Design and performance assessment of an in-house fabricated microreactor for enzyme-catalysed biodiesel synthesis

Abstract

The production of biodiesel from transesterification of oils is the currently available alternative to consumable fuels. The study focuses on the performance assessment of a microreactor platform for biodiesel production. Since a microreactor provides better mass transfer and a higher surface-to-volume ratio, this significantly contributes to augmenting the reaction rate of biodiesel production at reduced cost. In this study, firstly the T-shaped microfluidic reactor was fabricated on a PMMA (poly methyl methacrylate) sheet. Then, the kinetics of enzymatic transesterification reaction for biodiesel synthesis in this reactor was investigated. A combination of sunflower oil and methanol was used as feedstock in the presence of lipase from Thermomyces lanuginous as a catalyst and tert-butanol as a solvent. The reaction was carried out under ambient conditions (1 atm and 30 °C) where 285 v% of tert-butanol was loaded to oil with 4 wt% lipase, and the molar ratio of oil to methanol was 1:2.5. Gas Chromatography was used to measure the concentration of biodiesel. The kinetic parameters such as maximum reaction rate (V_max) and Michaelis –Menten constant (K_M) were found to be 0.725 (mol/L.min) and 0.092 mol/L, respectively. High V_max and low K_M values indicated the better performance of the reactor. In addition, a Lineweaver–Burk plot with excellent coefficient of determination (R²) of 99.5% supported the high performance of the microreactor.

Keywords:

• Biodiesel
• enzymatic reaction
• kinetic parameters
• microreactor

1. Introduction

In recent years, there has been a global trend towards replacing non-renewable energy resources. Biodiesel is one such renewable fuel resource that has high potential and plays a crucial role in the sustainability and reduction of greenhouse gases, due to the absence of sulfur and a reduction in the amount of toxic gases released (Demirbas, 2007). Biodiesel production found its way into industry in 1988 and since then, it has increased dramatically. For instance, in Europe it has reached a capacity of around nine million tons per year (Luque, Lin, Wilson, & Clark, 2016). Biodiesel can be produced from edible oils such as soybean, palm kernel and coconut, non-edible oils such as jatropha, camelina and rice bran, or from side stream while refining soap stock, acidulated soap stock and deodorizer distillates. The demand for oils or lipids for industrial purposes has reached up to 15% for different applications, such as surfactants, detergents and biofuels (Liu, Tan, Zhang, Yan, & Hameed, 2010; Luque et al., 2016).

Portability, better efficiency, lower sulfur content and aromatic reliability, sustainability and biodegradability of biodiesel are its main advantages over petroleum. The most significant aspects are that it is both renewable and environmentally friendly. Moreover, it can be naturally recycled, and microorganism processes can contribute significantly to its degradation and safer discharge (Zhang, 2012). However, in addition to the feature of being a clean source of energy, the fuel is also an economical one.

Traditional biodiesel production methods have a number of limitations, such as high cost, high consumption of energy, a long residence time and low efficiency. To alleviate these problems, a microreactor offers a novel and sustainable approach. Moreover, a microreactor provides a more efficient, more economical “reaction chamber” for the production of biodiesel than those of conventional macro-reactors (Bolivar, Valikhani, & Nidetzky, 2019). Continuous flow microreactors enhance both mass transfer and interaction between the phases, which are basic requirements in facilitating the production of biodiesel. Micro mixing configurations such as T-shaped and Y-shaped micro-channels are commonly used to enhance mixing performance in such applications (Natarajan et al., 2019; Santana, Silva, & Taranto, 2019).

2. Theory

2.1. Transesterification reaction

One of the most commonly used methods in synthesising biodiesel is the transesterification reaction. It is based on the process of exchanging organic groups between ester and alcohol. The raw materials required for the reaction are triglycerides, which can be obtained from animal or plant-based oils and fats and methanol, used due to its high biodiesel yield and feasible price, with the aid of a catalyst. The reaction process follows a three-step mechanism in which triglycerides convert to diglycerides, diglycerides into monoglycerides and monoglycerides into glycerol (Aarthy, Saravanan, Gowthaman, Rose, & Kamini, 2014), as follows:

2.2. Enzymatic catalyst

Generally, for any chemical process, catalysts are required due to several reasons: (i) they speed up the reaction, (ii) they help produce products with high yield, (iii) they allow the reaction to work at a much lower temperature, (iv) they can be used repeatedly and (v) they save energy. Therefore, selection of catalyst for a particular reaction is very important. For instance, several types of catalysts such as pure ZnO, nanorods of Hg doped ZnO, multi-walled carbon nanotubes, titanium dioxide, polyaniline (PANI)/ZnO, ZnO/CeO₂, ZnO/CuO, a composite of multi-walled carbon nanotube/tungsten oxide, ZnO/Ag/CdO, ZnO/γ-Mn₂O₃ and V₂O₅/ZnO nanocomposites are used for removing textile effluent by degradation reaction (Rajendran et al., 2016; Saleh & Gupta, 2011, 2012; Saravanan, Gupta, Mosquera, & Gracia, 2014; Saravanan, Gupta, Narayanan, & Stephen, 2013, 2014; Saravanan, Gupta, Prakash, Narayanan, & Stephen, 2013; Saravanan, Karthikeyan, et al., 2013; Saravanan, Thirumal, Gupta, Narayanan, & Stephen, 2013; Saravanan et al., 2015, 2016). Similarly, transesterification reaction is conducted in the presence of basic (KOH), acidic (HCl), or enzymatic (e.g., lipase) catalysts (Hossain, 2019; Zakir Hossain et al., 2016)_. Among these three methods, lipase catalysed transesterification has been considered as more favorable by many researchers, due to high biodiesel yields. It can also make the product separation easier and can be operated at low temperature and pressure leading to high conversion.

Further, lipases are considered to be the type of enzymes which are able to hydrolyse fat. They play a significant role in digestion and breaking the triglycerides in oils and lipids in the majority of living organisms. Lipases are classified according to their source of origin, such as animals, plants or microorganisms and their properties differ accordingly. One of the most commonly used lipases in industry is Candida Antarctica B, which is commercially known as Novozym435. This lipase, taken from fungi, has shown promising results for biodiesel production (Aarthy et al., 2014; Luque et al., 2016).

2.3. Microreactor technology

The main attractive features of microreactors are their efficient heat transfer, mass transfer and effective mixing, primarily due to their large surface-to-volume ratios (Suryawanshi, Gumfekar, Bhanvase, Sonawane, & Pimplapure, 2018). In addition, the laminar flow of reacted species is a key feature in microscale channels. This develops the efficiency of mixing in microreactors mainly because of the molecular diffusion that predominates in microchannels. In addition, bends in microreactor channels change the direction of the flow, which will lead to better mixing (Aoki, Hasebe, & Mae, 2004). Applications where a precise control of the molar ratio between reacted species necessitates having an efficient mixing within the system, which significantly contributes to reducing or eliminating side reactions.

A microreactor is also advantageous when scaling is needed to increase productivity. Scaling of microreactors is based on their numbering up to produce an assembly of microreactors instead of scaling up a pilot plant or an existing reactor in industry (Rizkin, Popovic, & Hartman, 2019). The numbering up approach offers profound advantages over the scaling up approach due to the fact that using identical microreactors in parallel produces higher quality than the scaling up of a pilot conventional reactor. Thus, the performance of microreactors remains unchanged through numbering up (Yao, Zhang, Du, Liu, & Yao, 2015), which facilitates the use of the same reactors in industry (He et al., 2020). In addition, taking into consideration the sustainability of energy sources, microfluidic systems are well aligned to target this aspect, especially because of its economical fabrication and availability (Kockmann, Gottsponer, Zimmermann, & Roberge, 2008).

Significant research has been documented for biodiesel production using microreactor systems; however, all of it is complex in design and requires emulsifier to enhance efficiency. This paper aims to construct a simple microreactor platform that does not need any emulsifier to synthesize biodiesel. The kinetics of the enzymatic reaction of biodiesel production has been studied using the Michaelis–Menten model. In addition, the rate of the reaction has been improved.

3. Experimental

3.1. Materials and method

The materials used in this study were T. lanuginous type Lipase solution (Cat. No. 2391 B), sunflower oil, methanol (99.8% purity, Sigma Aldrich, UK) and tert-butanol (from Research Lab Fine Chem Industries, India)

3.2. Reaction setup procedure

To start the experiment, a base was needed to build the reaction conditions³. While the reaction conditions were found to be 50 vol% tert-butanol to oil, 4 wt% lipase to oil and a 1:9 molar ratio of oil to methanol in a batch process, they were as follows in the microreactor: 285% vol% tert-butanol to oil, 4 wt% lipase to oil and 1:2.5 molar ratio of oil to methanol. Therefore, for the experiments, 35 mL of oil was selected as a base for the calculation of the rest of the chemicals: 100 mL of tert-butanol, 1.3 mL lipase and 3.75 mL of methanol. To achieve equal flowrates for the two inlet reservoirs, the chemicals were split evenly as much as possible, according to the details given in Table 1.

Table 1. Distribution of the chemicals per inlet of the microfluidic chip.

	Oil (mL)	Tert-butanol (mL)	Methanol (mL)	Lipase (mL)	Total volume (mL)
Inlet 1	17	51.7	0	1.3	70
Inlet 2	18	48.3	3.75	0	70

It is worth mentioning that dealing with tert-butanol at room temperature was a great challenge due to its low boiling point (T_bp=25.6 °C at atmospheric pressure). A lower temperature processing of tert-butanol blocks the microreactor channels. Therefore, the room temperature was controlled at 30 °C in the laboratory.

3.3. Concentration measurement

GC-MS was used to measure the concentration of the products. One of the methods used for gas chromatography quantitative analysis was the external standard calibration. This method requires preparing several standard solutions of methanol and solvent (tert-butanol in this work). Correlation between the peak area and the concentration of methanol was calculated to predict the concentration at any given peak area, according to the Beer-Lambert law principle.

3.4. GC-MS conditions and temperature program

The selected GC-MS column was Capillary DB-624 from Agilent Company. The GC-MS was run at the following operating conditions to get precise readings: The initial column temperature of the GC-MS was maintained at 70 °C for 3 min, then increased linearly at a rate of 25 °C/min. Once the temperature reached 250 °C, it was maintained for 5 min. All samples were injected manually instead of using the auto-injection setup, to guarantee a faster injection rate, hence, more accurate results can be obtained. The manual injection time was 15 sec in comparison with 2 min with the auto-injection setup.

3.5. Microreactor fabrication

An in-house T-shaped microreactor was fabricated. The AutoCAD software package (AutoCAD®, Autodesk 2017) was used to draw the design. An AutoCAD coding was written to mill the design on a poly methyl methacrylate (PMMA) sheet, using a laser beam operated at a velocity of 30 mm/s. The serpentine microchannel geometry was 400 μm width, 780 μm depth and 400 mm length as presented in (Figure 1(a)), which results in a total volume of 125 mm³. The length of the channel was chosen to produce efficient mixing, according to a previously published study (Yusuf, Redha, Al Meshal, & Shehab, 2018). Afterwards, the microreactor (Figure 1(b)) was physically cleaned, then sealed using polyester adhesive laminate (Plastic Art, UK), as shown in Figure 1(c). High pressure resistance materials were used as additional sealants to withstand high inlet pressure and thermodynamic resistance in the microchannel. Therefore, a 3 mm PVC laminated sheet (TMI. LLC, Pittsburgh, USA) was cured to secure the channels, followed by a 5 mm metal-based sheet, glued over the laminate, to avoid any leakage. A 2 mm diameter-drilling tool (Toolex Ltd) was used to form the inlets and outlets of the microreactor. Then, 2 mm² uninsulated bootlace ferrules (RS, Corby, UK) were joined to both inlets and outlets, as illustrated in Figure 1(d). The final assembly is presented in Figure 1(e) which demonstrates the front side of the fabricated microreactor.

Figure 1. The fabricated T-shaped microreactor: (a) Autocad 2D drawing, (b) the machined design on the PMMA sheet, (c) the chip sealed with a laminated sheet, (c) the back side of the chip after further sealing and ferrules connections and (d) the front side of the chip.

3.6. Flowrate setup

A gravitational flow was used to deliver the reactants. The maximum height of this setup produces around 700 µL/min. The inlet flowrate was varied upon changing the height of the input reservoirs (feed syringes), as presented in Figure 2.

Figure 2. Flow rate setup: (a) varying the height to change the inlet flow rates and (b) the inlet reservoirs (feed syringes).

4. Results and discussion

4.1. Flowrate measurement experiment

The range of flowrate in all experiments was 50–300 µL/min. A precise flowrate is a crucial requirement in this application, especially when dealing with microscale devices. Therefore, four runs per flowrate were measured to help improve the precision, as illustrated in Table 2.

Table 2. Flowrate measurements.

	Targeted vs. measured flow rates
Run No.	50 µL/min	100 µL/min	150 µL/min	200 µL/min	250 µL/min	300 µL/min
1	51	101	148	198	253	300
2	51	100	150	200	251	300
3	50	101	150	199	250	301
4	52	99	150	201	250	301
$\overline{\mathit{x}}\pm \mathit{\sigma }$	51$\pm$0.71	100.2$\pm$0.83	149.5$\pm$0.87	199.5$\pm$1.12	251$\pm$1.23	300.5$\pm$0.50

4.2. Calibration curve determination

A plot of concentrations vs. the GC-MS peak areas is presented in Figure 3. Three different concentration readings were taken per prepared standard solutions of methanol in tert-butanol. The best linear fit was found with R² = 0.998, as presented in the graph. Therefore, the concentration of methanol can be calculated using the following equation: (2) $$\left[\mathit{Methanol}\right]=0.005. \mathit{Peak} \mathit{Area}-535.31$$(2)

Figure 3. Calibration curve for methanol in tert-butanol solution.

4.3. Reaction kinetics evaluation

The rate of reaction was calculated using the following equation: (3) $$-rA=\frac{F_{A0}X}{V}$$(3) where −rA is the rate of reaction of methanol in (moles/ mL.min), X is the molar conversion in moles reacted/moles fed, F_A0 is the molar flowrate of methanol in (moles/time) and V is the volume of reactor in (mL) which is 125 mm³, as mentioned in the fabrication section.

In order to measure the kinetics, the rate expression will be calculated in terms of methanol concentration. Therefore, any decrease in methanol concentration will reflect the amount of biodiesel formed since methanol will be the limiting reactant. Eq. (3)(3) $$-rA=\frac{F_{A0}X}{V}$$(3) can be rewritten as follows: (4) $$\mathit{-rA}=\frac{\dot{V}({CA}_o-CA)}{V}$$(4)

where CA_o is the initial concentration of methanol in (mol/mL), CA is the final concentration of methanol in (mol/mL), $\dot{V}$ is the volumetric flowrate of the feed in (mL/min). The rate of reaction was therefore calculated at different final concentration of methanol, as illustrated in Figure 4 below:

Figure 4. The rate of reaction at different methanol concentration.

The reduction in methanol concentration in the experiments was an indication for biodiesel production, so the fabricated microreactor proved its suitability for synthesising biodiesel. In addition, the effectiveness of the enzyme used in this work contributes significantly to biodiesel formation. Thus, compared to other types of enzymes, a microbial enzyme is fast to synthesize biodiesel (Wancura et al., 2018), inexpensive and more stable, as agreed with Moazeni, Chen, and Zhang (2019). The enzyme substrate mechanism can be explained by the following reaction mechanism: (6) $$E+S\to ES$$(6) (7) $$ES\to E+S$$(7) (8) $$ES\to E+P$$(8)

In this mechanism, the substrate S will bind with the enzyme E to form enzyme-substrate complex, ES. The process of this bond formation is commonly referred to as enzymatic (or protein) crosslinking (Heck, Faccio, Richter, & Thöny-Meyer, 2013; Jus et al., 2011). The complex is then adsorbed by the surface of the enzyme for the reaction to take place. The product is then released from the enzyme while the enzyme is preserved for more substrates to bind in.

It must be noted that enzymes act apparently as adsorbents, similar to those of porous carbon, activated carbon, carbon nanotubes, fullerene, rich husk and waste carbon (obtained from fertilizer plant, rubber tire, sugar industry etc) and thereby hypothetically the enzyme-substrate reaction mechanisms should follow the adsorption-desorption phenomena (Ahmaruzzaman & Gupta, 2011; Gupta, Jain, Ali, Chandra, & Agarwal, 2002; Gupta, Nayak, Agarwal, & Tyagi, 2014; Gupta et al., 2016; Saleh & Gupta, 2014). There are various adsorption models available for various adsorption-desorption processes in the literature (Burakov et al., 2018; Dil et al., 2016; Ghaedi et al., 2015; Nekouei, Nekouei, Tyagi, & Gupta, 2015), especially adsorption isotherms (Langmuir, Freundlich, Temkin, Frumkin and Flory Huggins) and adsorption kinetics (pseudo fist order, pseudo second order). However, unlike the aforementioned processes, it has been proven that Michaelis-Menten model represents the enzyme-catalysed reaction mechanisms perfectly(Gunawan, Suhendra, Rizkia, & Hasanah, 2017). In this study, the experimental data fitted well also with Michaelis-Menten Model, which agrees with previous studies (Saleh & Gupta, 2011; Saravanan, Gupta, Narayanan, et al., 2013, 2014; Saravanan et al., 2015). This model is represented by the following formula: (9) $$\mathit{-rA}=\frac{V_{\mathit{max}}. \left(CA\right)}{K_M+\left(CA\right)}$$(9)

where –r_A is the rate of disappearance of the substrate in (mol/mL.min), V_max is the maximum reaction rate in (mol/mL.min), CA is methanol concentration in (mol/mL) and K_M is Michaelis constant in (mol/mL).

For plotting purposes, Eq. (9)(9) $$\mathit{-rA}=\frac{V_{\mathit{max}}. \left(CA\right)}{K_M+\left(CA\right)}$$(9) can be rewritten as follows to yield a linear relationship: (10) $$\frac{1}{–rA } = \frac{ 1}{V_{\mathit{max}}}+\frac{K_M}{V_{\mathit{max}}}\left(\frac{ 1}{CA}\right)$$(10)

Thus, a plot of $\frac{1}{–rA }$ versus $\frac{ 1}{CA}$ yields a straight line called Lineweaver–Burk plot, as illustrated in Figure 5 below:

Figure 5. Lineweaver–Burk plot for the enzymatic reaction.

V_max and K_M can be calculated from the intercept and the slope, respectively. V_max and K_M were found to be$7.245\times {10}^{-4}\frac{\mathit{mol}}{ml.\mathit{min}}$ and $9.187\times {10}^{-5} \frac{\mathit{mol}}{ml},$ respectively, with a better linearity (R²=0.9952) over a recently published study (Gojun et al., 2019). The final form of the Michaelis-Menten kinetic model can be written as follows: (11) $$\mathit{-rA}=\frac{ 7.245\times {10}^{-4} \left(CA\right)}{9.187\times {10}^{-5}+\left(CA\right)}$$(11)

The kinetics of the enzymatic transesterification of biodiesel from cooking oil has been compared with previously published works (Al-Zuhair, Dowaidar, & Kamal, 2009; Azócar, Navia, Beroiz, Jeison, & Ciudad, 2014; Krishna & Karanth, 2001; Yadav & Lathi, 2004), as illustrated in Table 3. The maximum reaction rate constant in comparison with other references is also illustrated in Figure 6. The fabricated microreactor showed a faster reaction rate when compared with those of some other researchers. This could be mainly attributed to the efficient mixing in micochannels and the appropriate residence time which can lead to higher rate of reaction. In addition, T-shaped microreactors with small microchannel diameters and serpentine pattern fabrication (Santana, Tortola, Silva, & Taranto, 2017; Verma, Prakash, Mehta, & Ghosh, 2019) is advantageous for achieving a good level of performance over conventional reactors (Franjo, Šalić, & Zelić, 2018; Natarajan et al., 2019; Shaaban, El-Shazly, & Elkady, 2015). The K_M value was found small in this study, which indicates lower dependency of methanol concentration on the transesterification in microreactors, which agrees with other researchers (Azócar et al., 2014; Halim & Harun Kamaruddin, 2008).

Figure 6. Representation of V_max obtained for current work compared to previously published works.

Table 3. Comparison between current work and previous work.

Michaelis-Menten Model constants	Previous work				Current work
	Zarejousheghani, Kariminia, and Khorasheh (2016)	Krishna and Karanth (2001)	Al-Zuhair et al. (2009)	Azócar et al. (2014)
Vmax (mol/L.min)	0.268	0.012	0.002	0.018	0.725
KM (mol/L)	3397	3060	0.11	1030	0.092

The use of the Michaelis–Menten model to study the kinetics of biodiesel synthesis based on transesterification approach with lipase from Thermomyces lanuginosus is limited in literature, as lately revealed (Gojun et al., 2019). Recent publications on the kinetics study of lipase-based biodiesel production using the Michaelis–Menten model is illustrated in Table 4. The current study shows a better reaction rate over Encinar, González, Sánchez, and Nogales-Delgado (2019), when a commercial lipase (Novozym 435) was used in the latter. Gojun et al. (2019) recently found a better reaction rate using the same lipase as the current study. However, the presented work shows better correlation with data fitting using the Michaelis–Menten model, over the aforementioned study. In addition, the current study is characterized by being less complex, as no emulsifiers were needed to synthesize biodiesel.

Table 4. Reaction rate comparison with references used Michaelis–Menten model to study the kinetics of lipase-based biodiesel production.

Lipase type	Microbial source	Vmax (mol/L.min)	Ref
Fungi	Thermomyces lanuginous	0.725	Current work
Bacteria and fungi: Novozym 435	Commercial	$6.38\times {10}^{-4}$	Encinar et al. (2019)
Fungi	Thermomyces lanuginous	1.397	Gojun et al. (2019)

In addition, it was found that the maximum % conversion achieved in this work was 68.90% at an inlet flowrate of 50 μL/min (per inlet). Aoki et al reported that a lower flow rate guarantees more residence times with better mixing in microreactors, which leads to a better conversion (Aoki et al., 2004). However, the operated flow rate range in this study was restricted by the laboratory facilities, especially while using gravitational flow.

5. Conclusions

The enzymatic transesterification of biodiesel reaction was successfully studied in an in-house fabricated microfluidic reactor of 0.125 mL total volume. The operating conditions for the experiment were at an ambient temperature of 30 °C with the following reactant conditions: 285% of tert-butanol loading, 4 wt% of lipases based on oil weight and oil-to-methanol ratio of 1:2.5. A superior reaction rate (V_max) of 0.725 mol/L.min, and lower Michaelis –Menten constant (K_M) of 0.092 mol/L demonstrated the better performance of the reactor in comparison to conventional reactors. While producing biodiesel in a microfluidic chip was quite challenging due to the viscous nature of oil and biodiesel, which increases the hydrodynamic resistances in microchannels, significant enhancement was shown in the rate of reaction. In addition, the fabricated platform provides several potential features over the macro scale production; it is a scale independent synthesis and economically favourable due to the low cost of the microreactor fabrication. Overall, this microfluidic system can be utilized as an effective pivotal tool with high performance for biodiesel production. Further advanced investigation is necessary to promote the fabricated microreactor performance with different operating temperatures and reactant concentrations, in order to maximize the biodiesel yield. Our current investigations are aligned in this direction.

Disclosure statement

No potential conflict of interest was reported by the author(s).