Hydrocracking of Waste Cooking Oil into Biofuel Using Mesoporous Silica from Parangtritis Beach Sand Synthesis by Sonochemistry


 High content of silica in beach sand can be synthesized into mesoporous silica (MS) using the sonochemistry method with dodecyl-amine (DDA) as a surfactant template. The preparation began by adding sodium silicate dropwise to DDA solution under the rotation speed of 120 rpm. The mixture was then added with H 2 SO 4 and sonicated at 37 Khz at a temperature of 35 ºC for 20 minutes. The product was then filtered, washed, and dried at 50 °C then calcined at 600 ºC for 5 h. After it was calcined, the sample was characterized by using FTIR, Surface Area Analyzer, XRD, SEM, and TEM as well as the acidity using pyridine vapor adsorption. The synthesized MS was then used as a catalyst in hydrocracking waste cooking oil in a semi-batch stainless steel reactor system at 450 ⁰C for 2 h, under 20 ml min -1 H2 flow rate. The hydrocracking product of the liquid fraction was analyzed using GC-MS. The results showed that the best performance of the MS1 was produced by using the DDA concentration of 0.1 M, had optimum acidity at 1.7 mmolg -1 , specific surface area of 233 m 2 g -1 , total pore volume of 0.4cc/g, and average pore diameter of 7.70 nm.The best MS1 catalyst produced liquid fraction with yield of 31.13 wt.% which consisted of 11.10 % diesel oil and 5.04 % gasoline.


Introduction
Environmental pollution remains global concerns. Increasing usage of cooking oil leaves a huge amount of waste cooking oil (WCO). This waste has low economic value and can contribute to increased environmental pollution. Therefore, coconut oil waste needs to be processed into other compounds that have higher economic value. One of the methods to process this waste is by cracking it into smaller molecules, hydrocarbon preferably, which can be used as biofuel.
Waste cooking oil can be a reliable source of energy [1]. The hydrocarbon produced from this process has similar compounds to those found in petroleum fuel, in which it has higher energy density, lower viscosity, and higher stability [2]. Conversion of waste coconut oil into hydrocarbon requires an acidic catalyst that has a large surface area to accommodate the reaction. Among many types of catalyst available, mesoporous silica has drawn many attentions for the past years because of its high porosity properties and better performance [3]. Mesoporous materials are de ned as material that has a pore size in the range of 2-50 nm. The mesoporous structure could resolve and avoid diffusion limitation of bulky molecules during the catalytic process; hence, a better catalytic activity can be expected [3]. There are many synthesis routes to achieve mesoporous silica material, i.e. hydrothermal, sol-gel. Previous literature states that alkaline medium could affect the control of mesopore size distribution using sol-gel methods, hence, making it easier to control the material porosity [4].
Mesoporous silica materials can be synthesized using surfactants as metal oxide pore templates, followed by surfactant removal [5]. Compared to the sol-gel and the hydrothermal method, the sonochemical method has one obvious advantage in which it is relatively faster [6]. Zhao et al. [3] studied the effect of alkaline media on the process of controlling the wide distribution of mesoporous sizes in porous bimetallic silica using the sol-gel method. The result suggests that the formation of mesoporous using the sol-gel method depends on the balance between SiO 4 and CTAB.
Mahardika et al. [7] investigated the transesteri cation of waste cooking oil using CaO/MCM-41 catalyst synthesized from Lapindo mud by the sonochemical method. The results showed that the MCM-41 material synthesized by the sonochemical method has a regular hexagonal structure and a uniform pore shape with a speci c surface area and large pore volume and pore diameter. The pore volume obtained from the sonochemical synthesis of MCM-41 catalyst was 0.37-0.68 cm 3 /g with a pore diameter of 2.5-3.5 nm. The superior properties of silica synthesized from the sonochemical method implies the immense potential that this method has. To increase the porosity properties and its performance, a good templating agent is strictly needed. Dodecyl amine is well known for its use in the synthesis of highly ordered mesoporous silica [8].
As one of the silica-containing materials, beach sand has silica composition, depending on its geographical location. Beach sands are one of the examples of carbonate sand which have a silicon dioxide content of 72-84 % [9]. Higher silica content could increase the economic value of beach sand if converted into a more valuable product. Silica and silicon dioxide could be used for any purposes, i.e. adsorbent, desiccant, lter media, and catalysator component. Silica-based material shows excellent promises in the adsorbent and catalyst industry if treated or converted into mesoporous material [10].
This research studied the usage of mesoporous silica in hydrocracking waste cooking oil. The silica used was extracted from the beach sand of Parangtritis Beach. The extracted silica was then used to synthesize silica and mesoporous silica. In mesoporous silica synthesis, dodecyl amine (DDA) was used as a template to create mesoporosity by using the sonochemistry method. The advantage of using the sonochemistry method in synthesizing the catalyst is that it does not require high temperature. Moreover, it takes a relatively short time to produce the desired catalyst [11]. The effect of template concentration was studied by analyzing the materials characteristics and its mesoporous silica activity with hydrocracking waste cooking oil.

Materials and Methods for Sample Preparation
The materials used in this study were Silica synthesis from Parangtritis beach sand; Dodecyl amine supplied by Fisher Scienti c, distilled water, HCl supplied by Mallinckrodt, NaOH supplied by VWR Chemicals, AgNO 3 , C 5 H 5 N and silica standard supplied by Sigma Aldrich.
Silica was extracted from Parangtritis beach sand which contains SiO 2 using the Re ux method [9].
Mesoporous silica was synthesized from previously extracted silica using the sonochemistry method with DDA surfactant as a template [10]. Powder SiO 2 was dissolved in a solution of 1.5 M NaOH and stirred at a temperature of 40 °C for 30 minutes to obtain soluble sodium silicate. As much as 0.5 g DDA was dissolved in a 25 mL solvent mixture of distilled water and ethanol with a ratio of 1:1, at a temperature of 40 °C for 30 minutes. Sodium silicate was added dropwise to DDA solution under rotation speed of 120 rpm at room temperature.
The mixture was left for one hour before it was added with 6M H 2 SO 4 to get a pH value of 5. The reaction mixture was stirred in Sono cator at 37 kHz and 35 ºC for 20 minutes. The solution was left under a static condition at room temperature for 1 hour. The product was then ltered and washed with distilled water until ltration reached the pH value of 6. Finally, it was dried at 50 °C for 4 hours. The dried product was then calcined at 600 ºC for 5 hours with a heating rate of 5 ºC /min to remove the surfactant template.

General procedure:
The functional group in silica, as well as the presence and disappearance of the DDA template from silica was observed and analyzed using Fourier-Transform Infrared Spectrometer (FTIR) using KBr disc technique. The acidity of sand, silica, and mesoporous silica were analyzed using gravimetric base adsorption. In this research, pyridine was used as base probe molecules. The acidity was calculated with the following formula, where W C5H5N is the weight of adsorbed pyridine vapor (g), W Y is the weight of mesopore silica and M C5H5N is the molecular weight of pyridine (79.01 g/mol). Pore size and volume were analyzed using N 2 gas sorption analysis which was carried out using Quanthachrome NOVAtouch. Adsorptiondesorption isotherms were measured by the multipoint method. The total surface areas were calculated by the BET method. BJH desorption model was used to provide pore size distribution. The mesoporous silica crystallinity was analyzed using X-ray diffraction (XRD) analysis and the analysis was performed using Rigaku Mini ex 600 with Cu Kα monochromatized radiation source (λ = 1.54 ºA), operated at 30 kV, 10 mA, scan speed 10°min -1 , and scan range 2-80°. Morphology of mesoporous silica was characterized with Scanning Electron Microscope (SEM) and Transmission Electron Microscope (TEM) to analyze its pores structure.
Hydrocracking waste cooking oil was done by putting 0.2 grams of mesoporous silica into the catalyst container and 10 grams of waste cooking oil into the oil container. Waste cooking oil sample and mesoporous silica were put into a hydrocracking reactor. The reactor was set at a temperature of 450 °C per 2 hours with a ow of hydrogen gas of 20 ml/min. The catalytic activity of mesoporous silica was evaluated in hydrocracking waste cooking oil by a reactor. The liquid produced by Hydrocracking waste cooking oil was analyzed by Gas Chromatography-Mass Spectrometer(GC-MS, Shimadzu QP2010S).

Silica
The silica was extracted from the beach sand analyzed using SEM-EDX. The extracted silica has the following characteristics. It has 21.19 % of Si, 34.22 % of O, 43.17 % of C, 1% of Br and 0.42% of Na. It has a surface area of 54.43 m 2 /g, a pore volume of 0.006 cc/g, and a pore diameter average of 3.13 nm (12). Commercial silica which was analyzed with SEM-EDX in contrast has a 40.91 % composition of Si and 59.09 % of O. It also has a characteristics surface area of 162.36 m 2 /g, a pore volume of 0.68 cc/g, and a pore diameter of 16.81 nm. Silica extracted from beach sand has a smaller surface area as it still contains impurities such as carbon, Br, Sodium (Na). Na comes from a solution of NaOH or NaCl contained in seawater salt.

Mesoporous Silica
The synthesized mesoporous silica (MS) was divided into three samples: MS1, MS2, and MS3. The rst sample, MS1, used DDA concentration of 0.1 M. MS2 and MS3 used DDA concentration of 0.2 M and 0.3 M respectively. After the samples were processed using the sonochemistry method, MS1 was immediately ltered and washed whilst MS2 and MS3 were left for 1 hour before ltered and washed until the samples free from Clor the pH value is neutral. The samples were calcinated at 600 ºC to remove its template and then characterized by FTIR before and after calcination, as shown in Figure 1.
The broadband at a wavenumber of 470 cm -1 showed bending vibration, indicating the presence of Si-O-Si. There was an in uentially strong intensity band at 1118.7 cm -1 which represented Si-OH stretching; this is due to the interaction of the O-H group originating from the water found on the silica surface through hydrogen bonds. It was generally known that 450 -1300 cm -1 range was the unique silica band [14]. The broadband at a wavenumber of 1620 cm -1 showed bending Si-OH as there were new bands at 1635 cm -1 and a strong band at 3448.7 cm -1 after the calcination, which represented water O-H bending and O-H stretching, respectively. The broadband at a wavenumber of 1465.9 cm -1 showed primer amine.
Spectra at wavenumber of 3425 cm -1 showed stretching vibrations of N-H primary amines. These bands indicate that the DDA template was still present in the MS. After the calcination process, the DDA template was expected to be removed. The DDA removal from MS was proven by the disappearance of 1465.96 cm -1 and 2924 cm -1 peaks.

Acidity Test
MS surface acidity was determined using the gravimetric acidity test with pyridine as a base probe molecule. The results are 0.2 mmol/g for sand, 0.96 mmol/g for silica and 1.7 mmol/g for MS1, MS2 2.64 and MS3 1.60 mmol/g. The higher acidity of MS showed its potential to be used as a catalyst that requires acid sites, such as hydrocracking. The greater the DDA concentration, the lower the acidity. This is due to the uneven distribution of the pores on the layer surface of the material [14].

Gas Sorption Analysis
The synthesized silica and mesoporous silica were analyzed using Gas Sorption Analyzer (GSA) to determine their porosity character. The isotherm graph and speci c surface area were calculated using the BET equation. Pore volume, average pore diameter, and pore size distribution were calculated using BJH. The results of characterization by GSA determine the classi cation of mesoporous materials based on their pore size according to the material classi cation [15]. The results of the characterization of GSA are in table 1. Table 1 The porosity property of silica and MS obtained from Parangtritis beach sand . Although silica without template was in mesoporous material range, its pore volume was shallow, compared to MS1, MS2 and MS3, indicating that low porosity formation occurred in that material. It was observed that the amount of template affects the porosity formation in MS synthesis. The more templates used, the higher the porosity properties, i.e., speci c surface area, pore volume. This phenomenon was in agreement with previous research [13]. Isotherm graph on both MS1, MS2, and MS3 was shown in Figure 2. All samples were con rmed to have a mesoporous structure, based on the presence of their hysteresis loop. These isotherm patterns are similar to the IUPAC type IV isotherm model, with the hysteresis loop similar to type H4. The hysteresis loop of type H4 was well known for materials that have narrow slit-like pores, particles with internal voids of irregular shape and broad size distribution [16], [17]. Figure 3 visualizes the mesoporous silica which has a reasonably wide pore size distribution. It is ranging from 1-10 nm, following the pore size range of the mesoporous which is 2-50 nm meaning that MS is mesoporous as expected. Besides that, there is also the distribution of pores in the micropore region. This indicates that the MS obtained is micropore and mesopore-sized, which is referred to as hierarchical mesoporous [18] [19].

Mesoporous Silica Crystallinity
MS crystallinity was examined using XRD, Figure 4. Excluding MS2, it was found that MS has an amorphous silica phase, where it was in agreement with previous studies [6], [17], [20], [21]. The diffractogram of MS2 showed an anomaly compared to the other series around 2θ and 15˚, where broadpeak could be seen. This phenomenon could be caused by the crystal formed in undecomposed surfactant found beneath the sample.

Characterization silica, MS with SEM and TEM
Characterization of MS with SEM-EDX and TEM was done to reveal the morphology of the synthesized mesoporous silica structure. SEM image was necessary to characterize the 3D surface morphology of mesoporous silica.
SEM is needed to determine the surface of the synthesized mesoporous silica. From gure 5, the mesoporous structure with the formation of small grains, the morphological structure is close to the MS structure synthesized It was seen from Figure 5. that extracted silica has a sizeable irregular chunk of silica.
The higher of DDA template concentration, there appears to be a blockage on the surface MS, this is probably due to the lack of evenness of the silicate in the template which is caused by the inappropriate sonochemistry frequency. On the other hand, it has much spherical silica that seems to be connected by an intercrystallite joint.
TEM images of MS were shown in Figure 6. The white dots on the TEM images represent the porosity that is present in the MS. The pore distribution data, Figure 3, could help to con rm that the white dot at the MS was a small mesopore that was detected by GSA. From SEM and TEM imaging, it could be concluded that dodecyl amine could act as a template, forming small particles that later could form a bigger aggregate through the intercrystallite joint, which has a small mesopore. This phenomenon was in agreement with previous literature [16].
TEM image is needed to ensure that the synthesized MS has pore gaps according to the results of the GSA analysis. The TEM image in Figure 6 appears to have clear pores like wormholes. The use of DDA as a non-ionic template made the pore structure of MS irregular. Pores with a wormhole-like structure were obtained [19,23]. At MS2, pore-blocking occurred and it caused the pore surface area to decrease, according to the results of the GSA analysis in table 4. Likewise, the higher the DDA concentration in MS3, the less regular the pore structure could get.

Catalytic Activity
The best synthesized mesoporous silica catalyst, MS1, was used for hydrocracking catalyst used in cooking oil. MS1 has a speci cally higher surface area than silica and MS2 and MS3. Activity MS1 has a liquid fraction of 31.13%. The surface area was higher and the location of the acid site could be responsible for this phenomenon, where the acid site of mesoporous silica could reside deep beneath the catalyst. While the feed and product could come in and out easily, the chance of the feed nding the active site before they exit the catalyst become slimmer. This hypothesis seems to be responsible for the relatively high amount of oxygenated product [22].  Table 3 The product distribution of hydrocracking waste cooking oil across the catalyst Under hydrocracking reaction, waste cooking oil was converted into liquid, gas, and coke, where the data could be seen in Table 3. Mesoporous silica showed superior performance in terms of producing liquid fraction but was still able to avoid over-cracking mechanism, indicated by a relatively low gas fraction produced. Over-cracking phenomenon marked by the increase of gas fraction, which is not considered to be the main product. MS1 had a better ability to produce more liquid fraction than the other catalyst. Its ability to avoid over-cracking could be owed to the good pore structure of the catalyst, where mesoporous on the surface of the catalyst enables the feed to enter the catalyst as well as releasing the product with ease. As a consequence of having higher porosity properties, mesoporous silica has a larger chance to accumulate more coke, that silica.
The ratio of alkane/alkene from the product of each catalyst re ected the ability of the catalyst to transport hydrogen atom into the intermediate molecules. Although MS1 has signi cantly higher acidity, determined by pyridine gas adsorption, compared to silica and sand, its alkane/alkene ratio was lower than silica. The location of the acid site could be responsible for this phenomenon, where the acid site of mesoporous silica could reside deep beneath the catalyst. While the feed and product could come in and out easily, the chance of the feed nding the active site before they exit the catalyst become slimmer. This hypothesis seems to be responsible for the relatively high amount of oxygenated product.
As the feed could come and go easily because of the mesopore structure, which is hypothesized to be responsible for low alkane/alkene ratio, the reaction frequency of this kind of system would be very low, leading to the high percentage of oxygenated product in the liquid fraction. The two phenomena, low alkane/alkene ratio and high oxygenated product, seems to be related to each other, where the two of them are the results of low reaction frequency, It means that many reactions take place.
Even though the mesoporous structure seems to bring disadvantage in the hydrocracking process of waste cooking oil, the mesoporous structure was responsible for the selectivity of diesel oil fraction, as presented in Fig 7. The product distribution of waste cooking oil catalyzed by MS1 turned into bimodal, where the compound with the highest yields is C 11 and C 17 . It seems that the presence of mesoporous structure in silica shifted the catalyst selectivity and also allowed the formation of diesel oil fraction by cutting less carbon chain. This task could only be done due to its mesoporous structure with dispersed acid sites. The main mechanism that turned fatty acid into hydrocarbon and its shorter hydrocarbon chain are decarboxylations, decarbonylation, and hydrodeoxygenation [23]. The reaction mechanism of hydrocracking waste cooking oil [24] can be seen in gure 8.
The Hydrodeoxygenation mechanism could be easily seen by the presence of C 18 hydrocarbon, where there is no carbon loss during the process [23]. It can be noticed that waste cooking oil does not contain any C 17 fatty acid, but the liquid product contains C 17 hydrocarbon in quite a large quantity. This phenomenon implies that either decarbonylation or decarboxylation occurred to C 18 fatty acid. This series of phenomena implies that the synthesized mesoporous silica has strong desorptive nature, making the product could exit the catalyst with more ease, which favors the formation of diesel oil fraction.

Conclusions
The mesoporous silica was successfully synthesized from Parangtritis Beach sand using the sonochemical method. High porosity properties of MS have optimal acidity at 1.7 mmol/g. The speci c surface area of 233 m 2/ g, pore volume 0.4 cc/g, average pore diameter in mesoporous range 7.70 nm. The presence of mesoporous silica was observed with SEM and TEM imaging. The mesopore structure of MS allows the formation and selectivity towards diesel oil fraction (C 13 -C 18 ), 11.10%, but this formation inevitably produces a high oxygenated product, 15.00 %.

Figure 8
Hydrocracking reaction mechanism of waste cooking oil