Reac Kinet Mech Cat DOI 10.1007/s11144-013-0596-8

Hydrogenolysis of glycerol to propylene glycol using heterogeneous catalysts in basic aqueous solutions Adriana Marinoiu • Claudia Cobzaru • Elena Carcadea • Catalin Capris • Vasile Tanislav Mircea Raceanu

•

Received: 29 March 2013 / Accepted: 18 June 2013 Ó Akade´miai Kiado´, Budapest, Hungary 2013

Abstract The efficient catalytic conversion of glycerol, the main by-product from the bio-diesel production, into valuable chemicals, can contribute to the promotion of bio-diesel utilization from economic and environmental points of view. Glycerol can be hydrogenated to propylene glycol using copper chromite catalyst with high yield and conversion under mild reaction conditions (20 bar, 200 °C, 8 h) and basic aqueous solutions. The effect of the added base type was discussed in this paper and the results indicated that different influences of alkali bases on the activity might be associated with the size of the metal cations. The highest conversion of glycerol and the highest selectivity to 1,2-propanediol were observed in presence of LiOH, whose additions proved to be effective. Keywords

Hydrogenolysis  Propylene glycol  Heterogeneous catalyst

Introduction It is generally accepted that the interest for the catalytic conversion of renewable feedstocks has been increased. In particular, glycerol has been identified as a promising alternative to petroleum and natural gas for the production of commodity chemicals [1–4]. The rapid growth in bio-diesel production from the last years, created a surplus of glycerol and, therefore, the opportunity for using renewable resources as feedstocks. A. Marinoiu (&)  E. Carcadea  C. Capris  V. Tanislav  M. Raceanu National R D Institute for Cryogenics and Isotopic Technologies (ICIT), 4 Uzinei Street, Rm Valcea, Romania e-mail: [email protected] C. Cobzaru Faculty of Industrial Chemistry, Technical University Iasi, 71, D. Mangeron Ave, 70050 Iasi, Romania

123

Reac Kinet Mech Cat

Glycerol, an available, cheap, non-toxic feedstock as sustainable resource, is one of the 12 chemicals from the top value added chemicals from biomass [5], which will serve as key feedstock in future bio-refineries and will generate important environmental benefits. The literature provides different alternatives to transform glycerol in various variable products. One of the most interesting possibilities is the conversion to propylene glycol by an eco-friendly process. For this purpose, many types of solid catalysts have been explored [6]. A variety of catalysts including palladium, platinum, nickel, copper have been used in the hydrogenolysis, but copper based catalysts are far the preferred ones [7–12]. Most of the works have focused on varying catalyst compositions and optimization of reaction parameters. The suitable copper chromite catalysts for use in the hydrogenolysis reaction are of copper–chromium oxide type. Their superiority is connected to the effects given by the mixing of copper with chromite, namely the increasing of the intrinsic catalyst activity of Cu, and the stability action of chromium for preventing sintering [13]. Previously, we studied the glycerol hydrogenolysis to propylene glycol using heterogeneous catalysts (copper chromite and nickel base). The results showed that the copper catalysts are less active than nickel and require longer hydrogenation times even when the hydrogenations are performed at higher temperatures and pressures [14]. Nickel, a known catalyst for simultaneous hydrolysis and hydrogenation of apolyols to polyhydric alcohols, become slightly more selective as temperature increases [15]. However we must note that for high pressures (which are specific hydrogenation conditions), Ni catalysts are not preferred because C–C bond breakages are expected. But, at moderate pressures, Ni base catalyst proved to be an adequate selective catalyst, as our research demonstrated recently. In this respect, our earlier experiments proved that Ni/Al2O3–SiO2 catalyst is an active catalyst for the hydrogenolysis of glycerol under low hydrogen pressures [16]. Another investigation devoted to this interesting field is presented in this paper and consist in the glycerol conversion under mild reaction conditions and basic aqueous solutions. Modifications of catalytic systems by additives often leads to significant changes in selectivities. Although it is known that additives can significantly affect the conversion and selectivity in this reaction, little research has been done regarding the effect of base additives on glycerol hydrogenolysis. Only some details are known regarding this objective, in particular using Ru/TiO2 in basic aqueous solution. Our aim was to explore the possibility of improving the intrinsic catalytic proprieties of copper based catalysts or nickel catalyst by adding a base component. The hydrogenolysis of higher polyols to form propylene glycol involve multiple steps [17]. The investigated literature shows that the polyol is first reversibly adsorbed and dehydrogenated by the catalyst, leading to a desorbed aldehyde species. The product of the dehydrogenation reaction can then be involved in a C–C cleavage by two pathways: either the retro-aldol mechanism and/or oxidation followed by decarboxylation or a C–O scission by dehydration. Both of these scissions pass through reaction intermediates.

123

Reac Kinet Mech Cat

The formation of glycerol products can be explained by the following mechanisms: hydro–dehydrogenation with irreversible adsorption of unsaturated species, retro-aldolization under the action of adsorbed hydroxil groups. This means favoring the selective cleavage of the C–O bond by hydrogen, without attacking the C–C bonds from the glycerol molecule. Modifications of metal catalytic systems by additives often leads to significant changes in selectivities. Several mechanisms have been proposed for the hydrogenolysis of polyols, and the most accepted one referring at glycerol hydrogenolysis in the presence of a base was suggested by Montassier et al. [18, 19]. According to the literature, the reaction route involves a reversible glycerol dehydrogenation to glyceraldehyde and then a dehydration and/or retro-aldolization of glyceraldehide and finally, the two precursors are hydrogenated to 1,2 propylene glycol and ethylene glycol. It is known the fact that modifications of metal catalytic systems by additives often leads to significant changes in selectivities. The study of the basic additives effect for studied reaction using two catalysts types (copper chromite and nickel base catalyst) is a novelty in actual hydrogenolysis context and represent an advancement of knowledge for the scientific community.

Experimental In our experimental study, we used glycerol (99.85 % w/w), commercially available from Oleo Chemicals Germany; hydrogen (99.99 % purity) purchased from Linde Gas Romania, copper chromite catalysts, commercially available from MERCK and Ni/Al2O3–SiO2 purchased from BASF. The copper chromite used in this study was in the powder form, had the surface ˚ . Elemental analysis gave area 10.5 m2/g and the metal particle size equal to 188 A the following composition: 43.89 % CuO, 42.09 % Cr2O3, 10.36 % BaO, 0.03 % CaO, 0.03 % Fe2O3, and 0.01 % Na2O. The characteristics of Ni/SiO2–Al2O3 were: ˚ metal pellet shape with 4 mm diameter, 10 mm length, 80 m2/g surface area, 86 A particle size. Elemental analysis for the BASF catalyst was: 17.73 % Ni; 5.7 % Cu, 1.7 % Mn, 45.2 % SiO2, 22.48 % Al2O3. The metal particle sizes of the catalysts were determined using X-ray diffraction (XRD) spectra recorded in Philips X diffractometer. The average crystallite sizes were calculated by authors from the full width at half maximum (FWHM) of the diffraction peaks using the Scherrer equation, D ¼ k  k=b cos a, where D means the particle diameter, k is a geometric factor (equal to 0.89), k is the X-ray wavelength, b is the FWHM diffraction peak and h is the diffraction angle. The surface area of the supported metal catalysts was measured by the porosimetry method by using a Quantachrome PoreMaster apparatus. The experiments were carried out discontinuously using a stainless steel 200 mL reactor, equipped with stirrer and electrical heater, designed specially for this work. A magnetic stirrer at a constant agitation speed was used to create a slurry reaction mixture. The speed of stirrer was constant throughout the reaction. The temperature

123

Reac Kinet Mech Cat

was monitored using a Pt-100 sensor, inserted into the autoclave and connected to the thermocontroller. The catalysts used in these reactions were activated prior the reaction at 10 bar hydrogen pressure, 300 °C reaction temperature for a period of 4 h, activation times. The hydrogenolysis reactions were conducted under the following conditions: catalyst/glycerol ratio wt. = 5 %, reaction temperature of 200 °C, pressure in the range of 10–20 bar, and a reaction time of 8 h. The autoclave containing the activated catalyst was loaded with the desired quantity of glycerol and certain amounts of base additives. Then, the reactor was flushed several times with hydrogen, to eliminate the trapped air completely. Subsequently, the autoclave was pressurized with hydrogen at room temperature. The reaction mixture was heated to the desired reaction temperature and kept under these conditions for 8 h. At the end of this period, the autoclave was cooled to ambient temperature, then brought to atmospheric pressure and opened to allow the reaction mass to be discharged and centrifuged for removing the catalyst. Products from liquid phase were analyzed with a Hewlett-Packard 6890 Plus gas chromatograph equipped with a flame ionization detector. Hewlett-Packard Chemstation software was used to collect and analyze the data. A capillary column HP 5, (with a stationary phase consisting of 5 % diphenyl and 95 % dimethyl siloxane copolymer, a length of 30 m; an inside diameter of 0.53 mm, and a film thickness of 1.5 lm) was used for the separation. The injector temperature was 300 °C, the detector temperature was 250 °C, while the column temperature was maintained at 45 °C for 5 min, increased to 220 °C with 30 °C/min, and then kept at 220 °C for another 5 min. Quantitative analyses of liquid phase and volatile products in gas phase were carried out using gas chromatography. The formulas that were used for yield, conversion and selectivity are mentioned. Propylene glycol selectivity is defined as the ratio of the number of moles of the propylene glycol product formation to that of the glycerol consumed in the reaction, taking into account the stoichiometric coefficient. The conversion of glycerol is defined as the ratio of number of moles of glycerol consumed in the reaction to the total moles of glycerol initially present. The yield of propylene glycol is defined as the ratio of the number of moles of propylene glycol produced to the theoretical number of moles of the propylene glycol.

Results and discussions Our previous work has highlighted that a possible way to valorize glycerol would be the increasing the selectivity for propylene glycol. Glycerol hydrogenolysis involves the dehydration of glycerol to acetol or 3-hydroxypropanal, followed by their hydrogenation to 1,2- and 1,3-propanediol, respectively [20]. Chai [21] have proposed a more detailed mechanism of hydrogenolysis according to which glycerol initially undergoes dehydration to 2-propene 1,2-diol or 1-propene 1,3-diol, and as a result of their re-arrangement, acetol and 3-hydroxypropanal are formed. The hydrogenation of acetol and 3-hydroxypropanal leads to formation of 1,2- and 1,3propanediol (Fig. 1). The by-products of glycerol hydrogenolysis are acetone,

123

Reac Kinet Mech Cat

Fig. 1 Glycerol hydrogenolysis mechanism

propionaldehyde and they are formed as a result of the dehydration of 1,2- and 1,3propanediol. Their hydrogenation leads to 2-propanol, 1-propanol and even to propane. Dehydration of propanols leads to the formation of propene. The main hydrogenolysis reactions are accompanied by cleavage of C–C bonds in glycerol and propanediol molecules. Thus, the by-products ethylene glycol, ethanol, methanol, methane, ethane and carbon monoxide could be formed. Hydrogenolysis of glycerol over heterogeneous catalysts can be regarded to be a structure sensitive reaction and for this reason the activity and selectivity depend on catalyst structure [22–24]. Earlier authors noted that addition of bases promotes selective hydrogenation. However, the effect is complex, and the change in selectivity depends on both the polyhydroxy compounds structure and the catalysts. Selectivity may change markedly with base, but some catalysts are much more sensitive than others to the effect of the base. The examined catalysts in this study were copper chromite and nickel base catalysts. Copper chromite is well known as an active catalyst for the hydrogenolysis of glycerol under low hydrogen pressure. The glycerol conversion, selectivity for propylene glycol and reaction yield are growing continuously with increasing reaction pressure, the increase being faster in the range 8–25 bar, as can be seen in Fig. 2. The likely reason is that the higher pressures shift equilibrium from acetol to propylene glycol. We investigated also the effect of reaction temperature in glycerol hydrogenolysis, as can be seen in Fig. 3. The presence of a base is expected to enhance the rate of glycerol hydrogenolysis. A sustainable base, as well as catalyst, may have an important influence on glycerol hydrogenolysis. The effect of additional base on the activity of copper chromite catalyst in glycerol conversion is illustrated in Fig. 4. It can be seen that the addition of hydroxides of Li, Na or K and carbonates of Li, Na, K led to a remarkable increase in glycerol conversion in almost all studied catalytic systems and a slight decrease in selectivity. The highest conversion of glycerol (60 %) and the highest selectivity to 1,2-propanediol were observed in the presence of LiOH. Interestingly, the selectivity for 1,2-propanediol changed slightly in all cases of added bases.

123

Reac Kinet Mech Cat

Fig. 2 Effect of reaction pressure in the glycerol hydrogenolysis. All reactions were performed under the following conditions: 8 h reaction time, 200 °C reaction temperature, 5 % wt copper chromite catalyst loadings

Fig. 3 Effect of reaction temperature in the glycerol hydrogenolysis. All reactions were performed in the following conditions: 8 h reaction time, 15 bar reaction pressure, 5 % wt copper chromite catalyst loadings

123

Reac Kinet Mech Cat

Fig. 4 Effect of basic promoters in glycerol hydrogenolysis in presence of copper chromite. The reactions were conducted under the following conditions: reaction temperature of 200 °C, 14–17 bar reaction pressure, 5 % mass ratio catalyst/glycerol, 8 h reaction time; 0.2 g basic compound was added in reaction (excepting the first experiment)

The distinct influences of alkali bases on the activity suggest that the alkali metal cations influence the glycerol hydrogenolysis. The alkali metal cations have affected the reaction in the following order: Li?, Na?, K?. This order may be connected with the cation size. This phenomenon can be explained as follows: the decreasing of crystallites leads to an increase in the number of the active sites and to an increase of stability of the catalyst surface from the tendency of prominent atoms to move into more stable positions. Although the selectivity decreased no matter which base was added, the glycerol conversion increased in all the cases comparative to the case when reaction was performed without base additives. Experimental data are in agreement with the observations of Feng et al. [25] for the Ru/TiO2. In that case, the authors concluded that catalysts containing various amount of bases additives gave a similar effect for glycerol hydrogenolysis on Ru/ TiO2. In this case, according to Feng, most likely mechanism in basic medium, involves a reversible dehydrogenation of glycerol to glyceraldehide, followed by dehydration and/or retro-aldolization of glyceraldehide to 2-hydroxyacrolein and finally the two precursors are hydrogenated to 1,2-propanediol respectively 1,3-propanediol. According to the cited reference, the base effect is related to the first stage of the hydrogenolysis reaction. In this respect, the hydrogen atom from C2 of glyceraldehyde is attacked by the OH group from base, making the primary hydroxyl of glyceraldehyde to be easily removed by dehydration, to form 2-hydroxy acrolein.

123

Reac Kinet Mech Cat

Therefore, the dehydrogenation–hydrogenation equilibrium will be shifted to the right, which implies an increase in the conversion of glycerol. The results of our work are similar to those from specialty literature and confirm the known mechanism proposed by Dasari [6]. This mechanism indicates that hydroxyacetone (acetol) was formed and is an intermediate of an alternative path for forming propylene glycol. Acetol is formed by dehydration of a glycerol molecule, which further reacts with hydrogen to form propylene glycol with 1 mol of water by-product. The alkaline effect is related of course to the effect of pH of the reaction medium. For this purpose, we studied glycerol hydrogenolysis regarding initial pH (adjusted by introducing of NaOH) and the results are presented in Fig. 5. The increasing of pH from 4.3 to 7.5 led to an increase in glycerol conversion from 31 to 41 %, which continued to increase to almost 53 % when we further increased the pH at 9.2. Selectivity to propylene remained almost constant in acidic to neutral but decreased in basic medium. It is interesting to note that although the reaction yield increases throughout the investigated pH range, the proportion of secondary reactions increases, leading to a decrease in product selectivity. So, good results are obtained in a weak base. Therefore, would be wise to introduce some basic compounds to increase the pH from the acidic to neutral medium in order to achieve a more efficient conversion of glycerol, using the same reaction time. The presence of lithium or sodium base significantly increases the conversion of glycerol. Base on the experimental results, the effect of base’s amount was studied using LiOH as additive. As we may see in Fig. 6, the glycerol conversion increases

Fig. 5 Effect of pH on glycerol hydrogenolysis. The reactions were conducted under the following conditions: reaction temperature of 200 °C, 14–17 bar reaction pressure, 5 % mass ratio catalyst/glycerol, 8 h reaction time, 0.2 g NaOH was added in reaction (excepting the first experiment)

123

Reac Kinet Mech Cat

Fig. 6 The effect of the amount of LiOH on the glycerol hydrogenolysis. Reaction conditions: 200 °C, 15 bar, 8 h, 5 % mass ratio catalyst/glycerol

as the LiOH amount increase from 0.1 to 1 g, obtaining a conversion of 75 % for 0.8 g LiOH, after which the conversion began to decrease. The yield for 1,2propanediol is independent of the base concentration when the LiOH dosage is more than 0.5 g. The experiments performed in order to optimize the amount of basic promoter showed that 0.5 g LiOH introduced into the reaction medium, ensures the best catalytic performances. Our last results proved that nickel is an adequate catalyst for more than 98 % selectivity to propylene glycol [16]. So, the experiments were expected also to continue with studying the effect of basic promotor in reaction medium. Under mild reaction conditions (200 °C, 15–25 bar), glycerol was hydrogenolyzed to propylene glycol in basic aqueous solutions and Ni/Al2O3–SiO2 catalyst, and a good selectivity and an acceptable conversion were obtained, as can be seen in Fig. 7. The glycerol conversion changed slightly for all added bases, which is due to the fact that the selectivity of the 1,2-propylene glycol is independent of base concentration in a certain range. However, keeping on increasing the base amount to 1 g only resulted in slight changes in the selectivity of the 1,2-propanediol. It can be noticed that introducing various basic substances in nickel-based catalytic system results in a high conversion of glycerol and higher yields than in their absence, but the selectivity to propylene glycol is much reduced (the decreasing is more pronounced than for copper chromite catalyst). This behavior is the result of glycerol transformation into different products. Therefore, the effect of basic promoters introduction is not beneficial in glycerol hydrogenolysis over nickel catalysts, regarding propylene glycol as the major product. In order to verify the effectiveness of the regeneration procedure in connection with the reusing of the catalyst for hydrogenolysis reaction of glycerol in basic

123

Reac Kinet Mech Cat

Fig. 7 Effect of basic promoters in glycerol hydrogenolysis in presence of nickel catalyst. The reactions were conducted under the following conditions: reaction temperature of 200 °C, 14–17 bar reaction pressure, 5 % mass ratio catalyst/glycerol, 8 h reaction time, 0.2 g basic compound was added in reaction. The first experiment was conducted in the absence of basic promoter

conditions, different treatments were applied for regenerating the used catalyst. The used catalysts were regenerated by different treatments (Table 1). The data obtained in this work show that the methanol washing seems to be the best regeneration procedure. In this paper, a process for the 1,2-propanediol using heterogeneous catalyst was studied. It is known that additives can significantly affect the activity and selectivity of this reaction.This study demonstrated the viability of using the copper chromite catalysts together with additional base in glycerol hydrogenolysis to propylene glycol. High conversions (up to 60 %) were achieved using copper–chromite catalyst, and operating under mild reaction conditions (20 bar, 200 °C, 8 h) and basic aqueous solutions. Table 1 Glycerol hydrogenolysis in the presence of regenerated catalysts Catalyst/washing solvent

Yield (%)

Conversion (%)

Selectivity (%)

Fresh catalyst

37

60

62

Used catalyst/water

8

15

53

Used catalyst/methanol

31

56

55

Used catalyst/ethanol

21

42

50

Used catalyst/2-propanol

18

29

58

All the catalysts (copper chromite) were reduced prior to hydrogenolysis reaction under the following conditions: reaction temperature of 300 °C, hydrogen pressure of 10 bar, 4 h reaction time

123

Reac Kinet Mech Cat

The effect of the added base type was discussed and the results indicated that different influences of alkali bases on the activity might be associated with the size of the metal cations. This motivation is supported by specialty literature. The highest conversion of glycerol (60 %) and the highest selectivity to 1,2propanediol were observed in the presence of LiOH. The effect of basic promotor amount on the hydrogenolysis reaction was investigated using LiOH as additive. As the amount of LiOH increased from 0.1 to 0.8 g, the conversion of glycerol exhibited an increase, reached a maximum for 0.8 g LiOH dosage, and after that the conversion began to decrease. This study is significant in the general context of glycerol hydrogenolysis, because the addition of LiOH in studied catalytic system proved effective. For this reason, we also expect this behavior to be efficient in other catalytic systems based on copper chromite and to be applied in the future researches. The described conditions should represent an advancement of knowledge for the scientific community, because the experimental data confirmed the fact that base additives could significantly affect the activity and selectivity of this reaction.

References 1. Bozell JJ (2008) Clean 36(8):641–647 2. JJ Bozell (2008) Biorefinery product opportunities from glycerol. In: Proceedings of a conference, Atlanta, Georgia, 12–13 Feb 2008 3. Bo¨hmer N, Roussie`re T, Kuba M, Schunk SA (2012) Comb Chem High Throughput Screen 15(2):123–135 4. Bauer F, Hulteberg C (2013) Biofuels Bioprod Biorefin 7(1):43–51 5. Top Value Added Chemicals from Biomass (2004) Volume I-results of screening for potential candidates from sugars and synthesis gas. http://www1.eere.energy.gov/biomass/pdfs/35523.pdf. Accessed Aug 2004 6. Dasari M, Kiatsimkul PP, Suterlin W, Suppes G (2005) Appl Catal A 281:225–231 7. Moulton K, Beal R (1974) US Patent 3.856.710 8. Suppes G, Sutterlin W, Dasari M (2005) US Patent Application 0244312 9. Wang S, Liu H (2007) Catal Lett 117:62–67 10. Casale B, Gomez AM (1993) US Patent 5.214.219 11. Huang L, Zhu YL, Zheng HY, Li YW, Zeng ZY (2008) J Chem Technol Biotechnol 83:1670–1675 12. Suppes G, Sutterlin W (2007) WO 053705 13. Chiu CW, Dasari MA, Suppes G, Sutterlin W (2006) AIChE J 52(10):3543–3548 14. Marinoiu A, Ionita G, Ga´spa´r CL, Cobzaru C, Marinescu D, Teodorescu C, Oprea S (2010) React Kinet Catal Lett 99:111–118 15. Paul Bloom, Archer Daniels Midland Company (2008) US Patent Application 0228014 A1 16. Marinoiu A, Ionita G, Gaspar CL, Cobzaru C, Oprea S (2009) React Kinet Catal Lett 97:315–320 17. Lahr DG, Shanks BH (2005) J Catal 232:386–394 18. Montassier C, Dumas JM, Granger P, Barbier J (1995) Appl Catal A 121(2):231–244 19. Sohounloue DK, Montassier C, Barbier J (1983) React Kinet Catal Lett 22:391 20. Miyazawa T, Kusunoki Y, Kunimori K, Tomishige K (2006) J Catal 240:213–221 21. Chai S-H, Wang H-P, Liang Y, Xu B-Q (2007) Green Chem 9:1130–1136 22. Casale B, Gomez AM (1994) US Patent 5276181 23. Lahr DG, Shanks BH (2003) Ind Eng Chem Res 42:5467 24. Maris EP, Ketchie WC, Murayama M, Davis RJ (2007) J Catal 251:281–294 25. Feng J, Wang J, Yhou Y, Fu H, Chen H, Li X (2007) Chem Lett 36(10):1274–1275

123