Microwave-assisted synthesis of trimethylolpropane triester (bio-lubricant) from camelina oil

Vegetable oils, whose hydrocarbon structure is very similar to that of petroleum products, are ideal renewable and sustainable alternatives to petroleum lubricants. Bio-lubricants are commonly synthesized by modifying the chemical structure of vegetable oils. In this study, microwave irradiation was applied to intensify the mass-transfer-limited transesterification reaction to produce trimethylolpropane triester (bio-lubricant) from camelina oil as a promising local energy crop. A rotatable RSM-BBD method was applied to find the optimal levels of experimental factors, namely reaction time (67.8 min), the catalyst concentration (1.4 wt%) and the molar ratio (3.5). In these optimal levels, the reaction yield of 94.3% was obtained with desirability of 0.975. The quadratic statistical model with a determination coefficient of 97.97%, a standard deviation of 0.91 and a variation coefficient of 1% was suggested as the most appropriate model by Design-Expert software. Finally, the physicochemical properties of the purified product were in accordance with the requirements of the ISO-VG22 base oil standard.

decade [7][8][9][10][11] . Transesterification reaction of TMPTE synthesis is a reversible time-consuming process, which in its progress, is highly dependent on the rate of mass and heat transfer. Therefore, some process intensification techniques are used, such as hydrodynamic cavitation [12][13][14][15] novel co-solvents 16 and ultrasonic irradiation 17,18 . Also, Aziz et al., in a review paper recommended the use of microwave technology in intensifying bio-lubricant production 19 .
The heating process in conventional transesterification reaction involves forced conduction and convection based on direct molecular collisions. Indeed, the temperature gradient between the internal reactor wall and the reactants as well as the conductivity of the reactor wall play key roles in the conventional heating process. Microwave radiation is unique due to the fact that the heating mechanism is generally controlled by direct radiation to a molecular level, based on the motion of charged protons and electrons. Therefore, it can reduce processing time, save energy and improve production yield, which is attributed to the intense heating at molecular level 20 . The microwave is part of the electromagnetic spectrum with a wavelength of 1 m to 1 mm, which corresponds to a frequency range of 300 MHz to 300 GHz. Common frequencies for home and industrial heatings are 915 MHz and 2.45 GHz. The two main mechanisms that produce heat are based on the polarization of the material and ion heating. The polarity and ionic properties of materials are affected by rapid changes in the electromagnetic field. In addition, microwave heating affects the internal energy and reduces the activation energy of the reaction, which is similar to the catalytic effect in accelerating the reaction. This friction and molecular collision improve the intermolecular mixing of the reactants 21 .
The Camelina sativa is a member of the Brassicaceae family ( Fig. 1) and has been shown in various experiments to have much lower water requirements and greater resistance to spring frosts than other oil plants, especially canola. Camellia oil is used in both edible and industrial types. The percentage of Camelina seed oil has been reported to be 29 to 41% 22 . Results of GC-Mass showed that Camellina oil is composed of palmitic acid (6.45%), stearic acid (2.56%), oleic acid (16.41%), eicosenoic acid (14.09%), erucic acid (3.21%), linoleic acid (21.03%) and linolenic acid (29.70%). "Soheil cultivar" is a dryland crop that requires less agricultural inputs than its competitors and is capable of growing in most regions of Iran. This cultivar has been released, propagated and cultivated for the first time at Biston Shafa Co., Razi University, Kermanshah, Iran [23][24][25] .
The Camelina sativa is suitable for temperate climates and poor soils that can withstand drought and heat. It grows rapidly and has a short life of 85 to 100 days. The higher the amount of unsaturated fatty acids, the lower the viscosity at room temperature and the stronger the frictional and abrasive properties of the oil. Also, oils with a high percentage of monounsaturated fatty acids (such as oleic acid) have a low coefficient of friction and wear rate and ensure greater oxidation stability at high and low temperatures 26,27 .
A review of the gap in the literature showed that microwave technology has not yet been introduced to intensify the bio-lubricant production process. This work addresses the intensification of the transesterification process of TMPTE production from Camelina oil (Soheil cultivar) as an indigenous cost-effective energy crop using microwave technology. The effect of some parameters in microwave-assisted bio-lubricant production has not yet been clarified; as the molar ratio of methyl ester to TMP, the weight ratio of catalyst to methyl ester and reaction time. Finally, the optimal parameters of the process are determined by Response Surface Methodology and Box-Behnken Design (RSM-BBD). Theis work seeks to bring new information and method to the domain of study for process intensification of bio-lubricant production. The general idea and purpose of this research is to prove the following hypotheses: The use of microwave method accelerates and improves the bio-lubricant production process and there is a direct relationship between microwave power and reaction time with the reaction efficiency of bio-lubricant production.  Table 1. we confirm that all methods applied on the Camelina plant were performed in accordance with the relevant institutional, national, and international guidelines and legislation. Methanol with a purity of 99.6% was purchased from Dr. Mojallali Company, Iran. Trimethylolpropane (2-ethyl-2-(hydroxymethyl)-1,3-propanediol), potassium carbonate, and potassium hydroxide (all with analytical grade) were purchased from Merck. The materials and equipment used in this research were listed in Table 2.
Equipment and experimental setup. The microwave oven used in this study was a Panasonic NN-ST34 with 25 L volume and 5 different power levels. Apart from light, loss of electrical power and other power consumptions, the device has a constant electromagnetic radiation power of 800 W and can generate five power levels by disconnecting and connecting the electromagnetic radiation in different cycles. The maximum total consumed power of the device was 1200 W (measured by a Lutron 6060 wattmeter). The Low mode state (irradiated power of 800 W) was selected as the most effective state for bio-lubricant production, otherwise, the material inside the reactor would overflow or the microwave device would shut down because of over power load.
During each experiment, the microwave power was constantly on the "Low" mode. In this mode, the 800-W electromagnetic waves continuously irradiated the reactants for 5 s at intervals of 30 s (5 s ON and 30 s OFF) 28 . A flexible Teflon rod (T shape) attached to the armature shaft was used to stir the material inside the Erlenmeyer (250 cc). The rotational speed of the shaft was kept in the range of 450 rpm. The experimental setup is depicted in Fig. 2.

TMPTE synthesis.
In the production process of TMP triester, Trimethylolpropane (TMP) polyhydric alcohol was used, which is a widely used polyol in the polymer industry. At this stage, potassium carbonate K 2 CO 3 was used as a catalyst due to its insolubility in alcohol, availability, cheap price, heterogeneity, ease of separation from the reaction medium and reusability in the bio-lubricant production process. The transesterification reaction is based on mass transfer, which takes place in two stages of methyl ester and bio-lubricant production. In the first stage, the reaction to produce methyl ester, oil and methanol are two separated phases, and it is necessary to transfer the mass to form an emulsion from two solutions. During the mass transfer, the triglyceride molecule reacts with methanol to produce glycerol and methyl ester. In the second stage, in order to produce bio-lubricant to break down the steroid group in methyl ester, an alcoholic group of TMP is used to produce new esters based on vegetable oil (Fig. 3). In both stages, stirring and microwave irradiation were used to increase the contact between the liquid phases, the mass and heat transfer rate. A rod made of refractory Teflon attached to the armature shaft was used to agitate the reactants inside the reactor vessel. The armature was installed on the outer body of the microwave device and Teflon tape was used to seal between the armature body and the vessel. During all experiments, the microwave was set to Low mode and stirring was performed continuously.   (Table 3).

Results and discussions
Product analysis. A 500 MHz nuclear-hydrogen magnetic resonance spectroscopy (H-NMR) BRUKER, was used to determine the reaction efficiency. This spectroscopy relies on the magnetic resonance of the atomic nucleus and is used to identify the structure of organic compounds. For analysis, 2 ml of produced bio-lubricant with 0.1 ml of di-chloroform (CDCl 3 ) was poured into a 5 ml tube and placed inside the device. The reaction yield of the reaction was calculated using Mestrenova 11.0.4 software and Eq. (1) 29 :   Table 3. Also, Fig. 4 illustrates the H-NMR spectrum of camelina oil, methyl ester and bio-lubricant samples. Which A, B and C are related to the coded levels of the independent variables of reaction time, potassium carbonate catalyst concentration (K 2 CO 3 ) and the molar ratio of methyl ester to TMP, respectively. At a glance, the effect of each of the independent variables separately on the reaction efficiency was shown in Fig. 6. The effect of variables A and B on the yield in the present system was direct and by increasing their actual or coded values, the yield first increased and then remained constant. Variable C or the molar ratio had an adverse effect on the yield and with increasing the molar ratio, it first decreased and then remained constant. However, according to the ANOVA results in Table 4, the proposed regression model was statistically significant in predicting the bio-lubricant production yield.
The Model F-value of 37.63 implies that the model is significant. There is only a 0.01% chance that an F-value this large could occur due to noise. The F-value of 1.42 for lack of fit, implies adequate model precision. In other words, there is a 36.02% chance that the fitted model could occur due to noise.
Based on the results of Table 4, the main effects of all variables, namely reaction time, catalyst concentration and molar ratio were statistically significant at 1% level. Also, the interaction effect of reaction time and catalyst concentration (AB) as well as time and molar ratio (AC) at 1% level was significant, but the interaction effect of catalyst concentration and molar ratio (BC) on the yield was not significant. The results of the analysis of variance showed that according to the values of the sum of squares and F-value, the most changes in the reaction yield were affected by the two main variables of reaction time and the amount of catalyst.  Effects of experimental factors. The interaction effect of reaction time and catalyst concentration on the reaction yield is shown in the form of a 3D surface diagram and the contour plot in Fig. 8. According to this figure, the reaction yield increased with increasing time and after reaching its maximum value, it remained constant and even decreased (at the highest values of catalyst concentration and reaction time). At higher time levels, due to the decomposition of TMP ester molecules to methyl ester, the reaction was reversed and the reaction yield decreased 31 . By keeping the molar ratio constant at the level of 4, the reaction yield reached its maxi-  In another study, optimal conditions of the process for maximum yield (96.4%), were K 2 CO 3 catalyst concentration of 1 wt%, the reaction time of 2 h, methyl ester to TMP molar ratio of 3.9, the temperature of 130 °C and vacuum pressure of 3.87 kPa 29 . Figure 9 shows the interaction effect of reaction time and molar ratio on the yield. As can be seen, increasing the time in different molar ratios increased the yield. Further increases in time kept this trend constant or led in the yield drop. Also, the reaction yield at lower molar ratios was higher than at elevated molar ratios. Although at low molar ratios the yield was higher in shorter reaction time, the trend increased with a gentle slope and after reaching its maximum in 60 min, it started to decline. At high molar ratios, the upward trend of the reaction yield continued with a steeper slope and remained almost constant during the reaction time of around 90 min. At this time, both curves corresponding to different molar ratios collided and intersected each other in approximately 70 min. It implies that to achieve maximum yield at higher molar ratios, the reaction time should be increased and at shorter reaction time, the molar ratio should be reduced. Also, by considering other effective conditions and parameters, optimal conditions of the variables can be obtained.

Process optimization.
Optimization means obtaining the conditions under which the answer to the problem is directed towards the predetermined goal. The present optimization aimed to obtain the maximum reaction efficiency with the least standard deviation. Thereby in the optimization process, the importance of maximizing yield had a factor of 5 (highest coefficient) and the importance of standard error minimization had a factor of 3. Considering the above conditions and solving the optimization problem by Design-Expert software, an optimal combination was obtained with a time of 67.8 min, a catalyst concentration of 1.4 wt% and a molar      www.nature.com/scientificreports/ ratio of 3.5. In this condition, the reaction yield will be 94.3% with desirability of 0.975 (Fig. 10). Of course, this combination of conditions is one of the solutions to the optimization problem that the software shows as the highest desirability and the first answer. Several other combinations in the same range could be offered with lower desirability as the answer to the optimization problem.
Analysis of physicochemical properties. At the end of the experimental steps, a sample of product synthesized under the above-mentioned optimal condition was purification and its physicochemical properties were compared with the ISO-VG standard ( Table 5). The results clearly showed that many of the obtained properties met the requirements of the standard. The pour point of the product (− 4) was a bit higher than the ISO-VG standard which can be attributed to the presence of unsaturated bonds in the fatty acids structure of camelina oil. However, in this case, adding a proper amount of pour point depressant additive is recommended.

Conclusions
In the present study, the feasibility of using camelina oil, as a promising industrial crop was investigated for biolubricant (TMP ester) production. In this regard, process intensification of the mass-transfer-limited transesterification reaction of methyl ester obtained from camelina oil was conducted using microwave irradiations. The RSM-BBD approach was also applied for optimization of the experimental factors and to find the optimal levels at which the highest transesterification reaction yield is obtained. The time and molar ratio had a direct effect on the reaction efficiency and increasing their values was an increasing trend. The molar ratio had an inverse effect and with its increase, the reaction efficiency was reduced. Indeed, setting the microwave power at 800 W can greatly reduce the conventional reaction time (10 min instead of 120 min). Thereby, it can be concluded that microwave irradiation is an efficient method for process intensification of bio-lubricant production. Moreover, Figure 10. Optimization of parameters affecting the bio-lubricant reaction.