Optimization of microwave-assisted etherification of glycerol to polyglycerols by sodium carbonate as catalyst

Accepted Manuscript Optimization of Microwave-assisted Etherification of Glycerol to Polyglycerols by Sodium Carbonate as Catalyst Pornpimol Bookong, Somsak Ruchirawat, Siwarutt Boonyarattanakalin PII: DOI: Reference:

S1385-8947(15)00515-X http://dx.doi.org/10.1016/j.cej.2015.04.033 CEJ 13522

To appear in:

Chemical Engineering Journal

Received Date: Revised Date: Accepted Date:

6 November 2014 20 March 2015 2 April 2015

Please cite this article as: P. Bookong, S. Ruchirawat, S. Boonyarattanakalin, Optimization of Microwave-assisted Etherification of Glycerol to Polyglycerols by Sodium Carbonate as Catalyst, Chemical Engineering Journal (2015), doi: http://dx.doi.org/10.1016/j.cej.2015.04.033

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Optimization of Microwave-assisted Etherification of Glycerol to Polyglycerols by Sodium Carbonate as Catalyst

Pornpimol Bookonga, Somsak Ruchirawatb,c, Siwarutt Boonyarattanakalina* a

School of Bio-Chemical Engineering and Technology, Sirindhorn International Institute of Technology, Thammasat University, Pathum Thani, Thailand b

Program in Chemical Biology, Chulabhorn Graduate Institute and the Center of Excellence on Environmental Health, Toxicology and Management of Chemicals, 54 Kamphaeng Phet 6, Talat Bang Khen, Lak Si, Bangkok 10210 c

Laboratory of Medicinal Chemistry, Chulabhorn Research Institute (CRI), 54 Kamphaeng Phet 6, Talat Bang Khen, Lak Si, Bangkok 10210

* Corresponding author. Tel.: +6-698-690-09~13 Ext. 2305; fax: +6-698-691-12~13; e-mail: [email protected], [email protected].

ABSTRACT

The process optimization of etherification of glycerol to polyglycerols by sodium carbonate as a catalyst using microwaves as a heat source in solvent free conditions has been investigated in this study. The regression models describing the linear correlations between reaction parameters and reaction outcomes were developed. The three reaction parameter studied are: reaction temperature (220 °C and 270 °C), catalyst concentration (1 %wt. and 3 %wt.), and reaction time (0.5, 1.0, 1.5, 2.0, and 3.0 h). The reaction temperature was found to have the most significant effect on the percent conversion of glycerol (X); combined yields of diglycerols, triglycerols, and tetraglycerols products (YDG+TG+TtG); selectivity toward pentaglycerols (SPG); and selectivity toward cyclic diglycerols (Sc-DG). Undesirable cyclic diglycerols formed at high temperature and the amount increased along with the longer reaction time. The higher catalyst concentration (3 wt%) provides higher amount of the desired products. Moreover, the longer reaction time also resulted in higher conversions and yields. For the optimized conditions, X, YDG+TG+TtG, and Sc-DG were predicted to be at 84%, 63%, and 9%, respectively, at the reaction temperature of 270 °C, catalyst concentration of 3 wt%, and reaction time of 1.0 h. Experimental verification of the predicted optimum conditions gave the actual responses of 93%, 70%, and 7% for X, YDG+TG+TtG, and Sc-DG, respectively, with small deviations from the predicted responses. The results indicated that the developed models were valid and accurate in describing the actual experimental data at any conditions within the range studied. The observation suggests that the sodium carbonate catalyst with microwave heat source have the potential to be used in glycerol conversions to polyglycerols.

HIGHLIGHTS • • • •

Process optimizations on the etherification of glycerol to polyglycerols are reported. Reaction temperature is the most significant variable for the optimized process. Microwave radiation shortens the reaction time for the etherification of glycerol to oligoglycerols. HPLC was used to determine the quantity of polyglycerol products without any product modifications.

Keywords: Glycerol; Polyglycerols; Microwave-Assisted Etherification; Optimization of Glycerol Etherification.

1.

Introduction

Biodiesel is a clean, renewable, and efficient alternative fuel to replace petroleum diesel which is predicted to be depleted in supply, and recently raises concerns on greenhouse gas concentrations [1]. Biodiesel can be produced by transesterification of animal or vegetable oils with alcohols, especially methanol. Glycerol is a main by-product from the transesterification process [2]. Glycerol supply exceeds its demand because of the recent dramatic growth of the biodiesel industry, resulting in the plunge of glycerol price [3]. Conversions of glycerol to other higher-valued chemicals are currently a focus of global research. Glycerol is a suitable starting material for a variety of chemical intermediates because of its nontoxic, edible, bio-sustainable, and biodegradable properties [4]. Cost-effective conversions of glycerol to valuable chemicals are necessary in order to improve the economy of the whole biodiesel production process [5, 6]. Glycerol can be used as a reactant in oxidation processes, fermentation process, acetylation processes with acetic acid, and acetalization processes with ketones. Valuable chemicals obtained from glycerol include glyceraldehyde, dihydroxyacetone, hydroxyl pyruvic acid glycolic acid, and glyceric acid [7], 1,3-propanediol [8], polyglycerol esters [9, 10], and oxygenated acetals and ketals [11-13]. Glycerol by-product from biodiesel production can be a source for the preparation of acidic catalysts for utilization in etherification reactions producing oxygenated additives for fuels [14]. Oxygenated components such as polyglycerols and polyglycerol ethers derived from etherification process are efficient chemical platforms [15]. Oligoglycerols, especially diglycerols and triglycerols (DG and TG, respectively) are of industrial interest because of their useful applications as components in cosmetics, polymers, food, and pharmaceutical industries [4]. Bigger oligoglycerol molecules such as pentaglycerols and the de-emulsifier cyclic diglycerols are not desired for food applications (JECFA, 1973; JECFA, 1989). The investigations of etherification of glycerol have been done by using both homogeneous and heterogeneous catalysts [16-21]. These studies were carried out by conventional heating which usually required longer reaction times. The current study reveals a key advantage of the heating by microwave radiation in shortening the reaction time required for the etherification. Previously, several homogeneous catalysts such as Cs2CO3, NaOH, Na2CO3, CsOH, and H2SO4 have been examined for the conversion of glycerol to polyglycerols. Homogeneous catalysts provide higher glycerol conversions but lower selectivity to DG and TG, when compared with heterogeneous catalysts [21]. Alkali hydroxides are generally more active than carbonate bases due to the more basic strength of the hydroxide, but are less soluble in glycerol than carbonates [21]. During the etherification of glycerol, the electrophilic carbinol carbons are attacked by either primary or secondary hydroxyls in intermolecular or even intramolecular fashion. Polyglycerols are formed through consecutive etherifications where the glycerol molecules are condensed in linear, branched, or cyclic fashions [22]. In general, heterogeneous catalysts hold the advantages of the ease in separation and reusability but they give poorer activity when compared to homogeneous catalysts. Some heterogeneous catalysts cannot be employed in industry because of the difficulty in production and high cost. Other disadvantages of heterogeneous catalysts include leaching of chemicals, solubility in polar solvent, low thermal stability, and high reaction temperature and long reaction time requirements [16-18]. For heterogeneous catalysts, especially, mesoporous materials are applied because they can improve selectivity though its limited pore size to favor low molecular weight oligomers, which are DG and TG [23]. Many efforts have been made to control the selectivity towards small oligoglycerols by investigating the effect of heterogeneous catalysts, such as impregnated CsMCM-41, grafted AlgSi-MCM-41, zeolite NaBeta, MgAl mixed oxides, and colloidal CaO [17, 24, 25]. Selectivity towards oligoglycerols tends to decrease at higher glycerol conversion because oligoglycerols are further converted into larger polyglycerols, in particular diglycerols and triglycerols are converted into larger polyglycerols. Reaction times required in most studies were longer than 24 h.

A study reported the application of microwave radiation as a heat source in the synthesis of polyglycerol from glycerol carbonate. It was found that the microwave radiation helps to improve the reaction yield and shorten the reaction time resulting in less energy consumption [26]. Only the polymerization of polymethylmethacrylate (PMMA), polymethylacrylate (PMA), and polystyrene (PS) at commercial scale applications were carried out by microwave heating [27]. This work aims to optimize the chemical conditions for the solvent-free microwave-assisted glycerol etherification by full factorial design at 2 levels to achieve the highest combined yields of small polyglycerols including diglycerols (DG), triglycerols (TG), and tetraglycerols (TtG). These three products can be used as additives in food applications. According to The Joint FAO/WHO Expert Committee on Food Additives (JECFA) specifications of polyglycerols composition, the polyglycerols in food additive can be the mixture of diglycerols, triglycerols, and tetraglycerols which must not be less than 70-75 %, and there should not be more than 10% heptaglycerols (or bigger) in the composition [28, 29]. In the present study, a sodium carbonate catalyst (Na2CO3) was the catalyst of choice because of its availability and high solubility in glycerol. Moreover, metal carbonate catalysts were chosen because: (i) base-catalyzed glycerol etherification pathway was reported to suppress the formation of undesirable cyclic oligomers and acrolein [23, 30]; (ii) carbonates were found to be more active than hydroxides because of their higher solubility in glycerol [31]. Microwave radiation was utilized as the heat source in order to overcome the typical long reaction time for the etherification of glycerol. The key reaction parameters investigated in the optimization process were reaction temperature, reaction time, and catalyst concentration. The polymer products were purified by column chromatography and characterized by nuclear magnetic resonance (NMR) (both 1H-NMR, 13 C-NMR), and electrospray ionization of high resolution mass spectrometry (ESI-HRMS). High performance liquid chromatography (HPLC) was used for qualitative and quantitative analyses. The etherification of glycerol and the possible products are shown in Fig. 1 and Fig. 2 respectively.

Fig. 1. Catalytic etherification of glycerol to polyglycerols.

Fig. 2. Possible isomers of diglycerols products formed during etherification of glycerol [21].

2.

Materials and methods

2.1 Reaction procedure and product analysis Specific amounts of Na2CO3 catalysts (Merck, Germany) were added to glycerol (10 mL) in a round bottom flask (100 mL). Glycerol etherification was conducted in a stirred microwave reactor, without additional solvent under various conditions as the followings: reaction temperature of 220 °C and 270 °C, catalyst concentrations of 1 %wt. and 3%wt., microwave power of 25 W, atmospheric pressure, and reaction time of 0.5 to 3 h. Sample collections (0.1 mL) were done at 0.5 h, 1.0 h, 1.5 h, 2.0 h, 2.5 h, and 3.0 h. In order to prevent the evaporated water from re-entering the system, a bump trap and a condenser (25 °C) were connected to the glass reactor. Microwave irradiation is provided by an open vessel mode of CEM Discover® SP operating at 2.45 GHz and 300 W maximum power output. Products from the glycerol etherification were qualitatively characterized by ESI-HRMS and thin layer chromatography (TLC). ESI- HRMS analyses were conducted in an MS instrument (Bruker, United States) with ionization of analytes by sodium cations. TLC was performed on normal phase silica gel (Merck, Germany) as the stationary phase, and a 10% H2O in acetonitrile solution as the mobile phase. The TLC plate was stained

with an aqueous solution of potassium permanganate (KMnO4, 0.96 % w/w) and potassium carbonate (K2CO3, 3.23 % w/w). The spots appeared on the TLC plates with different retention factor (Rf) for each oligoglycerol components. The Rf of standard samples of glycerol (Univar, United States), diglycerols (Tokyo Chemical Industry, Japan), and triglycerols (Aldrich, United States) are 0.41, 0.35, and 0.26, respectively. The composition and yield of the products from the glycerol etherification was determined by HPLC. Agilent 1260 Infinity series HPLC was used for analysis. Baseline separation of glycerol, DG, TG, TtG, pentaglycerols (PG), and octan-1-ol was achieved with a 25 cm × 4.6 mm ID, 5 µm VertiSepTM GES NH2 HPLC column at 30 °C. A solution of acetonitrile (85%) and water (15%) was used as a mobile phase at a flow rate of 1 mL/min. The compounds were detected by a refractive index detector (RID). Samples were dissolved in a solution equivalent to the mobile phase (1 % v/v), and 20 µL of solution was injected by automatic loop injector. Octan1-ol was used as internal standard (2.5 µL/ 1 mL of prepared sample) for the calibration curve. Samples were collected after the elution of each fraction of compounds and their identities were confirmed by ESI-MS. Glycerol conversion and selectivity towards DG, TG, TtG, and PG were quantified by HPLC; peak areas were normalized to represent the product compositions. Normalized peak areas were interpreted as normalized mass composition, and they were converted into normalized mole composition by dividing the molecular weight of the particular compounds. The following equations (Eqs. (1) - (6)) were then used to calculate the glycerol conversion (X) and selectivity towards diglycerols (SDG), triglycerols (STG), tetraglycerols (STtG), and pentaglycerols (SPG). n = n + 2n + 3n  + 4n  + 5n 

(1)

X = n − n ⁄n  × 100 %

(2)

S = n ⁄n − n ν, ⁄ν, ! × 100 %

(3)

S  = n  ⁄n − n ν," ⁄ν ," ! × 100 %

(4)

S  = n  ⁄n − n ν,# ⁄ν ,# ! × 100 %

(5)

S  = n  ⁄n − n ν,$ ⁄ν ,$ ! × 100 %

(6)

2C3H8O3 → C6H14O5 + H2O

(7)

3C3H8O3 → C9H20O7 + 2H2O

(8)

4C3H8O3 → C12H26O9 + 3H2O

(9)

5C3H8O3 → C15H32O11 + 4H2O

(10)

where G, DG, TG, TtG, and PG are abbreviations for glycerol, diglycerols, triglycerols, tetraglycerols, and pentaglycerols, respectively; nG, nDG, nTG, nTtG, nPG = normalized mole compositions of glycerol, diglycerols, triglycerols, tetraglycerols, and pentaglycerols, respectively; nG0 = normalized mole composition of initial amount of glycerol; X = percent glycerol conversion; SDG, STG, STtG, SPG = selectivity towards diglycerols, triglycerols, tetraglycerols, and pentaglycerols, respectively. According to the stoichiometries in chemical Eqs. 7, 8, 9, and 10, the stoichiometric coefficients (νG,i/νoligomer,i) in Eqs. 3, 4, 5, and 6 are 2, 3, 4, and 5, respectively. The percent yield of diglycerols, triglycerols and tetraglycerols (YDG+TG+TtG) can be calculated from the conversion and selectivity towards diglycerols, triglycerols, and tetraglycerols by Eq. (11). Y& &  = X × S + S  + S  /100%

2.2 Isolation and characterization of products

(11)

2.2.1 Column chromatography Polyglycerol products were carefully purified and isolated into a single fraction of polyglycerols with different sizes by a flash column chromatography packed with silica gel (230–400 mesh, Fluka Kieselgel). Mixtures of water and acetonitrile were used as the mobile phase system. The conditions were performed on normal phase silica gel (TLC Silica gel 60 F254, Merck, Germany) as the stationary phase with a 10% H2O in MeCN solution as the mobile phase. The retention times (Rf) of fractions for glycerol, diglycerols, triglycerols and tetraglycerols, obtained from our chemical reactions, are 0.41, 0.35, 0.26, and 0.20, respectively. After completion of separation, the eluent was removed in vacuo. The separated products from column chromatography were characterized by 1H- NMR and 13C -NMR, and ESI-HRMS. 2.2.2 High resolution mass spectrometry and NMR analyses All product compounds were identified by ESI- HRMS, 1H- NMR, and 13C–NMR. High resolution mass spectra (HRMS) were recorded with a micrOTOF instrument with source type of ESI. NMR spectra were recorded with a Bruker AVANCE III spectrometer (300 MHz for both 1H and 13C). Chemical shifts are expressed in ppm. Deuterium oxide (D2O) was used as a solvent for NMR analyses. 2.2.3 Design of experiment (DOE) A full factorial design was used to identify variables that have significant influence on the response of interest [32]. Our preliminary results suggested that the three reactions parameters, including reaction temperature (Tr), catalyst concentration ([Cat.]), and reaction time (tr) were the most crucial factors that greatly influence the etherification yield and selectivities. Two-level (Tr and [Cat.]) and six-level (tr) full factorial designs were employed in this study with three numerical factors that have significant effects on the etherification of glycerol to polyglycerols. Based on the preliminary results, Tr was 220 °C and 270 °C, and [Cat.] was 1 wt% and 3 wt%. The low and high values of Tr and [Cat.] were assigned as -1 and +1, respectively. The reaction time has more profound effects on the conversion and selectivities, and therefore, tr was studied in more detail at 6 levels as listed in Table 1.

Table 1 Experimental ranges and levels for the etherification variables.

The total number of experiments was 24 reactions. The responses considered were: glycerol conversion (X, %); combined yields of the desired diglycerols, triglycerols, and tetraglycerols (YDG+TG+TtG, %); selectivity towards pentaglycerols (SPG, %); and selectivity towards cyclic diglycerols (Sc-DG, %). Each reaction condition was carried out in triplicate and the average results are shown in Table 2. Statistical analysis of the responses was performed by least squares fitting using DOE PRO XL 2007 (Sigma zone) software. The mathematical models for the combined yields of desired products and glycerol conversion were fitted to the second-order interaction model as given in Eq. (12). This model is commonly used for twolevel and six-level full factorial design because they are not overwhelmingly complicated and they provide useful information, including main effects and interaction effects. Y = const. +β$ A + β/ B + β1 C + β$/ AB + β$1 AC + β/1 BC

(12)

where Y is dependent variables (X, YDG+TG+TtG, SPG, Sc-DG,). A, B, and C are the independent variables; β$ , β/ , and β1 are the linear coefficient; β$/ , β$1 , and β/1 are the second order interaction coefficients. After the model was verified, the experimental data analyses were performed based on the model equation.

Table 2 Experimental design and responses of interest for the etherification of glycerol to polyglycerols.

3.

Result and Discussion

3.1 Etherification of glycerol The results from the complete design matrix for the sodium carbonate etherification of glycerol are shown in Fig. 3. The X and YDG+TG+TtG are significantly increased when the temperature increased from 220 °C to higher than 270 °C. However, the higher temperature significantly generated cyclic-diglycerols (c-DG), whereas none of the c-DG was observed at a lower temperature. c-DG was also found to increase at the longer reaction time. The higher catalyst concentration (3 wt%) provides a larger percent of the desired products. Moreover, the longer reaction time also resulted in higher conversions and yields. At high Tr and long tr, the conversion of glycerol increased and led to the formation of non-desirable larger polyglycerols, resulting in a decrease of the combined yields of the desired oligoglycerols. Based on the highest possible yield and the criteria on the polyglycerol compositions set by JECFA, the specifications of polyglycerol compositions in food additives can be a mixture of DG, TG, and TtG which must not be less than 70-75 %, and there should not be more than 10% heptaglycerols (or bigger) in the composition [28, 29]. The best experimental conditions are 270 °C, microwave power of 25 W, 3 wt% Na2CO3 as a catalyst, at a reaction time of 1 h. The X, YDG+TG+TtG, SPG, and Sc-DG are 93, 70, 9, and 7 percent, respectively. These results were comparable with the work done by Barrault [21], where the reaction was carried out with MgO-SiO2 heterogeneous catalyst at 260 °C, 2 wt% catalyst, and 4 h of reaction time by a conventional heat source. The X and YDG+TG+TtG, were 83 and 80 %, respectively. The utilization of microwave radiation as the heat source and the homogenous Na2CO3 catalyst used in the current work led to the better glycerol conversion in shorter reaction times. However, the selectivity toward the total yield of DG, TG, and TtG was better in the case of the MgO-SiO2 heterogeneous catalyst. The shorter reaction time required in the current study allowed the utilization of the homogenous Na2CO3 which is known to be less selective under longer reaction times. Therefore, the benefits of using Na2CO3 , including lower cost, ready availability, and safety, are exploited in this study.

Fig. 3. Comparisons of the effect of the reaction temperature (220 °C and 270 °C); catalyst concentration (1 wt% and 3 wt%) at the microwave power of 25 W. The conditions were compared to see the conversion of glycerol, the combined yields of desired products, the selectivity toward pentaglycerols, and the selectivity toward cyclic-diglycerols.

Microwave radiation was proven to be a more effective heating method in the etherification of glycerol. The required reaction time to generate the desired polyglycerols by the etherification facilitated by the microwave radiation was much shorter than the etherification previously done by conventional heating [19]. Most of the previous investigations on the etherification were carried out by conventional heating and the typical reaction times were longer than 8 h [19]. In contrast, the microwave radiation applied in this study could generate significant amounts of the desired oligoglycerol within 30 min, and the desired reaction outcomes were reached within 2 h. In addition, our preliminary results (data not shown) were carried out for a side by side comparison between the etherifications of glycerol facilitated by microwave radiation and by conventional heating at 250

°C. The results evidently confirmed the advantage derived from the heating by microwave radiation. In order to achieve a glycerol conversion of at least 50%, the required heating time by conventional heating was at least 6 h. On the other hand, it took only 1 h for microwave heating. It should noted that the HPLC analytical method used in this study was limited to the commonly desired products which are c-DG, DG, TG, TtG, and PG compounds. Thus, a small portion of larger polyglycerols formed after 2.5 h at 270 ºC (Fig. 3) was not taken into account. The limitation led to the relative bigger standard deviations for the results from the reaction conditions at high temperature. 3.2 Development of regression model equation The correlation of responses and three reaction parameters in Table 2 were studied using multiple regression analysis, expressed in Eq. (12). After excluding some of the insignificant terms as identified by Fisher’s test, the obtained final equations are revealed below. X = 60.26 + 28.95A + 5.71B + 8.37C − 6.22AB

(13)

Y& &  = 44.29 + 14.96A + 6.30B + 3.44C − 3.97AB − 5.60AC

(14)

S  = 6.82 + 6.82A − 3.00B − 3.00AB

(15)

S9: = 6.52 + 6.52A + 1.48B + 4.69C + 1.48AB + 4.69AC + 3.28BC

(16)

The positive coefficients of the individual reaction parameters indicate the direct linear relationships between the reaction parameters and the responses. The antagonistic effects between the reaction parameters are reflected as the negative coefficients of the interacting parameters, while the positive values indicated the synergistic effect. Eqs. (13) - (16) suggested that the key reaction parameters including the reaction temperature, the catalyst concentration, and the reaction time have linear effects on the desired responses (X and Y& &  ) and also on the undesired Sc-DG. 3.3

Model adequacy check

The model accuracy is verified by the coefficients of correlation (R2). From the ANOVA, the values of R2 of X and YDG+TG+TtG were found to be 0.9414 and 0.7159, respectively. The R2 value of 0.9616 implies that 96.16% of the total variations in the responses were represented by the experimental variables studied. The high R2 values indicated the good correlations between calculated values obtained from the model and the experimental values. The obtained regression model accurately described the actual experimental data. The linear correlations between the reaction parameters (reaction temperature, catalyst concentration, and reaction time) and the reaction outcomes were captured by the developed models. When the values of R2 of X and YDG+TG+TtG were compared, as expected, the high coefficient of correlation was observed in the case of the glycerol conversion because X was linearly accumulated and cannot be reduced along the course of the reaction. However, the YDG+TG+TtG was somewhat deviated from the actual values because some amounts of DG, TG, and TtG were further converted into PG or larger polyglycerols, therefore, the R2 of YDG+TG+TtG was lower. The statistical analyses of variance for the regression model equations and significant coefficients are reported in Table 3. The higher computed F-value (269.18) than the theoretical F0.01 (4, 67) value of 3.61 indicated that the developed regression model of conversion of glycerol (X) was reliable, based on the 99% confidence level. For the other responses, the computed F-values were also higher than their theoretical F values.

Table 3 Analysis of variance (ANOVA) for the regression model equation and coefficients

3.4 Significance and effects of variables

The significance of each coefficient in single and combination factors was illustrated by Pareto charts (Fig. 4). The smaller p-value and the greater t-value accompanied the more significant coefficient [33, 34]. The data showed that the reaction temperature (A) was the most significant reaction parameter because its F-test value was the highest.

Fig. 4. The absolute coefficients comparison of one and two interaction factors of conversion (a) X, (b) YDG+TG+TtG, (c) SPG, and (d) Sc-DG.

The three dimensional response surface plots for the two interacting reaction parameters are shown in Fig. 5. The second interaction effect of catalyst concentration and reaction temperature and was observed on the glycerol conversion (X, Fig. 5a). The increase in the reaction temperature and the catalyst concentration resulted in dramatic increase of the glycerol conversion. Although, the high reaction temperature directly promoted the glycerol conversion, the reaction temperature should not be set higher than 270 °C to minimize the significant loss of glycerol through heat evaporation because the boiling point of glycerol is at 290 °C. In addition, the higher temperature would also accelerate the generation of the undesired c-DG and larger polyglycerols. The combined yield of desired products (YDG+TG+TtG) is higher at the higher amount of catalyst and the higher reaction temperature. However, at the high Tr, [Cat.], and tr, c-DG was significantly formed. The combination of a high amount of catalyst, longer reaction time, and higher temperature should be avoided because at these conditions, cyclic diglycerols were significantly formed.

Fig. 5. Graphical surface plot of variable second interaction of (a) temperature (Tr, °C) and catalyst conc. ([Cat.], wt%) at time (tr) = 1.75 h of X, (b) Tr (°C) and [Cat.] (wt%) at tr = 1.75 h of YDG+TG+TtG, (c) Tr, (°C) and tr (h) at [Cat.] = 2 wt% of YDG+TG+TtG., (d) Tr (°C) and [Cat.] (wt%) at tr = 1.75 h of SPG., (e) Tr (°C) and tr (h) at [Cat.] = 2 wt% of SPG., (f) [Cat.] (wt%) and tr (h) at Tr = 245 °C of Sc-DG.

3.5 Optimization process In order to identify the optimal conditions, the desired responses for the reaction outcomes were set as the followings: 1) X ≥ 70 %, 2) YDG+TG+TtG ≥ 65 %, 3) SPG < 10 %, and 4) Sc-DG < 10 %. Based on the response models (Eqs. 13 - 16), the optimum reaction conditions (25 W of microwave power) were predicted to be at the reaction temperature of 270 °C, catalyst concentration of 3 wt%, and reaction time of 1.0 h. The optimal conditions were predicted to yield 84%, 63%, and 9% of X, YDG+TG+TtG, and Sc-DG, respectively. The actual experimental outcomes were compared with the predicted responses to further evaluate the accuracy of the model. The X, YDG+TG+TtG, and Sc-DG for the experimental values were 93%, 70%, and 7%, respectively, indicating a 10.7%, 10.0%, and 17.6% errors between the observed and predicted values for X, YDG+TG+TtG, and Sc-DG, respectively. 3.6 Effect of reaction parameters The three reaction parameters including reaction temperature, catalyst loading, and reaction time, were investigated for their effects on the desired responses. Regarding the effect of the reaction temperature, it is important to balance between trying to achieve the highest possible glycerol conversion while avoiding the formation of the undesired c-DG and larger polyglycerols at the high reaction temperature. Therefore, the influence of reaction temperature at 220 °C and 270 °C on glycerol etherification was investigated. As shown in Fig. 3, the reaction conditions at 270 °C showed higher glycerol conversions than at 220 °C at the same [Cat.] and tr. The higher temperature increased the collision frequency of particles (glycerol, DG, TG, and TtG) and

the number of energetic particles, resulting in faster reaction rates and greater extents of the etherification. Consequently, the formation of the larger polyglycerols, such as pentaglycerols, was observed in higher amounts at 270 °C, especially at longer reaction times. The higher reaction temperature enhanced the consecutive etherifications of DG, TG, and TtG to the undesired polyglycerols, which are larger than tetraglycerols. The observations were in good agreements with the previous reports [35, 36]. Therefore, it is important to balance between the possible maximum glycerol conversion and the formation of the undesired larger polyglycerols by keeping a shorter reaction time, as suggested by the predicted optimal conditions. In addition to the better solubility of Na2CO3 in glycerol when compared to other metal hydroxides, each molecule of Na2CO3 can produce two hydroxide ions (-OH) which act as a strong basic catalyst for the etherification reaction (Fig. 6) [30]. The positive effect of the catalyst concentration was more pronounced at the lower temperature (220 °C) than at the higher temperature (270 °C), as the Na2CO3 amount was increased from 1 wt% to 3 wt%. Na2CO3 provides an alternative route for the reaction by removing a proton from glycerol to generate the glycerol alkoxide (3) which serves as a stronger nucleophile (Fig. 6). The reduction in activation energy by the Na2CO3 catalyst was more significant at the lower reaction temperature, while more energy was available at the high reaction temperature conditions so that the higher amount of catalyst became less critical.

Fig. 6. The catalytic activity of Na2CO3 catalyst on the etherification of glycerol to polyglycerols [30].

For the effect of the reaction time, it was found that that tr has a significant effect on the product profiles. An increase in the reaction time generally resulted in higher X and YDG+TG+TtG. As the etherification reaction was prolonged, more glycerol molecules were condensed leading to an increase in the conversion of glycerol [36]. However, at 270 °C, the increment of the desired X was accompanied by the undesired c-DG and larger polyglycerols, which can diminish the amount of YDG+TG+TtG already formed. The formation of both c-DG and larger polyglycerols can be minimized while maintaining the possible maximum glycerol conversion by keeping the reaction time shorter than 2 h. The reaction conditions with shorter reaction times also consume less energy, which can have a profound effect on the process economy. It should be also emphasized that a very high glycerol conversion may not be necessarily beneficial to the overall etherification process. A high glycerol conversion requires higher reaction temperatures and longer reaction time. At these conditions, the etherification of glycerol may not yield the desired polyglycerol formation, but could also lead to some other forms of side-products such as acrolein, due to the possible double dehydration of glycerol, in addition to the already mentioned c-DG and larger polyglycerols [1]. These byproducts are not desirable and would complicate the purification process of the obtained products, especially the larger polyglycerols with high boiling points.

3.7 High resolution mass and NMR spectra of the polyglycerol products NMR and HRMS were used to confirm the structural identities of the polyglycerol products by comparison with their standards. The mass and NMR spectra of the synthesized oligoglycerols were found to be the same as those of standard oligoglycerols. The HRMS of the synthesized polyglycerols from the conditions of reaction temperature of 270 °C, reaction time of 0.5 h, and catalyst concentration of 3 wt% showed peaks at m/z= 189.0731, 263.1091, and 337.1465, which were the same as the [M+Na]+peaks of diglycerol standard at m/z= 189.0731, triglycerol standard at m/z= 263.1091, and tetraglycerol standard at m/z = 337.1465. The 1H- NMR, and 13C –NMR spectra of the standard polyglycerols and the identified polyglycerols obtained from our own experiments were closely similar. The chemical shifts were assigned based on the published literature [37]. 1H-NMR of the oligoglycerol products showed highly overlapped peaks at δ = 3.40-3.85 (data not shown). Multiple peaks with various chemical shifts are from different proton signals of many isomers of

each oligoglycerols. The overall splitting patterns of the 1H-NMR in both standard and the synthetic polyglycerols are similar. The 13C-NMR spectra of polyglycerol products from the reaction show the same spectra as the 13C-NMR of standards DG, TG, and TtG (Fig. 7). For the 13C-NMR, there are four main ranges of chemical shifts (δ) as follows: 1) 60-63 ppm, 2) 70-71 ppm, 3) 72-73 ppm, and 4) 80-81 ppm for the corresponding signals of the types of carbons as follows: 1) –CH2OH, 2) –CH–OH, 3) –CH2–O–, and 4) –CH–O–, respectively [37]. At a chemical shift range of 60 - 63 ppm, the carbons of –CH2OH are indicated as carbons a and e of the diglycerol molecules (Fig. 7). At the chemical shifts of 70-73 ppm, the carbons labeled as c and f are more downfield than carbon b because the c and f carbons are connected with the oxygen atom in an ether functional group (–O–), which is more electron withdrawing than the functional group –OH, which is attached to the carbon b. Fig. 7. 13C-NMR spectra with assigned carbon signals for (a) polyglycerol products obtained from the etherification of glycerol in the reaction conditions of 220 °C, 3 h, 3 wt% catalyst; (b) polyglycerol products obtained from the etherification of glycerol in the reaction conditions of 270 °C, 0.5 h, and 3 wt% catalyst; (c) DG standard; (d) TG standard; and (e) TtG standard.

4.

Conclusion

This study has demonstrated the feasibility of converting the less valuable glycerol into the valueadded polyglycerols successfully by a solvent free etherification of glycerol facilitated by microwave radiation. The introduction of the microwave radiation could generate the desired polyglycerols in much shorter reaction times than conventional heating employed in previous reports. The investigations have revealed that the reaction temperature has the most significant effect on the etherification of glycerol to polyglycerols. The optimization suggested the optimal reaction parameters as follows: reaction temperature at 270 °C, catalyst concentration at 3 wt%, and reaction time of 1.0 h, which experimentally produced the highest conversion of glycerol, the highest combined yields of desired products, and acceptable selectivity toward cyclic diglycerols at 93%, 70%, and 7%, respectively. The developed models which describe the linear correlations between the reaction outcomes and the reaction parameters were proved to be sufficiently accurate in representing the actual experimental data and also in predicting the experimental results at any conditions within the studied ranges. The results suggest that the sodium carbonate catalyst with microwave radiation as a heat source has the potential to be used in the glycerol conversion to polyglycerols.

5.

Acknowledgments

This research was supported by the Thailand Research Fund (TRF, Grant # RSA5580059), the National Research University Project of Thailand Office of Higher Education Commission, and the PTT Research & Technology Institute. Pornpimol Bookong is a recipient of a scholarship from the Joint Graduate School of Energy and Environment, Thailand. We thank Chulabhorn Research Institute for chemicals and equipment.

6.

References

[1] J.M. Clacens, Y. Pouilloux, J. Barrault, Selective etherification of glycerol to polyglycerols over impregnated basic MCM-41 type mesoporous catalysts, Appl. Catal. A: Gen. 227 (2002) 181-190. [2] M.A. Olutoye, B.H. Hameed, Synthesis of fatty acid methyl ester from used vegetable cooking oil by solid reusable Mg1−x Zn 1+xO2 catalyst, Bioresource Technol. 102 (2011) 3819-3826. [3] M.S. Khayoon, B.H. Hameed, Acetylation of glycerol to biofuel additives over sulfated activated carbon catalyst, Bioresource Technol. 102 (2011) 9229-9235.

[4] A. Behr, J. Eilting, K. Irawadi, J. Leschinski, F. Lindner, Improved utilisation of renewable resources: New important derivatives of glycerol, Green Chem. 10 (2008) 13-30. [5] J.A. Melero, G. Vicente, M. Paniagua, G. Morales, P. Munoz, Etherification of biodiesel-derived glycerol with ethanol for fuel formulation over sulfonic modified catalysts, Bioresource Technol. 103 (2012) 142-151. [6] N. Rahmat, A.Z. Abdullah, A.R. Mohamed, Recent progress on innovative and potential technologies for glycerol transformation into fuel additives: A critical review, Renew. Sust. Energ. Rev. 14 (2010) 987-1000. [7] V. Augugliaro, H.A.H. El Nazer, V. Loddo, A. Mele, G. Palmisano, L. Palmisano, S. Yurdakal, Partial photocatalytic oxidation of glycerol in TiO2 water suspensions, Catal. Today 151 (2010) 21-28. [8] H. Tokumoto, M. Tanaka, Novel anaerobic digestion induced by bacterial components for value-added byproducts from high-loading glycerol, Bioresource Technol. 107 (2012) 327-332. [9] V.L.C. Gonçalves, B.P. Pinto, J.C. Silva, C.J.A. Mota, Acetylation of glycerol catalyzed by different solid acids, Catal. Today 133–135 (2008) 673-677. [10] M. Balaraju, P. Nikhitha, K. Jagadeeswaraiah, K. Srilatha, P.S. Sai Prasad, N. Lingaiah, Acetylation of glycerol to synthesize bioadditives over niobic acid supported tungstophosphoric acid catalysts, Fuel Process. Technol. 91 (2010) 249-253. [11] G. Vicente, J.A. Melero, G. Morales, M. Paniagua, E. Martin, Acetalisation of bio-glycerol with acetone to produce solketal over sulfonic mesostructured silicas, Green Chem. 12 (2010) 899-907. [12] S.B. Umbarkar, T.V. Kotbagi, A.V. Biradar, R. Pasricha, J. Chanale, M.K. Dongare, A.-S. Mamede, C. Lancelot, E. Payen, Acetalization of glycerol using mesoporous MoO3/SiO2 solid acid catalyst, J. Mol. Catal. A. Chem. 310 (2009) 150-158. [13] P.H.R. Silva, V.L.C. Gonçalves, C.J.A. Mota, Glycerol acetals as anti-freezing additives for biodiesel, Bioresource Technol. 101 (2010) 6225-6229. [14] M. Gonçalves, M. Mantovani, W.A. Carvalho, R. Rodrigues, D. Mandelli, J. Silvestre Albero, Biodiesel wastes: An abundant and promising source for the preparation of acidic catalysts for utilization in etherification reaction, Chem. Eng. J. 256 (2014) 468-474. [15] J.A. Melero, G. Vicente, G. Morales, M. Paniagua, J. Bustamante, Oxygenated compounds derived from glycerol for biodiesel formulation: Influence on EN 14214 quality parameters, Fuel 89 (2010) 2011-2018. [16] M. Ayoub, A.Z. Abdullah, Diglycerol synthesis via solvent-free selective glycerol etherification process over lithium-modified clay catalyst, Chem. Eng. J. 225 (2013) 784-789. [17] C. García-Sancho, R. Moreno-Tost, J.M. Mérida-Robles, J. Santamaría-González, A. Jiménez-López, P.M. Torres, Etherification of glycerol to polyglycerols over MgAl mixed oxides, Catal. Today 167 (2011) 84-90. [18] Z. Gholami, A.Z. Abdullah, K.-T. Lee, Glycerol etherification to polyglycerols using Ca1+xAl1−x LaxO3 composite catalysts in a solventless medium, J. Taiwan Inst. Chem. E. 44 (2013) 117-122. [19] M. Richter, Y. Krisnandi, R. Eckelt, A. Martin, Homogeneously catalyzed batch reactor glycerol etherification by CsHCO3, Catal Commun. 9 (2008) 2112-2116. [20] E. Pérez-Barrado, M.C. Pujol, M. Aguiló, J. Llorca, Y. Cesteros, F. Díaz, J. Pallarès, L.F. Marsal, P. Salagre, Influence of acid–base properties of calcined MgAl and CaAl layered double hydroxides on the catalytic glycerol etherification to short-chain polyglycerols, Chem. Eng. J. 264 (2015) 547-556. [21] M.V. Sivaiah, S. Robles-Manuel, S. Valange, J. Barrault, Recent developments in acid and base-catalyzed etherification of glycerol to polyglycerols, Catal. Today 198 (2012) 305-313. [22] A. Corma, S. Iborra, A. Velty, Chemical Routes for the Transformation of Biomass into Chemicals, Chem. Rev. 107 (2007) 2411-2502. [23] J. Barrault, J. Clacens, Y. Pouilloux, Selective oligomerization of glycerol over mesoporous catalysts, Topics in Catal. 27 (2004) 137-142. [24] A.M. Ruppert, J.D. Meeldijk, B.W. Kuipers, B.H. Erné, B.M. Weckhuysen, Glycerol Etherification over Highly Active CaO‐Based Materials: New Mechanistic Aspects and Related Colloidal Particle Formation, Chem. - A European J. 14 (2008) 2016-2024. [25] Y.K. Krisnandi, R. Eckelt, M. Schneider, A. Martin, M. Richter, Glycerol Upgrading over Zeolites by Batch‐Reactor Liquid‐Phase Oligomerization: Heterogeneous versus Homogeneous Reaction, ChemSusChem. 1 (2008) 835-844. [26] I. Kamal, D. Stéphane, F. Emmanuel, G. Christine, L. Alain, G. Philippe, Microwave-assisted synthesis of polyglycerol from glycerol carbonate, J. Appl. Polym. Sci. 120 (2011) 2354–2360. [27] A. Stankiewicz, Microwave Equipment for Polymerization Reactions and Polymer Processing, European Roadmap of Process Intensification, http://www.efce.info/efce_media/Appendix+1+_+Process+Intensification+Technologies+Description+and+Rev iew_tcm24_258504-p-868.pdf (accessed 5th November 2014). [28] JECFA, Polyglycerol esters of fatty acid 1989, http://www.fao.org/ag/agn/jecfaadditives/specs/Monograph1/Additive-317.pdf (accessed 1st August 2014) [29] JECFA, polyglycerol esters of interesterified ricinoleic acid. 1992 ed.; FNP 52 1973. http://www.fao.org/ag/agn/jecfa-additives/specs/Monograph1/Additive-318.pdf (accessed 1st August 2014).

[30] S. Salehpour, M.A. Dubé, Towards the Sustainable Production of Higher Molecular Weight Polyglycerol, Macromol. Chem. Phys. 212 (2011) 1284-1293. [31] A. Martin, M. Richter, Oligomerization of glycerol a critical review, European J. Lipid Sci. Tech. 113 (2011) 100-117. [32] R. Carlson, J.E. Carlson, Data Handling in Science and Technology: Design and optimization in organic synthesis, Third ed., Elsevier, 2005. [33] C. Chang, P. Cen, X. Ma, Levulinic acid production from wheat straw, Bioresource Technol. 98 (2007) 1448-1453. [34] W.N.N. Wan Omar, N.A. Saidina Amin, Optimization of heterogeneous biodiesel production from waste cooking palm oil via response surface methodology, Biomass Bioenerg. 35 (2011) 1329-1338. [35] J.M. Clacens, Y. Pouilloux, J. Barrault, C. Linares, M. Goldwasser, Mesoporous basic catalysts: comparison with alkaline exchange zeolites (basicity and porosity). Application to the selective etherification of glycerol to polyglycerols, Stud. Surf. Sci. Catal. 118 (1998) 895-902. [36] M. Ayoub, M.S. Khayoon, A.Z. Abdullah, Synthesis of oxygenated fuel additives via the solventless etherification of glycerol, Bioresource Technol. 112 (2012) 308-312. [37] S. Cassel, C. Debaig, T. Benvegnu, P. Chaimbault, M. Lafosse, D. Plusquellec, P. Rollin, Original synthesis of linear, branched and cyclic oligoglycerol standards, European J. Org. Chem. 2001 (2001) 875-896.

List of all symbols and abbreviations JECFA − the Joint FAO/WHO Expert Committee on Food Additives DG − Diglycerols TG − Triglycerols TtG − Tetraglycerols RSM − Response surface methodology 1

H-NMR − 1H Nuclear magnetic resonance spectroscopy

13

C–NMR − 13 C Nuclear magnetic resonance spectroscopy

ESI-HRMS − Electrospray ionization of high resolution mass spectrometry HPLC − High performance liquid chromatography PG − Pentaglycerols c-DG − Cyclic diglycerols X − Glycerol conversion YDG+TG+TtG − Combined yields of diglycerols, triglycerols and tetraglycerols Sc-DG − Selectivity towards cyclic diglycerols SPG − Selectivity towards pentaglycerols Rf − Retention times Tr − Temperature [Cat.] − Catalyst concentration tr − Reaction time DF – Degree of freedom

SS – Sum of squares MS – Mean of square

Table 1 Experimental range and levels for the etherification variables.

Variables

Coded

Range and level

Reaction temperature (°C)

A

220 (-1) and 270 (+1)

Catalyst concentration (wt%)

B

1 (-1) and 3 (+1)

Reaction time (h)

C

0.5 (-1), 1.0 (-0.6), 1.5 (-0.2), 2.0 (+0.2), 2.5 (+0.6), and 3 (+1)

Table 2 Experimental design and responses of interest for the etherification of glycerol to polyglycerols. Factors

Exp no. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

A

B

C

X

220 220 220 220 220 220 220 220 220 220 220 220 270 270 270 270 270 270 270 270 270 270 270 270

1 1 1 1 1 1 3 3 3 3 3 3 1 1 1 1 1 1 3 3 3 3 3 3

0.5 1.0 1.5 2.0 2.5 3.0 0.5 1.0 1.5 2.0 2.5 3.0 0.5 1.0 1.5 2.0 2.5 3.0 0.5 1.0 1.5 2.0 2.5 3.0

10.40 13.95 20.13 19.68 26.07 26.03 26.79 31.79 48.01 51.66 48.83 52.38 75.99 85.26 91.40 96.03 93.53 96.12 74.02 92.83 96.89 96.20 85.05 87.18

Average responses of interest (%) (3 replications) YDG+TG+TtG SPG Sc-DG 10.40 13.93 20.13 19.70 25.60 24.63 26.00 29.90 43.57 46.47 47.20 44.47 56.53 59.13 52.40 66.83 59.27 47.33 58.17 70.13 65.17 53.30 64.97 57.77

0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 15.18 11.44 28.14 12.50 13.00 37.53 0.00 9.06 16.56 13.33 6.91 0.00

0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 6.42 7.18 9.62 12.94 16.05 8.33 5.21 6.50 7.54 12.63 28.48 35.60

Table 3 Analysis of variance (ANOVA) for the regression model equation and coefficients Source X model

SS 67829.70

DF 4

MS 16957.400

F 269.18

P

Source SS DF MS F 21549.40 5 4309.900 33.26 YDG+TG+TtG model Temperature 60341.20 1 60341.200 1303.28 0.000 Temperature 16110.13 1 16110.125 113.06 (°C) (°C) Catalyst 2347.10 1 2347.100 50.70 0.000 Catalyst 2857.68 1 2857.680 20.06 conc. conc. (%wt.) (%wt.) Time (h) 3390.90 5 678.200 14.65 0.000 Time (h) 899.79 5 179.958 1.26 AB 2788.50 1 2788.500 60.23 0.000 AB 1132.88 1 1132.880 7.95 AC 1278.74 5 255.748 1.79 Residual 8552.6 66 129.6 4220.7 67 63.0 Residual Total 72050.50 71 Total 30101.9 71 4641.40 3 1547.100 14.99 6 868.600 32.02 SPG model Sc-DG model 5211.70 Temperature 3348.07 1 3348.074 33.36 0.0000 Temperature 3061.37 1 3061.369 113.74 (°C) (°C) Catalyst 646.68 1 646.681 6.44 0.0141 Catalyst 156.84 1 156.844 5.83 conc. conc. (%wt.) (%wt.) AB 646.68 1 646.681 6.44 0.0141 Time (h) 825.94 5 165.188 6.14 Residual 7020.6 68 103.2 AB 156.84 1 156.844 5.83 Total 11662.0 71 AC 825.94 5 165.188 6.14 BC 521.55 5 104.310 3.88 1763.3 65 27.1 Residual Total 6975.0 71 DF: degree of freedom; SS: sum of squares; MS: mean of square; F: probability distribution; P: probability > F

-

P

0.0000 0.0000 0.2936 0.0067 0.1298 0.0000 0.0193

0.0001 0.0193 0.0001 0.0046 -

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

HIGHLIGHTS • • •

Process optimizations on the etherification of glycerol to polyglycerols are reported. Reaction temperature is the most significant variable for the optimized process. Microwave radiation shortens the reaction time for the etherification of glycerol to oligoglycerols.