Simultaneous development of biodiesel synthesis and fuel quality via continuous supercritical process with reactive co-solvent

Fuel 237 (2019) 117–125

Contents lists available at ScienceDirect

Fuel journal homepage: www.elsevier.com/locate/fuel

Full Length Article

Simultaneous development of biodiesel synthesis and fuel quality via continuous supercritical process with reactive co-solvent

T

Nattee Akkarawatkhoositha, Amaraporn Kaewchadab, Attasak Jareea,

⁎

a

Center of Excellence on Petrochemical and Materials Technology, Department of Chemical Engineering, Faculty of Engineering, Kasetsart University, Chatuchak, Bangkok 10900, Thailand b Department of Agro-Industrial, Food and Environmental Technology, King Mongkut’s University of Technology North Bangkok, Pracharat 1 Road, Wongsawang, Bansue, Bangkok 10800, Thailand

ARTICLE INFO

ABSTRACT

Keywords: Biodiesel Supercritical Reactive co-solvent Biodiesel fuel quality Microreactor

This work addresses the simple and effective technique to simultaneously improve biodiesel synthesis and fuel quality. The co-solvent technology was applied to fulfill this purpose. Experiments were performed through the transesterification of palm oil and ethanol in a microtube reactor under the supercritical conditions. The isopropanol was added in the system and played the roles as a source of reactant to produce isopropyl esters and cosolvent to enhance the homogeneity of the mixture. Upon the application of iso-propanol, high yield of biodiesel was attained under the milder conditions than those reported in the literature in terms of reaction temperature, pressure, residence time, and ethanol-to-oil molar ratio. The influence of operating conditions on the %Ester was investigated and optimization of %Ester was carried out via response surface methodology. The improvement of biodiesel quality was achieved particularly the cloud and pour points which were better quality than that of the conventional biodiesel production.

1. Introduction

requirements remain the major hurdles such as high temperature, high pressure, and high molar ratio of alcohol-to-oil in order to achieve high conversion of oil. The supercritical process has been continuously developed to reduce these limitations. For instance, two-step supercritical process, known as Saka-Dadan process [5], is firstly performed with the hydrolysis of triglyceride in subcritical water followed by esterification of fatty acid in supercritical alcohol. This process greatly reduced the requirements for reaction temperature, pressure, and molar ratio of alcohol to oil. The addition of catalyst in the system such as Cs2.5PW12O40 [6], ZnO [7], and CH3ONa [8] is another effective option to reduce the extreme supercritical operation conditions. However, these methods are complicated and require expensive instruments [9]. Therefore, the use of non-reactive solvent known as co-solvent has been proposed to deal with these problems. Several non-reactive solvents were studied including both gas and liquid phases such as propane [10], hexane [11], and carbon dioxide [12]. All reports suggested that the addition of co-solvent in the system could increase the solubility of triglyceride and alcohol, improving the mass transfer and the overall production rate of biodiesel. The advance of biodiesel synthesis by applying the co-solvent technology has been extensively reported. Iso-propanol is one of the best candidates for co-solvent to enhance the performance of biodiesel

At present, most commercial biodiesel production is performed via liquid alkali-catalyzed transesterification in a stirred-batch reactor. However, the use of this technique is uneconomical and environmentally harmful due to the high-quality feedstock requirement, un-reusable catalyst, and large wastewater generation [1]. The application of solid catalysts is perceived as one of the possible means to overcome these challenges. Biodiesel production can be both economically viable and environmentally friendly due to the reduced amount of wastewater from purification, reusability of solid catalyst, and simple separation of catalyst from the product [2]. At present, the main barriers for this route to develop the industrial-scale process include the requirements of high-quality feedstock and long residence time as well as the low catalyst activity and stability [3]. These problems are due to the side reaction (saponification) and mass-transfer limitation [4]. One of the promising approaches to overcome these problems is the synthesis of biodiesel through the catalyst-free transesterification reaction at high temperature and high pressure, which is known as a supercritical process. Although short residence time for the supercritical process is satisfying for the industrial-scale biodiesel production, other

⁎

Corresponding author at: Department of Chemical Engineering, Faculty of Engineering, Kasetsart University, Chatuchak, Bangkok 10900, Thailand. E-mail address: [email protected] (A. Jaree).

https://doi.org/10.1016/j.fuel.2018.09.077 Received 10 July 2018; Received in revised form 30 August 2018; Accepted 16 September 2018 0016-2361/ © 2018 Elsevier Ltd. All rights reserved.

Fuel 237 (2019) 117–125

N. Akkarawatkhoosith et al.

synthesis under subcritical conditions [13,14]. Chueluecha et al. [14] presented that the addition of iso-propanol as a co-solvent could improve the yield of biodiesel as well as improve the energy savings. The simple separation of iso-propanol from the biodiesel and alcohol was demonstrated. Besides, due to the mild operating temperature (below 65 °C), iso-propanol is a non-reactive co-solvent. However, iso-propanol can react with triglycerides to form biodiesel at high temperature (usually higher than iso-propanol’s boiling point [15]) and sufficiently long residence time. In this case, it is regarded as a reactive co-solvent. Despite the advanced biodiesel production, the development of biodiesel synthesis under supercritical conditions by the application of reactive co-solvent has never been studied. One major issue of biodiesel derived from vegetable oil is their coldflow properties (cloud point, pour point, and cold filter plugging point). The cold-flow properties of biodiesel obtained from both subcritical and supercritical short-chain alcohol processes are much poorer than that of diesel fuel [16,17]; thus limiting the use especially in freezing conditions. Therefore, the cold-flow properties of biodiesel should be improved. Instead of the common short-chain alcohol, the application of long-chain alcohol is a rather simple way to overcome this problem. Isopropanol is the long-chain alcohol and can be used as a reactant to convert triglyceride into biodiesel (isopropyl esters). As shown in Table 1, the cold-flow properties of isopropyl esters are better than those of the other esters. Though the improved cold flow properties by using iso-propanol can be clearly noticeable, the product becomes more viscous due to the higher molecular weights. Furthermore, the conversion of triglyceride with iso-propanol is lower than that of the other ones [18]. Apart from providing high conversion rate, ethyl esters (as fuel) exhibit good kinematic viscosity and mild cold flow properties. Consequently, the addition of iso-propanol as a co-reactant may be applied to improve the fuel quality of biodiesel especially the cold flow properties. This work dealt with the use of a simple and effective means to enhance both biodiesel synthesis and biodiesel fuel quality simultaneously by using a continuous process under supercritical co-reactant conditions. One one hand, iso-propanol was used as a solvent to reduce the mass-transfer resistance across the boundary layer of reactants (oil and ethanol). At the same time, iso-propanol was also used as a reactant to produce biodiesel. The effects of reaction temperature, molar ratio of ethanol-to-oil, weight ratio of iso-propanol-to-oil, and residence time on the purity of biodiesel were investigated in this work. Moreover, the optimization of the operating conditions was performed via response surface methodology (RSM). The efficiency of biodiesel synthesis and biodiesel fuel properties were determined and compared to the literature data.

2. Materials and methods 2.1. Materials Refined palm oil, as a source of triglycerides, was purchased from the local market. (Morakot palm oil, manufactured by Morakot Industries PCL., Thailand). Ethanol (AR grade, ≥99.9%) was purchased from Merck company. Iso-propanol (HPLC grade, ≥99.9%) as a co-reactant was supplied by RCI Labscan company. HPLC grade of acetone (≥99.8%) and acetonitrile (≥99.9%) for high performance liquid chromatography analysis were obtained from RCI Labscan and Honeywell company, respectively. The physico-chemical properties of raw materials are shown in Table 2. Table 2 Physico-chemical properties of raw materials.

No. 2 diesel fuela

Methyl estersa,c

Ethyl estersb,c

Isopropyl estersa,c

Cetane number Net heat combustion (Btu/lb) Density (60 °C) Viscosity (40 °C, mm2/s) Cloud point (°C) Pour point (°C)

42.2 18,235 0.85 2.89 −18 −30

50.4 16,072 0.87 4.59 −2 −6

48.2 17,200 0.88d 4.41 1 −4

51.5 16,155 0.87 5.26 −9 −12

a b c d

Unit

Refined palm oil

Ethanol

Iso-propanol

Molecular weight Densitya Viscosityb Critical Temperature Critical pressure Acentric factor

kg/kmol kg/m3 Mm2/s °C MPa –

848.2 0.885 41.5 912c 0.74c –

46.7 0.785 1.1 241 6.3 0.644

60.1 0.785 2.1 236 4.9 0.655

a b c

at 25 °C. at 25 °C. Cunico et al. [21].

2.2. Transesterification The continuous transesterification reaction of palm oil and co-reactant was carried out in a stainless steel microtube reactor (1/16″ OD × 0.012″ W/T) under supercritical alcohols (ethanol and iso-propanol). First, iso-propanol was mixed with palm oil to obtain a homogeneous solution at the desired weight ratio (based on oil weight). Then, the mixture (palm oil and iso-propanol) and ethanol were separately fed into a T-way micromixer (0.02″ thru hole) via two HPLC pumps (2150 HPLC pump, LKB Bromma) at different volumetric flow rates based on the ethanol-to-oil molar ratio and residence time. Prior to entering the micromixer, the feedstocks were preheated to the desired reaction temperature. After that, the mixture was introduced into the reaction zone (reactor volume; 1.85 mL) where the reaction took place. The microtube reactor was placed inside a convection oven to maintain the reaction temperature. A back-pressure regulator, installed at the outlet of the microtube reactor located outside of the convection oven, was used to control the pressure of this system. The product stream exiting the microtube was rapidly cooled down in order to stop the reaction. The outlet product was collected for purification and analysis. The sample product was purified by rinsing with deionized water and was centrifuged to obtain the purified product. To ensure the steady state condition, the product sample was collected after six folds of residence time was elapsed. The experimental setup for transesterification supercritical process is shown in Fig. 1.

Table 1 Physico-chemical properties of biodiesel and diesel fuels from the literature. Properties

Properties

Wang et al. [19]. Encinar et al. [20]. Soybean oil. at 25 °C.

118

Fuel 237 (2019) 117–125

N. Akkarawatkhoosith et al.

Fig. 1. Schematic diagram of an experimental apparatus.

2.3. Product analysis

Table 3 Variables, levels of variable and constrains used for Box-Behnken design.

The content of fatty acid esters was determined by using high performance liquid chromatography with refractive index detector (RI; model RI-101, Shodex). The C18 column (250 mm × 4.6 mm, 5 µm particle size, Advanced Chromatography Technologies) was used to separate the components at the temperature of 40 °C. The mixture solution of acetone and acetonitrile at the ratio of 70:30 (v/v) was used as a mobile phase. The peaks were identified by comparing the chromatogram of sample and that of the standard compounds. The content of fatty acid esters was evaluated based on the report of Chueluecha et al. [14]. The amount of mono-, di-, and tri-glycerides in the biodiesel product was analyzed by gas chromatography with flame ionization (GC-FID, Agilent CP-3800). The separation was performed on a DB-5HT capillary column (30 m in length × 0.32 mm in diameter × 0.5 μm in film thickness). The injection and detector temperatures were both held at 350 °C. The initial temperature of the oven was 50 °C for 1 min and increased to 180 °C with the heating rate of 15 °C/min. Next, the oven temperature was continually increased with the heating rate of 7 °C/ min to 230 °C. Finally, the temperature was increased to 380 °C with the heating rate of 30 °C/min and held at this temperature for 10 min. The flow rate of carrier gas was constant at 5 mL/min. 1 μL of the sample was injected with a split ratio of 80:1.

Variables

Symbol

Independent variables Reaction temperature Residence time Ethanol-to-oil molar ratio Iso-propanol-to-oil weight ratio Dependent variable Biodiesel purity

Unit

Levels

Constrains

−1

0

1

X1 X2 X3

°C min mol/mol

325 3.7 15:1

350 5.3 25:1

375 7 35:1

In the range In the range In the range

X4

w/w

0.1:1

0.55:1

1:1

In the range

Y

%

Optimize

Table 4 Experimental design matrix and experimental results of the response. Run

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27

2.4. Experimental design Response surface methodology, based on three-level and four-factor Box-Behnken design (BBD), was employed to investigate the main and interaction effects of the operating conditions on the purity of biodiesel (dependent variable or “response”), which is one of the important parameters to control the biodiesel quality. In this work, four major variables (independence variable) affecting the transesterification reaction include the reaction temperature (x1), residence time (x2), ethanol-to-oil molar ratio (x3), and iso-propanol-to-oil weight ratio (x4). Three different levels; low (−1), medium (0), and high (+1) were applied for each variable. Table 3 summarizes the independent and dependent variables, levels and constrains used for the design of experiments. A total of 27 experiments according to the Box-Behnken design was created and results were statistically analyzed by using Minitab statistical software (version 16). All runs were performed in a randomized order with two duplications. As well, the optimization of operating conditions was carried out to obtain the maximum response. The total BBD experimental design matrix is shown in Table 4.

119

Independent variables

Response (Y, %)

Error (%)

X1

X2

X3

X4

Experimental

Predicted

−1 0 0 0 0 −1 1 0 0 0 0 0 0 1 1 −1 1 0 0 0 1 −1 1 −1 −1 0 0

1 0 1 0 0 0 1 1 1 0 −1 −1 0 0 −1 −1 0 1 −1 0 0 0 0 0 0 −1 0

0 1 0 0 −1 0 0 1 0 0 −1 1 0 1 0 0 0 −1 0 −1 −1 1 0 0 −1 0 1

0 1 −1 0 −1 1 0 0 1 0 0 0 0 0 0 0 1 0 −1 1 0 0 −1 −1 0 1 −1

71.2 93.2 96.2 81.6 90.0 67.4 95.5 94.0 97.2 80.8 79.6 92.6 81.3 96.5 96.8 58.0 96.6 93.4 89.5 86.7 89.0 64.3 94.3 65.5 62.5 90.9 93.6

72.3 94.5 96.5 81.2 88.6 66.3 95.1 93.1 96.8 81.2 80.8 92.7 81.2 97.0 95.7 58.3 96.4 93.7 89.6 87.7 88.7 64.3 95.7 66.0 61.7 90.3 92.6

1.53 1.44 0.31 0.45 1.53 1.60 0.41 0.94 0.45 0.54 1.52 0.07 0.08 0.53 1.12 0.53 0.19 0.35 0.11 1.10 0.31 0.06 1.52 0.83 1.32 0.65 1.06

Fuel 237 (2019) 117–125

N. Akkarawatkhoosith et al.

2.5. Statistical analysis

Ester increased with increasing in residence time reaching 95% with the residence time of 7.2 min. This result was in line with those observed by Rathore et al. [26] and Farobie et al. [27]. It also indicates that the influence of residence time became insignificant when the reaction approached equilibrium as shown with the dotted line in Fig. 3, i.e., the %Ester was nearly constant when the residence time exceeded 4.4 min at the reaction temperature of 350 °C. In addition, the long-term exposure (> 10 min) at high temperatures (≥375 °C) can cause the decomposition of products [28,29]. Therefore, to obtain the optimized conditions, the residence time was studied in the range of 3.7 to 7 min. A set of experiments was designed to investigate whether the decomposition of product (biodiesel) occurred in our system. The feed was biodiesel (B100) and no alcohol was introduced into the reactor. The temperature and residence time were 350–375 °C and 3.2–8 min, respectively. The HPLC analysis of product suggested that thermal decomposition of biodiesel in this system could be neglected. Next, the influence of reactive co-solvent (iso-propanol) was examined in terms of co-solvent-to-oil weight ratio. The weight ratio in the range of 0.1:1–1:1 was varied with different reaction temperatures (250 and 350 °C) while the residence time and ethanol-to-oil molar ratio were kept constant. Fig. 4 shows that, for temperatures up to

The relationship between a set of independent variables and the response was estimated by a regression model via the Minitab statistical software. The relationship was expressed in the form of polynomial regression model as shown below: 4

Y=

0

+

4 i Xi

i=1

+

2 iXi

i=1

3

4

+

ij Xi Xj i=1 j=i+1

(1)

where Y is the dependent variable, Xi and Xj are the independent variables and β0, βi, βj, βii, and βij are regression coefficients. Analysis of variance (ANOVA) was used to validate the performance of the model. The accuracy of the regression equation was verified via the lack of fit, pure error, and coefficient of determination (R2). To study the statistical significance of main and interaction effects of variables, the P-value with the significance level of 95% was applied. 3. Results and discussion 3.1. Experimental range screening The screening of operating variable levels was carried out to ensure the appropriate range of variables which could be appropriately used to study the behavior of the system and to perform the global optimization. The constraint of variable level (high and low levels) was selected by considering several factors including the limits of our laboratory equipment, quality of product, and the applicability for industrial-scale production. In this work, one-factor-at-a-time method was used to screen the influence of operating variables. First, the reaction temperature in the range of 220–375 °C was used to study the influence of reaction temperature on the purity of biodiesel (%Ester) while other variables including ethanol-to-oil molar ratio, iso-propanol-to-oil weight ratio, and residence time were held constant. The results presented in Fig. 2 apparently indicate that the %Ester was greatly influenced by the reaction temperature. The %Ester increased from 0.5% to 85.3% as the reaction temperature increased from 220 to 350 °C under the constant ethanol-to-oil molar ratio of 35:1, iso-propanol-to-oil weight ratio of 1:1, and residence time of 4 min (see solid line in Fig. 2). This was due to the endothermic reaction in which the equilibrium constant increases with increasing temperature [22]. The %Ester became stable with further increase in reaction temperature due to the near-equilibrium conditions. In addition, elevated temperatures beyond 400 °C can significantly affect the biodiesel composition as new products are formed via thermal decomposition of the substrates [23]. The same phenomenon could be observed when the constant amount of ethanol-to-oil molar ratio and residence time were adjusted as represented by the dash line and dotted line in Fig. 2. High purity of % Ester was obtained when the reaction temperature exceeded 325 °C. Moreover, it is noted that the influence of reaction temperature under supercritical conditions was more evident than the influence on %Ester under subcritical conditions. This result was in accordance with the work of Demirba [24]. Therefore, the lower and upper edges of reaction temperature window were 325 °C and 375 °C, respectively. The influence of residence time on the %Ester was then examined. In this work, the density of reactants was calculated based on the PengRobinson equation of state (PR-EOS) [25]. The influence of residence time was investigated in the range of 1.3–20.3 min. The results are shown in Fig. 3. At 250 °C (dash line), which was close to the critical temperature of solvents (ethanol and iso-propanol), no ester was formed under the short residence time (5.98 min) and %Ester of 11.3% was obtained under prolonged residence time of 20.3 min. Different from that of near critical temperature with the same operating conditions, a profound impact of reaction temperature under supercritical conditions (350 °C) was observed with the biodiesel purity of 35.2% with a much shorter residence time of 1.8 min as shown at solid line in Fig. 3. The %

Fig. 2. Influence of reaction temperature on the %Ester under various conditions; WR: iso-propanol-to-oil weight ratio, MR: ethanol-to-oil molar ratio, and RT: residence time (min).

Fig. 3. Influence of residence time on the %Ester under various conditions; T: reaction temperature (°C), MR: ethanol-to-oil molar ratio, and WR: iso-propanol-to-oil weight ratio.

120

Fuel 237 (2019) 117–125

N. Akkarawatkhoosith et al.

Fig. 4. Influence of iso-propanol-to-oil weight ratio on the %Ester under various conditions; T: reaction temperature (°C), MR: ethanol-to-oil molar ratio, and RT: residence time (min).

Fig. 6. Influence of ethanol-to-oil molar ratio on the %Ester under various conditions; T: reaction temperature (°C), WR: iso-propanol-to-oil weight ratio, and RT: residence time (min).

350 °C, increasing of co-solvent-to-oil weight ratio could enhance the % Ester over the entire range of weight ratio study. The %Ester obtained at 350 °C was 76.0% with the co-solvent-to-oil weight ratio elevated to unity. It is probable that better mixing condition of oil and alcohols was achieved leading to the improved interfacial mass transfer [30]. However, at high reaction temperature (375 °C), the %Ester was dropped when the large amount of iso-propanol was applied. This was due to the fact that iso-propanol is less active to react with triglyceride compared to ethanol [31]. The activity of single solvent (ethanol or iso-propanol) and co-solvent (mixture of ethanol and iso-propanol) was verified via several experiments as shown in Fig. 5. The results show that the ethanol-iso-propanol mixture exhibited the highest activity, followed by the individual ethanol and iso-propanol, respectively. Hence, the positive synergistic effect of ethanol-iso-propanol mixture was observed. The co-solvent-to-oil weight ratio in the range of 0.1:1 to 1:1 was also used to further optimize the operating conditions. The influence of ethanol-to-oil molar ratio was then studied in the range of 15:1 to 45:1. The residence time, iso-propanol-to-oil weight ratio and reaction temperature were held constant. As shown in Fig. 6, the use of high ethanol-to-oil molar ratio could improve the %Ester, i.e.,

at the temperature of 350 °C, the %Ester increased from 47% to 76% when the of ethanol-to-oil molar ratio increased from 15:1 to 25:1. The reason is that the interfacial area of oil-alcohol mixture was relatively larger than that with low ethanol-to-oil molar ratio [4]. The positive influence on the of ethanol-to-oil molar ratio was in accordance with the work of He et al. [32] and Weng et al. [33]. No significant enhancement of %Ester was observed when the ethanol-to-oil molar ratio was further increased from 35:1 to 45:1. Therefore, the optimal ethanol-to-oil molar ratio should be between 15:1 and 35:1. The appropriate ranges of parameters as previously mentioned can be summarized in Table 3. This was used to optimize the operating variables in terms of %Ester through the BBD method. The interaction effects between variables were also investigated. 3.2. Main and interaction effects of variables The statistical significance of the main and interaction effects of variables on the mean %Ester were evaluated via the P-value analysis with 95% confidence level. The statistical significance of the main effect of all variables on the mean %Ester were identified according to the analysis. The effect of individual variable on the %Ester was similarly found as mentioned in the Section 3.1. The interaction effect of reaction temperature with other variables on the mean %Ester is shown in Fig. 7. The low (−) and high (+) levels of reaction temperature were used to

Fig. 5. Activity of each solvent substrate for transesterification reaction under various conditions: Experimental runs; (1) alcohol-to-oil molar ratio of 35:1, reaction temperature of 350 °C, residence time of 3 min, (2) alcohol-to-oil molar ratio of 27:1, reaction temperature of 360 °C, residence time of 5.4 min, (3) alcohol-to-oil molar ratio of 20:1, reaction temperature of 375 °C, residence time of 7 min.

Fig. 7. Interaction effect of variables on the mean %Ester (WT: iso-propanol-tooil weight ratio, MR: ethanol-to-oil molar ratio, and RT: residence time (min)). 121

Fuel 237 (2019) 117–125

N. Akkarawatkhoosith et al.

evaluate the interaction effect (325 °C and 375 °C). Note that, if a pair of parallel lines (approximately equal slope) was obtained, a statistical interaction effect does not exist. Apparently, the interaction effect was observed for the residence time-reaction temperature correlation. The interaction effect of this pair was also supported by the AVONA with the P-value of less than 0.05 (see Table 5) indicating the significant interaction effect. At low level of reaction temperature (325 °C), the mean %Ester decreased with decreasing the residence time from 7 min to 5.3 min. The opposite trend was found when the residence time was further decreased possibly due to the enhanced mixing of the reactants. On the contrary, at high level of reaction temperature (375 °C), a slight increase of mean %Ester was observed with increasing the residence time. This was owing to the fast reaction and diffusion at high temperature weakening the mixing effect. The different slopes on the plot between ethanol-to-oil molar ratio and %Ester at high and low reaction temperatures were also observed in Fig. 7 indicating the interaction effect between reaction temperature and ethanol-to-oil molar ratio (P-value = 0.03). At low-level of reaction temperature, the mean %Ester increased with increasing ethanol-to-oil molar ratio from 15:1 to 25:1 as a result of large interfacial area generated [23]. However, further increase in ethanol-to-oil molar ratio caused much dilution in the system and significant shortening of contact time. On the other hand, a slight reduction in %Ester was found at high-level region with the increase of ethanol-to-oil molar ratio from 15:1 to 25:1 and approached a plateau for the ethanol-to-oil molar ratio beyond 25:1. Apparently, the influence of ethanol-to-oil molar ratio was subdued at high reaction temperature and the ethanol-to-oil molar ratio of 15:1 was preferable. The result was in line with the work of Trentin et al. [4] who investigated the influence of ethanol-to-oil molar ratio via supercritical ethanol system with addition of CO2 as a co-solvent and found that ethanol-to-oil molar ratio of 20:1 was the most suitable condition for their system. The interaction effect of reaction temperature and iso-propanol-tooil weight ratio was also observed as shown in Fig. 7 and Table 5 with the P-value of less than 0.05. At low reaction temperature (325 °C), the increase in the mean %Ester was observed with increasing iso-propanolto-oil weight ratio from 0.1:1 to 0.55:1 and the increasing rate of % Ester was decreased when the further iso-propanol was added. It might be possible that the larger size of iso-propanol molecules (compared to ethanol) impeded the interaction with all three chains of triglycerides leading to the requirement of longer residence time [16]. Another reason is that the diffusion of ethanol molecules was not facilitated by the excess presence of excess iso-propanol despite the enhanced

homogeneity. At high reaction temperature (375 °C), the mean %Ester apparently reached a plateau with the iso-propanol-to-oil weight ratio of 0.55:1 as the reaction was almost complete. It can be inferred from the result that the ethanol-to-oil molar ratio influence was dominant at low reaction temperature and became suppressed at high reaction temperature. The interaction effects of the other variable pairs on the mean % Ester, such as interaction between iso-propanol-to-oil weight ratio and residence time, were not significantly observed, i.e., the slopes of the interaction response of variable pairs was roughly the same as also supported by P-value > 0.05 (see Table 5). 3.3. Optimization and validation The optimization of operating conditions including reaction temperature, ethanol-to-oil molar ratio, iso-propanol-to-oil weight ratio, and residence time on the %Ester was carried out through RSM using BBD method. The boundary of these variables was established following the results previously mentioned in Section 3.1. The desirability function with the desirability value of unity was used to reach the proposed response. The predicted maximum %Ester (100%) was projected under the conditions with reaction temperature of 375 °C, iso-propanol-to-oil weight ratio of 0.1:1, ethanol-to-oil molar ratio of 23:1, and residence time of 7 min. However, %Ester of 100% might not be necessary since the requirement of biodiesel fuel quality according to ASTM D6751-02 and EN 14214 is only 96.5%. Reducing the %Ester to 98.5% would be more practical in terms of the manufacturing cost and operating conditions than those of 100%. The predicted operating conditions for 98.5% of %Ester were as follows: residence time of 4.7 min, ethanol-tooil molar ratio of 20:1 (alcohol-to-oil molar ratio of 25:1), iso-propanolto-oil weight ratio of 0.2:1, and reaction temperature of 360 °C. The accuracy of predictions was confirmed through the experiments under the predicted operating conditions. The result indicated that at the predicted operating conditions for the maximum %Ester (100%), the % Ester obtained from the experiment was 99.2%. The actual %Ester of 98% was achieved under the selected operating conditions (%Ester of 98.5% from model prediction). Hence, very small discrepancy between experimental and predicted values was observed. The correlation between operation variables and %Ester could be expressed in the form of polynomial equation as presented in Eq. (2). This equation signifies that the %Ester (Y) was a function of reaction temperature (X1), residence time (X2), ethanol-to-oil molar ratio (X3), and iso-propanol-to-oil weight ratio (X4). The statistical analysis (ANOVA) was performed to determine the quality of model prediction as summarized in Table 5. The P-value for the lack of fit was 0.098, which was higher than 0.05, representing the statistically insignificant lack of fit. The sum of squares of pure error was 0.33 reflected the decent reproducibility of the experimental values. The coefficient of determination (R2) of 0.99 was obtained implying that the actual experimental data fitted well with the model equation.

Table 5 Analysis of variance for polynomial regression model. Source

Degrees of freedom

Sum of squares

Mean squares

P-value

Remark

X1 X2 X3 X4 X1X2 X1X3 X1X4 X2X3 X2X4 X3X4 X12 X22 X32 X42 Regression Residual error Lack of fit Pure error R2

1 1 1 1 1 1 1 1 1 1 1 1 1 1 14 12 10 2 0.99

218.79 168.36 34.36 6.65 53.23 8.12 38.90 0.04 0.05 2.10 226.78 175.83 54.76 219.59 3926.33 16.07 15.74 0.33

218.79 168.36 34.36 6.65 53.23 8.12 38.90 0.04 0.05 2.10 226.78 175.83 54.76 219.59 280.45 1.34 1.57 0.16

0.000 0.000 0.000 0.046 0.000 0.030 0.000 0.866 0.851 0.234 0.000 0.000 0.000 0.000 0.000

Significant Significant Significant Significant Significant Significant Significant Not significant Not significant Not significant Significant Significant Significant Significant Significant

0.098

Not significant

Y = - 1069.15 + 7.29X1 - 60.37X2 - 4.42X3 - 0.42X 4 + 0.09X1X2 - 0.01X1X3 - 0.19X1X 4 - 0.01X12 + 2.11X22 + 0.03X32

(2)

3.4. Characterization of biodiesel fuel The properties of biodiesel fuel obtained at the optimal conditions by using iso-propanol as the reactive co-solvent were compared with those obtained by using individual ethanol or iso-propanol as presented in Table 6. The cloud and pour points were significantly improved when iso-propanol was applied as reactive co-solvent. Compared to the conventional biodiesel fuel obtained from ethanol or methanol, excellent quality of biodiesel fuel from our process was clearly observed with lower value of cloud and pour points. The cold-flow properties of our biodiesel fuel were also better than those reported in the literature 122

Fuel 237 (2019) 117–125

N. Akkarawatkhoosith et al.

(see Fig. 8). The cold-flow properties (i.e., cloud and pour points) were important parameters when the fuel was used in the low temperature environment which can cause the crystallization of fuel. Low cold-flow properties mean that the fuel could be more resistant for the crystallization. However, the viscosity of biodiesel obtained via reactive cosolvent was slightly increased due to the content of isopropyl esters in biodiesel. Note that, the presence of isopropyl ester in the biodiesel should be carefully monitored because of its high viscosity [35]. Moreover, these properties of our synthesized biodiesel were in accordance with the EN 14214 and ASTM D6751-02 specifications. According to these results, it is clear that the addition of reactive co-solvent in the system could enhance both biodiesel synthesis and biodiesel fuel quality.

Table 6 Biodiesel properties obtained with reactive co-solvent in comparison to biodiesel obtained with single solvent and from literature. Properties

Unit

FAEEa FAEEa [36] FAEEb [36] FAEEc [12] FAEE/ FAIEd

Viscosity at 40 °C Density at 20 °C Cloud point Pour point

mm2/s kg/m3 °C °C

4.5 0.84 16 7

a b c d

4.9 0.87 – 6

4.1 0.86 – 9

5.4 0.88 – 13

4.9 0.85 8 4

Obtained under conventional condition. Obtained under supercritical condition. Obtained under supercritical condition with hexane. Obtained under supercritical condition with reactive co-solvent.

3.5. Performance comparison of biodiesel synthesis [12,34], i.e., the pour point of 13 °C was found through the work of Muppaneni et al. [11] who produced biodiesel from palm oil under the supercritical ethanol with hexane as a co-solvent. The upgraded coldflow properties can be explained by the presence of isopropyl ester as a cold flow modifier in biodiesel fuel as indicated in the chromatogram

The optimal operating conditions and biodiesel yield were used to compare the performance of biodiesel synthesis with those reported in the literature as summarized in Table 7. This comparison reveals the advantages of the application of reactive co-solvent over those with only ethanol. Apparently, the very short resident time for biodiesel

Fig. 8. Chromatogram of biodiesel fuel obtained at optimal operating condition. Table 1 Physico-chemical properties of biodiesel and diesel fuels from the literature.

Table 7 Comparison of biodiesel synthesis performance with literature data. Reactor volume (cm3)

Reactor type a

Batch reactor Batch reactor Tubular reactorb Spiral reactor Tubular reactor Tubular reactor Microtube reactor Microtube reactorc Microtube reactord a b c d e f g h

100 11 88 3.7 66 260 36.5 37.9 1.85

Temp (°C) 300 349 325 350 350 345 325 325 360

Pressure (MPa) – 20 20 20 20 16 20 20 8

Ratioe 33 33 40 40 40 40 40 20 25

Hexane used as a co-solvent at hexane-to-oil of 0.4 (v/v). CO2 used as a co-solvent at CO2-to-oil of 0.05:1 (w/w). CO2 used as a co-solvent at CO2-to-oil of 0.2:1 (w/w). Iso-propanol used as a reactive co-solvent at iso-propanol-to-oil of 0.2:1 (w/w). Alcohol-to-oil molar ratio. %Ester. Yield. Conversion. 123

Time (min) 20 30 110 30.45 15 28 45 50 4.7

Esterf/Yieldg/Conversionh (%) g

91 79.2g 76g 98g 80h 91f 70g 80g 98f/97g

Ref. [12] [36] [37] [38] [39] [40] [41] [4] This work

Fuel 237 (2019) 117–125

N. Akkarawatkhoosith et al.

synthesis was the highlight of this work. The reduced severity of operating conditions including pressure and alcohol-to-oil molar ratio was also evident; however, the requirement of reaction temperature was slightly elevated in order to meet the international standard on the purity of biodiesel (96.5% of %Ester). In our work, the reaction pressure of 8 MPa was applied, which was relatively lower than the average value in the literature (20 MPa). To obtain the high yield under this low pressure, other parameters including residence time, reaction temperature, and alcohol-to-oil molar ratio were adjusted. Note that, operating at relatively low pressure could help decrease the operating cost and equipment cost for biodiesel production. It can be seen from the Table 7 that, for the cases of continuous reactor, the reaction temperature was the significant factor affecting the biodiesel yield, i.e., a slight increase in reaction temperature could greatly reduce the requirement of residence time. According to Table 7, although the addition of non-reactive cosolvent (hexane and CO2) could improve the biodiesel synthesis, one could further enhance the synthesis by the addition of reactive co-solvent (iso-propanol). For instance, based on a microtube reactor, the addition of CO2 in a system could obtain the higher yield with a decrease in residence time and alcohol-to-oil molar ratio compared to that of without co-solvent. Nevertheless, in our system, the addition of reactive co-solvent could improve the efficiency of biodiesel production as indicated by very short residence time and relatively low pressure. Moreover, the type of reactor is one of the important aspects to develop the biodiesel synthesis. As shown in Table 7, the performance of microreactor is much better than that of tubular and batch reactors. This is due to the outstanding characteristics of microreactor such as high surface-to-volume ratio, excellent heat and mass transfer, and narrow residence time distribution. Finally, this work has demonstrated that both microreactor and reactive co-solvent can be simultaneously applied as one of the promising techniques for enhancing the biodiesel synthesis.

[4] Trentin CM, Lima AP, Alkimim IP, Silva CD, Castilhos FD, Mazutti MA, et al. Continuous production of soybean biodiesel with compressed ethanol in a microtube reactor using carbon dioxide as co-solvent. Fuel Process Technol 2011;92:952–8. [5] Kusdiana D, Saka S. Two-step preparation for catalyst-free biodiesel fuel production: Hydrolysis and methyl esterification. Appl Biochem Biotechnol 2004;115:781–91. [6] Shin HY, An SH, Sheikh R, Park YH, Bae SY. Transesterification of used vegetable oils with a Cs-doped heteropolyacid catalyst in supercritical methanol. Fuel 2012;96(96):572–8. [7] Kim M, Lee HK, Yoo SJ, Youn YS, Shin YH, Lee YW. Simultaneous synthesis of biodiesel and zinc oxide nanoparticles using supercritical methanol. Fuel 2013;109(109):279–84. [8] Zeng D, Yang L, Fang T. Process optimization, kinetic and thermodynamic studies on biodiesel production by supercritical methanol transesterification with CH3ONa catalyst. Fuel 2017;203:739–48. [9] Boer KD, Bahri PA. Supercritical methanol for fatty acid methyl ester production: A review. Biomass Bioenergy 2011;35:983–91. [10] Cao W, Han H, Zhang J. Preparation of biodiesel from soybean oil using supercritical methanol and co-solvent. Fuel 2005;84:347–51. [11] Muppaneni T, Reddy HK, Ponnusamy S, Prafulla D, Sun Y, Dailey P, Deng S. Optimization of biodiesel production from palm oil under supercritical ethanol conditions using hexane as co-solvent: a response surface methodology approach. Fuel 2013;107:633–40. [12] Han H, Cao W, Zhang J. Preparation of biodiesel from soybean oil using supercritical methanol and CO2 as co-solvent. Process Biochem 2005;40:3148–51. [13] Roschat W, Siritanon T, Kaewpuang T, Yoosuk B, Promarak V. Economical and green biodiesel production process using river snail shells-derived heterogeneous catalyst and co-solvent method. Bioresour Technol 2016;209:343–50. [14] Chueluecha N, Kaewchada A, Jaree A. Biodiesel synthesis using heterogeneous catalyst in a packed microchannel. Energy Convers Manage 2017;141:145–54. [15] Huang R, Cheng J, Qiu Y, Li T, Zhou J, Cen K. Using renewable ethanol and isopropanol for lipid transesterification in wet microalgae cells to produce biodiesel with low crystallization temperature. Energy Convers Manage 2015;105:791–7. [16] Lapuerta M, Fernandez JR, Rodriguez DF, Camino RP. Cold flow and filterability properties of n-butanol and ethanol blends with diesel and biodiesel fuels. Fuel 2018;224:552–9. [17] Sierra-Cantor JF, Guerrero-Fajardo CA. Methods for improving the cold flow properties of biodiesel with high saturated fatty acids content: a review. Renew Sustain Energy Rev 2017;72:774–90. [18] Hanh HD, Dong NT, Okitsu K, Nishimura R, Maeda Y. Biodiesel production through transesterification of triolein with various alcohols in an ultrasonic field. Renewable Energy 2009;34:766–8. [19] Wang PS, Tat ME, Gerpen JV. The production of fatty acid isopropyl esters and their use as a diesel engine fuel. JAOCS 2005;82:1–5. [20] Encinar JM, Gonzalez JF, Rodriguez JJ, Tejedor A. Biodiesel fuels from vegetable oils: transesterification of Cynara cardunculus L. oils with ethanol. Energy Fuels 2002;16:443–50. [21] Cunico MLP, Ceriani R, Guirardello R. Estimation of physical properties of vegetable oils and biodiesel using group contribution. Chem Eng Trans 2013;32:535–40. [22] Sun Y, Ponnusamy S, Muppaneni T, Reddy HK, Patil PD, Li C, et al. Optimization of high-energy density biodiesel production from camelina sativa oil under supercritical 1-butanol conditions. Fuel 2014;135(135):522–9. [23] Kusdiana D, Saka S. Kinetics of transesterification in rapeseed oil to biodiesel fuel as treated in supercritical methanol. Fuel 2001;80:693–8. [24] Demirbas A. Biodiesel fuels from vegetable oils via catalytic and non-catalytic supercritical alcohol transesterifications and other methods: a survey. Energy Convers Manage 2003;44:2093–109. [25] Kwak TY, Mansoori GA. Van der waals mixing rules for cubic equations of state. Applications for supercritical fluid extraction modelling. Chem Eng Sci 1986;41:1303–9. [26] Rathore V, Madras G. Synthesis of biodiesel from edible and non-edible oils in supercritical alcohols and enzymatic synthesis in supercritical carbon dioxide. Fuel 2007;86:2650–9. [27] Farobie O, Leow ZYM, Samanmulya T, Matsumura Y. New insights in biodiesel production using supercritical 1-propanol. Energy Convers Manage 2016;124:212–8. [28] Shin HY, Lim SM, Bae SY, Oh SC. Thermal decomposition and stability of fatty acid methyl esters in supercritical methanol. J Anal Appl Pyrolysis 2011;92:332–8. [29] Quesada-Medina J, Olivares-Carrillo P. Evidence of thermal decomposition of fatty acid methyl esters during the synthesis of biodiesel with supercritical methanol. J Supercritical Fluids 2011;56:56–63. [30] Pereda S, Bottini SB, Brignole EA. Gas-liquid reactions under supercritical conditions-phase equilibria and thermodynamic modeling. Fluid Phase Equilib 2002;194–197:493–9. [31] Cardoso CC, Celante VG, Castro EVRD, Pasa VMD. Comparison of the properties of special biofuels from palm oil and its fractions synthesized with various alcohols. Fuel 2014;135:406–12. [32] He H, Tao W, Zhu S. Continuous production of biodiesel from vegetable oil using supercritical methanol process. Fuel 2007;86:442–7. [33] Wang C, Zhou J, Chen W, Wang W, Wu Y, Zhang J, et al. Effect of weak acids as a catalyst on the transesterification of soybean oil in supercritical methanol. Energy Fuels 2008;22:3479–83. [34] Sawangkeaw R, Teeravitud S, Bunyakiat K, Ngamprasertsith S. Biofuel production from palm oil with supercritical alcohols: effects of the alcohol to oil molar ratios on the biofuel chemical composition and properties. Bioresour Technol 2011;102:10704–10.

4. Conclusion Iso-propanol was applied to enhance the biodiesel synthesis in a microtube reactor under supercritical conditions. Iso-propanol in our work played two important roles as a source of reactant to produce isopropyl esters and as a co-solvent to enhance the homogeneity of reacting mixture. The improvement of biodiesel synthesis in terms of the operating parameters was achieved especially the substantial shortening of residence time which is one of the highlights of this work. The influence of operating conditions, including reaction temperature, iso-propanol-to-oil weight ratio, ethanol-to-oil molar ratio, and residence time, on the %Ester were observed and interpreted through the prediction model. The optimal conditions were found at the reaction temperature of 360 °C, reaction pressure of 8 MPa, iso-propanol-to-oil weight ratio of 0.2:1, alcohol-to-oil molar ratio of 25:1, and residence time of 4.6 min. The biodiesel fuel quality in terms of cloud and pour point was significantly improved by the presence of isopropyl ester in biodiesel. The superior efficiency of biodiesel synthesis in our work was clearly presented when compared to the literature. Acknowledgements This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. References [1] Stephen JL, Periyasamy B. Innovative developments in biofuels production from organic waste materials: a review. Fuel 2018;214:623–33. [2] Leung DYC, Wu X, Leung MKH. A review on biodiesel production using catalyzed transesterification. Appl Energy 2010;87:1083–95. [3] Atadashi IM, Aroua MK, Aziz AA. High quality biodiesel and its diesel engine application: a review. Renew Sustain Energy Rev 2010;14:1999–2008.

124

Fuel 237 (2019) 117–125

N. Akkarawatkhoosith et al. [35] Refaat AA. Correlation between the chemical structure of biodiesel and its physical properties. Int J Environ Sci Technol 2009;6:677–94. [36] Gui MM, Lee KT, Bhatia S. Supercritical ethanol technology for the production of biodiesel: process optimization studies. J Supercrit Fluids 2009;49:286–92. [37] Bertoldi C, Silva C, Bernardon JP, Corazza ML, Cardozo-Filho L, Oliveira JV, et al. Continuous production of biodiesel from soybean oil in supercritical ethanol and carbon dioxide as co-solvent. Energy Fuels 2009;23:5165–72. [38] Farobie O, Sasanami K, Matsumura Y. A novel spiral reactor for biodiesel production in supercritical ethanol. Appl Energy 2015;147:20–9.

[39] Silva C, Weschenfelder TA, Rovani S, Corazza FC, Corazza ML, Dariva C, et al. Continuous production of fatty acid ethyl esters from soybean oil in compressed ethanol. Ind Eng Chem Res 2007;46:5304–9. [40] Velez A, Soto G, Hegel P, Mabe G, Pereda S. Continuous production of fatty acid ethyl esters from sunflower oil using supercritical ethanol. Fuel 2012;97:703–9. [41] Silva CD, Castilhos FD, Oliveira JV, Filho LC. Continuous production of soybean biodiesel with compressed ethanol in a microtube reactor. Fuel Process Technol 2010;91:1274–81.

125