Physicochemical, Performance, Combustion and Emission Characteristics of Melaleuca Cajuputi Oil-Refined Palm Oil Hybrid Biofuel Blend

To reduce the economic impact caused by the fossil fuel crisis and avoid relying on existing biofuels, it is important to seek locally available and renewable biofuel throughout the year. In the present work, a new light biofuel—Melaleuca Cajuputi oil (MCO)—was introduced to blend with refined palm oil (RPO). The physicochemical properties, combustion characteristics, engine performance, and exhaust emissions were comprehensively examined. It was found that the higher the percentage of MCO, the lower the viscosity and density of the blends obtained. Calorific value (CV) was increased with the increase of MCO fraction in the blend. Regression analysis has suggested that the blend of 32% (v/v) of RPO and 68% (v/v) of MCO (RPO32MCO68) is optimal to obtain viscosity and density in accordance with ASTM 6751/EN 14214 standards. The experimental results show that the in-cylinder pressure, brake torque, and brake power of the optimal blend were slightly lower than those of baseline diesel fuel. Brake specific fuel consumption (BSFC), carbon monoxide (CO), and unburnt hydrocarbon (HC) were found to be slightly higher compared to diesel fuel. Notably, nitrogen oxides (NOx) and smoke opacity were found to be decreased over the entire range of the test. Overall, the optimal blend of RPO32MCO68 has shown a decent result which marks it as a potential viable source of biofuel.


Introduction
The rising demand for global energy, depletion of fossil fuel, and concern over environmental issues has led to an intensive search for renewable and reliable biofuel. The fluctuation of crude oil price and currency exchange rate could affect the energy supply and economics of those countries that solely depend on imported oil from others. To reduce the economic impact caused by the fuel crisis, it is important to seek biofuel which is available locally and renewable throughout the year. Thus, continuous effort in the search for new energy resources is very important to ensure energy security for the future.
Over the past decades, straight vegetable oils (SVOs) have gained interest among researchers as a potential fossil fuel replacement. Their comparable properties to diesel, nontoxicity, and the absence of sulfur content make them an attractive alternative biofuel. At present, the European Union is the major producer of rapeseed and sunflower oils, while soybean, palm, and coconut oils are largely produced in the United States, South East Asia, and Philippines, respectively [1]. At present, over 95% of biodiesel produced globally is from edible vegetable oil [2].
(also spelt cajuput or cajeput) is possibly referring to the Indonesian name for the tree [25,26]. Melaleuca Cajuputi oil (MCO) is categorized as an essential oil commercially used as a home remedy for stomach aches, headaches, colds, and relief of itching caused by insect or mosquito bites [26]. Trees of these species can typically reach up to 25 m in height. Figure 1 shows the tree and leaves of Melaleuca cajuputi while Figure 2 shows the neat MCO and refined palm oil (RPO). The world production of MCO is estimated to be more than 600,000 kg annually where Indonesia and Vietnam produce more than half of the total oil. Indonesia is estimated to produce more than 350,000 kg of oil annually, and Vietnam produces an average of 100,000 kg of oil per year [27,28]. This oil is extracted from the leaves and twigs mostly through water-steam distillation. This extraction method is cheap, simple, and not hazardous. An alternative process, like supercritical fluid extraction, is seldom used because it is more expensive compared to the water-steam distillation process. Figure 3 shows the schematic diagram of the water-steam distillation method to extract MCO.
This oil is a complex mixture of volatile components extracted from the leaves and twigs through water-steam distillation. MCO can be divided into two main groups which are monoterpenes (monoterpenes hydrocarbon and oxygenated monoterpenes) and sesquiterpenes (sesquiterpenes hydrocarbon and oxygenated sesquiterpenes) [29]. As previously reported, the major constituent of MCO is 1,8-cineole comprising 30% to 70% of the oil [25,30]. In addition, MCO comprises lipophilic compound; this lipophilicity property enhances MCO solubility in SVOs and a wide range of liquid hydrocarbons without the presence of a surfactant [26].
Refined palm oil (RPO) is a product from the fractionation process of crude palm oil, which has better flow properties compared to crude oil. In general, crude palm oil is a semi-solid fat at temperatures lower than 20 • C. Palm oil was chosen in this study because of its abundantly available feedstock and some of its properties are similar to diesel fuel [31]. In comparison, palm oil yields the highest oil compared to soybean, sunflower, rapeseed, or even other non-edible vegetable oils. Palm oil can produce an average of 4 to 5 tons of oil every year for every hectare of land [32]. The high yield of palm oil is making it a promising feedstock for an alternative biofuel. Despite numerous research conducted on non-edible oils, those non-edible oils are still unable to compete with palm oil in terms of oil yielded.
Energies 2018, 11, x 3 of 20 remedy for stomach aches, headaches, colds, and relief of itching caused by insect or mosquito bites [26]. Trees of these species can typically reach up to 25 m in height. Figure 1 shows the tree and leaves of Melaleuca cajuputi while Figure 2 shows the neat MCO and refined palm oil (RPO). The world production of MCO is estimated to be more than 600,000 kg annually where Indonesia and Vietnam produce more than half of the total oil. Indonesia is estimated to produce more than 350,000 kg of oil annually, and Vietnam produces an average of 100,000 kg of oil per year [27,28]. This oil is extracted from the leaves and twigs mostly through water-steam distillation. This extraction method is cheap, simple, and not hazardous. An alternative process, like supercritical fluid extraction, is seldom used because it is more expensive compared to the water-steam distillation process. Figure 3 shows the schematic diagram of the water-steam distillation method to extract MCO. This oil is a complex mixture of volatile components extracted from the leaves and twigs through water-steam distillation. MCO can be divided into two main groups which are monoterpenes (monoterpenes hydrocarbon and oxygenated monoterpenes) and sesquiterpenes (sesquiterpenes hydrocarbon and oxygenated sesquiterpenes) [29]. As previously reported, the major constituent of MCO is 1,8-cineole comprising 30% to 70% of the oil [25,30]. In addition, MCO comprises lipophilic compound; this lipophilicity property enhances MCO solubility in SVOs and a wide range of liquid hydrocarbons without the presence of a surfactant [26].
Refined palm oil (RPO) is a product from the fractionation process of crude palm oil, which has better flow properties compared to crude oil. In general, crude palm oil is a semi-solid fat at temperatures lower than 20 °C. Palm oil was chosen in this study because of its abundantly available feedstock and some of its properties are similar to diesel fuel [31]. In comparison, palm oil yields the highest oil compared to soybean, sunflower, rapeseed, or even other non-edible vegetable oils. Palm oil can produce an average of 4 to 5 tons of oil every year for every hectare of land [32]. The high yield of palm oil is making it a promising feedstock for an alternative biofuel. Despite numerous research conducted on non-edible oils, those non-edible oils are still unable to compete with palm oil in terms of oil yielded.

Materials
Melaleuca Cajuputi oil (MCO) used in this study was purchased from a local supplier who imports it from Indonesia. In Indonesia, most of the Cajuputi trees are found in the Maluku islands. MCO is highly volatile and has a clear greenish color. Meanwhile, the refined palm oil (RPO) used was purchased from the local market.

Hybrid Biofuel Preparation
Hybrid biofuel of RPO-MCO blends was prepared on a volume basis (v/v). Two samples of straight oil of 100% RPO and 100% MCO, as well as three blends at a mixing ratio of 25% (MCO25RPO75), 50% (MCO50RPO50), and 75% (MCO75RPO25) MCO with RPO, were prepared. Due to MCO's lipophilicity property, it is able to blend with RPO at any ratio without the presence of a surfactant [26]. This advantage leads to simplicity of the blend at minimum cost and fast preparation. Each blend was homogenously mixed with a magnetic stirrer. The gravitational stability test [33] was conducted for a month to observe any signs of phase separation. A stable blend is important to ensure the quality of the fuel and make it commercially usable.

Physicochemical Properties Analyses
The kinematic viscosity of RPO, MCO, and their blends were measured according to American Society for Testing and Materials (ASTM) D445. Calibrated glass capillary viscometer (Cannon-

Materials
Melaleuca Cajuputi oil (MCO) used in this study was purchased from a local supplier who imports it from Indonesia. In Indonesia, most of the Cajuputi trees are found in the Maluku islands. MCO is highly volatile and has a clear greenish color. Meanwhile, the refined palm oil (RPO) used was purchased from the local market.

Hybrid Biofuel Preparation
Hybrid biofuel of RPO-MCO blends was prepared on a volume basis (v/v). Two samples of straight oil of 100% RPO and 100% MCO, as well as three blends at a mixing ratio of 25% (MCO25RPO75), 50% (MCO50RPO50), and 75% (MCO75RPO25) MCO with RPO, were prepared. Due to MCO's lipophilicity property, it is able to blend with RPO at any ratio without the presence of a surfactant [26]. This advantage leads to simplicity of the blend at minimum cost and fast preparation. Each blend was homogenously mixed with a magnetic stirrer. The gravitational stability test [33] was conducted for a month to observe any signs of phase separation. A stable blend is important to ensure the quality of the fuel and make it commercially usable.

Physicochemical Properties Analyses
The kinematic viscosity of RPO, MCO, and their blends were measured according to American Society for Testing and Materials (ASTM) D445. Calibrated glass capillary viscometer (Cannon-

Materials
Melaleuca Cajuputi oil (MCO) used in this study was purchased from a local supplier who imports it from Indonesia. In Indonesia, most of the Cajuputi trees are found in the Maluku islands. MCO is highly volatile and has a clear greenish color. Meanwhile, the refined palm oil (RPO) used was purchased from the local market.

Hybrid Biofuel Preparation
Hybrid biofuel of RPO-MCO blends was prepared on a volume basis (v/v). Two samples of straight oil of 100% RPO and 100% MCO, as well as three blends at a mixing ratio of 25% (MCO25RPO75), 50% (MCO50RPO50), and 75% (MCO75RPO25) MCO with RPO, were prepared. Due to MCO's lipophilicity property, it is able to blend with RPO at any ratio without the presence of a surfactant [26]. This advantage leads to simplicity of the blend at minimum cost and fast preparation. Each blend was homogenously mixed with a magnetic stirrer. The gravitational stability test [33] was conducted for a month to observe any signs of phase separation. A stable blend is important to ensure the quality of the fuel and make it commercially usable.

Physicochemical Properties Analyses
The kinematic viscosity of RPO, MCO, and their blends were measured according to American Society for Testing and Materials (ASTM) D445. Calibrated glass capillary viscometer (Cannon-Fenske The density was obtained by dividing mass over the volume of the biofuel sample. The volume was measured using a 25 mL Gay-Lussac pycnometer at an ambient temperature (27 • C) and weighed using a high precision Shimadzu ATX224 analytical balance. The pycnometer used was certified according to ISO 3507.
The calorific value was measured according to ASTM D240. An adiabatic oxygen bomb calorimeter (Nenken 1013-B, Japan) was used to measure the energy available from the biofuel.
RPO fatty acids and MCO constituents were analyzed using the Perkin Elmer, Clarus 600 gas chromatography mass spectrometry (GC-MS). The GC-MS used helium gas as a carrier with a column size of 30 m × 250 µm. The initial temperature was 50 • C and then increased to 230 • C at 3 • C/min. The chemical compositions were confirmed by comparing their retention times and mass spectra with data from library databases and published literature.
The flash point value was measured according to ASTM D3828. A closed cup flash tester (K16500, Koehler instrument company, USA) was used to measure the flash point value of the MCO.
The acid value of MCO was measured according to ASTM D664. A 785 DMP Titrino (Metrohm, Switzerland) was used to measure the amount of acidic substance in the oil. Generally, a high acid value will cause corrosion to the fuel system and parts of the internal combustion engine.
The water content of MCO was measured according to European Standards (EN) 12937 and carried out on a 787 Karl Fisher titrator (Metrohm, Switzerland). High water content will likely promote the corrosion of the engine parts and injection system.
Boiling point was determined using the Melting Point M-565 (BUCHI, Switzerland). The measurement method was performed according to the manufacturer's standard.

Hybrid Biofuel Optimization and Selection
In the present study, a high viscosity RPO and low viscosity MCO were used to formulate a hybrid biofuel, aiming to study its potential as a new source of biofuel. This hybrid biofuel is totally 100% neat oil, which does not include any biodiesel or fossil fuel in its preparation. Somehow, at present, there is no specific biofuel standard for such a blend. Therefore, in this work, the newly developed hybrid biofuel will be benchmarked against the ASTM D6751/EN 14214 standards as guidelines. It is renowned that RPO is the promising biofuel alternative to fossil diesel. However, the main reasons why RPO cannot directly be used as diesel engine fuel are its relatively high viscosity and density, which lead to poor atomization, low volatility, and excessive carbon deposit inside the combustion chamber in the long term. Thus, in this study, only the selected and optimized hybrid biofuel blend will be chosen to be tested on a diesel engine to investigate its performance and emission characteristics.
Obviously, viscosity and density have enormous influence on fuel spray characteristics, fuel atomization, flow properties, engine performance, and exhaust emissions [34][35][36]. Therefore, in this study, these two key properties were thoroughly optimized to meet the ASTM D6751/EN 14214 standards before the fuel was introduced to the engine.
The data of physicochemical properties were analyzed using the regression method to generate a mathematical model that presented the correlation between blending ratio and its effect on viscosity or density. The mathematical model obtained was used to predict the optimum blend proportion for a desirable viscosity and density according to ASTM D6751/EN 14214. This optimization method was very beneficial to expedite the experiment and reduce cost and waste of materials compared to the trial and error experiment.

Experimental Setup and Measurement Procedure
In this study, a single-cylinder air-cooled four-stroke diesel engine manufactured by Yanmar (Italy) was used for the experiment. The specifications of the engine are tabulated in Table 1. The engine was coupled to the 20 kW eddy current dynamometer to measure the brake torque and brake power. The fuel mass consumption was measured using an electronic scale with a 0.1 gram accuracy. In order to measure the air flow rate, a mass air flow sensor (hot wire type) was used. In addition, the exhaust gas temperature was measured using a K-type thermocouple.
The in-cylinder pressure was measured using a Kistler 6051A piezoelectric pressure transducer mounted at the top of the cylinder head and the signal was transmitted to a Kistler 5027A12 charge amplifier. Furthermore, the crankshaft rotation angle was measured using an incremental type rotary encoder.
In this study, the exhaust emissions and smoke opacity were measured using a SPTC Autocheck (Korea) emissions and smoke analyzer. The emissions analyzer measured nitrogen oxides (NO x ) at a range of 0-5000 ppm, hydrocarbon (HC) at a range of 0-10,000 ppm, carbon monoxide (CO) at a range of 0-10% vol., carbon dioxide (CO 2 ) at a range of 0-20% vol., and oxygen (O 2 ) at a range of 0-25% vol. Meanwhile, the smoke analyzer measured smoke opacity at a range of 0-100% with a resolution of 0.1%.
In this work, only optimized hybrid biofuel which has the viscosity and density that meet ASTM D6751/EN 14214 standards was selected for the engine test and compared with the baseline diesel fuel. The experiment was performed at the wide-open throttle (WOT) with engine speed ranging from 2000 to 3500 rpm. Before the experiment started, the engine was initially running for 10 minutes to allow the engine to attain the optimum operating condition. For in-cylinder pressure measurement, 100 continuous engine cycles were recorded, and the average pressure trace was calculated. The schematic diagram of the experimental setup is presented in Figure 4. power. The fuel mass consumption was measured using an electronic scale with a 0.1 gram accuracy. In order to measure the air flow rate, a mass air flow sensor (hot wire type) was used. In addition, the exhaust gas temperature was measured using a K-type thermocouple. The in-cylinder pressure was measured using a Kistler 6051A piezoelectric pressure transducer mounted at the top of the cylinder head and the signal was transmitted to a Kistler 5027A12 charge amplifier. Furthermore, the crankshaft rotation angle was measured using an incremental type rotary encoder.
In this study, the exhaust emissions and smoke opacity were measured using a SPTC Autocheck (Korea) emissions and smoke analyzer. The emissions analyzer measured nitrogen oxides (NOx) at a range of 0-5000 ppm, hydrocarbon (HC) at a range of 0-10,000 ppm, carbon monoxide (CO) at a range of 0-10% vol., carbon dioxide (CO2) at a range of 0-20% vol., and oxygen (O2) at a range of 0-25% vol. Meanwhile, the smoke analyzer measured smoke opacity at a range of 0-100% with a resolution of 0.1%.
In this work, only optimized hybrid biofuel which has the viscosity and density that meet ASTM D6751/EN 14214 standards was selected for the engine test and compared with the baseline diesel fuel. The experiment was performed at the wide-open throttle (WOT) with engine speed ranging from 2000 to 3500 rpm. Before the experiment started, the engine was initially running for 10 minutes to allow the engine to attain the optimum operating condition. For in-cylinder pressure measurement, 100 continuous engine cycles were recorded, and the average pressure trace was calculated. The schematic diagram of the experimental setup is presented in Figure 4.

Gravitational Stability Test
All the blends were kept in bottles and stored at ambient temperature for a month. The blends were visually observed, and no sign of separation occurred during the test period. The blends also remained as one-phase clear and transparent liquid. Thus, the blends were considered highly stable and suitable for long-term storage and transportation. Figure 5 shows the appearance of blended fuels after being stored for one month.

Gravitational Stability Test
All the blends were kept in bottles and stored at ambient temperature for a month. The blends were visually observed, and no sign of separation occurred during the test period. The blends also remained as one-phase clear and transparent liquid. Thus, the blends were considered highly stable and suitable for long-term storage and transportation. Figure 5 shows the appearance of blended fuels after being stored for one month.   Table 2 lists the relative content of the main composition from MCO expressed as the percentage from the total area obtained from the gas chromatography mass spectrometry (GC-MS) analysis. Overall, there were 35 compounds (96.6% of the total oil) identified. Hydrocarbon is the major group at 64.8% and oxygenated product at 34.6%. Under oxygenated products, seven alcohols were identified to represent 6.8% of the total oil. The presence of alcohol is important for its antibacterial properties [37], and to enhance the volatility of the oil. The high compound of hydrocarbons (HC) and the presence of alcohols are possibly the reasons for the high calorific value (CV) of MCO compared to that of diesel fuel. The high compound of HC also indicated that MCO is a potential source of biofuel [38].

Chemical Constituent of Melaleuca Cajuputi oil
Individually, the main components of MCO are 1,8-cineole (27.7%), alpha-pinene (26.1%), and tetratetracontane (25.9%). The cyclic ether 1,8-cineole, has high stability and low chemical reactivity which make it resistant to polymerization, oxidation, and thermal decomposition [22]. It is reported that the blend of eucalyptus oil (which is rich in 1,8-cineole) either with gasoline or biodiesel has enhanced the performance and exhaust emissions of the engine [21,22,39]. Thus, sole or blended MCO with other fuels is expected to provide reliable engine performance and exhaust emissions.    Table 2 lists the relative content of the main composition from MCO expressed as the percentage from the total area obtained from the gas chromatography mass spectrometry (GC-MS) analysis. Overall, there were 35 compounds (96.6% of the total oil) identified. Hydrocarbon is the major group at 64.8% and oxygenated product at 34.6%. Under oxygenated products, seven alcohols were identified to represent 6.8% of the total oil. The presence of alcohol is important for its antibacterial properties [37], and to enhance the volatility of the oil. The high compound of hydrocarbons (HC) and the presence of alcohols are possibly the reasons for the high calorific value (CV) of MCO compared to that of diesel fuel. The high compound of HC also indicated that MCO is a potential source of biofuel [38].

Gravitational Stability Test
All the blends were kept in bottles and stored at ambient temperature for a month. The blends were visually observed, and no sign of separation occurred during the test period. The blends also remained as one-phase clear and transparent liquid. Thus, the blends were considered highly stable and suitable for long-term storage and transportation. Figure 5 shows the appearance of blended fuels after being stored for one month.   Table 2 lists the relative content of the main composition from MCO expressed as the percentage from the total area obtained from the gas chromatography mass spectrometry (GC-MS) analysis. Overall, there were 35 compounds (96.6% of the total oil) identified. Hydrocarbon is the major group at 64.8% and oxygenated product at 34.6%. Under oxygenated products, seven alcohols were identified to represent 6.8% of the total oil. The presence of alcohol is important for its antibacterial properties [37], and to enhance the volatility of the oil. The high compound of hydrocarbons (HC) and the presence of alcohols are possibly the reasons for the high calorific value (CV) of MCO compared to that of diesel fuel. The high compound of HC also indicated that MCO is a potential source of biofuel [38].

Chemical Constituent of Melaleuca Cajuputi oil
Individually, the main components of MCO are 1,8-cineole (27.7%), alpha-pinene (26.1%), and tetratetracontane (25.9%). The cyclic ether 1,8-cineole, has high stability and low chemical reactivity which make it resistant to polymerization, oxidation, and thermal decomposition [22]. It is reported that the blend of eucalyptus oil (which is rich in 1,8-cineole) either with gasoline or biodiesel has enhanced the performance and exhaust emissions of the engine [21,22,39]. Thus, sole or blended MCO with other fuels is expected to provide reliable engine performance and exhaust emissions.    Individually, the main components of MCO are 1,8-cineole (27.7%), alpha-pinene (26.1%), and tetratetracontane (25.9%). The cyclic ether 1,8-cineole, has high stability and low chemical reactivity which make it resistant to polymerization, oxidation, and thermal decomposition [22]. It is reported that the blend of eucalyptus oil (which is rich in 1,8-cineole) either with gasoline or biodiesel has enhanced the performance and exhaust emissions of the engine [21,22,39]. Thus, sole or blended MCO with other fuels is expected to provide reliable engine performance and exhaust emissions.

Basic Physicochemical Properties of RPO-MCO Blends
The properties of biofuel are important to ensure the suitability of the oil before being introduced to the engine. The basic properties of MCO, RPO, and their blends are presented in Table 3. The kinematic viscosity of MCO was 2.2 mm 2 /s, which is comparable to diesel fuel, yet significantly lower compared to those of vegetable oils. Mostly, vegetable oils either edible or non-edible have viscosity in the range of 25-40 mm 2 /s. The low viscosity suggests that MCO will offer a good flow property, especially during cold weather. MCO density was determined at 869 kg/m 3 which was within the range of diesel. Viscosity and density are the well-known factors that have significant effects on fuel atomization and injection spray characteristics. MCO, having low viscosity and density, is expected to provide a decent quality of spray characteristics and enhance fuel atomization for a better air fuel mixture. Notably, the amount of energy (CV) obtained from MCO is comparable to that of diesel fuel. High CV is important to provide good engine thermal efficiency and reduce fuel consumption. The flash point of MCO was 50 • C, which is mostly similar to diesel fuel. Its low flash point value indicated that MCO is a volatile oil, which is important for fuel atomization. However, water content and acid value of the oil were high compared to the ASTM 6751 standard. The high moisture and acid content could lead to corrosion of certain parts of the engine and fuel system.
On the other hand, the kinematic viscosity of RPO was significantly higher compared to the maximum limit of ASTM D6751/EN 14214. The high viscosity of RPO comes from triglycerides that contain three fatty acids and one glycerol. The fatty acid composition of RPO is presented Table 4. The major compositions identified were oleic and palmitic acid at 49% and 32.6%, respectively. Specifically, high content of oleic acid contributes to a good oxidative stability [40]. However, high content of oleic acid contributes to high density of RPO [41]. The high viscosity and density are the main reasons that limit its potential as engine fuel. High viscosity and density will increase the fuel impingement with cylinder wall and piston due to longer fuel spray penetration. Impingement would lead to the formation of carbon deposit and higher unburnt hydrocarbon emissions.
Notably, the strategy to blend high viscosity RPO with low viscosity MCO successfully reduced the viscosity and density of hybrid biofuel. The higher the fraction of MCO, the lower the viscosity and density of the blends observed. The blend of RPO25MCO75 was found to have viscosity and density levels within the ASTM/EN standards. In addition, the calorific value of the blend was also close to diesel, which is important to provide good combustion efficiency. It was observed that all the blends exhibited lower flash points, similar to the diesel range. The flash point value describes the tendency of the fuel to initiate flash and start the flame. The higher the MCO fraction in the blends, the lower the flash point values obtained. The blends also obtained lower boiling points which could enhance the atomization and evaporation of the fuel.  Even though the blend of RPO25MCO75 successfully achieved the viscosity and density within the ASTM D6751/EN 14214 standards, the present work targeted to obtain the blend that has the highest RPO percentage but still obtain viscosity and density levels within the allowable limits. Thus, further optimization analysis was carried out to select the optimum blend ratio.

Optimization of Hybrid Biofuel
To predict and select the optimum blend ratio, the regression analysis method was employed. This method developed the mathematical model of correlation between the blends' fraction and their effect on the viscosity and the density. Figure 7 shows the correlation between the MCO fraction and viscosity obtained from the experimental data. From the regression analysis, the following mathematical equation was proposed: where v is kinematic viscosity (mm 2 /s) and x is volume fraction of MCO in the blend. The exponential regression obtained high R 2 value of 0.9996 which shows a good fit of regression trend line with the measured data. Meanwhile, the correlation between MCO fraction and density is shown in Figure 8. From the tabulated data, linear regression with R 2 of 0.9967 is the best fit to describe the correlation of MCO fraction and density of the blends. The mathematical equation of the correlation is given by: where ρ is density (kg/m 3 ) and x is volume fraction of MCO in the blend. The optimization was performed based on the mathematical equations obtained from the regression analysis. The goal of the optimization was set to maximize the fraction of RPO in the blend and the desirable viscosity and density within the ASTM D6751/EN 14214 requirements. From the evaluation of Equations (1) and (2), the optimum blend ratio of RPO32MCO68 was suggested. The predicted kinematic viscosity and density obtained are 5.7 mm 2 /s and 881.44 kg/m 3 , respectively. To evaluate the accuracy of the predicted values obtained, the measured data were compared to the calculated data. The comparison and the discrepancy between measured and calculated data were tabulated in Table 5. Notably, the discrepancy is less than 5% which indicates that the mathematical models obtained were reliable to predict the viscosity and density of the RPO-MCO blends. Therefore, the blend ratio of RPO32MCO68 was selected to further analyze its performance and emissions on a diesel engine.

Combustion Analysis
The comparison of in-cylinder pressure between optimized hybrid biofuel (RPO32MCO68) and baseline diesel fuel at various engine speeds is presented in Figure 9. The pressure curves plotted were the average of 100 consecutive combustion cycles. It can be observed that an engine running on hybrid biofuel had a slightly lower peak pressure at low and medium engine speeds. However, at high engine speeds, the peak pressure differences between hybrid biofuel and diesel became larger.

Combustion Analysis
The comparison of in-cylinder pressure between optimized hybrid biofuel (RPO32MCO68) and baseline diesel fuel at various engine speeds is presented in Figure 9. The pressure curves plotted were the average of 100 consecutive combustion cycles. It can be observed that an engine running on hybrid biofuel had a slightly lower peak pressure at low and medium engine speeds. However, at high engine speeds, the peak pressure differences between hybrid biofuel and diesel became larger.

Combustion Analysis
The comparison of in-cylinder pressure between optimized hybrid biofuel (RPO32MCO68) and baseline diesel fuel at various engine speeds is presented in Figure 9. The pressure curves plotted were the average of 100 consecutive combustion cycles. It can be observed that an engine running on hybrid biofuel had a slightly lower peak pressure at low and medium engine speeds. However, at high engine speeds, the peak pressure differences between hybrid biofuel and diesel became larger. At the engine speed of 2000 rpm, peak cylinder pressure of the RPO32MCO68 blend occurred earlier than diesel fuel. Notably, the peak cylinder pressure of the blend was found to be 80.0 bar at 367.4 • after top dead center (ATDC), while the peak cylinder pressure obtained for baseline diesel fuel was 82.4 bar at 369.5 • ATDC. The pressure difference was about 2.9% as compared to diesel fuel. At 2500 rpm, it was found that the peak pressure for baseline diesel was 76.0 bar at 369.6 • ATDC, meanwhile, the RPO32MCO68 biofuel blend obtained peak pressure of 73.1 bar at 368.1 • ATDC. The drop in peak pressure was about 3.8%. As the engine speed increased to 3000 rpm, the peak cylinder pressure for baseline diesel was 71.6 bar at 371 • ATDC and 69.7 bar at 369.5 • for the RPO32MCO68 biofuel blend. The pressure reduction was about 2.7%. Upon further increase of engine speed to 3500 rpm, the peak cylinder pressure for baseline diesel was recorded at 69.1 bar at 370.3 • ATDC, while the peak cylinder pressure for the RPO32MCO68 blend was recorded at 61.7 bar at 371.6 • ATDC. The difference in pressure was about 10.7%. At all ranges of testing, the in-cylinder pressure of hybrid biofuel (RPO32MCO68) is slightly lower as compared to baseline diesel fuel. This is probably due to lower calorific value of hybrid biofuel which leads to lower heat release during the combustion process. In addition, hybrid biofuel has a slightly higher viscosity and density compared to diesel fuel, and has contributed to poor fuel atomization and low combustion efficiency, which reduces the in-cylinder peak pressure. Generally, for both fuels, the peak cylinder pressure decreases as the engine speed increases. From the analysis, it was found that the RPO32MCO68 blend has shown remarkable peak cylinder pressure close to the baseline diesel. Taking into consideration that this blend involved only two neat biofuels, the results obtained are quite impressive.
At the engine speed of 2000 rpm, peak cylinder pressure of the RPO32MCO68 blend occurred earlier than diesel fuel. Notably, the peak cylinder pressure of the blend was found to be 80.0 bar at 367.4° after top dead center (ATDC), while the peak cylinder pressure obtained for baseline diesel fuel was 82.4 bar at 369.5° ATDC. The pressure difference was about 2.9% as compared to diesel fuel. At 2500 rpm, it was found that the peak pressure for baseline diesel was 76.0 bar at 369.6° ATDC, meanwhile, the RPO32MCO68 biofuel blend obtained peak pressure of 73.1 bar at 368.1° ATDC. The drop in peak pressure was about 3.8%. As the engine speed increased to 3000 rpm, the peak cylinder pressure for baseline diesel was 71.6 bar at 371° ATDC and 69.7 bar at 369.5° for the RPO32MCO68 biofuel blend. The pressure reduction was about 2.7%. Upon further increase of engine speed to 3500 rpm, the peak cylinder pressure for baseline diesel was recorded at 69.1 bar at 370.3° ATDC, while the peak cylinder pressure for the RPO32MCO68 blend was recorded at 61.7 bar at 371.6° ATDC. The difference in pressure was about 10.7%. At all ranges of testing, the in-cylinder pressure of hybrid biofuel (RPO32MCO68) is slightly lower as compared to baseline diesel fuel. This is probably due to lower calorific value of hybrid biofuel which leads to lower heat release during the combustion process. In addition, hybrid biofuel has a slightly higher viscosity and density compared to diesel fuel, and has contributed to poor fuel atomization and low combustion efficiency, which reduces the in-cylinder peak pressure. Generally, for both fuels, the peak cylinder pressure decreases as the engine speed increases. From the analysis, it was found that the RPO32MCO68 blend has shown remarkable peak cylinder pressure close to the baseline diesel. Taking into consideration that this blend involved only two neat biofuels, the results obtained are quite impressive.

Performance Analysis
3.6.1. Engine Torque Figure 10 shows the engine torque variation as a function of the engine speed for the RPO32MCO68 hybrid biofuel blend and baseline diesel fuel. The torque values for both fuels decreased with the increase in the engine speed. The maximum torque was at 2000 rpm and then gradually decreased as the engine speed increased afterwards. The range of test for the RPO32MCO68 blend demonstrated a comparable engine torque to that of diesel fuel except for at 3500 rpm. At maximum engine speed, torque for RPO32MCO68 was found to be 21.4% lower compared to baseline diesel fuel. Lower torque at maximum speed is associated with the low peak pressure obtained during the combustion process. This can be related to the fact that the RPO32MCO68 blend has lower CV compared to the baseline diesel. In addition, the RPO32MCO68 blend also had slightly higher viscosity which affected fuel atomization, thus leading to low combustion efficiency and in-cylinder peak pressure [42]. 3.6. Performance Analysis 3.6.1. Engine Torque Figure 10 shows the engine torque variation as a function of the engine speed for the RPO32MCO68 hybrid biofuel blend and baseline diesel fuel. The torque values for both fuels decreased with the increase in the engine speed. The maximum torque was at 2000 rpm and then gradually decreased as the engine speed increased afterwards. The range of test for the RPO32MCO68 blend demonstrated a comparable engine torque to that of diesel fuel except for at 3500 rpm. At maximum engine speed, torque for RPO32MCO68 was found to be 21.4% lower compared to baseline diesel fuel. Lower torque at maximum speed is associated with the low peak pressure obtained during the combustion process. This can be related to the fact that the RPO32MCO68 blend has lower CV compared to the baseline diesel. In addition, the RPO32MCO68 blend also had slightly higher viscosity which affected fuel atomization, thus leading to low combustion efficiency and in-cylinder peak pressure [42]. Energies 2018, 11, x 14 of 20 Figure 10. Comparison of engine torque as a function of engine speeds.

Engine Brake Power
The engine brake power as a function of engine speed for all tested fuels is presented in Figure  11. At the whole range of speed, the brake power produced from the RPO32MCO68 biofuel was slightly lower compared to baseline diesel fuel. The brake power reductions were 3.42%, 4.2%, 2.62%, and 20.6% at 2000, 2500, 3000, and 3500 rpm, respectively. Notably, the largest brake power reduction for the RPO32MCO68 blend was recorded at the maximum engine speed. The possible explanation for this result is related with the lower engine torque as discussed earlier. Despite lower brake power at 3500 rpm, interestingly, the RPO32MCO68 blend recorded a comparable brake power compared to the baseline fuel at lower engine speed. Figure 11. Comparison of engine power as a function of engine speeds. Figure 12 compares the experimental data of brake specific fuel consumption (BSFC) for RPO32MCO68 and baseline diesel fuel. BSFC is an important parameter to compare the fuel efficiency of different tested fuels. It is apparent that the RPO32MCO68 blend recorded a slightly higher BSFC than diesel fuel at the entire range of engine speeds. The BSFC increments were 3.3%, 6.6%, 7.5%, and 6.5% at 2000, 2500, 3000, and 3500 rpm, respectively. Comparing the two results, it can be seen that there was no significant increase in RPO32MCO68 BSFC as compared to baseline diesel fuel where the maximum increment was only 7.5%. Higher BSFC obtained by the RPO32MCO68 blend was mainly due to lower CV where more fuel was required to combust in order to retain the same engine speed.

Engine Brake Power
The engine brake power as a function of engine speed for all tested fuels is presented in Figure 11. At the whole range of speed, the brake power produced from the RPO32MCO68 biofuel was slightly lower compared to baseline diesel fuel. The brake power reductions were 3.42%, 4.2%, 2.62%, and 20.6% at 2000, 2500, 3000, and 3500 rpm, respectively. Notably, the largest brake power reduction for the RPO32MCO68 blend was recorded at the maximum engine speed. The possible explanation for this result is related with the lower engine torque as discussed earlier. Despite lower brake power at 3500 rpm, interestingly, the RPO32MCO68 blend recorded a comparable brake power compared to the baseline fuel at lower engine speed.

Engine Brake Power
The engine brake power as a function of engine speed for all tested fuels is presented in Figure  11. At the whole range of speed, the brake power produced from the RPO32MCO68 biofuel was slightly lower compared to baseline diesel fuel. The brake power reductions were 3.42%, 4.2%, 2.62%, and 20.6% at 2000, 2500, 3000, and 3500 rpm, respectively. Notably, the largest brake power reduction for the RPO32MCO68 blend was recorded at the maximum engine speed. The possible explanation for this result is related with the lower engine torque as discussed earlier. Despite lower brake power at 3500 rpm, interestingly, the RPO32MCO68 blend recorded a comparable brake power compared to the baseline fuel at lower engine speed. Figure 11. Comparison of engine power as a function of engine speeds. Figure 12 compares the experimental data of brake specific fuel consumption (BSFC) for RPO32MCO68 and baseline diesel fuel. BSFC is an important parameter to compare the fuel efficiency of different tested fuels. It is apparent that the RPO32MCO68 blend recorded a slightly higher BSFC than diesel fuel at the entire range of engine speeds. The BSFC increments were 3.3%, 6.6%, 7.5%, and 6.5% at 2000, 2500, 3000, and 3500 rpm, respectively. Comparing the two results, it can be seen that there was no significant increase in RPO32MCO68 BSFC as compared to baseline diesel fuel where the maximum increment was only 7.5%. Higher BSFC obtained by the RPO32MCO68 blend was mainly due to lower CV where more fuel was required to combust in order to retain the same engine speed.  Figure 12 compares the experimental data of brake specific fuel consumption (BSFC) for RPO32MCO68 and baseline diesel fuel. BSFC is an important parameter to compare the fuel efficiency of different tested fuels. It is apparent that the RPO32MCO68 blend recorded a slightly higher BSFC than diesel fuel at the entire range of engine speeds. The BSFC increments were 3.3%, 6.6%, 7.5%, and 6.5% at 2000, 2500, 3000, and 3500 rpm, respectively. Comparing the two results, it can be seen that there was no significant increase in RPO32MCO68 BSFC as compared to baseline diesel fuel where the maximum increment was only 7.5%. Higher BSFC obtained by the RPO32MCO68 blend was mainly due to lower CV where more fuel was required to combust in order to retain the same engine speed.  Figure 13 shows the variation of NOx emission for the RPO32MCO68 blend in comparison to baseline diesel fuel. High NOx formation was mainly related to the high combustion temperature. The graph shows that both fuels exhibited a similar trend where the NOX was low at low engine speed, significantly increased at medium speed, and gradually decreased toward high engine speed. Notably, at the entire range of test, the RPO32MCO68 blend produced lower NOX as compared to baseline diesel fuel. The NOX reductions were 3.9%, 3.5%, 7.7%, and 17.3% at 2000, 2500, 3000, and 3500 rpm respectively. This reduction was possibly because of lower heat released during the combustion process due to low CV [43]. Furthermore, slightly high viscosity of the blend also contributed to low combustion temperature in comparison to baseline diesel.

CO Emission
The CO emission as a function of engine speed for tested fuel is presented in Figure 14. CO emission is mainly formed when there is lack of oxygen to react with the carbon to form CO2. As can be seen in Figure 14, at all range of engine speeds, the CO emission of the RPO32MCO68 blend slightly surpassed the baseline diesel fuel. CO was high at low speed and progressively decreased as the engine speed increased. Relatively high CO at low engine speed was mainly due to the engine operating at rich air-fuel mixture, where more fuel was injected but there was less oxygen to react. Moreover, incomplete combustion due to non-homogenized air-fuel mixture also led to the formation  Figure 13 shows the variation of NO x emission for the RPO32MCO68 blend in comparison to baseline diesel fuel. High NOx formation was mainly related to the high combustion temperature. The graph shows that both fuels exhibited a similar trend where the NO X was low at low engine speed, significantly increased at medium speed, and gradually decreased toward high engine speed. Notably, at the entire range of test, the RPO32MCO68 blend produced lower NO X as compared to baseline diesel fuel. The NO X reductions were 3.9%, 3.5%, 7.7%, and 17.3% at 2000, 2500, 3000, and 3500 rpm respectively. This reduction was possibly because of lower heat released during the combustion process due to low CV [43]. Furthermore, slightly high viscosity of the blend also contributed to low combustion temperature in comparison to baseline diesel.  Figure 13 shows the variation of NOx emission for the RPO32MCO68 blend in comparison to baseline diesel fuel. High NOx formation was mainly related to the high combustion temperature. The graph shows that both fuels exhibited a similar trend where the NOX was low at low engine speed, significantly increased at medium speed, and gradually decreased toward high engine speed. Notably, at the entire range of test, the RPO32MCO68 blend produced lower NOX as compared to baseline diesel fuel. The NOX reductions were 3.9%, 3.5%, 7.7%, and 17.3% at 2000, 2500, 3000, and 3500 rpm respectively. This reduction was possibly because of lower heat released during the combustion process due to low CV [43]. Furthermore, slightly high viscosity of the blend also contributed to low combustion temperature in comparison to baseline diesel.

CO Emission
The CO emission as a function of engine speed for tested fuel is presented in Figure 14. CO emission is mainly formed when there is lack of oxygen to react with the carbon to form CO2. As can be seen in Figure 14, at all range of engine speeds, the CO emission of the RPO32MCO68 blend slightly surpassed the baseline diesel fuel. CO was high at low speed and progressively decreased as the engine speed increased. Relatively high CO at low engine speed was mainly due to the engine operating at rich air-fuel mixture, where more fuel was injected but there was less oxygen to react. Moreover, incomplete combustion due to non-homogenized air-fuel mixture also led to the formation

CO Emission
The CO emission as a function of engine speed for tested fuel is presented in Figure 14. CO emission is mainly formed when there is lack of oxygen to react with the carbon to form CO 2 . As can be seen in Figure 14, at all range of engine speeds, the CO emission of the RPO32MCO68 blend slightly surpassed the baseline diesel fuel. CO was high at low speed and progressively decreased as the engine speed increased. Relatively high CO at low engine speed was mainly due to the engine operating at rich air-fuel mixture, where more fuel was injected but there was less oxygen to react. Moreover, incomplete combustion due to non-homogenized air-fuel mixture also led to the formation of CO [43]. As the engine speed increased, more oxygen was introduced into the combustion chamber, which allowed the engine to operate at lean mixture, thus reducing CO emission. of CO [43]. As the engine speed increased, more oxygen was introduced into the combustion chamber, which allowed the engine to operate at lean mixture, thus reducing CO emission.

Figure 14.
Comparison of carbon monoxide emission as a function of engine speeds.

HC Emission
HC emission is the product of incomplete combustion due to fuel properties, fuel spray characteristics, and engine operating conditions. Figure 15 shows the HC emission variation as a function of the engine speed for the RPO32MCO68 hybrid biofuel blend and baseline diesel fuel. At the entire range of engine speeds, the HC emission of the RPO32MCO68 blend was higher compared to the baseline diesel fuel. The emission was higher at 2000 rpm but gradually decreased as engine speed increased. A possible explanation for these results may be due to high viscosity and poor fuel atomization of blended fuel. Poor fuel atomization formed a non-uniform air-fuel mixture, thus causing some of the fuel being unable to combust. Another possible explanation is the occurrence of fuel impingement with the cylinder wall and piston head due to excessive spray penetration [6]. High viscosity and density caused large droplets and longer spray penetration compared to baseline diesel fuel. In addition, blended fuel contains lower cetane numbers than diesel fuel, causing longer ignition delay and a shorter combustion period, thus leading to higher HC emission being produced.

Smoke Emission
The variation in smoke opacity over the engine speed is shown in Figure 16. Overall, the RPO32MCO68 blend exhibited lower smoke opacity as compared to the baseline diesel fuel. It was observed that the smoke opacity was highest at 2000 rpm and gradually decreased as the engine speed increased. At 2000 rpm, the RPO32MCO68 biofuel blend showed similar smoke opacity to the

HC Emission
HC emission is the product of incomplete combustion due to fuel properties, fuel spray characteristics, and engine operating conditions. Figure 15 shows the HC emission variation as a function of the engine speed for the RPO32MCO68 hybrid biofuel blend and baseline diesel fuel. At the entire range of engine speeds, the HC emission of the RPO32MCO68 blend was higher compared to the baseline diesel fuel. The emission was higher at 2000 rpm but gradually decreased as engine speed increased. A possible explanation for these results may be due to high viscosity and poor fuel atomization of blended fuel. Poor fuel atomization formed a non-uniform air-fuel mixture, thus causing some of the fuel being unable to combust. Another possible explanation is the occurrence of fuel impingement with the cylinder wall and piston head due to excessive spray penetration [6]. High viscosity and density caused large droplets and longer spray penetration compared to baseline diesel fuel. In addition, blended fuel contains lower cetane numbers than diesel fuel, causing longer ignition delay and a shorter combustion period, thus leading to higher HC emission being produced. of CO [43]. As the engine speed increased, more oxygen was introduced into the combustion chamber, which allowed the engine to operate at lean mixture, thus reducing CO emission.

Figure 14.
Comparison of carbon monoxide emission as a function of engine speeds.

HC Emission
HC emission is the product of incomplete combustion due to fuel properties, fuel spray characteristics, and engine operating conditions. Figure 15 shows the HC emission variation as a function of the engine speed for the RPO32MCO68 hybrid biofuel blend and baseline diesel fuel. At the entire range of engine speeds, the HC emission of the RPO32MCO68 blend was higher compared to the baseline diesel fuel. The emission was higher at 2000 rpm but gradually decreased as engine speed increased. A possible explanation for these results may be due to high viscosity and poor fuel atomization of blended fuel. Poor fuel atomization formed a non-uniform air-fuel mixture, thus causing some of the fuel being unable to combust. Another possible explanation is the occurrence of fuel impingement with the cylinder wall and piston head due to excessive spray penetration [6]. High viscosity and density caused large droplets and longer spray penetration compared to baseline diesel fuel. In addition, blended fuel contains lower cetane numbers than diesel fuel, causing longer ignition delay and a shorter combustion period, thus leading to higher HC emission being produced.

Smoke Emission
The variation in smoke opacity over the engine speed is shown in Figure 16. Overall, the RPO32MCO68 blend exhibited lower smoke opacity as compared to the baseline diesel fuel. It was observed that the smoke opacity was highest at 2000 rpm and gradually decreased as the engine speed increased. At 2000 rpm, the RPO32MCO68 biofuel blend showed similar smoke opacity to the

Smoke Emission
The variation in smoke opacity over the engine speed is shown in Figure 16. Overall, the RPO32MCO68 blend exhibited lower smoke opacity as compared to the baseline diesel fuel. It was observed that the smoke opacity was highest at 2000 rpm and gradually decreased as the engine speed increased. At 2000 rpm, the RPO32MCO68 biofuel blend showed similar smoke opacity to the diesel fuel. However, the smoke opacity reduced by 10.4%, 29.4%, and 64.5% at 2500, 3000, and 3500 rpm, respectively. Lower smoke opacity was possibly due to the presence of oxygen content in the blend that allowed a more homogenous mixture for better combustion efficiency. Another possible explanation for this is that blended fuel contained less sulfur, which is believed to help reduce smoke opacity [44]. Meanwhile, diesel fuels containing aromatic hydrocarbon have a greater tendency to produce more smoke (soot) [45].
Energies 2018, 11, x 17 of 20 diesel fuel. However, the smoke opacity reduced by 10.4%, 29.4%, and 64.5% at 2500, 3000, and 3500 rpm, respectively. Lower smoke opacity was possibly due to the presence of oxygen content in the blend that allowed a more homogenous mixture for better combustion efficiency. Another possible explanation for this is that blended fuel contained less sulfur, which is believed to help reduce smoke opacity [44]. Meanwhile, diesel fuels containing aromatic hydrocarbon have a greater tendency to produce more smoke (soot) [45].