Novel iron sand-derived α-Fe2O3/CaO2 bifunctional catalyst for waste cooking oil-based biodiesel production.

The main aim of this work was to develop a heterogeneous Fe2O3/CaO2 bifunctional catalyst prepared from iron sand and 3 different CaO2 sources (CaCO3, Ca (OH)2, and limestone) using wet impregnation and calcination methods for biodiesel production. The effects of different CaO2 sources and Fe/Ca ratio in the catalyst were investigated to provide insight into the catalyst character and biodiesel yield. X-ray diffraction, X-ray fluorescence, and scanning electron microscopy analyses were used to characterize the catalyst. CaCO3 was concluded as the best CaO2 source, while the best Fe/Ca configuration was found to be 1:4, giving the highest biodiesel yield (97.0401%) with no diglycerides. Greater addition of Fe loading would result in an amorphous structure, and all catalysts were relatively crystalline. Fe was concluded to favor the esterification reaction and biodiesel formation, while CaO2 was seen to favor the transesterification reaction and fatty acid methyl ester (FAME) formation. The catalyst mechanism was also established in this study, where esterification of free fatty acid (FFA) and glycerol took place on the acid site to produce diglyceride and transesterification of triglyceride by methanol occurred on the basic site.

Introduction

Energy is not only one of the major aspects of the national economy, but also an important aspect of life. According to BP Statistical Review of World Energy (2021), world energy consumption fell by 4.5% on 2020 due to the COVID-19 pandemic affecting the economy output and activity. Despite the fall, the renewable energy demand rose to 9.7% that indicates the world’s commitment to reduce emission as established on The Paris Agreement and the opportunity to further develop the area. Biofuel, particularly biodiesel, is one of the most used energies in Indonesia with numerous advantages, such as renewable, biodegradable, and lower emission level, and promotes complete combustion in engines. It is commonly produced through a transesterification process using an acid or base catalyst and has strict restrictions on the free fatty acid (FFA) and water content (<2%) to avoid saponification (de Araújo et al. 2013). It is known that esterification process has helped with these problems, but alas the 2 different steps require different catalysts, are complex, and would inevitably raise the production cost (Pruszko 2020; Pan et al. 2017). To overcome these concerns, a heterogeneous bifunctional catalyst that could facilitate a simultaneous esterification-transesterification reaction could be used. Past researches on the bifunctional catalysts were SO4/Mg-Al-Fe3O4 prepared from FeCl2.4H2O, FeCl3.6H2O precursors via co-precipitation method and followed by encapsulation and sulfation for transesterification of waste cooking oil on 95°C and 300-min operation conditions (Gardy et al. 2019); CaO/Ag for transesterification of soybean oil on 72°C and 90-min operation conditions prepared by impregnation of calcium methacrylate to AgNO3 solution (Zhu et al. 2021); and CaO/La2O3 prepared from Ca (NO3)2.4H2O and La(NO3)3.6H2O via co-precipitation method for transesterification of jatropha oil with 160°C and 3-h reaction temperature and time (Lee et al. 2015) where all those catalysts showed 97-99% biodiesel conversion. However, the complex catalyst preparation and synthetic precursors used hindered the possibility of the catalyst for sustainable biodiesel production. Hence, sustainable and suitable compounds are needed to ensure sustainability.

A heterogeneous base catalyst such as CaO is one of the best and most used catalysts in biodiesel production due to its high catalytic and basicity levels, moderate condition operability, and high biodiesel conversion (Faruque et al. 2020). Numerous studies had reported excellent CaO catalyst activity, including Kouzu et al. (2008) who conducted transesterification of biodiesel from soybean oil and obtaining a 99% conversion with 120 min and 65°C reaction time and temperature; Viola et al. (2012) obtaining a 93% biodiesel conversion in 80 min and 65°C from transesterification of used cooking oil; Colombo et al. (2017) who conducted transesterification of soybean oil in a recycling reactor and achieved a 100% biodiesel conversion with 75-min reaction time and 3%w/w catalyst mass; Huang et al. (2021) achieving 97.8% biodiesel yield with 2 h and 65°C reaction time and temperature from transesterification of soybean oil. However, the possible suspension formation and high glycerin viscosity in the product consequently made CaO to be mostly used as a support catalyst (Helwani et al. 2020). CaO2 could be synthesized from many resources, namely CaCO3, Ca(OH)2, waste (eggshells, fish bones), or natural rocks (limestone) that guaranteed raw material sustainability. On the other hand, Fe2O3 also has the potential to be developed along with CaO2 as a bifunctional catalyst. Zhang et al. (2021) have investigated the effects of Fe2O3 catalyst on the combustion and emission characteristics in diesel engine and noted a significant NO and NOx conversions of 72.3% and 76.1% respectively in a diesel fueled engine. The NOx emission is an important aspect in supporting an environmentally friendly fuel alternative. It has been strictly regulated in the Euro VI standard, requiring NOx emission reduction by 95% in heavy-duty diesel engines (Zhang et al. 2022a) since 90% of the NOx emission was known to originate from fossil fuel combustion (70% caused from diesel engines), further highlighting the potential of Fe2O3 as a catalyst (Zhang et al. 2022b). Fe2O3 could be synthesized from iron sand which is an abundant natural resource in Indonesia. Chemical co-precipitation using an acid and base solution, followed by calcination, was the commonly used methods for synthesis. Widayat et al. (2019) have successfully obtained the best surface area and pore volume of Fe2O3 at 700°C calcination temperature. Fatmaliana et al. (2020) have also successfully obtained a 96.58% Fe2O3 content using these methods, which was further supported by Widodo et al. (2020) whom obtained 100% Fe2O3 phase after calcination at 900°C temperature. In addition to its sustainability and cheap price, the utilization of iron sand is still limited to metal industries and buildings’ raw material (Ardiani et al. 2020) that allows a room for further development. Moreover, there has been no research that studied and compared iron sand-derived bifunctional catalyst Fe2O3/CaO2 with different CaO2 sources (CaCO3, Ca(OH)2, and limestone) for biodiesel production. This research aims to develop a Fe2O3/CaO2 bifunctional catalyst using wet impregnation method and calcination for biodiesel production with waste cooking oil (WCO) raw material through simultaneous esterification-transesterification process.

Materials and methods

This study followed the research method by Widayat et al. (2019) with several modifications. In particular, modifications in the preparation of α-Fe2O3/CaO2 bifunctional catalyst, simultaneous esterification-transesterification method in biodiesel production, and the centrifugation of biodiesel product to separate the impurities.

Preparation of α-Fe2O3 and CaO2

Figure 1 presents the preparation steps of α-Fe2O3/CaO2 bifunctional catalyst. α-Fe2O3 was prepared using iron sand (obtained from Adipala Beach, Cilacap, Central Java, Indonesia) by chemical co-precipitation and calcination methods. Iron sand was subjected to magnetic separation pre-treatment to separate the iron ores (Fe3O4) from the impurities. The chemical co-precipitation began with adding Fe3O4 to 12 M HCl (37%, Merck, Germany), stirred using a magnetic stirrer (MS7-H550-S, DLAB, China) at 70°C for 50 min. Afterward, the mixture was filtered and added with polyethylene glycol 4000 (PEG 4000) with a 5:1 volume ratio before stirred for 1 h at 70°C. NH4OH (25%, Merck, Germany) was added next, and the mixture was stirred again until a precipitate was formed. The precipitate was then washed and vacuum filtered (WP6122050, Merck Millipore, Germany) to obtain the filtrate. The filtrate was dried in the vacuum oven (DZF-6020, MTI Corporation, USA) for 2 h at 120°C before calcined using a tube furnace (STF54454C, Lindberg/Blue M, Asheville NC, USA) at 700°C for 2 h to change the Fe3O4 phase into α-Fe2O3 and remove the remaining unwashed impurities.

Fig. 1

Flowchart of α-Fe2O3/CaO2 bifunctional catalyst preparation

To prepare CaO2, CaCO3 (99%, Merck, Germany) was first ground to homogenize the particle size using mortar and pestle. The homogenized powder was then calcined using a tube furnace (STF54454C, Lindberg/Blue M, Asheville NC, USA) at 950°C for 2 h to obtain CaO2. These steps were repeated with Ca(OH)2 (96%, Merck, Germany), and limestone (obtained from Padalarang, Bandung, West Java, Indonesia).

Preparation and characterization of α-Fe2O3/CaO2 bifunctional catalyst

The α-Fe2O3/CaO2 bifunctional catalyst was prepared by wet impregnation and calcination methods. First, the previously prepared α-Fe2O3 powder was mixed to each of the CaO2 powder obtained from CaCO3, Ca(OH)2, and limestone with varying weight ratios (2:1, 1:1, 1:2, 1:3, 1:4, and 1:5 %w/w). The mixture was diluted with demineralized water (100 %w/w CaO) and stirred using a magnetic stirrer (MS7-H550-S, DLAB, China) at 60°C for 13 h. The mixture was then filtered to obtain the filtrate, and later dried in the vacuum oven (DZF-6020, MTI Corporation, USA) at 120°C for 2 h before calcined in a tube furnace (STF54454C, Lindberg/Blue M, Asheville NC, USA) at 700°C for 2 h.

X-ray diffraction (XRD), X-ray fluorescence (XRF), and scanning electron microscopy (SEM) characterizations were carried out to determine the surface morphology, composition, and crystallinity of the prepared catalysts. The XRD analysis was performed using the Shimadzu XRD-7000 equipment with patterns measured at 30 mA and 40 kV under Cu-Kα radiation at k = 1.54 Å. The diffraction pattern was observed at 2θ angles from 10 to 90° with a 0.5-s step time. The X’Pert HighScore software (PANalyticle copyright 2022) was used to determine the components through the obtained peak intensities. The XRF analysis was carried out using the Rigaku Supermini 200 equipped with a Benchtop tube under a sequential wavelength-dispersive XRF (WDXRF) spectrometer. Meanwhile, the SEM analysis was carried out using the SEM JEOL JSM6510LA equipment at 20 kV and 1.00000 nA probe current. A detailed specification of equipment used in the catalyst preparation is shown in Table 1.

Table 1

Specification of equipment used in α-Fe2O3/CaO2 bifunctional catalyst preparation

Magnetic stirrer (MS7-H550-S, DLAB, China)
Speed range	0-1500 rpm
Heating temperature range	25-550°C, increment 5°C
Temperature control, display accuracy	±10°C, ± 1°C
Vacuum oven (DZF-6020, MTI Corporation, USA)
Vacuum level	< 50 mTorr (at 200°C)
Temperature range	30-200°C
Temperature control accuracy	± 1°C
Temperature uniformity	± 5°C at 100°C
Tube furnace (STF54454C, Lindberg/Blue M, Asheville NC, USA)
Heating temperature range	25-1700°C
Process tube diameter	3″ (76.2 mm)
Heated length	24″ (609.6 mm)
Controller accuracy	±0.1%
XRD (Shimadzu XRD-7000S, Japan)
Goniometer radius	200-275 mm
Operation angle range	−12 to 164° (2q), −6 to 82° (cs), −6 to 132° (qd)
Operating speed	0.1-50°/min (qs, qd), 0.1-100° (2q)
Detector preset time	0.1-1000 s
XRF (Rigaku Supermini 200, Japan)
Sample rotation	60 rpm (50/60 Hz)
2θ scanning speed	Max. 1200°/min
Step scanning	0.002-1.0°
Temperature control	35°C ± 0.3°C
SEM (JEOL JSM6510LA, Japan)
Magnification	×5 to ×300,000
Electron gun accelerating voltage	0.5-30 kV (55 steps)
Maximum specimen size	150 mm

Catalytic performance test

The catalytic performance of the prepared catalysts was tested by biodiesel production. Waste cooking oil (WCO) obtained from household and restaurant wastes in Tembalang, Central Java, Indonesia was used as the raw material. Prior to the biodiesel production, WCO was pre-treated to remove the water and impurities by heating it to 100°C for 40 min and filtrating. The filtrate was then determined its %FFA content using the acid-base titration. First, 5 ml of methanol (99.98%, Merck, Germany) was added to 5 ml of WCO and added with 2 drops of Phenolphthalein indicator. The mixture was titrated with 0.1 N NaOH until a pink color persisted. The following equation was used to calculate the FFA content:

The simultaneous esterification-transesterification process was carried out to produce biodiesel. The pre-treated WCO and methanol (1:15 molar ratio) was poured into a three-necked flask. A 1 %w/w catalyst was also added to the mixture, then stirred at 65°C for 3 h. A Liebig condenser was set into the three-necked flask and was connected to a cooler (CCP5-15, DLAB, China) to maintain the cooling water at 19-20°C. The mixture was taken 10 ml every 15 min and centrifuged (3500 rpm, 15 min) to determine the formation mechanism of the biodiesel, with 5 ml subjected to gas chromatography mass spectrometry (GCMS) analysis and the other 5 ml subjected to FFA titration. After the simultaneous esterification-transesterification process ended, the remaining product was poured into a separating funnel and left for 12 h where the methanol would rise to the top layer, biodiesel in the middle layer, and the catalyst settling on the bottom layer. The middle layer was collected and washed in the separating funnel with warm water (60-70°C) until the spent water was transparent, indicating no impurities left. The product was centrifuged (DMO 412, Scilogex, USA) at 3500 rpm for 15 min as the final separation process and biodiesel were collected.

Characterization of biodiesel product

The biodiesel product obtained in this study was subjected to yield, density, viscosity, %FFA, and GCMS analyses to determine its properties and suitability to both the Indonesian and European standards. The density and viscosity were measured using 25 ml pycnometer and Ostwald viscometer respectively. The GCMS analysis to determine the biodiesel components was carried out using the QP 2010S Shimadzu, DB-1 column equipment, while the yield of the biodiesel and fatty acid methyl ester content (%FAME) were calculated using the following equation:

Results and discussion

Characterization of α-Fe2O3/CaO2 bifunctional catalyst

Effect of different CaO2 sources to α-Fe2O3/CaO2 catalyst

Based on Fig. 2, it could be seen that the α-Fe2O3/CaO2 catalyst synthesized from all 3 CaO sources (CaCO3, Ca(OH)2, and limestone) have common peaks of CaO2 (portlandite) and Fe2O3 phases. The CaO2 peaks were detected at 2𝛳 = 18.228°, 34.358°, 47.547°, 51.161°, and 54.764°, while peaks belonging to Fe2O3 were detected at 2𝛳 = 29.726°, 33.963°, 50.988°, 51.226°, and 54.869°. The presence of both the CaO2 and Fe2O3 peaks on all of the synthesized α-Fe2O3/CaO2 catalysts indicated that the α-Fe2O3 metal dopant has been successfully impregnated to the CaO2 catalyst. Both the CaCO3 and Ca(OH)2-derived catalysts have a monoclinic crystal structure, while the limestone-derived catalyst has a cubic crystal structure. The peak intensity correlates with the degree of crystallinity of the catalyst. The higher the peak intensity, the higher the degree of crystallinity is. This was reflected in Fig. 2, and the Ca(OH)2-derived catalyst has the highest degree of crystallinity (68.8628%) that indicates it being the most crystalline, while the CaCO3-derived catalyst has the lowest (54.1226%). The low crystallinity in the CaCO3-derived catalyst could be explained by the high Fe2O3 phase present that acted as defects in the structure, hence disrupting the order of the crystal system and consequently lowering the crystallinity. Nevertheless, all the three catalysts were concluded to be relatively crystalline judged by the relatively high degree of crystallinity.

Fig. 2

XRD result of α-Fe2O3/CaO2 bifunctional catalyst from 3 different CaO2 sources

The CaO2 phase formed in this study was due to the presence of water vapor during the calcination of CaCO3, Ca(OH)2, and limestone (mainly composed of CaCO3) as an open tube furnace was used, allowing ambient air to flow through. The presence of water vapor would form an intermediate (CaHCO3.OH) that later decomposes as Ca(OH)2 and finally forming CaO2. Giammaria and Lefferts (2019) that had studied this pathway also noted that the formation of the intermediate from CaCO3 was faster than the formation of CaO, explaining the favorable formation of CaO2. The pathway is illustrated in Fig. 3. Another factor affecting the formation of this CaO2 phase was the calcination temperature of the catalyst, which was only 700°C. Numerous studies have stated that the CaO phase was formed on calcination temperatures higher than 900°C (Mohadi et al. 2016; Malau and Adinugraha 2020). At the temperature of 700°C, the decomposition of CaCO3, Ca(OH)2, and limestone (composed mainly of CaCO3) was incomplete, hence forming the CaO2 phase as there was insufficient energy to break the oxygen bond.

Fig. 3

CaO2 formation pathway (modified from Giammaria and Lefferts 2019)

The phases detected on the XRD results were cross-checked with XRF analysis to accurately determine the composition of each α-Fe2O3/CaO2 catalysts. Based on Table 2, Ca(OH)2-derived α-Fe2O3/CaO2 catalyst has the highest Fe2O3 composition, followed with limestone-derived and lastly CaCO3-derived catalyst with very small differences (±0.1 to 0.5). These findings were in line with the peak intensities shown from the XRD results: Ca(OH)2 showed the highest intensity, followed by limestone and lastly CaCO3. The presence of MgO as an impurity in limestone-derived α-Fe2O3/CaO2 catalyst was not an unusual phenomenon as limestones were known to have impurities of magnesium carbonates and dolomites (CaMg(CO3)2) (Fairbridge et al. 1967).

Table 2

XRF result of α-Fe2O3/CaO2 catalyst from 3 different CaO2 sources

Components	Composition (%mass)
	CaCO3	Ca(OH)2	Limestone
Fe2O3	17.3153	17.8523	17.7001
CaO	57.5325	57.2912	54.5526
Al2O3	0.5190	0.3186	0.2741
SiO2	0.9950	0.6521	0.6314
TiO2	1.3199	1.5439	1.4172
V2O5	0.1995	0.1208	0.1780
MnO	0.1296	0.1329	0.1305
MgO	n.d*	n.d*	0.2333

*n.d not detected

The synthesized α-Fe2O3/CaO2 catalyst surface morphology was also analyzed via SEM characterization as shown in Fig. 4. Based on the XRD and XRF results, it was gathered that impregnation has been successfully carried out in the catalyst. Impregnation process would deposit the metal load on the surface of the catalyst instead of the catalyst pore, and this phenomenon could be observed in the SEM results (Fig. 4), where well-distributed small circular particles spread on the surface of the larger irregular shaped bulk, proving that the metal load (α-Fe2O3) has been successfully deposited and simultaneously supported the conclusion gathered from the previous XRD and XRF results. It could also be seen that the Ca(OH)2-derived α-Fe2O3/CaO2 catalyst has the most defined shape of a monoclinic structure, while the limestone-derived catalyst has a relatively defined cubic shape, supporting the XRD results that conclude Ca(OH)2 has the highest crystallinity.

Fig. 4

SEM result of α-Fe2O3/CaO2 catalysts from 3 different CaO2 sources: A CaCO3, B Ca(OH)2, and C limestone on 10,000× magnification

Similarly, the least defined shape on CaCO3-derived catalyst shown in Fig. 4(A) also supported the XRD result previously mentioned. The CaCO3-derived catalyst gave the lowest peak intensity (Fig. 2) consequently giving lower crystallinity and structure arrangement. Agglomeration on the limestone-derived catalyst could also be seen on Fig. 4(C) which could be explained by the presence of water during the calcination process (Fig. 3) that inevitably bind the catalyst particles together by Van der Waals forces on the catalyst surface, known as the sintering effect commonly found in heterogeneous catalysts (Forzatti and Lietti 1999). The high amount of α-Fe2O3 metal dopant detected by the XRF analysis (Table 2) in the limestone-derived catalyst also contributed to the big particle size, as metal dopants will promote the crystal growth rate (Kasim et al. 2014).

Effect of Fe/Ca ratio to α-Fe2O3/CaO2 catalyst

Based on Fig. 5, the 2 major peaks detected were CaO2 and Fe2O3 that indicates a successful impregnation. The CaO2 peaks were detected at 2𝛳 = 18.228°, 34.358°, 47.547°, and 51.161°, while the Fe2O3 peaks were detected at 2𝛳 = 33.963°, 51.226°, and 54.869°. It could also be clearly observed that greater addition of α-Fe2O3 to the catalyst (Fe/Ca ratio 2:1) will give a very amorphous structure, as indicated by the lack of distinguishable sharp peaks in the diffractogram. There was also no Fe2O3 phase and peaks detected on this variable, even though the α-Fe2O3/CaO2 was added twice the amount of CaO. Mos et al. (2018) stated that copper (Cu) radiation in XRD analysis is not suitable to be used in characterizing Fe2O3 peaks, although majority of studies nowadays still used it to characterize iron oxides containing samples and could provide a decent analysis. This incompatibility resulted in a high background noise that could mask the low intensity peaks given by the Fe2O3 phase. Moreover, the fact that the addition of α-Fe2O3 gave an amorphous structure to the catalyst also further hid the possible α-Fe2O3 peaks present, since the weak peaks could be easily masked by the broad hump of the amorphous pattern. Mardwita et al. (2020) had also obtained similar condition in their alumina-supported cobalt and chromium catalyst, and explained that these undetected peaks were due to the metal particles highly dispersed in the catalyst surface, consequently pushing alumina to be dominant and the only phase able to be detected by the XRD radiation.

Fig. 5

XRD result of CaCO3-derived α-Fe2O3/CaO2 catalyst with different Fe/Ca ratio

The elemental composition of the CaCO3-derived α-Fe2O3/CaO2 catalyst was investigated through XRF analysis and presented in Table 3. It could be seen that the α-Fe2O3 composition decreases as the Fe/Ca ratio decreases, and that the α-Fe2O3 and CaO2 compositions relatively reflect the ratio. The high amount of the Fe2O3 particles detected in the Fe/Ca ratio 2:1 further supported the conclusion drawn from the XRD result (Fig. 5, Fe/Ca ratio 2:1) that the undetected Fe2O3 peaks were due to the Fe particles very highly dispersed throughout the catalyst surface.

Table 3

XRF result of CaCO3-derived α-Fe2O3/CaO2 catalyst with different Fe/Ca ratio

Component	Composition (%mass)
	2:1	1:1	1:2	1:3	1:4	1:5
Fe2O3	46.5210	37.6119	17.3153	14.7596	14.0586	11.6086
CaO	34.0959	40.2651	57.5325	57.5465	57.5688	60.4373
Al2O3	0.9191	0.5194	0.5190	0.3843	0.3916	0.4364
SiO2	1.0155	0.8946	0.9950	0.9464	0.9205	0.9223
TiO2	3.6866	2.9273	1.3199	1.1800	0.9737	1.0770
V2O5	0.4386	0.3648	0.1995	0.1282	n.d	n.d
MnO	0.3395	0.2545	0.1296	0.1065	0.0892	0.0760

*n.d not detected

Based on Fig. 6, it could be observed that the α-Fe2O3 particles were present in all of the catalyst surfaces, corroborating the XRD and XRF results presented in Fig. 5 and Table 3 respectively that the impregnation process has been successful. Sintering effect was clearly observed in the results, starting from the Fe/Ca ratio 1:2 (Fig. 6(C)) to the most prevalent agglomerate formed in Fe/Ca ratio 1:5 (Fig. 6(F)). It was established that CaO could readily absorb water in ambient atmosphere; hence, with higher CaO composition, there would be a higher chance of sintering effect taking place.

Fig. 6

SEM result of CaCO3-derived α-Fe2O3/CaO2 catalyst with different Fe/Ca ratio: A 2:1, B 1:1, C 1:2, D 1:3, E 1:4, and F 1:5 on 10,000× magnification

Characterization of the biodiesel product

Effect of different CaO2 sources to the biodiesel and FAME yield

The successfully synthesized α-Fe2O3/CaO2 catalysts from CaCO3, Ca(OH)2, and limestone sources were subjected to biodiesel production to determine their catalytic performance. Based on Fig. 7, the limestone-derived α-Fe2O3/CaO2 catalyst has the highest biodiesel yield (77.1672%), followed by the Ca(OH)2 source (68.7972%), and lastly the CaCO3 source (62.3761%). In this study, biodiesel yield referred to the yield of product containing FAME, diglycerides, and monoglycerides; hence, to determine the pure FAME content (without the incompletely converted triglycerides), the FAME yield was also calculated. The limestone-derived catalyst gave the highest FAME yield (62.0656%), followed closely by CaCO3 (61.5902%) and lastly Ca(OH)2 (32.7062%). Based on the XRF result (Table 2), limestone-derived catalyst has the lowest CaO composition but a high α-Fe2O3 composition with only a very small difference (±0.1) with the Ca(OH)2-derived catalyst. This indicates that the limestone-derived catalyst favored the esterification reaction catalyzed by α-Fe2O3 rather than the transesterification reaction catalyzed by the CaO component to form biodiesel. Similarly, the Ca(OH)2-derived catalyst that has the highest CaO composition (Table 2) has a lower biodiesel yield due to the dominant transesterification reaction led by the CaO component. The transesterification reaction caused soap formation since the waste cooking oil used as the raw material contained high FFA, further explaining the very low FAME yield. Likewise, the CaCO3-derived catalyst that has the lowest α-Fe2O3 component (Table 2) has the lowest biodiesel yield (62.3761%) since there would be less available active metal species for the esterification reaction to form biodiesel.

Fig. 7

Effect of different CaO2 sources to the biodiesel and FAME yield

Based on Table 4, the only biodiesel product that has met both the Indonesian and European standards (SNI 7182:2015 and EN 1421:2008) respectively was the biodiesel produced using the CaCO3-derived catalyst. This catalyst also gave biodiesel with the lowest FFA content (2.2534%), indicating that the catalyst was the most successful in converting the FFA in the waste cooking oil to biodiesel. Therefore, it could be concluded that among the 3 CaO sources investigated in this study, CaCO3 was the best CaO source to be combined with the metal dopant α-Fe2O3 in giving the best catalytic performance in biodiesel production despite it giving the lowest biodiesel yield (Fig. 7).

Table 4

Effect of different CaO2 sources to the physicochemical characteristics of biodiesel

Parameters	CaO2 sources			WCO	Indonesian standard (SNI 7182:2015)	European standard (EN 1421:2008)
	CaCO3	Ca(OH)2	Limestone
Density (g/ml)	0.8830	0.9448	0.9144	1.0028	0.85-0.89	0.86-0.90
Kinematic viscosity (cSt)	4.3880	4.3790	4.3870	5.4330	2.3-6.0	3.5-5.0
FFA (%)	2.2534	4.2817	2.7699	6.8164	-	-

The trend of FAME formation was investigated using GCMS analysis in this study (Fig. 8). Based on Fig. 9, the Ca(OH)2-derived α-Fe2O3/CaO2 catalyst formed the highest FAME (hexadecanoic acid, methyl ester and 9-octadecenoic acid, methyl ester) on the 60th min mark, which was the fastest among the catalysts. It was followed by the limestone-derived α-Fe2O3/CaO2 catalyst which experienced the highest FAME formation on the 90th min mark. Meanwhile, the CaCO3-derived α-Fe2O3/CaO2 catalyst gave the highest FAME formation on the 180th min mark.

Fig. 8

GCMS results of biodiesel product from different CaO2 sources

Fig. 9

Formation of FAME components through the course of reaction time

The CaCO3 and limestone-derived catalyst showed higher levels of the diglyceride hexadecanoic acid, 2-hydroxy-1,3-propanediyl ester (an incomplete form of FAME), which meant that these catalysts have lower selectivity and thus favored the formation of biodiesel instead of FAME. Meanwhile, the Ca(OH)2-derived catalyst showed a complete formation of the 9-octadecenoic acid methyl ester and only a small amount of diglyceride formed, indicating that this catalyst was more selective towards FAME formation and thus more suitable in biodiesel conversion. Nevertheless, all the catalysts showed good conversion of the free fatty acid 9-octadecenoic acid to the FAME 9-octadecenoic acid methyl ester as shown on their higher yield compared to the other components.

Simultaneous esterification-transesterification of the bifunctional α-Fe2O3/CaO2 catalyst

To justify the role of esterification in the synthesized bifunctional α-Fe2O3/CaO2 catalyst, the trend of FFA conversion through the course of reaction time is investigated and depicted in Fig. 10. The bifunctional α-Fe2O3/CaO2 catalysts from 3 different CaO sources were shown to successfully convert the FFA in the WCO feedstock, as seen by the decreasing FFA trend from 6.8164 to 2.2534% (CaCO3), 4.2871% (Ca(OH)2), and 2.7699% (limestone). The CaCO3-derived catalyst has the lowest FFA content left by the end of the 3-h reaction time, indicating a higher conversion rate than the other catalysts. Despite the generally downwards trend of the conversion, there was still an increase in the FFA content; which could be attributed to the reaction between the triglycerides and water that was the byproduct of the FFA esterification by the acid site of the bifunctional catalyst.

Fig. 10

FFA trend through the course of reaction time

Meanwhile, the bifunctional catalyst was already proven able to carry out the transesterification reaction by the formation of FAME and diglycerides as seen in the GCMS results (Fig. 9). The FFA conversion trend in Fig. 9 was shown to correlate with the FAME formation in Fig. 8 where decreasing FFA gave increasing FAME yield, which was congruent with the esterification and transesterification reactions. The raise in diglycerides was also consistent with the raise in the FFA, thus giving evidence to the reaction between the triglyceride and water that yield not only FFA, but also diglycerides. In conclusion, the bifunctional α-Fe2O3/CaO2 catalyst synthesized in this study was capable of carrying the simultaneous esterification-transesterification reaction as seen by the evidences.

Effect of Fe/Ca ratio to the biodiesel and FAME yield

Based on Fig. 11, the highest biodiesel yield (97.0401%) was achieved on Fe/Ca ratio 1:4. It was established that the synthesized catalyst favored the esterification reaction and that the Fe2O3 component was more dominant than the CaO2. This conclusion was corroborated by the very low biodiesel yield (54.3709%) given when the Fe ratio was the lowest (1:5), as there was not enough Fe active species to carry out the esterification of the relatively high FFA content in the WCO, thus giving a low yield. The Ca leaching to the methanol phase of the biodiesel system could also contribute to this low yield, which has been proven by numerous studies including Mansir et al. (2017) and Colombo et al. (2017). It was also noteworthy that when the Fe ratio was double (2:1), the FAME yield was lower than the Fe/Ca ratio 1:5, indicating that despite the catalyst favoring the esterification reaction, the CaO2 component that was responsible for the transesterification reaction was also important and created a synergistic effect in the biodiesel conversion.

Fig. 11

Effect of Fe/Ca ratio to the biodiesel and FAME yield

To further understand the phenomena, the FAME components and physicochemical characteristics of the biodiesel are investigated and shown in Tables 5 and 6 respectively. Based on Table 5, the 1:4 Fe/Ca ratio that gave the best biodiesel and FAME yield was shown to only have the complete methyl ester components (hexadecanoic acid methyl ester and 9-octadecenoic acid methyl ester) with no diglycerides, indicating that this ratio was the best configuration. Meanwhile, the small number of diglycerides and high number of complete FAME in the 1:5 Fe/Ca ratio explained that the Fe2O3 component in the catalyst favored the formation of diglycerides, while the CaO component favored the formation of methyl esters. This phenomenon could be seen in the other ratios, including the 2:1 Fe/Ca ratio which gave the highest diglyceride components among the other.

Table 5

Major FAME components in biodiesel produced from varying Fe/Ca ratios

FAME component	Fe/Ca ratio
	2:1	1:1	1:2	1:3	1:4	1:5
Hexadecanoic acid, methyl ester	7.75%	34.56%	32.09%	31.16%	31.78%	31.71%
9-Octadecenoic acid, methyl ester	12.17%	60.67%	47.32%	63.38%	63.65%	45.72%
Hexadecanoic acid, 2-hydroxy-1,3-propanediyl ester	13.39%	0.64%	13.58%	n.d	n.d	3.53%
9-octadecenoic acid, 2-hydroxy-1,3-propanediyl ester	1.94%	n.d	3.01%	n.d	n.d	0.41%

*n.d not detected

Table 6

Effect of Fe/Ca ratio to the physicochemical characteristics of biodiesel

Parameters	Fe/Ca ratio						Indonesian standard (SNI 7182:2015)	European standard (EN 1421:2008)
	2:1	1:1	1:2	1:3	1:4	1:5
Density (g/ml)	0.8788	0.8747	0.8830	0.8420	0.8805	0.8815	0.85-0.89	0.86-0.90
Kinematic viscosity (cSt)	4.7387	4.6331	4.3882	4.2076	4.4318	4.7027	2.3-6.0	3.5-5.0
FFA (%)	1.4084	0.8450	2.2534	1.1267	0.4131	2.8167	-	-
Diglycerides (%)	15.33	0.64	16.59	-	-	3.94	<0.24	<0.2
Methyl ester (%)	34.29	95.89	79.41	98.20	99.73	82.08	>96.5	>96.5

The favorable formation of diglycerides instead of methyl esters in the biodiesel affected the kinematic viscosity property. Diglyceride is an incomplete methyl ester with 2 unbroken triglyceride chains that would have a heavier molecular weight, raising the kinematic viscosity of the biodiesel. Based on Table 6, all Fe/Ca ratios except 1:3 and 1:4 have higher viscosities as there was the diglyceride component. Nevertheless, the viscosities of all the biodiesel still passed the European and Indonesian standards. The diglyceride content is also regulated and included in the total glycerol parameter, and per the standards only the 1:3 and 1:4 Fe/Ca ratio had passed. Meanwhile, the densities of all the ratios had passed both standards, except the 1:3 Fe/Ca ratio; and the lowest FFA content of the 1:4 Fe/Ca ratio showed the best conversion of FFA to methyl esters among the variables. In conclusion, the physicochemical results (Table 6) supported the conclusion that 1:4 Fe/Ca ratio was the best catalyst configuration.

Predicted mechanism of the bifunctional α-Fe2O3/CaO2 catalyst

Based on the GCMS results and physicochemical properties of the synthesized CaCO3-derived bifunctional α-Fe2O3/CaO2 catalyst, it was concluded that the catalyst was able to catalyze simultaneous esterification-transesterification reaction, with α-Fe2O3 component as the acid site favoring diglyceride formation and the CaO2 component as the basic site favoring the formation of methyl esters. It was known that the diglyceride compound could be formed due to the esterification between FFA and glycerol with an acid catalyst (Kong et al. 2019). Considering that there was no glycerol layer present on the decantation process, the mechanism of the bifunctional α-Fe2O3/CaO2 catalyst was predicted with the assumption that the glycerol formed in the transesterification reaction was completely converted (Fig. 12). The mechanism started with the adsorption of the carbonyl group in the FFA to the acid site and the adsorption of the methanol to the basic site of the catalyst. These adsorptions would produce carbocation and oxygen anion respectively. Next, nucleophilic attack will happen on the surface of the catalyst: the carbocation with the second hydroxyl group in glycerol (due to the GCMS result detecting the component fatty acid- 2-hydroxy-1,3-propanediyl ester) on the acid site while the oxygen anion in the hydroxyl group of methanol with the carbonyl group in the triglyceride on the basic site. The breaking down of the 3 fatty acid chains of the triglyceride (transesterification reaction) in the basic site will yield 3 FAME molecules and glycerol as the byproduct. Meanwhile, the esterification of glycerol with FFA on the acid site will produce water and the diglyceride compound. Finally, desorption of the products will take place and the catalytic cycle will begin again.

Fig. 12

Predicted mechanism of the bifunctional α-Fe2O3/CaO2 catalyst

Conclusion

The bifunctional α-Fe2O3/CaO2 catalysts were successfully synthesized from 3 different CaO2 sources, with Ca(OH)2-derived catalyst having the highest crystallinity (68.8628%) and CaCO3-derived catalyst the lowest (54.1226%). It was observed that the greater addition of α-Fe2O3 component to the CaCO3-derived catalyst would give a very amorphous structure. Meanwhile, the limestone-derived α-Fe2O3/CaO2 catalyst gave the highest biodiesel yield (77.1672%), but biodiesel obtained from CaCO3-derived catalyst was the only product that met the Indonesian and European standards. The best Fe/Ca configuration was concluded to be 1:4 as it gave the highest biodiesel yield (97.0401%) with complete methyl ester conversion and no diglycerides, indicating high purity biodiesel with negligible impurities which would help reduce the wear and tear in engines. It was also discovered that the α-Fe2O3 component favored the esterification reaction and formation of diglycerides, while the CaO2 favored the transesterification reaction and formation of methyl esters (FAME). Based on these discoveries, the catalyst mechanism was predicted that esterification of FFA and glycerol taking place on the acid site will produce diglyceride, while transesterification of triglyceride by methanol took place on the basic site. These discoveries would significantly increase the efficiency in biodiesel production process as the cause for incomplete methyl ester conversion was identified, consequently giving insight to industries for a way to produce biodiesel without any by-products and impurities. Moreover, the elimination of glycerol or diglyceride separation process would also save the cost of water and electricity, hence improving the economics of production in biodiesel industries.