*3.2. Effect of Coolant Temperature in SAC*

The trend line in Figure 3 shows a slight decrease in trend, from −4 ◦C to −8 ◦C, until it drops downs steeply, at −10 ◦C, until −12 ◦C. The highest percentage of 100% purity is at the highest temperature, which is at −4 ◦C; meanwhile, at −10 ◦C, the FAME yield purity obtained is the lowest, which is at 60.68%. The crystallization temperature indicated by the onset temperature of this experiment is found to be at 9.6 ◦C and expected to end (as estimated) by the end set temperature of −7.42 ◦C. Hence, biodiesel is predicted to crystallize during conducting this experiment as all the parameters are lower than the crystallization temperature. As the coolant temperatures of −10 ◦C and −12 ◦C are much lower than the end set temperature, it is expected that this operating condition would yield a low purity of FAME. In comparison to biodiesel produced from solvent 1-butanol, the study mentioned that the highest biodiesel purity was achieved, 99.375%, when the coolant temperature was set at 12.7 ◦C [12]. This is because their onset and endset temperatures obtained from their DSC analysis were 9.45 ◦C and −5.18 ◦C, respectively, with a peak temperature of 8.4 ◦C. Nevertheless, this study was able to achieve even higher biodiesel purity which is 100% at a coolant temperature of −4 ◦C [12]. Hence, it is concluded that the use of 2-MeTHF as a solvent for SAC is able to produce higher biodiesel purity than 1-butanol despite the coolant temperature used.

In reference to the FAME purity versus coolant temperature graph, a higher yield is obtained at a temperature farther than the end set temperature and closer to the crystallization temperature. This can be explained by Ahmad et al. [18], who explained that FAME is more likely to be trapped within the solid layer developed by glycerol and other contaminants when the such temperature is approaching the crystallization point. When the heat transfer rate is slower at higher coolant temperatures, the solid can form in a more orderly pattern, leaving the pure methyl ester to concentrate in the solution [19]. The solid development rate is larger at lower temperatures of coolant, resulting in more methyl ester retention into

contaminating solids. This can be further proven by research from Yahya et al. [20], who stated that the rate of ice crystals or solid development is governed by the temperature of the coolant.

#### *3.3. Effect of Cooling Time in SAC*

The trend line from Figure 4 shows the FAME yield to be increasing from 5 min to 10 min, which is from 86.72% to 99.99% purity. From 10 min until 20 min, the FAME yield is found to decrease slightly before it increases at 25 min with the FAME yield of 99.17%. The highest purity obtained is at 10 min with 99.99% of FAME purity which can be considered to be pure biodiesel. Considering the result obtained from DSC, the crystallization temperature obtained is at 9.6 ◦C, and the end set temperature is at −7.42 ◦C. The experiment is carried out at a temperature close to the end set temperature, which is −8 °C. As the experiment is conducted at a temperature much lower than the crystallization temperature, the solid layer from the contaminants is expected to be formed in the inner vessel. In comparison to biodiesel produced from solvent 1-butanol, the study mentioned that the highest biodiesel purity was achieved, 99.375%, when the cooling time was set to 35 min [12], which was longer than the optimum cooling time found in this study. The highest biodiesel purity found in this study is 99.99%, at a cooling time of 10 min, with which even higher biodiesel purity was obtained at a shorter cooling time compared to the study with 1-butanol. Hence, it is concluded that the use of 2-MeTHF as a solvent for SAC is able to produce higher biodiesel purity than 1-butanol despite the cooling time used.

In addition, it can be examined from Table 2 that during 5 and 10 min of cooling time, a white layer of glycerol is formed. Clear yellowish liquid biodiesel can also be seen formed in the vessel. Subsequently, during 15, 20 and 25 min of crystallization time, the liquid layer of biodiesel appears to be cloudy. The glycerol layer also appears to be thick over time. During 25 min of crystallization time, the layer of glycerol can be observed to be the thickest, resulting in a small volume of biodiesel formed. A larger yield of pure methyl ester was attained by using a prolonged cooling period [19]. In addition, as stated by Ahmad and Samsuri [11], for crystallization to occur, a longer crystallization time is preferable. However, as the FAME purity drops after an increasing amount of time, it can also be deduced that an increase in cooling time would also cause the growth of solid from the methyl ester to be reduced. The authors also stated that this may have been caused due to the saturation of the solute in the liquid phase inducing contamination of the solid. Consequently, the best range for cooling time is from 10 to 15 min, as proven by the purity of FAME at a constant temperature of −8 ◦C. Thus, this cooling time is not too prolonged for the separating process to take place.
