limits its catalytic effect. Consequently, higher catalyst charging, and increased reaction time are required to complete the transesterification reaction, which leads to an extra production cost. The catalytic performance of HC can be enhanced by expanding the surface area and elevating the basicity of the catalyst. 13,14 An increase in surface area can be achieved by reducing the catalyst particle size. For example, manufacturing a nano-sized catalyst can significantly improve the catalytic fruition of an HC. [15][16][17] Therefore, using nano-sized heterogeneous catalysts is one of the best ways to solve the transesterification process's high energy utilization and low diffusion rate problems. Also, heterogeneous catalyst can be utilized in petroleum refining to increase feedstock conversion and product selectivity. 18 Heterogeneous catalysts (HC) being used in transesterification reactions include calcium oxide (CaO), [19][20][21] magnesium oxide (MgO), 22 zeolite (KI/KIO3/H), 23 zirconia, 24,25 zinc oxide (ZnO), 26 and hydrocalcite catalyst. 27 The price of synthetic heterogeneous catalysts is high, so it is not economical to acquire them. 28 Therefore, HC derived from crab shells, mollusk shells, snail shells, animal teeth, capiz shells, oyster shells, animal bones, mussel shells, eggshells, abalone shells, and fish bones can solve the problem of the high price of heterogeneous catalysts for the transesterification process. 29 Available literature shows that catalyst preparation from animal bones and teeth, fish bones, snail, and mollusk shells require simple technology. 30,31 Cow teeth (CT) can be converted to organic fertilizers, handicrafts, bone meal, and calcium-based catalysts. 28,32 Hydroxyapatite (HA) and Ca 10 (PO 4 ) 6 (OH) 2 can also be produced from CT. 33 The animal teeth consist of enamel, tooth bulk, and outer surface. The enamel has the highest percentage of HA of about 92-96 wt%. 34,35 The significant component of CT is HA, which contains a large quantity of calcium and phosphate. 35,36 HA is a good material for forming heterogeneous catalysts. 33 Using animal bones and teeth for producing heterogeneous catalysts proffers a solution to the high cost of sourcing synthetic heterogeneous catalysts. 28,35,37 Several research works on HC obtained from animal bones have been reported in the literature. 38 Ayodeji et al. 1 reported biodiesel production from soybean using calcium oxide and cow bone catalysts. Sitompul and Mohadi 37 reported the production of calcium oxide from cattle bones as a catalyst in converting waste cooking oil to biodiesel. Jazie et al. 39 worked on transesterification of peanut and rapeseed oils using animal bone wastes as the cost-effective catalyst. Obadiah et al. 40 utilized HC derived from animal bone wastes in the transesterification of palm oil to biodiesel. However, the production of fatty acid alkyl ester using CaO derived from CT as HC has not been considered and reported to the best of our understanding.
The consumption of beef is increasing every day in many countries such as Nigeria due to rapid rate of population growth. The beef is majorly source from cow, and this has made cow production increased significantly. However, a major part of the cow's body that are not edible such as bones, teeth and horn are categorized as waste and thus discarded. In Nigeria, huge amounts of cow teeth are generated yearly as a byproduct from cow processing to produce meat and with no current use, they litter the environment causing environmental pollution. 41 This therefore served as our motivation to use this byproduct as a precursor for solid catalyst production, because it contains elements and compounds that confers catalytic activity on it. 42 Therefore, the major driver behind this research was to synthesize a new heterogeneous solid base CT-derived nano-catalyst (CTNC) and to evaluate its potential performance in the FAME production from EWFVO under different transesterification process conditions (process reaction time, methanol: esterified WFVO molar ratio, and CTNC charging). The process variables were optimized using RSM. The CT and CTNC were characterized for their oxide content, elemental composition, pore volume, morphological structures, and specific surface area. The CT was washed rigorously with distilled water to eliminate sand and other undesirable materials. The washed CT was then boiled for 4 h to eradicate the fats, specks of dirt, and tissues attached to it. 44 After cooling at ambient or room temperature, it was sun-dried for 7 days before being crushed mechanically with a grinder into smaller particles. It was oven-dried at 110 • C to constant weight. The oven-dried CT was milled and sieved to achieve a 120 μm particle size. 28 The heterogeneous nano-catalyst was synthesized from CT via the three-step method of calcination, hydration-dehydration (or hydrothermal-dehydration), and recalcination 46 (Figure 1). In the first step, the milled CT was thermally calcined in a lab-scale muffle furnace at 900 • C for 3 h. 45 In the second step, the calcined CT samples were hydrated at 60 • C for 6 h. After that, it was strained and oven-dried at 105 • C for 24 h. 47,48 While in the third step, the calcined-hydrothermal-dehydrated CT samples were further re-calcined at 900 • C for 3 h for conversion of the formed Ca(OH) 2 into a highly porous nano-sized catalyst particles, 46,49 which was referred to as cow teeth nano-catalyst (CTNC). The CTNC samples were cooled in a desiccator and kept in a sealed container to circumvent the catalyst from carbon dioxide (CO 2 ) and humidity before use.

CT and CTNC physicochemical characterization
The percentage compositions of the oxides present in the CT and CTNC were identified and quantified using the XRF spectrometer. 45 The major compounds present in the CT and CTNC were determined using the Mac science XRD system with CuKα radiation (k = 1.54 Å) operated at 40 mA and 40 kV and X-ray source (λ = 0.15 nm), 50 while the generated diffractogram was analyzed using Mac science software. The Ultima III theta-theta goniometer was used. The morphological structures and percentage of elements present in the CT and CTNC were determined using SEM and EDS, respectively. The morphological structures and catalysts' size were determined using TESCAN-VEGA SBH operating at 5 kV. The magnification varied from 100× to 10,000×, and VEGA3 software was used to obtain and process the images. The specific surface area of the CT and CTNC was determined through the adsorption-desorption of nitrogen at the boiling point of 77 K by utilizing the Nova Quantachrome surface analyzer. 31 The surface area was measured using F I G U R E 1 Flow diagram for the synthesis of heterogeneous nano-sized catalyst Brunauer-Emmett-Teller (BET) equation, and the pore size distribution was determined using Barrett-Joyner-Halenda (BJH) method.

De-acidification of WFVO by esterification reaction
De-acidification of the WFVO by esterification was done 5 before the transesterification process, as a result of high percentage of FFA in the WFVO. The esterification reaction was done following Aworanti et al. 3 method. WFVO (100 ml) was heated to 60 • C and mixed thoroughly with 0.14 ml of H 2 SO 4 and 55 ml of CH 3 OH in a 500 ml-three-necked round bottom flask at 800 rpm for 1 h. The solution settled for 2 h in a separation funnel to facilitate the esterification process and finally to achieve two distinct liquid phase layers (methanol-water fraction as the top layer and esterified WFVO (EWFVO) as the bottom layer). The bottom layer was collected and then purified using a rotary evaporator. The reaction scheme for the esterification process is given below: where, RCOOCH 3 stands for methyl ester.

Box-Behnken experimental design for the transesterification of esterified WFVO
A Box-Behnken experimental design (BBD) of the RSM was utilized to evaluate the impact of three unconventional process operating variables (reaction time, CH 3 OH-to-EWFVO molar ratio (MTOR), and CTNC concentration) on the dependent response variable (FAME yield or EWFVOME), as well as to optimize the process operating conditions. The ranges of the three unconventional process variables are presented in Table 1.
The transesterification reaction was done following Aworanti et al. 3 method. The transesterification process was performed in a 500 ml-three-necked round bottom flask. A water-cooled condenser was fitted in the middle neck, a thermometer in one of the side necks and a magnetic stirrer in the third neck. The round bottom flask set up was placed on a digital magnetic stirring hot plate. Weighed EWFVO (100 ml) was poured into the round bottom flask and heated to 100 • C. Afterward, a measured amount of CH 3 OH (9:1-15:1, w/w) was added and mixed with the heated EWFVO for 20 min before being dosed with the desired amount of CTNC (3 − 5 wt% of oil). The mixture of EWFVO and CH 3 OH was then heated at 65 • C and agitated at 600 rpm for a reaction time that ranged from 1-3 h. After the completion of the reaction, the mixture was centrifuged at 1000 rpm for 10 min to eliminate the CTNC. The supernatant was decanted through a filter paper into a separator funnel and then allowed to settle for 24 h to achieve a clearer FAME or EWFVOME and glycerol. The EWFVOME was later washed with warm distilled water to remove the remnant of CH 3 OH, glycerol, and the CTNC. The EWFVOME produced was dehydrated through oven-drying to eliminate the moisture. The transesterification reaction scheme or path is given as follows: The physicochemical properties of the EWFVOME were checked and compared with ASTM D6751 and EN 14214, respectively. Equation 1 was used for the evaluation of FAME yield. 3 FAME Yield(%) = Weight of FAME produced Weight of EWFVO used × 100 (1)

Statistical modeling and analysis
The FAME yield experimental values were inputted to the second-order polynomial Equation (2). Design-Expert software version 11.0 (Stat-Ease Inc., Minneapolis, USA) was used.
Y is the predicted response in this polynomial equation, X i and X j are independent variables, and o is the intercept term. The i , ii , and ij represent the linear, quadratic, and interaction coefficients. 51 However, reaction time, methanol to oil ratio, and CTNC charging are represented as A, B, and C, respectively. Thus, the second-order polynomial Equation (3) is presented as: The adequacy and goodness of fit of the developed model were evaluated using the regression analyses (coefficient of determination (R 2 ), adjusted R 2 and predicted R 2 , lack of fit, coefficient of variation, and adequate precision) and the analysis of variance (ANOVA). Evaluation of the model terms in the developed model was done using the probability value (i.e., the p-value). The polynomial equation was plotted in three-dimensional response surface graphs to show the interactive effects of variables. Diagnostics Plots and model graphs were obtained with the help of Design-Expert software version 11.0 (Stat-Ease Inc., Minneapolis, USA).

Optimization of operating conditions and FAME yield
In optimizing the transesterification process and FAME yield, the desirability index (DI) function of the numerical optimization tool of RSM in the Design-Expert software was applied. The DI is a multivariate optimization method employed to reveal the desirability of the various responses. 52 DI is represented by Equation (4): DI values range from 0 to 1. The value of 0 typifies the least desirable, while the value of 1 indicates the most desirable. The ultimate goal of optimization is to maximize the DI value. In optimization, the goals and objectives (i.e., target criteria) for the dependent and independent factors are inserted. In performing the optimization in this study, the target benchmark for the unconventional factors was set at the ranges of the values being studied. In contrast, the goal of the dependent (response) factors was to maximize the value of FAME yield ( Table 2). The predicted experimental process or operating conditions and FAME yield with the highest DI (i.e., desirability) were selected.

Validation of predicted optimum transesterification conditions and yield
To validate the predicted optimum transesterification process conditions and FAME yield, the transesterification of EWFVO and CH 3 OH in the presence of CTNC was experimentally conducted again at these predicted optimum conditions. The experimentally obtained value for the response or dependent factor (i.e., FAME yield) was recorded. This value was, however, compared with the predicted value to check the model's validity. The percentage error (%E) between the actual-experimental value (Q e ) and the predicted value (Q p ) was gotten from Equation (5) 53 :

Physical and chemical characterization of the WFVO and FAME
The physicochemical characterization of the WFVO and FAME produced were done according to the American Standards Test Methods (ASTM) procedures. The physicochemical properties analyzed were the flash point, pour point, density, kinematic viscosity, AV, iodine number (IN), and FFA (%). The fatty acids profile of the FAME or biodiesel was analyzed and identified with a Trace Gas Chromatography Ultra gas chromatograph (Thermo Electron Corporation, USA) furnished with a flame ionization sensing set-up. The temperature program was 180 • C for 2 min, 8 • C/min up to 240 • C, and a holding time of 8 min. Nitrogen was used as a carrier gas at a 1.0 ml/ min flow rate.

Determination of pour point
The pour point of the FAME sample was determined in accordance with the procedure of ASTM D97-11. The sample was put in the freezer at about −4 • C and gradually heated in a heating mantle to melt. The temperature at which the sample began to pour was recorded as the pour point.

Determination of flash point
The flash point of the FAME sample was determined according to the ASTM D 93-08. The sample was heated in an agitated container to determine the flash point that is, a flame was passed over the liquid surface. The temperature at which the vapor started to ignite was recorded as the flash point.

Determination of density
Density determination of the FAME sample was conducted based on ASTM D4052. The sample density at 32 • C was determined by gravimetric analysis, where 25 ml of sample was weighed using a glass cylinder, and the mass of the sample was obtained by electronic weighing balance. With the use of Equation (6), the density ( ) was calculated.

Determination of viscosity
The FAME sample kinematic viscosity determination was performed in accordance with the ASTM-D445. The viscosity was weighed by utilizing a falling-ball viscometer. The kinematic viscosity was determined by noting the time needed for a ball to fall under gravity through a sample-filled tube that is inclined at an angle. An average time of 10 tests was obtained and recorded. Thus, the dynamic viscosity and kinematic viscosity were calculated by applying Equations (7a) and (7b).

2.9.5
Determination of acid value AV (acid value) is the number of potassium hydroxide (KOH) milligrams that is needed to neutralize all the acid present in 1 g of the sample. 5 The AV of the FAME sample was determined based on ASTM D664. Acid value was determined by weighing a known amount (1 g) of sample into a flask and dissolving in a known volume (10 ml) of the ethanol-toluene mixture. Phenolphthalein (two drops) as an indicator was added to the solution, and the solution was then titrated against a standardized KOH solution. The titration was brought to an end when the solution turned into pink color (i.e., endpoint). The AV was then calculated with the use of Equation (8): Where: 56.1 = Molecular weight of KOH solution (g/mol), V KOH = Volume of KOH solution consumed during titration (ml), C sKOH = KOH solution concentration, M = weight (g) of the sample.

2.9.6
Determination of iodine number The FAME sample IN (iodine number) was worked on in accordance with the ASTM standard-5554. The IN value was obtained by weighing 0.5 g of sample and introduced into a 250 ml-Erlenmeyer flask. Chloroform solution (10 ml) and 25 ml of iodine-bromide reagent (Hanus solution) were well blended, introduced into the flask, and then kept in a dark room for 30 min. Ten milliliters of 15% potassium iodide (KI) solution were added to the mixture. The mixture was titrated with a solution of 0.1 N sodium thiosulphate (Na 2 S 2 O 3 ) and a drop of 1% starch was added as an indicator. The endpoint was reached when a clear solution was gotten. The IN was evaluated using Equation (9): Where, a = Solution volume required for the sample titration, b = Solution volume needed for blank titration, and W = Weight of sample.

Determination of free fatty acid
The FFA of the FAME sample was carried out according to the method published by Aworanti et al. 11 The FFA was determined by weighing 5 g of sample and then dissolved in 25 ml of propanol, after which five drops of phenolphthalein indicator were added to the sample propanol solution. The mixed solution was thereafter titrated with 0.1 N KOH solution until the solution turned to pink color. The FFA was computed using Equation (10): V = Volume of KOH solution (ml) required for sample titration, M = Molecular weight of oleic acid (g/mole), N = KOH solution normality (g/L) and W = Weight of sample (g).

Determination of water content
The FAME sample's water content was determined using a moisture titrator (Karl-Fisher, MKC-610 model, KEM Co. Ltd, Kyoto, Japan).

2.9.9
Determination of cloud point The cloud point of the FAME sample was carried out according to the method published by Eyankware et al. 54 The cloud point was determined by filling the test jar with FAME and closed tightly by the cork carrying the thermometer and placed into a bath of crushed ice. The test jar was removed from the jacket quickly without disturbing the specimen. Inspection for cloud point was done and jacket replaced. The determination was done without exceeding the time duration of 3 s. Since cloud point is the temperature of a liquid specimen when the smallest observable cluster of hydrocarbon crystals first occurs upon cooling under prescribed conditions, the observations made showed that cloud point was found to be approximately 1 • C. At this point, cloudiness is observed at the bottom of the test jar, which is confirmed by continued cooling.

Determination of heating value
The Heating value, of the FAME were measured in a bomb calorimeter according to ASTM D14214 standard method. An oxygen-bomb was pressurized to 3 Mpa with an oxygen container. The bomb was fired automatically after the jacket and a bucket temperature equilibrates to within acceptable accuracy of each other. 55 2.9.11 Determination of cetane value The cetane value was determined according to the norm ASTM D-14214. The distillation curve was made using the equipment Precision PS Scientific series (Chicago, USA), series 10Z9, a thermometer PG ERTCO ASTM 400 C (Vermont Hill, USA) and a chronometer OAKTON (Ontario, Canada).

XRF analysis
The data from XRF spectrometry of CT and CTNC are presented in Table 3. XRF analysis showed that the significant primary percentage elemental composition of inorganic oxides found in the CT and CTNC were CaO (59.96% and 70.13%), followed by P 2 O 5 (37.53% and 28.26%) which accounted for over 95% of the total oxides. Minor composition of the catalyst includes metal oxides such as MgO, SrO, ZnO, ZrO 2 , MnO, and TiO 2 . A Similar trend of results has been reported for cow bone and cow bone catalyst 2 as well as for ostrich bone, 43 except that the cow bone and cow bone catalyst did not contain SiO 2 and MgO while the ostrich bone did not contain TiO 2 , ZrO 2 and MnO (Table 3). These inorganic oxides have been confirmed to be effective transesterification materials. 56 CaO is a significant component considered the best base catalyst in FAME production because of its reduced toxicity, high basic strength, and simple reactions with water. [57][58][59] The basic oxide (MgO) will enhance the catalyst's essential strength. 60 Figure 2B shows that the CTNC also displayed sharp diffraction spectra or XRD patterns with no significant difference or variation from the uncalcined CT. A similar observation was reported for raw ostrich bones and calcined ostrich bones 43 and for raw beef bones and calcined/hydrothermal bone bones. 62 63,64 In addition, the XRD patterns displayed by CTNC are similar to the XRD patterns exhibited by the nano CaO catalyst synthesized from duck eggshell. 64 Based on the XRD analysis, the principal component of the CTNC is CaO. Applying the Debye Scherrer method in Equation (11), 63 the crystallite size diameter (D [nm]) of the CaO nano-catalyst was evaluated for each 2θ value, and the average or mean D was obtained to be 44.75 nm.
Where (is in radian) is the full width at half-maximum (FWHM) or half-width, and is the position of the maximum diffraction peak (the Bragg angle). K is a term for the constant shape factor (with a value of 0.94), and stands for the X-ray wavelength (=0.154178 nm for CuKα). The average D (crystallite diameter or mean particle size) obtained for the CaO nano-catalyst in this present study was compared to the value of 43.96 nm reported by Reference 63; the average particle size attained for CaO nano-catalyst synthesized from waste goat bone.

SEM-EDS analysis
The surface morphological structures obtained from the SEM that were respectively performed on CT and CTNC are presented in Figure 3A,B. Figure 3A,B shows the SEM micrographs of CT (i.e., uncalcined CT) and CTNC (i.e., calcined CT) samples. The micrographs indicate dissimilarity in the morphological structure of CT and CTNC. Figure 3A revealed that the surface morphology of the CT particles or grains is an amorphous, un-homogeneous and non-porous structure with the particles densely aggregated together into rough clusters having small or less surface area but still accommodating CaCO 3 . The F I G U R E 3 The morphology structure of (A) CT (B) CTNC surface of the CTNC ( Figure 3B) indicated an alteration in the morphological structure of the CT. The surface morphology was observed to be uniformly distributed with no accumulation. It displayed a honeycomb-like porous microstructure with a higher surface area due to particle size reduction and higher crystallinity.
The SEM analysis of the CTNC also showed that there is maximum particle size reduction which makes the CTNC be of nano-particle size. The tiny size of the grains and clump could offer higher specific surface areas. 65 This change in morphology could result from the fact that the CT's CaCO 3 was transformed into calcium oxide during the first calcination, making the materials to be more crystalline and less heterogeneous in morphology. This implied that the first calcination increased the porosity and non-aggregation of the material structure. The increased porosity may be due to the elimination of organic components related to the CT. 55 This is the desired property that is required for a faster reaction. In addition, this morphological and structural change was also due to the water refluxing, further dehydration, and second calcination (i.e., re-calcination). The formation of calcium hydroxide and its decomposition release more water vapor molecules and produce more CaO, lowering the cluster's degree and creating higher porosity, increased surface area, higher crystallinity, and increased catalytic activity. 17 Re-calcination after hydration-dehydration also helps to homogenize its textural properties, as indicated in the work of Sitompul and Mohadi. 24 Nisar et al. 66 and Chingakham et al. 62 reported that animal bone subjected to calcination at a higher temperature displayed particle size reduction with higher surface area. Figure 4 revealed the EDS analysis of inorganic elements (in percentage) present in (A) CT (uncalcined CT) and (B) CTNC (i.e., re-calcined CT).
The inorganic elements found to be present in the CT and CTNC include calcium, sodium, oxygen, carbon, magnesium, silicon, aluminum, phosphorus, potassium, and iron ( Figure 4). Together, these elements formed the HA in the milled CT. 43 The presence of trace elements in the HA could be advantageous in accelerating the catalytic action of the CTNC. 43 The milled or uncalcined CT contained a significant amount of calcium (50.85 wt%), oxygen (25.72 wt%), and phosphorus (8.84%), while CTNC also contained a significant amount of calcium (66.73 wt%), oxygen (15.99 wt%) and a trace amount of phosphorus (0.58%). They are sore reactive metals responsible for the catalytic action of the CTNC. 59,67 According to the reports of Mitaphonna et al. 28 Sitompul and Mohadi, 37 and Khan et al. 43 phosphorus, calcium, and oxygen were the only inorganic elements found in a larger percentage in raw cattle bones/calcined cattle bones, raw Aceh cow bones/calcined Aceh cow bones, and calcined ostrich bone, respectively. The percentage of calcium in uncalcined CT increased from 50.85 to 66.73 wt% in CTNC, while the percentage of phosphorus and oxygen decreased from 8.84% to 0.58% and 25.72 to 15.99 wt%, respectively, in CTNC ( Figure 4A,B). The CTNC had a higher percentage of calcium and a lower percentage of oxygen as well as a trace amount of phosphorus in comparison with uncalcined CT. This is due F I G U R E 4 EDS structure of (A) CT and (B) CTNC to the effect of calcination, hydration-dehydration and re-calcination of the CT. Since calcium and oxygen are the highest amounts of inorganic elements in CTNC, it thus indicates that calcination, hydration-dehydration and re-calcination emanated in the formation of calcium oxide in the CTNC.

BET analysis
The BET analysis results are presented in Table 4.
The results show that the specific surface area and external surface area of CT are 105.35 and 102.24 m 2 /g, while they are 124.770 and 112.35 m 2 /g for CTNC, respectively. It can be deduced that the CTNC had the highest surface area, allowing for simple diffusion of reactants into the internal areas of the catalyst. As reported in the literature, the greater the surface area of a solid catalyst, the stronger its catalytic action. 47,68 Similarly, the hydration-dehydration treatment and re-calcination of CTC at 900 • C played an essential part in upgrading the catalyst standard due to the multiplication in the surface area of CTNC. The pore volumes are 0.125 and 0.303 cm 3 /g for CT and CTNC, respectively. The pore size (diameter) showed that CTNC is a micropore catalyst. Thus the CTNC was envisaged to have high catalytic action because of its elevated surface area. 59,68,69 The BET-specific surface area and pore volume respectively obtained for CT and CTNC in this study (Table 4) are relatively higher than the values obtained by Jazie et al. 39 Manalu et al. 70 and Yan et al. 71 for calcined goat bone (90.6 m 2 /g and 0.051 cm 3 /g), calcined cow bone (1.78 m 2 /g and 0.0009 cm 3 /g), and calcined pig bone (104.6 m 2 /g and 0.104 cm 3 /g), respectively.

Biodiesel (FAME) yield, statistical modeling, and analysis
Based on the Box-Behnken experimental design, the experimental FAME yields were obtained as a response to the design applied variables ( Table 5). The biodiesel or FAME content obtained was in the range of 80.94% to 99.50%. The minimum biodiesel or FAME yield was found to be 80.94% obtained at a process operating variable of 3 h (reaction time), 4 g (catalyst charging), and 9:1 (CH 3 OH-to-EWFVO molar ratio (MTOR)). The maximum biodiesel or FAME yield was obtained to be 99.50% attained at operating conditions of reaction time (2 h), catalyst charging (4 g), and methanol: oil ratio (12:1).
In terms of uncoded variables, the multiple regression analysis of the experimental value yielded the following second-order polynomial in Equation (12): The coefficients obtained for each parameter in terms of the mathematical equation of coded factors to predict the biodiesel yield is as shown in Equation (13): A positive sign in front of the terms in the above equations represents the synergistic effects of increased biodiesel yield, whereas a negative sign indicates the antagonistic effect. 72 The RSM's ANOVA and regression analysis (Table 6) revealed that the three linear terms (A, B, and C), quadratic terms (A 2 , B 2 , and C 2 ), and all cross products or interaction terms (AB, AC, and BC) were all significant model terms at the 95% confidence levels (p = 0.05).
The model F value of 722.79 with a low probability value (p-value less than 0.0001) indicates that the regression model is highly significant. The R 2 assesses the goodness of the model's fit, with a value greater than 0.90 considered desirable. 31 The independent variables are responsible for the R 2 value of 0.9982, which means that the model can explain 99.82% of TA B L E 4 The values of BET surface area and pore size of CT and CTNC Sample Specific surface area (m 2 /g) External surface area (m 2 /g) Average pore diameter (nm) Pore volume (cm 3   the variation in FAME content, and the model cannot explain only 0.18% of the total variation. It indicates the fitness of the model. The predicted R 2 of 0.9751 agreed with the adjusted R 2 of 0.9958, indicating that the regression model could be used to analyze response trends. The F-value of 6.37 for lack of fit indicates that the lack of fit was significant compared to the pure error. As a result, there is only a 5.28% chance that a significant lack of fit F-value could occur due to noise factors such as human or experimental errors. 73 The non-significant lack of fit demonstrates that the model is significant and that the model equation is adequate for predicting FAME yield under any variable combinations. 74 The low coefficient of variation (CV = 0.48) demonstrates the dependability of the experiments. 75 The signal-to-noise ratio is measured with adequate precision. An adequate precision of 53.59 indicates an adequate signal, as a ratio greater than 4 was required to navigate the design space excellently. 31 Thus, adequate precision obtained for the model shows that the model has good indicators acceptable for the process optimization. The model Equation (12) was utilized to predict the FAME yield, and the outcomes are presented in Table 5 above. The outcome indicated that the differences between the actual and predicted values are slight, as depicted in Figure 5.

Effect of the transesterification process conditions on FAME yield
The numerical coefficients of the generated model Equation (12) and the ANOVA presented in Table 7 indicated that both reaction time, CTNC charging, and MTOR (CH 3 OH-to-EWFVO molar ratio) had positive significant linear effects on the FAME yield. The linear effects of the reaction time, CTNC loading, and MTOR (CH 3 OH-to-EWFVO) on FAME yield are displayed in Figure 6A-C. It was noticed that the FAME yield elevated with increasing reaction time until the reaction time was 2 h and thereafter started to decline ( Figure 6A). The decline in yield after 2 h may result from backward hydrolysis   66 Chingakham et al. 62 and Erchamo et al. 46 for biodiesel produced from jatropha oil, honge oil, and waste cooking oil, using KOH-calcined waste bone catalyst, waste animal bone-derived catalyst and eggshell-based CaO nano-catalyst, respectively. Similarly, Figure 6B shows that the yield of FAME elevated with the increase in the CTNC loading until the loading was up to 4 g and above this level of loading quantity, the yield decreased. The increase in yield may be accredited to an increase in the catalyst's active sites as the catalyst loading was relatively improved. 63 The decrease in yield may be due to mixing and poor reactants' diffusional problems induced by the increased viscosity of the reactants and catalyst mixture due to higher catalyst loading. 77,78 In addition, the decrease may be adduced to soap formation, favored by higher catalyst loading above the optimum. 63,79 This observation concurred with the reports made by Nisar et al. 66 Chingakham et al. 62 and Mamo and Mekonnen 63 for the biodiesel production from jatropha oil, honge oil, and Scenedesmus algal oil using KOH-calcined waste bone catalyst, animal waste bone-derived catalyst and goat bone-nano-catalyst, respectively. Furthermore, the FAME yield also increased when the MTOR changed from ratio 9:1 to ratio 12:1. The FAME yield declined above this ratio. This decline in yield may be attributed generally to the liquefying of generated glycerol in the excess methanol leading to the obstruction of the methanol reactions with the oil and catalyst, thereby hindering the separation of the glycerol resulting in a backward reversible reaction. 43,46 A corresponding increase in FAME yield with elevating MTOR up to an optimum value after which the yield decreased has been reported by Nisar et al. 66 Chingakham et al. 62 Mamo and Mekonnen, 63 Khan et al. 43 and Erchamo et al. 46 for biodiesel synthesis from jatropha oil, honge oil, Scenedesmus algal oil and waste cooking oil using KOH-calcined waste bone catalyst, waste animal bone-derived catalyst, goat bone-nano-catalyst, ostrich bones-derived heterogeneous catalyst and eggshell-based CaO nano-catalyst, respectively. Figure 7A-C show the interaction effects of reaction time, CTNC loading, and MTOR on FAME yield. The interaction effect of reaction time and CTNC charging on FAME yield is depicted in Figure 7A. The positive numerical coefficient of the model term AB (i.e., the interaction between reaction time and CTNC loading) and the ANOVA results presented in Table 7 showed that the interaction between reaction time and CTNC loading had a significant (p ≤ 0.05) positive effect on the FAME yield. This is illustrated in Figure 7A. The curvilinear shape of the plot revealed that, as the reaction time and CTNC loading increased from 1 to 2 h and 3 to 4 g, respectively, at constant MTOR (12:1 CH 3 OH-to-EWFVO molar ratio), there was an increase in the FAME yield. Further increases above these values caused the biodiesel yield to decrease. This showed that the catalyst loading was already at a critical point (4 g) at which any further addition did not favor the reaction because high catalyst concentration makes the reaction mixture gummy and could withstand the mass transfer resistance to the reaction system. 80 A similar observation was reported by Aworanti et al. 11 for the production of biodiesel from waste cooking oil using commercial CaO. Figure 7B delineates the interaction effect of reaction time and MTOR (CH 3 OH-to-EWFVO molar ratio) on the FAME yield. The interaction between reaction time and MTOR had a significant (p ≤ 0.05) adverse effect on the FAME yield (Table 7 and Equation [4]). Figure 7B indicated that at constant CTNC loading (4 g), the FAME yield elevated significantly as the reaction time increased from 1 to 2 h with a mutual increase in MTOR from 9:1 to 12:1. Any further increase beyond these values led to a reduction in the FAME yield. A similar observation was reported for biodiesel production from microalgal (Spirulina platensis) oil utilizing the β-strontium silicate as HC. 81 The response surface plot of the FAME yield as an outcome of the interaction between CTNC charging and MTOR is presented in Figure 7C. The negative numerical coefficient of BC (interaction between CTNC loading and MTOR), as well as the outcome of the ANOVA, presented in Table 7, specified that the interaction between CTNC loading and MTOR had a significant (p ≤ 0.05) adverse effect on the yield of FAME. As depicted in Figure 7C, the FAME yield increased as the CTNC loading increased from 3 to 4 g with a simultaneous increase of MTOR from 9:1 to 12:1. However, further increment in the CTNC loading above 4 g, as well as MTOR above 12:1, resulted in the decrease of the FAME yield. Singh et al. 81 reported a similar observation for biodiesel production from microalgal (Spirulina platensis) oil utilizing the β-strontium silicate as HC.

Optimization of transesterification process conditions and verification
Optimum transesterification process variables for FAME production were predicted using the numerical optimization tool of RSM in Design Expert (6.08) statistical software based on the DI function. The predicted optimum values for the reaction time, CTNC loading, and MTOR were obtained to be 2.00 h, 3.75 g, and 12:1, which are required to achieve FAME or biodiesel production with a maximum or optimum yield of 99.65% (Table 7). The FAME yield desirability resulted in an overall DI of 1.000. A validation experiment was conducted at these theoretical optimum transesterification process variables to corroborate the maximum FAME yield. As obtained from the verification experiment, the maximum FAME yield was found to be 98.7% (Table 7). The calculated %E (percentage errors) between the theoretical or predicted FAME yield and the actual validated experimental FAME yield was −0.96%, which implied that the predicted value was closer to the actual verified value, thereby indicating that there are no significant differences between the values. Some research works have investigated the effects of catalysts acquired from different waste bones in the transesterification of oils. Selected research works by Ayodeji et al. 1 Volli et al. 44 Khan et al. 43 and Anand et al. 49 were compared with this study as presented in Table 8.
From Table 8, it could be noticed that the highest FAME yield obtained at the optimum transesterification process conditions in this study was comparatively higher than the 92.2%, 94%, 90.56% and 82.3% biodiesel yields obtained by Ayodeji et al. 1 Volli et al. 44 Khan et al. 43 and Anand et al. 49 for catalyst acquired from cow bone, chicken bone, ostrich bone and goat bone, respectively.

Methyl esters composition of FAME and its fuel properties
The major and minor methyl esters composition of FAME (biodiesel) determined are presented in Table 9. Table 9 revealed that the FAMEs consisted of methyl esters having fatty acids with carbon atoms in C 4 to C 24 . In addition, olenic acid, palmitic acid, and butyric acid are the major methyl esters in the FAMEs. The minor methyl esters include octanoic acid, linolenic acid, lauric acid, myristic acid, lignoceric acid, benenic acid, tricosanoic acid, and ecosatrienoic acid. The fuel properties of FAME are listed in Table 10.

3.5.1
Acid value and FFA AV is one of the significant biodiesel properties required for check of quality. AV indicates the level of FFAs occurrence in biodiesel. The higher the AV, the lower the biodiesel quality. 37 High AV can result in sediment or deposit accumulation TA B L E 9 Methyl esters composition of FAME (biodiesel) in the diesel fuel system, clogging the fuel filter and causing engine corrosion. 11 AV that is lower than 0.5 mg KOH/g is ideal for vehicle fuel. Table 10 shows that the AV for the produced biodiesel or FAME is 0.23 mg KOH/g while the %FFA present in the biodiesel is 0.11%. According to ASTM D6751, the maximum AV is less than 0.8 mg KOH/g, while the FFA found in FAME is 0.8 mg/KOH. The results, therefore, divulge that the FAME has lower AV and FFAs values and is in concordance with the ASTM standard. The AV obtained in this study is comparatively lower than the values of 0.30 mg KOH/g obtained by Anand et al. 49 for biodiesel produced from WCO using catalyst synthesized from goat bone. Also, the AV obtained in this research work is smaller than the values obtained by Aworanti et al. 3,11 3.5.2 Density and kinematic viscosity Density provides the needed details on how the diesel fuel will perform. The high value of density indicates that there is the presence of some impurities in the biodiesel. 11 The density value obtained for the biodiesel produced is 0.885 g/cm 3 and was observed to meet the range of densities specified for biodiesel by EN14214 and ASTM D6751. The density obtained in this study is relatively similar to the values of 0.88 g/cm 3 , obtained by Anand et al. 49 for biodiesel produced from WCO using catalysts synthesized from goat bone. The Density gotten in this work is equal to the values of 0.0884 g/cm 3 obtained by Ajala et al. 31 for FAME generated from WCO using catalysts synthesized from cattle bones, while the density in this study is equal to the values of 0.863 g/cm 3 obtained by Aworanti et al. 11 Viscosity is a fluid's resistance to the flow rate of an mm-sized capillary. High viscosity produces oxidized polymeric compounds, leading to sediments and gum formation that clog engine filters. 22 The kinematic viscosity value achieved for the biodiesel is 4.10 mm 2 /s, which falls within the specification value provided by ASTM D-6751 for quality FAME. The kinematic viscosity obtained in this study is relatively similar to the values of 3.53 mm 2 /s obtained by Anand et al. 49 for biodiesel produced from WCO using catalysts synthesized from goat bone. The kinematic viscosity gotten in this work is lower than the values of 5.25 mm 2 /s obtained by Amenaghawon et al. 57 for biodiesel production generated from waste vegetable oil using functionalized cow horn catalysts.

Pour point and flash point
Pour point is taken as the lowest temperature at which the liquid oil or fuel is observed not to flow when subjected to cooling under given conditions or as the lowest temperature at which wax begins to be seen when the oil is cooled. 11 This is an important parameter used to indicate fuels' minimum or lowest temperature performance. 82 The pour point obtained for the biodiesel produced is −5 • C (Table 10), and this value is within the specified standard range adduced by ASTM D6751. The pour point attained for the FAME produced in this research is relatively higher than the values of −11 • C achieved by Anand et al. 49 for FAME production from waste cooking oil using catalysts produced from goat bone. Flashpoint is a parameter that specifies the minimum temperature at which the diesel fuel vaporizes to release ignitable vapors above the fuel, which can ignite when it comes in contact with a flame. 82 It measures the liquid oil or fuel volatility and ease of ignition. 83 Flashpoint is utilized as an index to characterize fuels' fire risk or hazards. The higher the flash point, the safer it is to handle the liquid oil. 82 The flash point achieved for the biodiesel produced is 165 • C (Table 10), and this value falls within the range specified by ASTM D6751 and EN 14214 for standard fuel. This indicates that the biodiesel's storage will be safe and stable under ambient conditions. The flash point gotten for the FAME produced in this study is comparatively higher than the values of 114 • C gotten by Anand et al. 49 for FAME production from waste cooking oil using catalysts generated from goat bone and the values of 158 • C obtained by Amenaghawon et al. 57 for biodiesel production generated from waste vegetable oil using functionalized cow horn catalysts.

3.5.4
Iodine number and water content The IN (iodine number) measured for the produced biodiesel is 2.1 g I 2 /100 g ( Table 10). The obtained value was found to be less than the value of 120, which is in accordance with the EN 14214 standard. High IN leads to polymerization and deposit formation in the injector's nozzle and piston rings and thus results in engine failure. The iodine number obtained in this study is relatively similar to the values of 2.210 g I 2 /100 g obtained by Aworanti et al. 49 for biodiesel produced from WFCO using homogeneous catalysts. Also, it was observed that the result obtained in this study is lower than the value of 84.21 g I 2 /100g obtained by Amenaghawon et al. 59 for biodiesel production generated from waste vegetable oil using functionalized cow horn catalysts. No water was found in the FAME (i.e., water content is nil), which met the water content requirement specified by ASTM D6751 and EN 14214, respectively. The result is similar to the result obtained by Ajala et al. 31 and Amenaghawon et al. 59

Cloud point
Cloud point is the temperature at which a cloud of wax crystals first appears in the oil when it is cooled. The cloud point for the produced biodiesel is 4 • C (Table 10) and this value falls within the range specified by ASTM D6751 and EN 14214 for standard fuel. The Cloud point obtained for the FAME produced in this study is comparatively higher than the values reported by Lang et al. 84 which says that the cloud point of ethyl esters of linseed oil, canola, sunflower and rapeseed oil were −2, −1, −1, and −2 • C, respectively. The cloud points of ethyl esters reported by Lang et al. 84 were approximately lower than those FAME obtained in this study. This indicates that the ethyl esters marginally perform better in cold temperatures than the FAME obtained in this study.

Heating value
The heating value is an important property defining the energy content and thereby efficiency of fuels. The heating value for the produced biodiesel is 36.6 MJ/kg (Table 10) and this value falls within the range specified by ASTM D6751 and EN 14214 for standard fuel. The heating values (HVs) of biodiesels are relatively high. The HVs of biodiesel (39 to 43.33 MJ/kg) is slightly lower than that of diesel (49.65 MJ/kg). The oxygen content of biodiesel improves the combustion process and decreases its oxidation potential. The structural oxygen content of a fuel improves its combustion efficiency due to an increase in the homogeneity of oxygen with the fuel during combustion. Because of this the combustion efficiency of biodiesel is higher than that of petrodiesel and the combustion efficiency of methanol and ethanol is higher than that of diesel. 55 3.5.7 Cetane number Cetane numbers are the indicators of ignition properties of diesel fuel. The higher the cetane number, the more efficient the ignition is. The cetane number measured for the produced biodiesel is 53 ( Table 10). The obtained value was found to be in the range of >51, which is in accordance with the EN 14214 standard. The higher cetane number indicates the higher engine performance of FAME, resulting in lower emission of all pollutants other than oxides of nitrogen (NOx). 85 The cetane index is frequently influenced by the structure of fatty acids, the saturated fatty acid content, and the length of the carbon chains in a fuel. Longer carbon chains associated with a higher saturated fatty acid content result in a higher cetane index. 85 Hilber et al. 86 reported the cetane number of methyl esters of rapeseed oil, soybean oil, palm oil, lard and beef tallow to be 58, 53, 65, 65 and 75 respectively. The beef tallow has the highest cetane number which means it has huge amount of saturated fatty acids, the increase in the saturated fatty acids content positively enhanced the cetane number of biodiesel. 85 The oxidative stability of biodiesel fuels also increases due to the presence of higher amount of saturated fatty acids. However, the drawback of a higher amount of saturated fatty acid content in biodiesel fuel is that the cold filter plugging point occurs at a higher temperature.

CONCLUSIONS
In this present study, a novel green and low-cost or cost-effective heterogeneous nano-catalyst for transesterification reactions was synthesized from cow teeth via a three-step method (calcination, hydration-dehydration, and re-calcination) characterized by XRF, XRD, SEM-EDS, and BET. The possibility of CT being used as an HC for FAME production under different transesterification process conditions (Catalyst loading, reaction time, and methanol/oil molar ratio) using RSM was evaluated. RSM was used for modeling and process conditions optimization. From the results obtained, the following conclusions were made: 1. A three-step method, viz. calcination, hydration-hydration, and re-calcination, can result in the synthesized catalyst (from animal teeth) with porous microstructure, higher surface area, and pore volume as well as higher crystallinity and reduced particle size, which leads to better or increased catalytic activity. 2. The synthesized catalyst (CTNC) consisted principally of CaO with a mean particle size of 44.75 nm, a specific surface area of 124.77 m 2 /g, and a pore volume of 0.303 cm 3 /g. 3. Catalyst loading, reaction time, and methanol/oil molar ratio significantly affect the transesterification of WFVO with a relatively high acid value of 2.5 mg KOH/ g oil, implying changes in these operating variables will affect the FAME yield. 4. Optimum transesterification process conditions of reaction time (2 h), methanol to oil molar ratio (12:1), and CTNC loading (3.75 g) will be needed to produce an optimum FAME yield of 98.70%. The FAME yield can be adequately predicted by applying a second-order quadratic polynomial regression model. 5. The fuel properties of FAME produced under optimal conditions meet the recommended and specified fuel quality standards provided respectively by ASTM D6751 and EN 14214.
Therefore, the utilization of waste cow teeth in a sustainable manner not only assists with managing our waste but will help in the evolution of a cost-effective catalyst to sustainably produce FAME.