**2. Materials and Methods**

#### *2.1. FFA Determination and Pre-Treatment (Acid Esterification) of JCO*

Typically, biodiesel feedstocks that have a FFA value below 2% can be converted into biodiesel through a single step process known as transesterification. However, if the FFA content is > 2% than the additional process (i.e., esterification) is required. Jatropha oil was produced in Ghana and collected from a UK supplier. Isopropanol, H2SO4, methanol, phenolphthalein 1% and potassium hydroxide (KOH) were purchased from Sigma-Aldrich (Dorset, UK). An acid base titration method (using KOH) was used to determine the FFA content present in JCO. Alcohol to oil ratio, temperature and reaction time are important parameters for esterification process [38,39]. The formulae used by Heroor and Bharadwaj [38] was used to calculate the FFA content (in %) of the jatropha oil and biodiesel. The most favorable esterification process ascertained for the jatropha oil were: using a methanol to oil (molar) ratio of 6:1, with a H2SO4 catalyst concentration of 0.5%, reaction time of 45 min whilst maintaining the temperature of 40 ± 5 ◦C [40]. Esterification of jatropha oil was conducted using 5% H2SO4 and 20% methanol at a steady temperature of 65 ◦C [41]. Tiwari et al. [42] reported that for esterification of jatropha oil, a methanol to oil ratio of 0.28:1 (*v*/*v*) should be used with a catalyst concentration of 1.43%, along with a reaction time of 88 min at 60 ◦C temperature. Based on the above literature and laboratory trials, following methods were applied for esterification of JCO:


#### *2.2. Transesterification*

The FFA content in pre-treated JCO (i.e., esterified JCO) was measured again by the titration procedure in order to make sure that the FFA value is below 2%. Literature reported effective transesterification of JCO by using a methanol to oil (molar) ratio of 9:1, and with a KOH catalyst of 0.5% [40]. In another study, the recommended amount of catalyst suggested is between 0.1% to 1% (*w*/*w*) of oils [43]. Furthermore, another literature reported that the molar ratio of methanol to oil at 6:1, NaOH concentration of 0.7% (*w*/*w*), and reaction temperature of 65 ◦C was found to be effective for the transesterification of JCO [41]. The optimum parameters for the transesterification of JCO were ascertained by implementing: (i) a methanol to oil ratio of 0.16 (*v*/*v*), and (ii) a constant temperature of 60 ◦C. The KOH was used as a catalyst.

#### *2.3. Addition of Nanoparticles to Neat Jatropha Biodiesel*

Jatropha biodiesel was produced in the lab (Sections 2.1 and 2.2); nanoparticles Al2O3 and CeO2, and surfactant Triton X-100 were purchased from Sigma-Aldrich. A GT Sonic ultrasonicator (Shenzhen, China) was used for mixing the nanoparticles in biodiesel-surfactant mixture. Quantities of 50 ppm and 100 ppm of nanoparticles were added to neat jatropha biodiesel (J100). The following method was adapted for nanoparticles addition into neat biodiesel:


• Biodiesel-nanoparticles mixtures were placed in an ultrasonicator (frequency at 40 kHz and water at 45 ◦C) for a duration of 45 min. After that the samples were left for 72 h at room temperature to see the stability of the mixture

**Figure 1.** Fuel samples (from left to right): Diesel, J100, J100A100, J100A50, J100C100 and J100C50.

#### *2.4. Characterisation of Fuel Samples*

Crucial properties that are vital to the fuel's performance in engine and exhaust emissions were measured in the lab. A Parr 6100 bomb calorimeter (Parr Instrument Company, Moline, IL, USA) was used to measure the higher heating value (HHV) in accordance to ASTM-D240 standard. Canon Fenski u-tube viscometers (CANON Instrument Company, State College, PA, USA) and thermostatic water bath (±0.1 ◦C) was used to measure the kinematic viscosities according to ISO 3104 having an accuracy of ±0.22%. Viscosity was measured at temperatures of 40 ◦C and 22 ◦C. The densities were measured using a hydrometer in accordance to ISO 3675 standard. Flash point temperatures were measured using a Setaflash series 3 plus closed cup flash point tester (model 33000-0, STAN-HOPE SETA, UK). The test methods used were in compliance with DIN EN 22719, a part of the EN14214 standard. The acid value was measured using the same technique used for FFA measurement [30].

### *2.5. Engine Testing*

A model LPWSBio3 three cylinder engine manufactured by Lister Petter (Teignmouth, UK) was used in the investigation (Table 1). An eddy current Froude AG80HS dynamometer (Froude Ltd., Worcester, UK) was used to measure and adjust the engine load and speed (Figure 2). The torque and speed accuracies of the dynamometer are ±0.4 Nm and ±1 rpm respectively. A five-gas emission analyser BEA 850 (Robert Bosch Ltd., Middlesex, UK) and smoke opacity meter (Bosch RTM 430) was used to analyse the exhaust gas components (CO, CO2, NOx, O2 and UHC) and to measure the smoke intensity respectively. The resolution for CO, CO2, NOx, O2 and UHC measurements are 0.001% vol., 0.01% vol., 1 ppm vol., 0.01% vol. and 1 ppm vol. respectively. The absorption coefficient resolution for the smoke meter is 0.01 m<sup>−</sup>1. Ratio of air to fuel was also measured using the same emission analyser. A LabVIEW data acquisition system was used to log the temperatures at different locations of the engine. Combustion characteristics were evaluated using a Kistler combustion analyser. A pressure sensor (6125C11, Kistler Instruments Ltd., London, UK) and charge amplifier (Kistler 5064B11) was used to measure pressure inside the cylinder. Another pressure sensor (Kistler 4065A500A0) and amplifier (Kistler 4618A0) was used to measure the fuel line injection pressure. An optical encoder (Kistler 2614A) was used for detection of the crank angle position. The amplifiers and the encoder electronics were connected to the 'KiBox' (Kistler, model 2893AK8) for data logging. The KiBoxCockpit software (Kistler Instruments Ltd., London, UK) was used to measure and analyse various combustion parameters such as in-cylinder pressure, P-V diagram, heat release rate, combustion duration etc.

**Table 1.** Specifications of the engine used in the experiment.

**Figure 2.** Schematic diagram of the engine test rig.

Fossil diesel, neat jatropha biodiesel (J100) and neat jatropha biodiesel with aluminium nanoparticles were tested in the engine. The engine was first operated with fossil diesel, then switched to neat jatropha biodiesel, and then finally operated with jatropha biodiesel-nanoparticle blend. The engine was operated at constant speed of 1500 rpm. For each test fuel, the engine was tested at six (6) different loads starting from low to full engine load. Once all data were measured and recorded at one load, the engine was ramped up to the next load using the dynamometer. At the end of test, the engine was switched back to fossil diesel and operated for about 15 min before stopping the engine.

#### **3. Results and Discussion**

#### *3.1. Nanoparticles Addition and Fuel Characteristics*

Through visual observation it was found that the Al2O3 nanoparticles had fully dissolved into their biodiesels and there was no sedimentation present. On the other hand, CeO2 nanoparticles did not dissolved completely and some sedimentation was seen at the bottom of the container for both 50 ppm and 100 ppm doses. Hence, blend containing Al2O3 nanoparticles was chosen for engine testing due to having better diffusion characteristics. The failure of the CeO2 not mixing fully was perhaps due to the type of surfactant used. Other surfactant such as Span 80 and Tween 80 might help to blend CeO2 nanoparticles with J100 fuel [44].

The results given in Table 2 demonstrate that the properties of J100 with or without nanoparticles mostly comply with the EN14214 standard. The properties of the J100 biodiesel was similar to the properties reported in the literature [41]. Density of J100 and jatropha-nanoadditive blends were higher than that of fossil diesel. Greater fuel density would allow for more fuel to be pumped via the fuel line, and greater mass of the fuel can be stored in a tank [40]. The EN14214 standard states that all biodiesel fuels must have acid value lower than 0.50 mg KOH/g, it had been observed that all fuel samples except JCO were able to achieve this value. Due to the addition of both surfactant and nanoparticles, the heating values of the nanoadditive fuel blends samples were slightly lower than that of neat jatropha biodiesel.


**Table 2.** Measured properties of the test fuels.

#### *3.2. Engine Performance and Emission Characteristics*

Pure fossil diesel, J100 and J100A100 fuels were tested in the engine. In general, the bsfc for fossil diesel operation was found to be lower than those obtained for J100 and J100A100 fuels (Figure 3). The reason why J100 and J100A100 had higher BSFC values was owed to the fact they both had higher density values and lower calorific values when compared to corresponding values for fossil diesel.

**Figure 3.** BSFC vs. engine load.

At about 3.8 kW engine load, the BSFC of the fossil diesel was 6% lower than that of J100A100 fuel. However, on average, a difference of 3% in BSFC was observed between J100 and J100A100 fuels. At 9.75 kW (100%) load, the BSFC of J100A100 was about 13% higher than that of fossil diesel; on the other hand, the BSFC of J100 was 4.5% lower when compared to J100A100 fuel (Figure 3). On the contrary, it was observed that the brake specific energy consumption (BSEC) value of the J100A100 fuel was lower than that of fossil diesel throughout all load range (Figure 4). At full load, BSEC of the J100A100 fuel was found to be decreased by approximately 6%. This explains that when the engine was operated with J100A100 fuel, comparatively less energy was required to produce the same power output as compared to fossil diesel operation.

**Figure 4.** BSEC vs. engine load.

The brake thermal efficiency (BTE) of J100 and J100A100 fuels were higher than the corresponding values observed for fossil diesel (Figure 5). At 20% engine load (1.9 kW), the BTE values for fossil diesel, J100 and J100A100 fuels were respectively 11%, 14.52% and 13.65%. A higher BTE value for the biodiesel-nanoparticle blend might be attributed to a higher oxygen content present in the biodiesel and higher reactivity of the fuel mixture due to the nanoadditives [14,18]. On average, the BTE values of J100A100 fuel was about 3% higher than that of fossil diesel. At higher loads, the thermal efficiency of J100 was observed to be slightly higher than those obtained for J100A100 fuel. Higher viscosity of J100A100 fuel might have caused this characteristic. At higher loads, the volume of carbon monoxide (CO) produced by the biodiesel-nanoparticle blend was found to be higher than that of fossil diesel (Figure 6); on the contrary, opposite characteristic was observed at lower loads. It was believed that higher BSFC value and higher oxygen content in the J100A100 fuel caused higher CO emissions at higher loads. Similarly, at higher loads, the CO2 emission of J100A100 fuel was found to be slightly higher than fossil diesel due to the higher BSFC value and higher oxygen content (Figure 7). The amount of UHC produced by the biodiesel blend was found to be lower than that of regular diesel (Figure 8). This was due to the fact that in the case of biodiesel-nanoparticles blend, more complete combustion took place inside the cylinder. It was observed that better combustion characteristics of the nanoadditive blend led to higher BTE (Figure 5). The J100A100 blend had the aid of an increased catalytic effect which helped in improving the overall combustion [14].

**Figure 5.** Brake thermal efficiency (BTE) vs. engine load.

**Figure 6.** CO emission vs. engine load.

**Figure 7.** CO2 emission vs. engine load.

**Figure 8.** UHC emission vs. engine load.

It was observed that the J100 and J100-nanoparticle blend produced a greater amount of nitrogen oxide (NOx) emissions when compared to fossil diesel (Figure 9). Biodiesels intrinsically containing a greater amount of double bond molecules caused a higher adiabatic flame temperature which in turn leads to a greater concentration of NOx emissions. In the case of nanoparticle blends, an increase in combustion temperature caused due to a greater rate of reaction and conversion of the oxygen present in the nanoparticle blend led to an increased rate of NOx emissions [33]. Under most engine loads, the NOx emissions of J100 were higher than those of J100A1000 fuel. The reason for this was due to the catalytic nature of the nanoparticles present in the J100A1000 fuel, the nanoadditives broken down hydrocarbon compounds before they were able to become fully formed products [23]. The smoke opacity values for both pure biodiesel and biodiesel-nanoparticle blend were found to be much lower than those observed for fossil diesel (Figure 10). Oxygen content of the test fuel plays a significant role in the formation of smoke [45]. Better combustion due to higher oxygen content caused lower smoke levels in the case of nanoparticle-biodiesel blend [46,47]. When comparing between J100 and J100A100 fuels, it was observed that J100A100 had a lower smoke opacity values because of the greater amount of oxygen present (aided by the Al2O3 nanoadditive) in the J100A100 fuel.

**Ϭ Ϭ͘ϭ Ϭ͘Ϯ Ϭ͘ϯ Ϭ͘ϰ ϭ͘ϵ ϯ͘ϴ ϱ͘ϳ ϲ͘ϲϱ ϳ͘ϲ ϵ͘ϳϱ >ŽĂĚ;ŬtͿ ŝĞƐĞů :ϭϬϬ :ϭϬϬϭϬϬ**

**Figure 10.** Smoke opacity vs. engine load.

#### *3.3. Combustion Characteristics*

Figures 11 and 12 shows in-cylinder pressures for all test fuels at 60% and 100% loads respectively. In general, increase in the in-cylinder pressures were observed with the increase in engine loads (Figures 11 and 12). At 60% engine load (Figure 11), the charge temperature was low, lower charge temperature lengthened the ignition delay period [48]. At 100% load (Figure 12), more fuel was injected into the chamber which caused the gas and wall temperatures to increase, this in turn reduced the ignition delay period. The peak in-cylinder pressures were occurred almost at the same crank angle location for all fuels. The fossil diesel gave highest peak in-cylinder pressure at 60% load; however, at 100% load, the peak in-cylinder pressures were almost equal for all fuels.

The heat release rates for both 60% and 100% engine loads are shown in Figures 13 and 14. At 60% load (Figure 13), the peak of heat release rate for fossil diesel was much higher than other fuels. The J100A100 fuel produced lowest peak heat release rate; this was caused due to the longer ignition delay period of J100A100 fuel when compared to fossil diesel and J100 fuels. The lower viscosity and better volatility traits of pure diesel fuel enabled to produce highest peak heat release rate at 60% engine load [17]. However, at 100% load (Figure 14), peaks for both pure diesel and J100A100 fuels were almost the same; this combustion characteristic suggested that at higher temperatures, the nanoadditive fuel blend were not as volatile, and not enough fuel mixture is formed in the premixed burning phase. The total heat release values at 60% and 100% loads are demonstrated in Figures 15 and 16. At 60% load (Figure 15), amongst all fuels, J100A100 gave highest amount of total heat release; which suggested that once burning started, nanoadditive fuel burnt quite quickly relative to other fuels. The oxygen donated by the nanoparticles aided accelerated burning of the J100A100 fuel [49]. It was observed that the total heat release values for all test fuels were almost same at 100% load (Figure 16). This might be attributed to the absorption of heat by high heat capacity gases such as CO2, whose concentration increases with the increase of engine load. These gases absorb a fragment of the amount of total heat release which caused the J100 and its blend to follow a same trend [50], as demonstrated in Figure 16.

**Figure 11.** Cylinder pressure at 60% load.

**ƌĂŶŬŶŐůĞ;ŽͿ**

**Figure 13.** Heat release rate at 60% load.

**Figure 14.** Heat release rate at 100% load.

**Figure 15.** Total heat release at 60% load.

**Figure 16.** Total heat release at 100% load.

#### **4. Conclusions and Recommendations**

Jatropha biodiesel was produced using both esterification and transesterification processes. The addition of Al2O3 and CeO2 nanoparticles to pure jatropha biodiesel was carried out by using an ultrasonicator device and surfactant Triton X-100. It was observed that the surfactant decreased the calorific values of the fuel blends. The CeO2 nanoparticles failed to fully amalgamate with the jatropah biodiesel. For successful amalgamation of the CeO2 nanoparticles into J100 fuel, other surfactant such as Span 80 and Tween 80 might be effective instead of Triton X-100. The J1000A100 blend was observed as having promising characteristics that would aid in the fuel's performance when tested in the IC engine. Neat jatropha biodiesel and neat fossil diesel fuels were used as a control. The findings on engine performance, combustion and emission characteristics results are summarised below:


To conclude, overall J100-Al2O3 nanoadditive fuel performed better when compared to fossil diesel and J100 fuels. A reduction in NOx and UHC emission as well as smoke opacity, and an overall increase in BTE were observed when compared to J100 fuel. Effect of nanoparticles in the environment is yet to be investigated and recommended as a future work. Measurement of cetane number and oxygen content in the biodiesel-nanoparticle blends are other important items for further studies.

**Author Contributions:** Conceptualization, A.K.H. and A.H.; methodology, A.K.H. and A.H.; validation, A.K.H. and A.H.; formal analysis, A.K.H. and A.H.; investigation, A.K.H. and A.H.; resources, A.K.H. and A.H.; data curation, A.K.H. and A.H.; writing—original draft preparation, A.K.H. and A.H.; writing—review and editing, A.K.H.; supervision, A.K.H.; project administration, A.K.H.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.
