A versatile method for obtaining new oxygenated fuel

3 downloads 0 Views 1MB Size Report
Jul 27, 1995 - flow properties, reduce particulate emission and gum formation, etc. (Trifoi et al., 2016). Solketal, obtained from glycerol condensation with.
Industrial Crops & Products 113 (2018) 288–297

Contents lists available at ScienceDirect

Industrial Crops & Products journal homepage: www.elsevier.com/locate/indcrop

A versatile method for obtaining new oxygenated fuel components from biomass

T



Emil Stepana, Cristina-Emanuela Enascutaa, , Elena-Emilia Oprescua,b, Elena Radua, Gabriel Vasilievicia, Adrian Radua, Rusandica Stoicaa, Sanda Veleaa, Alina Nicolescuc,d, Vasile Lavrice a

National Research & Development Institute for Chemistry and Petrochemistry ICECHIM, 202 Splaiul Independentei St., 060021, Bucharest, Romania Petroleum-Gas University of Ploiesti, 39 Bucharest Blv., 100680, Ploiesti, Romania c “Petru Poni” Institute of Macromolecular Chemistry, Romanian Academy, 41A Grigore Ghica Voda St., 700487, Iasi, Romania d “C. D. Nenitescu” Centre of Organic Chemistry, Romanian Academy, Splaiul Independentei 202B, 060023, Bucharest, Romania e University “POLITEHNICA” of Bucharest, Faculty of Applied Chemistry and Materials Science, 1-7 Polizu St., 011061, Bucharest, Romania b

A R T I C L E I N F O

A B S T R A C T

Keywords: Glycerol acetal ester Furfurylideneglycerol ester Ethyl levulinate glycerol ketal Reactive distillation Oxygenated fuel additive

Glycerol acetals/ketals and their esters, produced from renewable raw materials, are highly valuable compounds, being used as oxygenated fuel additives and ecological solvents. A new method for obtaining glycerol acetal/ketal esters, difficult to synthesize using classical techniques, was developed. This method is based on the reversible reaction of 1,2-O-isopropylidene-glycerol esters (IPGEs) obtained from glycerin, by-product from biodiesel production, with low volatility aldehydes/ketones (e.g. furfural and ethyl levulinate, both obtained from lignocellulosic biomass), in the presence of a heterogeneous acid catalyst. To circumvent reaching equilibrium, the continuous removal of acetone (Ac) from the reaction mixture was done, the former being reused in the synthesis of IPGEs. The method viability was assessed by synthesis and characterization of eight new compounds from two classes: furfurylideneglycerol esters (FGEs) and ethyl levulinate glycerol ketal esters (ELGKEs). A detailed kinetic study was done throughout an experimental program, first developed in Asia 330, a modular range flow chemistry system, then in a continuous-flow process at micropilot scale.

1. Introduction Biomass is an attractive renewable resource seen as a sustainable alternative to produce liquid transportation fuels, diminishing thus fossil fuels need. The production of ecological fuel components derived from renewable sources increased significantly in recent years due to the harsh environmental policy promoted in many developed countries. Accordingly, the production of biodiesel and oleochemicals is booming, more than two million tons of glycerol consistently reach the market every year, (Ciriminna et al., 2014). In order to overcome this issue of large glycerol quantities, the researchers’ attention has been focused on upgrading glycerol to more valuable chemicals, i.e. glycerol acetals/ketals. Different types of glycerol acetals and ketals have been identified to have particular qualities as fuel additives/components: improve the octane number and cold flow properties, reduce particulate emission and gum formation, etc. (Trifoi et al., 2016). Solketal, obtained from glycerol condensation with acetone, has a low solubility in diesel fuel, disadvantage common to



Corresponding author. E-mail address: [email protected] (C.-E. Enascuta).

https://doi.org/10.1016/j.indcrop.2018.01.059 Received 9 July 2017; Received in revised form 10 January 2018; Accepted 22 January 2018 0926-6690/ © 2018 Elsevier B.V. All rights reserved.

many glycerol acetals/ketals coming from low molecular weight aldehydes/ketones. This drawback was removed by transesterification with monocarboxylic acid methyl esters (Stepan et al., 2017a,b), the resulted IPGEs being miscible with diesel fuel and biodiesel. The reaction between (2,2-dimethyl-1,3-dioxolan-4-yl) methanol (solketal) and acetic anhydride in triethylamine solution gives a mixture of (2,2-dimethyl1,3-dioxolan-4-yl) methyl acetate and triacetin. The introduction of an acetyl group in the free OH of solketal is an effective solution to get a viscosity improvement that also meets the requirements of EN 14214 Standard, for flash point and oxidation stability (Garcia et al., 2008). Thus, the potential for a greener, solvent-free, heterogeneous catalyzed process of solketal and glycerol formal acetylation was proved. High conversions (72–95%) and selectivities (86–99%) to the desired products have resulted from using acetic anhydride as the acetylation reagent and an equimolar ratio (Dodson et al., 2014). Another group of compounds which has received heavy attention in recent years are the biomass derived furans, furfural and hydroxymethylfurfural. They are prepared through the dehydration of sugar

Industrial Crops & Products 113 (2018) 288–297

E. Stepan et al.

ketal ester and acetone, thus avoiding side reactions of transesterifications, due to alcohol, as in case of classical transacetalation. Thermodynamic limitations were overcome by reactive distillation, removing continuously and very fast the acetone from the liquid phase. The viability of the method was assessed by synthesizing eight new compounds belonging to two classes: FGEs and ELGKEs. These new compounds can be used as oxygenated fuel additives and ecological solvents. A detailed two steps kinetic study was done throughout an experimental program. The first step was the study of the process in Asia 330, a modular range flow chemistry system. The second step was to render the process continuous, at micropilot scale. Tungstophosphoric acid catalyst supported on γ-alumina extrudates was prepared, characterized and used. The mathematical model of the reactive distillation process, based upon the mass balances of the implied chemical species together with the intrinsic kinetics of the catalytic reactions, was built. Kinetic constants of the mathematical model and thermodynamic equilibrium constant were found through regression analysis.

residues and are being regarded as suitable platform chemicals for replacing numerous petroleum based products (Wegenhart et al., 2012). US Department of Energy’s report included glycerol and levulinic acid in “Top 12” most important bio-based chemicals in the world (Werpy and Petersen, 2004), together with furfural (Bozell and Petersen, 2010), the latter based on advances in technology. Acetalization of glycerol with furfural gives a mixture of the cis and trans isomers of five- and six-membered furan 1,3-dioxacyclanes: 2furyl-4-hydroxymethyl-1,3-dioxolane and 5-hydroxy-2-furyl-1,3-dioxane (Stepan et al., 2017a,b). The cis- and trans-5-hydroxy-2-furyl-1,3dioxanes were isolated by column chromatography, and their stereochemical structures were established by IR and NMR spectroscopy (Gromachevskaya et al., 2004). However, these products are poorly soluble in hydrocarbons, therefore they are not used as oxygenated fuel component. Ethyl levulinate glycerol ketal was obtained with a high yield and selectivity at 80–110 °C, and 30 mmHg, using sulfuric acid as catalyst (Selifonov et al., 2013). Good dissolving power makes alkyl levulinate acetals indispensable solvents in various detergents, lotions, and paints formulations (Maximov et al., 2015). Blocking of free hydroxyl group present in the structures of furfural acetal of glycerol and of ethyl levulinate glycerol ketal by esterification would significantly increase the solubility in hydrocarbons, enabling their use as oxygenated fuel component. One of the difficulties in the preparation of acetals of levulinic acid from polyols and levulinic acid (or its esters), in the presence of acid catalysts, is that the latter trigger not only the acetalization reaction of the keto group, but also the esterification reaction of the carboxyl group (or transesterification of the ester group) (Maximov et al., 2015). Addition of glycerol ketal propionate improves diesel blend properties like pour point, viscosity and flash point, the best performance being achieved with 10 wt.% 2-methyl-2-ethyl-1,3-dioxolane-4-methylpropionate (Oprescu et al., 2013). Butan-2-one glycerol ketal was synthesized using a solid superacid catalyst, SO42− −/TiO2-La2O3. Free hydroxyl groups have been blocked by transesterification with methyl hexanoate. (2-Ethyl-2-methyl-1,3-dioxolan-4-yl) methyl hexanoate diesel blend slightly improves engine performance and reduces HC, CO and smoke emission by 26%, 18.9% and 19.2%, respectively (Oprescu et al., 2014). Fatty acid formal glycerol ester (FAGE) was produced by a transketalization-transesterification combined process from crude glycerol and/or waste oils to be blended with diesel fuel in concentrations up to 20% (Lapuerta et al., 2015a). Tested under the New European Driving Cycle, diesel automotive engines showed a substantial reduction (around 20–40%, with the highest reduction at warm engine temperature) of soot and particulate matter when fueled with FAGE blends (Lapuerta et al., 2015b). Most of the studies related to glycerol processing showed that it is still difficult to obtain good selectivity in the desired products at high glycerol conversion, due to the wide hydroxylic functionalization of the triol glycerol molecule of similar reactivity, the unknown reaction conditions or the lack of optimum catalysts (Bagheri et al., 2015). According to the authors’ best knowledge, there are no information in literature about the characteristics of FGEs and ELGKEs, nor about new methods of their synthesis, based on the reaction of glycerol acetal/ketal ester with aldehyde or ketone, resulting in a new glycerol acetal/ketal ester and another aldehyde/ketone. The purpose of this paper is to report a versatile method for obtaining new glycerol acetal/ketal esters, difficult to synthesize using classical methods, by a reversible reaction of IPGEs, obtained from glycerin, by-product from biodiesel manufacture, with low volatility aldehydes/ketones, (i.e. furfural and ethyl levulinate) obtained from lignocellulosic biomass, using a heterogeneous acid catalyst. Acid-catalyzed acetal or ketal exchange was conducted by adapting the classical transacetalation, i.e. reaction of acetal/ketal with alcohol, resulting a new acetal/ketal and a new alcohol. Our method is based on the reaction of ketal ester with aldehyde or ketone, resulting a new acetal/

2. Experimental 2.1. Chemicals and materials Tungstophosphoric acid hydrate was purchased from Merck, ethyl levulinate min. 98%, DL-1,2-isopropylideneglycerol min. 98%, methyl propionate 99%, methyl butyrate 99%, methyl valerate 99%, methyl hexanoate 99%, methyl octanoate 99% were supplied by SIGMAALDRICH, and γ-alumina extrudates from Sasol. The sunflower husks have been obtained from a high-oleic variety of sunflower from Dow AgroSciences breeding company. The crops were cultivated in South of Romania Plane, in Călărași County. The seed were processed to obtain sunflower oil by a classical technology; the by-product, sunflower husks, a waste, have been used as raw material to obtain furfural. The sunflower seed husks had the following composition: 28.62% hemicellulose, 23.78% lignin, 34.71% cellulose, 4.58% lipids, 3.72% protein, 2,71% water, 1.88% ash. 2.2. Furfural acquisition and characterization Furfural was obtained by hydrolysis of hemicellulosic fractions from sunflower seed husks. In a typical reaction, 50 g of sunflower seed husks (ground to about 0.5 mm diameter), 250 g of 20% sulfuric acid and 250 g of toluene were placed in a high-pressure reactor BR-1000. The mixture was stirred and maintained at 185 °C for 90 min, after that, the solid residue was removed by filtration. The toluene solution was separated, washed with water and then distilled. Crude furfural was purified by vacuum distillation (70 °C/20 mmHg). Furfural yield was 81% based on hemicellulose content of sunflower seed husks. Quantitative analysis of furfural was performed by using GC–MS/ MS TRIPLE QUAD (Agilent 7890 A). Water content in furfural was determined using Karl Fischer specific reagents for aldehydes: 188002 Aquqstat – CombiTitrant 2, as titrant and 188008 Aquastar – CombiSolvent, as solvent. 2.3. Syntheses and characterizations of IPGEs The IPGEs synthesis method was described into a previous paper by Stepan et al. (2017a,b). 1,2-O-isopropylideneglycerol was treated with stoichiometric amount of monocarboxylic aliphatic acid (C3, C4, C5, C6, C8) methyl esters and potassium alkoxide at 85–135 °C, for 180–420 min. Methanol resulted as by product was collected. The catalyst was removed by filtration and raw IPGEs were purified by vacuum distillation. The following IPGEs were synthesized and characterized by GC–MS: 1,2-O-isopropylideneglycerol propanoate (IPPr), 1,2-O-isopropylideneglycerol butanoate (IPBu), 1,2-O-isopropylideneglycerol pentanoate (IPPe), 1,2-O-isopropylideneglycerol hexanoate (IPHe), 1,2289

Industrial Crops & Products 113 (2018) 288–297

E. Stepan et al.

from the reactor, withdrawn at specific intervals of the total reaction time of 360 min.

O-isopropylideneglycerol octanoate (IPOc) and 1,2-O-isopropylideneglycerol 2-ethylhexanoate (IPEHe). Quantitative analysis of IPGEs was performed by using GC Clarus 500 (Perkin Elmer) with DB-WAX capillary column (30 m length, 0.25 mm internal diameter, 0.25 μm film thickness) equipped with flame ionization detector. The response factors (RF) were calculated using synthetic mixtures, prepared at the expected level of the reagents, contaminants and IPGEs, from standards obtained by us.

2.7. Synthesis of ELGKEs using a flow chemistry system Syntheses of ELGKEs were conducted as reported in 2.6, putting in contact IPGEs (IPGPr and IPGHe) and ethyl levulinate (EL) in molar ratio of 1:1, at reaction temperatures in the range of 80–150 °C (see Fig. 2 for the generic chemical reaction).

2.4. Preparation of the catalyst 2.8. Synthesis of FGEs using a micropilot plant

Several heterogeneous catalysts (e.g. strongly acidic polymeric catalyst Purolite CT 275, SO42−/TiO2-La2O3 strongly acidic catalyst) were tested in some preliminary studies, but good results were obtained only when tungstophosphoric acid catalyst supported on γ-alumina extrudates was used. This catalyst was prepared by pore-filling method, following Pizzio et al. (1998). 4 mL water solution containing 120 g/L tungstophosphoric acid was added on 1 g support, to obtain 40 wt.% H3PW12O40/γ-Al2O3 after a contact time of 48 h.

The discontinuous process was implemented as a continuous process at micropilot scale, its simplified flow diagram being presented in Fig. 3. IPEs and Fu in molar ratio of 1:2 were introduced in the stirring vessel (SV), to prepare the solution of reactants. This solution was fed with the metering pump (P1) in the fixed-bed reactor (R – an adjusted Parr continuous flow tubular reactor, electrically heated), previously loaded with H3PW12O40/γ-Al2O3 extrudates, with a flow corresponding to the desired residence time of the mixture in the reactor. The reaction temperature was accurately maintained at the desired value of 12 °C. Ac resulted by reactive distillation, escaped from reactor (R) as vapors which were condensed in (C) and then continuously removed from the system, by placing a vacuum pump (VP) at the outlet of the condenser (C). In the reactor (R), as well, the vertical movement of the Ac vapor bubbles through the interstitial volume of the fixed-bed ensures such a hydrodynamic that the mass transfer resistances could be disregarded. The liquid phase, containing Fu and the reaction products, enters in the storage vessel (V), from which it fed the short path distiller (KDL-5 from UIC GmbH Germany) – Fu was continuously removed at 60 °C and 15 mbar. To prevent secondary reactions (i.e. oxidation of Fu and products), the plant was purged with argon.

2.5. Catalyst characterization The acid catalyst was characterized by X-ray diffraction. The XRD patterns were collected using a Rigaku SmartLab equipment, operating at 45 kV and 200 mA, using Cu Kα radiation in parallel beam configuration (2θ/θ scan mode), from 3 to 80 2θ°. The source of the X-rays was a Cu tube (λ = 0.15418 nm) operating at 40 kV and 30 mA. Counts were collected from 10° to 80° with a step size of 0.02 and a rate of 5°/ min. Thermogravimetric analysis (TGA) of the catalysts was done with a TGA/SDTA 851Mettler Toledo in the temperature range 20–700 °C, at a heating rate of 10 °C/min, under nitrogen flow. The FT-IR spectra were recorded at room temperature with a Spectrometer FT-IR Tensor 27-Bruker in the range 4000–400 cm−1 by using KBr pellets. The specific surface areas were measured by BET analysis on a Quantachrome NovaWin instrument at liquid nitrogen temperature. Prior to the analysis, samples were degassed at 150 °C.

2.9. Characterization of products Variation in time of the composition of reaction mixture yielding FGEs and ELGKEs, was studied using GC Clarus 500 (Perkin Elmer) with DB-WAX capillary column (30 m length, 0.25 mm internal diameter, 0.25 μm film thickness) equipped with flame ionization detector. The response factors (RF) were calculated using synthetic mixtures, typically prepared at the theoretically expected level of the reagents and new reaction products. Confirmation of the IPGEs, FGEs and ELGKEs structures was performed by gas chromatography coupled with mass spectrometry (GC/ MS), using a GC–MS/MS TRIPLE QUAD (Agilent 7890 A) with DB-5MS capillary column (30 m length, 0.25 mm internal diameter, 0.25 μm film thickness) and helium as carrier gas at 1 mL min−1 constant flow. The NMR spectra of FGEs and ELGKEs have been recorded on a Bruker Avance DRX 400 instrument operating at 400.1 and 100.6 MHz for 1H and 13C. Chemical shifts are reported in δ units (ppm) and were referenced to internal TMS for 1H chemical shifts and to the internal deuterated solvent for 13C chemical shifts (CDCl3 referenced at 77.0 ppm). Unambiguous 1D NMR signal assignments were made based on 2D NMR homo- and hetero-correlation. H,H-COSY, H,C-HSQC and H,C-HMBC experiments were recorded using standard pulse sequences in the version with z-gradients, as delivered by Bruker with TopSpin 1.3

2.6. Syntheses of FGEs using a flow chemistry system Syntheses of FGEs were performed in Asia 330 (Syrris Company), a modular range flow chemistry system. Solid Phase Reactor Size 3 (5.6 mL) was loaded with H3PW12O40/γ-Al2O3 extrudates and accurately heated in the range of 80–150 °C. Solutions consisting of IPEs and Fu in molar ratio of 1:1 and 1:2 (IPEs and Fu are miscible) were introduced in the reactor, then heated at 80, 100, 120 and 150 °C. The following IPEs were used in experiments: IPPr, IPBu, IPPe, IPHe, IPOc and IPEHe, which gave the corresponding FGEs: furfurylideneglycerol propanoate (FGPr), furfurylideneglycerol butanoate (FGBu), furfurylideneglycerol pentanoate (FGPe), furfurylideneglycerol hexanoate (FGHe), furfurylideneglycerol octanoate (FGOc), furfurylideneglycerol 2-ethylhexanoate (FGEh) – see Fig. 1 for the generic chemical reaction. Ac resulted in the reaction was continuously removed by reactive distillation at the top of the reactor. Ac boiling ensured such hydrodynamic conditions that the hypothesis of no mass transfer resistances holds true. The catalytic conversions and FGEs yields were determined periodically by gas chromatography (GC) of sampling small aliquots

Fig. 1. Reaction pathway of FGEs (R = Pr, Bu, Pe, He, Oc, Eh).

290

Industrial Crops & Products 113 (2018) 288–297

E. Stepan et al.

Fig. 2. Reaction pathway of ELGKEs (R = Pr, He).

Fig. 3. The simplified flow diagram of the continuous process to obtain FGEs.

Al2O3, as reported also by Kim et al. (2006). The XRD pattern of the impregnated tungstophosphoric acid on γAl2O3 (Fig. 5), showed characteristic XRD pattern of γ-Al2O3 support, indicating no crystalline structure, due to highly dispersed species of tungstophosphoric acid on the support (Karim et al., 2011). This assertion is supported by the EDX-TEM analysis data (Fig. 6) which show a homogeneous distribution of W and P species on γ-Al2O3. The quantitative determination of interest elements by XRF technique suggests that H3PW12O40/γ-Al2O3 catalyst contains 50.1% Al, 49.35% W and 0.55% P which represents almost 40 wt.% of H3PW12O40 promoted on γ-Al2O3 support. The nitrogen adsorption–desorption isotherms of γ-Al2O3 and H3PW12O40/γ-Al2O3 show type IV adsorption isotherm (Fig. 7) with a hysteresis loop which can be classified as close to H3 hysteresis (Khalil, 2007). This is consistent with high specific surface-area and mesoporous nature of the samples. The surface area, pore volumes, pore diameter and acid site concentration for γ-Al2O3 and H3PW12O40/γ-Al2O3 are presented in Table 1. The values for surface area, pore volume and pore diameter of H3PW12O40/γ-Al2O3 decreased compared with γ-Al2O3, possibly due to the entering of H3PW12O40 species into the mesopores of the support. As expected, the total acidity increases for catalyst up to 1.712 mEq diethyl amine/g with respect to support due to the presence of impregnated tungstophosphoric acid Patel and Brahmkhatri (2013). The catalyst distribution of acid sites strength calculated using thermal desorption of diethyl amine method is the following: 11.73% strong acid sites, 39.75% medium acid sites and 48.52% weak acid sites, respectively. The TGA analysis of γ-Al2O3 support exhibit a gradual weight loss, which is associated to the release of physically adsorbed water. The TGA profiles for H3PW12O40 and H3PW12O40/γ-Al2O3 (Fig. 8) present three distinct weight loss patterns in the following temperature ranges of 25–80 °C, 80–210 °C and 210–480 °C, respectively, which correspond

PL10 spectrometer control and processing software. Differential scanning calorimetry (DSC) method was used to measure the boiling temperature of FGEs and ELGKEs. The measurements were carried out with DSC 823 system from Mettler Toledo with nitrogen flow of 60 mL min−1. From the resulting thermograms, the measured boiling temperatures were converted to normal boiling points (BP) by Sidney Young equation (103 – OECD Guidelines for the Testing of Chemicals, 1995). DSC method was used to measure the pour points of FGEs and ELGKEs. The measurements were carried out with DSC Q2000 from TA Instruments, with helium flow of 30 mL min−1, in cooling and in heating mode. 3. Results and discussion 3.1. Catalyst characterization The FT-IR spectra of tungstophosphoric acid (H3PW12O40), γ-alumina (γ-Al2O3) and impregnated tungstophosphoric acid on γ-Al2O3 (H3PW12O40/γ-Al2O3) are presented in Fig. 4. The FT-IR spectra of γ-Al2O3 support show two shoulders at 750 and 565 cm−1 assigned to vibration of aluminum ions in octahedral and tetrahedral environments, as reported also by Karim et al. (2011). As can be seen in Fig. 4 the tungstophosphoric acid exhibits four bands at 1081, 987, 893 and 798 cm−1, respectively, corresponding to the characteristic vibrations of Keggin structure (PeO symmetric stretch, W]O asymmetric stretch, WeOeW inter- and intraoctahedral stretches, respectively), however these peaks decrease in intensity for impregnated tungstophosphoric acid on γ-Al2O3 (H3PW12O40/γ-Al2O3), being masked by the support bands (Pizzio et al., 2003). Two shoulders peaks at 878 and 809 cm−1, corresponding to W]O and WeOeW stretching vibrations, confirm the presence of these groups in the prepared catalyst. The slight shift to the smaller wavelength of PeO peak indicates a strong interaction between tungstophosphoric acid and γ291

Industrial Crops & Products 113 (2018) 288–297

E. Stepan et al.

Fig. 4. FT-IR spectra of H3PW12O40, γ-Al2O3 and H3PW12O40/γ-Al2O3.

Fig. 7. Nitrogen adsorption/desorption isotherms for γ-Al2O3 and H3PW12O40/γ-Al2O3. Fig. 5. The XRD patterns of H3PW12O40, γ-Al2O3 and H3PW12O40/γ-Al2O3.

Table 1 The specific BET surface area, pore diameter, pore volume and acid sites amount for γAl2O3 and H3PW12O40/γ-Al2O3.

to loss of non-coordinated water, loss of water of crystallization of Keggin type anion, and loss of water of condensation, respectively (Kamalakar et al., 2006). 3.2. Effects of reaction conditions in synthesis of FGEs in the flow chemistry system The influence of reaction parameters upon synthesis of the FGEs was studied, namely for FGPr, FGBu, FGPe, FGHe, FGOc and FGEh.

Sample

Surface area (m2/g)

Pore volume (cm3/g)

Pore diameter (nm)

Acid sites conc. (mEq diethyl amine/g)

γ-Al2O3 H3PW12O40/γAl2O3

262.7 219.7

0.752 0.493

11.46 8.99

0.766 1.712

content in furfural less than 0.2% did not affect either the yield or product characteristics. At higher water content, the hydrolysis of the ketals in the presence of the acid catalyst becomes important. We only used fresh distilled furfural prepared by us. In the case of long-term storage, furfural degrades, thus its redistillation becomes necessary.

3.2.1. Influence of reactants characteristics The freshly distilled furfural has the following characteristics: assay (GC–MS/MS): 99 ± 0.3%; water (K.F.) = 0.18 ± 0.2%. The water

Fig. 6. EDX mapping and P and W element distribution from H3PW12O40/γ-Al2O3.

292

Industrial Crops & Products 113 (2018) 288–297

E. Stepan et al.

Fig. 8. TGA profiles of γ-Al2O3, H3PW12O40, H3PW12O40/γ-Al2O3.

Regarding IPGEs (assay (GC-FID) = 98.2 ± 0.9%), the impurities in low concentrations did not affect the synthesis of FGEs.

3.2.2. Influence of temperature, initial molar ratio of reactants and reaction time To achieve high yields of FGEs, the reaction equilibrium was displaced by continuous removal of acetone by reactive distillation. The temperature and molar ratio of Fu to IPGEs were varied according to Section 2.4, the results being reported in Fig. 9 (stoichiometric feed ratio) and Fig. 10 (2:1 feed ratio). As shown in Fig. 9, at stoichiometric feed ratio and 80 °C, FGPr yield increases continuously, faster for the first 60 min when reaching pseudo-equilibrium, and then slower till the end of the process (see Table 2 for specific process related values). The same pattern is valid for the working temperature of 120 °C, except the shorter time for the faster continuous increase of FGPr yield till reaching pseudo-equilibrium, 30 min. Here too, the continuous increase slows down for the rest of the reaction time, till the FGPr yield reaches a plateau. At 150 °C, there is still a sharp increase of FGPr yield, in only 15 min the latter reaching pseudo-equilibrium. After this moment, there is a continuous decrease of FGPr yield till the end of the reaction time.

Fig. 10. Effect of reaction temperature on FGPr yield, at Fu/IPGPr molar ratio of 2:1.

Table 2 The FGPr yields at pseudo-equilibrium and after process completion. Fu/IPGPr molar ratio 2:1

Temperature, °C

Pseudo-equilibrium Time, min

FGPr yield, %

Final (360 min) FGPr yield, %

1

80 120 150

60 30 15

51 61.1 60.2

61.8 70 51.1

2

100 120 150

20 15 10

63.1 70.4 80.3

70 86.7 78

The FGPr synthesis process is endothermic, therefore as the working temperature increases, so does the equilibrium conversion, raising the current yield, till a pseudo-equilibrium state is reached. The process is far more complicate, due to at least two other steps concurrent to synthesis, namely the decomposition of the product, faster at higher temperatures, and the withdrawal of the Ac by reactive distillation, which promotes higher conversions by shifting the thermodynamic equilibrium to products, thus a possible increase in the product yield.

Fig. 9. Effect of reaction temperature on FGPr yield, at stoichiometric feed ratio.

293

Industrial Crops & Products 113 (2018) 288–297

E. Stepan et al.

At the beginning of the synthesis process, the latter steps (decomposition and withdrawal) are less important, since their rates are lower than the main reaction rate. Still, their influence is more pronounced at lower temperature that at higher ones, dictating the pseudo-equilibrium reached at the end of the fast increasing FGPr yield period. After that, a competition establishes between the decomposition, which tends to decrease the yield, and the withdrawal of Ac, having an opposite effect. At 80 °C, FGPr yield raises faster than at 120 °C after reaching pseudoequilibrium (see Fig. 9), the former increasing continuously, while the latter flattening. In the latter case, the decomposition and the withdrawal rates are as such that their effects cancel on each other on the last half of the reaction time, while in the former case, the withdrawal is faster than the decomposition, thus insuring a steady increase of the FGPr yield. At 150 °C, the rate controlling step becomes decomposition, after reaching the pseudo-equilibrium, having higher rates than the withdrawal (see Fig. 9). Although more FGPr gets produced due to chemical equilibrium displacement towards products by reactive distillation, even more is decomposed due to the increase of the decomposition rate owing to the temperature increase. Therefore, the FGPr yield starts decreasing continuously. At Fu/IPGPr molar ratio of 2:1, FGPr yields are higher than at stoichiometric ratio, as shown in Fig. 10 and Table 2. Still, the patterns of the process are the same, except for the FGPr yield at 120 °C, which has a slightly decreasing profile for the last half of the reaction time, instead of reaching a plateau (see the corresponding profile in Fig. 9). The higher synthesis ensures a higher FGPr concentration after a shorter time from the beginning of the reaction, thus decomposition can act longer, with detrimental effects upon the FGPr yield, which starts declining towards the end of the process. The benefits of using an excess concentration of Fu are hindered by the necessity of its separating and recycling, although it can be easily removed from the crude FGPr by distillation in vacuo. In one experiment, FGPr synthesis was performed at 25 °C, and Fu/ IPGPr molar ratio of 2:1. After 24 h, FGPr yield was too small to be of interest. Figs. 1 and 12 show the reduction of the yield consecutive to the temperature increase, due to the higher rates of the degradation reactions.

Fig. 12. Comparison of the evolution of the yields to following FGEs: FGPr, FGBu, FGPe, FGOc, at 150 °C and Fu/IPGPr molar ratio of 2:1.

of furfural over the Brønsted acid centers. Then, nucleophilic attack occurs via oxygen of IPGEs, followed by ring opening and re-cyclization via intramolecular displacement of the acetone. 3.4. Intrinsic kinetics of the process To develop an intrinsic kinetic model, some notations should be used, to ease following the procedure. The generic IPEs are denoted with A, Fu is denoted with B, FGEs with R and Ac with S, therefore the main reaction happening at the active sites could be abstracted as: kf

A + B ⇄ R + S↑

(1)

kr

Here, the upper arrow means that Ac is leaving the liquid phase as vapors, displacing the thermodynamic equilibrium toward products. The intrinsic reaction rate for (1) could be derived using the LHHW paradigm. According to the reaction mechanism, the elementary steps are (X stands for the free active site):

3.3. Reaction mechanism

k1

B + X ⇄ BX

(2)

k−1

The mechanism of transacetalation of Fu with IPGEs is presented in Fig. 13. The proposed scheme is in accordance with Moraes et al. (2001). The first step consists in protonation of carbonyl function group

k2

BX + A ⇄ ABX

(3)

k−2 k3

ABX ⇄ X + R + S ↑

(4)

k−3

The best candidate to be rate limiting is the last elementary step, which includes the simultaneous transformation in products and desorption. Applying the LHHW paradigm (detailed in the Appendix A Supplementary data), the intrinsic reaction rate, at the catalytic sites, is:

(

k⋅ CA⋅CB − r=

CR ⋅ CS KC

)

1 + K C1⋅CB⋅(1 + K C 2⋅CA)

(5)

where the kinetic parameters are: k, KC1 and KC2, while KC, the equilibrium constant, is a thermodynamic parameter. The temperature dependency of the constant k is of Arrhenius type, while KC1, KC2 and KC are considered independent of temperature, to simplify the regression analysis. Since the porosity of the catalyst is practically absent and the upward vapor movement through the liquid generates enough turbulence to consider that there is no external mass transfer resistance, the bulk liquid phase concentrations could be safely used in computing the reaction rate (1). In the liquid phase, there could happen, also, two important side

Fig. 11. Comparison of the evolution of the yields to following FGEs: FGPr, FGBu, FGPe, FGPHe, FGOc, FGEh, at 120 °C and Fu/IPGPr molar ratio of 2:1.

294

Industrial Crops & Products 113 (2018) 288–297

E. Stepan et al.

Fig. 13. Reaction mechanism of Fu transacetalation with IPGEs for FGEs obtainment.

reactions, which degrade one of the reactants and the desired product: kS1

A ⎯⎯→ P + S ↑, kS2

R ⎯⎯⎯→ P + B,

rS1 = kS1⋅CA

(6)

rS 2 = kS 2⋅CR

(7)

Here, P is a generic side and stable product, which lumps all the unwanted byproducts, while kS1 and kS2 are, also, temperature dependent, described by an Arrhenius relationship. During the catalytic process, Ac is continuously removed from the liquid phase as vapors. The kinetic of this process may be approximated as first order with respect to Ac:

rrem = krem⋅CS

(8)

where krem is the first order rate constant of the removal process, which is the last parameter of the model. Fig. 14. The effect of repeated use of catalyst.

3.5. The mathematical model 3.8. Effects of reaction conditions in synthesis of ELGKEs in flow chemistry systems

The mathematical model of the discontinuous process, based upon the intrinsic kinetics and the mass balances for the involved species, considering the volume variation due to ketone removal and chemical composition change, is presented in Appendix A Supplementary data.

The effect of temperature in the range of 120–150 °C, and of reaction time in synthesis of the following ELGKEs: ELGKPr and ELGKHe was studied, using the stoichiometric feed ratio. High conversions of ELGKEs were achieved, displacing the equilibrium by continuous removal of acetone, by reactive distillation. An excess of ELGKEs was disregarded; ELGKEs having high boiling temperatures, their removal through distillation is difficult. Influence of reaction temperature on ELGKPr yield is shown in Fig. 15. At 120 °C, ELGKPr yield continuously increases, first rapidly in 60 min reaching 58.8%, and then slowly towards 72.3%, during the next 290 min. At 140 °C ELGKPr yield increased rapidly in 30 min reaching 66.4%, then remained almost constant. At 150 °C, the yield profile presents a slower decrease, after reaching 67.8%, due to the side reactions (Fig. 16). The influence of reaction temperature on ELGKHe yield follows the same pattern as in the case of ELGKPr, but with somehow lower yields. The maximum yield of 61.1% was reached after 300 min at 120 °C, while a value close to 58.6% was reached after 25 min at 150 °C.

3.6. Regression analysis The kinetic constants of the mathematical model, k 0, E R , kS10, ES1 R , kS 20, ES2 R , K C1, K C3 and krem together with the thermodynamic equilibrium constant, KC, were found through regression analysis. The resulted regressed values and the goodness of fineness, together with some relevant figures are presented in Appendix A Supplementary data. 3.7. Recycling of catalyst To confirm the reusability of catalyst, recycling experiments over ten runs were conducted, similar to those presented in Section 2.6. The molar ratio between IPEs and Fu was 1:2, at 120 °C and 2 h. FGPr yield vs. recycle run number is shown in Fig. 14. Fig. 14 shows a slow decrease in the catalyst activity, translated into the reduction of the FGPr yield from 87.6% to 75.4%. This seems to result of some morphological changes, due to the deposition on the catalyst surface of Fu oxidation products and oligomers. The contact of Fu wetting the acidic catalyst surface with atmospheric oxygen, during its handling for the next batch process, can form oligomers and oxidation products. A longer lifetime of the catalyst proved to be in a continuous-flow process (see Section 3.4).

3.9. Synthesis of furfurylideneglycerol esters by continuous-flow processing in micropilot plant The analysis of the experimental findings regarding the synthesis of FGPr (see Section 3.2 for details) emphasizes that, for a Fu/IPGPr molar ratio of 2:1 and a reaction temperature of 120 °C, the best performance can be reached after a reaction time of 120 min. Thus, a residence time 295

Industrial Crops & Products 113 (2018) 288–297

E. Stepan et al.

Table 3 NBP and PP (in cooling and heating) of FGEs and ELGKEs. No.

FGEs and ELGKEs

NBP (°C), at 99.8 kPa

PP (°C), in cooling

PP (°C), in heating

1 2 3 4 5 6 7 8

FGPr FGBu FGPe FGHe FGOc FGEh ELGKPr ELGKHe

287.1 308.3 312.5 324.5 358.5 334.0 300.1 344.7

−72.7 −79.6 −81.7 −86.9 −84.8 −79.4 −84.7 −90

−70.8 −77.4 −79.8 −85.4 −83.5 −78.1 −82.1 −88.2

(see Table 3 for details). The tested samples undergo a transition involving a change in enthalpy. Table 3 reveals that NBP of FGEs increased from 287.1 °C to 358.5 °C as the length of hydrocarbon chain increased, the ELGKEs presenting same behavior. PP of FGEs decreased from −72.7 °C (−70.8 °C) to −86.9 °C (−85.4 °C) as the length of hydrocarbon chain increased, in succession FGPr, FGBu, FGPe, FGHe, exception being FGEh and FGOc. PP of ELGKEs decreased as the length of hydrocarbon chain increased. DSC thermograms of FGEs and ELGKEs, in cooling and in heating were presented in Figs. S10–S13, in Appendix A Supplementary data. The 1H NMR and 13C NMR spectra of FGHe, FGBu, FGPR, ELGKPr and ELGKHe were recorded in compliance with Section 2.9 and were presented in Appendix A Supplementary data. All compounds are mixtures of cis-trans isomers in approximately 1:1 ratio.

Fig. 15. The effect of reaction temperature on ELGKPr yield, at stoichiometric feed ratio.

4. Conclusions A new method for obtaining glycerol acetal/ketal esters has been developed, and its viability has been proven by synthesizing eight new compounds, having structures of FGEs and ELGKEs. Due to versatility of method, it is possible to replace Ac from structures of IPGEs with both aldehydes (e.g. Fu) or ketones (e.g. EL). The catalyst, prepared and characterized by means of FT-IR, TG-DTA, XRD, EDX-TEM, and BET had good activity, stability and reusability during experiments, better results being obtained in continuous-flow processing. According to the experiments in the discontinuous system for syntheses of FGEs, the best result (86.7% yield) was obtained for FGPr, using Fu/IPGPr molar ratio of 2:1, 120 °C reaction temperature and 120 min reaction time. Better yields were obtained, e.g. 89.3% for FGPr, in case of continuous-flow processing in micropilot plant, in similar reaction conditions and a residence time of 2 h. The reusability of catalyst increased as well when used in the continuous process; after 24 h of continuous operation, the FGPr yield decreased by only 5.3%. Based upon the experimental results, a new intrinsic kinetic model was proposed, for the catalytic site elementary chemical reactions, which was used to build a complex mathematical model of the discontinuous synthesis of FGEs and ELGKEs. The predictions of this model are in fair agreement with the experimental findings, on a rather large variation domain for the parameters. As already reported, the mass transfer limitations on the liquid side was disregarded due to the turbulences induced by the ascending movement of the Ac vapors. Another phenomenon, unaccounted for, which could enhance the predictive capacity of the model, is the slow deactivation of the catalyst. Recycling the heterogeneous catalyst and Ac, and avoiding the use of solvents as reaction medium should increase the eco-friendliness of this method.

Fig. 16. The effect of reaction temperature on ELGKHe yield, at stoichiometric feed ratio.

of 2 h was chosen for the continuous process presented in Section 2.7, keeping the other two parameters at the aforementioned values. Several advantages were emphasized, against the discontinuous synthesis of FGPr: a) Ac was faster removed from the reactor due to the vacuum, preventing secondary reactions, b) the contact with the oxygen from air was avoided, which was quasi inevitable in the discontinuous process – more, the micropilot plant was purged with argon, prior to process start-up, to prevent these secondary reactions (e.g. oxidation of Fu and products). Accordingly, increased yields were obtained, e.g. 89.3% for FGPr. The reusability of catalyst increased as well, after 24 h of continuous operation, the FGPr yield decreased by only 5.3%.

3.10. Characterization of products The main components in crude and purified FGEs and ELGEs were determined by GC method, according to Section 2.9. Purified FGEs and ELGEs are a complex mixture of isomers, i.e., the cis and trans forms of substituted 1,3-dioxolanes and the cis and trans forms of substituted 1,3dioxanes, the share of substituted 1,3-dioxanes in the mixture being of max. 4%. Mass spectra of representative isomers (cis forms of substituted 1,3-dioxolanes) of FGEs and ELGEs are presented in Figs. S2–S9 (Appendix A Supplementary data). The characteristic molecular ion peaks M, confirmed the structures of FGEs and ELGEs isomers. The normal boiling points (NBP) and pour points (PP) of FGEs and ELGKEs were measured by the DSC method, described in Section 2.9

Acknowledgement The authors gratefully acknowledge the financial support of the UEFISCDI, Romania, in the framework of National Partnership Program, financing contract no. 65/2014. 296

Industrial Crops & Products 113 (2018) 288–297

E. Stepan et al.

Appendix A. Supplementary data

Biomass Bioenergy 76, 130–140. Lapuerta, M., Rodríguez-Fernandez, J., García-Contreras, R., 2015b. Effect of a glycerolderived advanced biofuel-FAGE (fatty acid formal glycerol ester)-on the emissions of a diesel engine tested under the New European Driving Cycle. Energy 93, 568–579. Maximov, A.L., Nekhaev, A.I., Ramazanov, D.N., 2015. Ethers and acetals promising petrochemicals from renewable sources. Pet. Chem. 55, 1–21. Moraes, L.A.B., Mendes, M.A., Sparrapan, R., Eberlin, M.N., 2001. Transacetalization with gaseous carboxonium and carbosulfonium ions. J. Am. Soc. Mass Spectrom. 12, 14–22. Oprescu, E.E., Stepan, E., Dragomir, R.E., Radu, A., Rosca, P., 2013. Synthesis and testing of glycerol ketals as components for diesel fuel. Fuel Process. Technol. 110, 214–217. Oprescu, E.E., Dragomir, R.E., Radu, E., Radu, A., Velea, S., Bolocan, I., Stepan, E., Rosca, P., 2014. Performance and emission characteristics of diesel engine powered with diesel-glycerol derivatives blends. Fuel Process. Technol. 126, 460–468. Patel, A., Brahmkhatri, V., 2013. Kinetic study of oleic acid esterification over 12-tungstophosphoric acid catalyst anchored to different mesoporous silica supports. Fuel Process. Technol. 113, 141–149. Pizzio, L.R., Caceres, C.V., Blanco, M.N., 1998. Acid catalysts prepared by impregnation of tungstophosphoric acid solutions on different supports. Appl. Catal. A 167, 283–294. Pizzio, L.R., Vázquez, P.G., Cáceres, C.V., Blanco, M.N., 2003. Supported Keggin type heteropolycompounds for ecofriendly reactions. Appl. Catal. A 256, 125–139. Selifonov, S., Rothstein, S.D., Mullen, B.D., 2013. Method of making ketals and acetals, US Patent 8604223. Stepan, E., Oprescu, E.E, Radu, A., Enăşcuţă, C.E., 2017a. Process for obtaining acetals and ketals of glycerol, RO Patent 128997. Stepan, E., Pleşu, V., Neamţu, C., Iancu, P., 2017b. Process for preparing 1,2-O-isopropylidene-glycerol esters, RO Patent 129837. Trifoi, A.R., Agachi, P.S., Pap, T., 2016. Glycerol acetals and ketals as possible diesel additives. A review of their synthesis protocols. Renew. Sustain. Energy Rev. 62, 804–814. Wegenhart, B.L., Shou, L., Thom, M., Stanley, D., Abu-Omar, M.M., 2012. Solvent-free methods for making acetals derived from glycerol and furfural and their use as a biodiesel fuel component. ACS Catal. 2, 2524–2530. Werpy, T., Petersen, G., 2004. Top Value Added Chemicals from Biomass, vol. 1 US Department of Energy (USDOE). 103 – OECD Guidelines for the Testing of Chemicals (adopted by the Council on 27th July 1995).

Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.indcrop.2018.01.059. References Bagheri, S., Julkapli, N.M., Yehye, W.A., 2015. Catalytic conversion of biodiesel derived raw glycerol to value added products. Renew. Sustain. Energy Rev. 41, 113–127. Bozell, J., Petersen, G., 2010. Technology development for the production of biobased products from biorefinery carbohydrates the US Department of Energy’s Top 10 revisited. Green Chem. 12, 539–554. Ciriminna, R., Della Pina, C., Rossi, M., Pagliaro, M., 2014. Understanding the glycerol market. Eur. J. Lipid Sci. Technol. 116, 1432–1439. Dodson, J.R., Leite, T.d.C.M., Pontes, N.S., Pinto, B.P., Mota, C.J., 2014. Green acetylation of solketal and glycerol formal by heterogeneous acid catalysts to form a biodiesel fuel additive. ChemSusChem 7, 2728–2734. Garcia, E., Laca, M., Perez, E., Garrido, A., Peinado, J., 2008. New class of acetal derived from glycerin as a biodiesel fuel component. Energy Fuels 22, 4274–4280. Gromachevskaya, E.V., Kvitkovsky, F.V., Usova, E., Kulnevich, V.G., 2004. Investigation in the area of furan acetal compounds. 13*. Synthesis and structure of 1, 3-dioxacyclanes based on furfural and glycerol. Chem. Heterocycl. Compd. 40, 979–985. Kamalakar, G., Komura, K., Sugi, Y., 2006. Tungstophosphoric acid supported on MCM41 mesoporous silicate: an efficient catalyst for the di-tert-butylation of cresols with tert-butanol in supercritical carbon dioxide. Appl. Catal. A 310, 155–163. Karim, M.R., Rahman, M.A., Miah, M.A.J., Ahmad, H., Yanagisawa, M., Ito, M., 2011. Synthesis of γ-alumina articles and surface characterization. Open Colloid Sci. J. 4, 32–36. Khalil, K.M.S., 2007. Synthesis and characterization of mesoporous ceria/alumina nanocomposite materials via mixing of the corresponding ceria and alumina gel precursors. J. Colloid Interface Sci. 307, 172–180. Kim, P., Kim, H., Yi, J., Song, I.K., 2006. Preparation, characterization, and catalytic activity of H3PW12O40 heteropolyacid catalyst supported on mesoporous γ-Al2O3. Stud. Surf. Sci. Catal. 159, 265–268. Lapuerta, M., Rodríguez-Fernandez, J., Estevez, C., Bayarri, N., 2015a. Properties of fatty acid glycerol formal ester (FAGE) for use as a component in blends for diesel engines.

297