1. LITERATURE
3.2 S YNTHESIS OF VALUE - ADDED GLYCEROL DERIVATIVES
3.2.1 Synthesis of triacetin
Effect of catalysts on reaction selectivity and triacetin yield
The esterification reaction of glycerol with acetic acid was carried out without catalyst or in presence of sulfuric- and phosphoric acid starting from pure glycerol and partly purified glycerol, still containing inorganic salts. The formed reaction water was removed by straight-forward distillation in absence of entraining solvents. The comparison of the selectivity values and triacetin yields for esterification of glycerol with and without catalysts at reaction temperature 115-145 °C is shown in Table 3.2.1.1.
Table 3.2.1.1 Comparison of the selectivity values and triacetin yields for esterification of glycerol with and without catalysts at reaction temperature 115-145 °C in case of straight-forward distillation (molar ratio of acetic acid/glycerol: 5:1) XG%: glycerol conversion, YTA%: triacetin yield
Raw
material Catalyst MA Selectivity % DA TA XG % YTA %
Pure glycerol - 18.0 57.4 23.6 100.0 23.6
Glycerol
cont. salt - 24.7 56.0 14.6 91.6 13.3
Pure glycerol H2SO4 1.8 30.9 67.4 100.0 67.4
Glycerol
cont. salt H2SO4 0.0 59.0 40.2 98.4 39.5
Crude
glycerol H2SO4 Consistent, dark-brown tar, could not be analysed
Pure glycerol H3PO4 5.0 40.1 54.4 99.0 53.9
Glycerol
cont. salt H3PO4 0.0 68.0 30.9 97.6 30.2
Better glycerol conversions and triacetin yields were achieved when using pure glycerol, rather than using partly purified glycerol, as it was predicted before conducting the experiments. Minor by-product formation was observed during the reactions.
The best results, namely a glycerol conversion of 100 % and a triacetin selectivity of 67.4 % were obtained in presence of H2SO4 as catalyst using pure glycerol as raw material.
When the starting glycerol was containing salts the yield of triacetin dropped down to ~ 39.5 % by H2SO4-catalysis and ~ 30.2 % in presence of H3PO4. The experimental data for the selectivity and yield of triacetin demonstrate that acetylation of glycerol with acetic
acid can be carried out without a catalyst, but acetic acid is not strong enough to catalyze the reaction properly to yield pure tiacetin, only.
Crude glycerol without any purification steps was also tested as raw material for a triacetin synthesis. A dark brown tar was received as reaction product, which could not be analysed. Therefore, crude glycerol without any purifications is not recommended for triacetin synthesis!
Effect of acetic acid excess on reaction selectivity and triacetin yield
Due to the fact that the boiling point of acetic acid (118 °C) and boiling point of water (100 °C) are close to each other, a loss of acetic acid to the end of the water removing during the reaction is significant because it is volatile together with water. This problem can be eliminated by utilizing a structured packed distillation column to separate water and/or by help of entraining solvent to remove water azeotropically at lower temperature.
Using a Raschig-ring packed laboratory scale distillation column for water separation and a higher molar excess of acetic acid, better reaction selectivity results and triacetin yield were achieved when salt-containing glycerol is used and H2SO4 as catalyst.
Table 3.2.1.2 Effect of efficient water separation by packed distillation column and higher molar excess (AA/G) of acetic acid (AA) on selectivity and triacetin yield of esterification of glycerol (G) starting from partly purified glycerol. XG%: glycerol conversion, YTA%:
triacetin yield Raw
material Catal.
Molar excess
(AA/G) Procedure
Selectivity % XG
%
YTA
%
MA DA TA
Glycerol
cont. salt H2SO4 1.67 Without packed
column 0.0 59.0 40.2 98.4 39.5 Glycerol
cont. salt H2SO4 2.67 Raschig column
distillation 0.0 36.0 64.0 100.0 64.0
Effect of azeotropic water removal by entraining solvent
To make the removal of water easier an entraining solvent was added into the reaction media. Entraining solvents should satisfy the following criterias:
• Low cost
• No azeotrope formation with acetic acid
• Formation of a heterogeneous azeotrope with water
• High water content of the azeotrope
• Normal boiling point of the azeotrope should be lower than < 90 °C
For azeotropic water removal n-hexane, methyl-isobutyl-ketone and toluene were found to be the best candidates, which fulfill all the necessary criteria (see above). However, in case of MIBK a significant by-product ((2-isobutyl-2-methyl-1,3-dioxolan-4-yl)methanol) formation was observed (see sub chapter 3.2.4). Table 3.2.1.3 shows that n-nexane and toluene pushed the reaction towards the triacetin formation, triacetin was achieved in almost quantitative yield.
Table 3.2.1.3 Effect of entraining solvents on reaction selectivity and triacetin yield of esterification of glycerol starting from pure and partly purified glycerol (molar ratio of acetic acid/glycerol: 5:1; reaction temperature was 100-120 °C in case of toluene as entraining solvent and 68-74 °C in case of n-hexane as entraining solvent)
Raw
material Catal. Entraining solvent
Selectivity % XG % YTA %
MA DA TA
Pure
glycerol H2SO4 n-hexane 0.0 0.0 100.0 100.0 100.0 Glycerol
cont. salt H2SO4 n-hexane 0.0 2.0 98.0 100.0 98.0 Pure
glycerol H2SO4 toluene 0.0 0.0 100.0 100.0 100.0 Glycerol
cont. salt H2SO4 toluene 0.0 1.0 99.0 100.0 99.0 When starting from partly purified glycerol the best selectivity and triacetin yield results were obtained with toluene as entraining agent. This could be explained by the higher boiling point of solvent-water azeotrope which efforts a higher reaction temperature.
Effects of heterogenous catalysts on reaction selectivity and triacetin yield
Despite of the fact that most ion-exchange resins are not stable at temperatures above 140 °C (Zhang et al., 2001), which limits their application to reactions that require higher temperatures, Amberlyst 15 and 36 were investigated as heterogeneous catalysts for the esterification of glycerol by acetic acid in the present work. Physical characteristics of ion-exchange resin catalysts are summarized in Table 3.2.1.4 (Klepacova et al., 2005).
Table 3.2.1.4 Characteristics of ion-exchange resins (Amberlyst 15 and 36) Catalyst
The use of entraining agents for the water removal is of great help not to exceed the maximal operation temperature of ion-exchange resins in esterification reactions. Together with ion-exchange resins as esterification catalysts MIBK and n-hexane were tested as entraining agents (see Table 3.2.1.5). When MIBK was used as entraining agent the formation of (2-isobutyl-2-methyl-1,3-dioxolan-4-yl)methanol as by-product was observed, just as in case of homogenous catalysis, but to a lesser degree.
It can be seen from Table 3.2.1.5, that the catalysts Amberlyst A15 and A36 result the same yield of triacetin (~92%) when MIBK was used as entraining solvent. When n-hexane was used as entraining solvent the yields of triacetin dropped significantly (Amberlyst A15: ~ 40 %; Amberlyst 36: ~ 75%). This can be explained by the lower reaction temperature (68-74 °C).
Table 3.2.1.5 Comparison the effects of Amberlyst 15 and Amberlyst 36 catalysts on reaction selectivity and triacetin yield of esterification of glycerol (molar ratio of acetic acid/glycerol: 5:1; reaction temperature was 110-120 °C in case of MIBK as entraining solvent and 68-74 °C in case of n-hexane as entraining solvent)
Raw
material Catal. Entraining solvent
*Significant by-product formation was observed
Moreover, in case of pure glycerol and presence of n-hexane Amberlyst A15 gave better selectivity results than Amberlyst A36 which may mean that the optimal working temperature of Amberlyst A15 is lower than that of the A36.
When partly purified glycerol is used as raw material ion-exchange resins are not as effective as in case of pure glycerol. This could be explained by the deactivation of active sites of the ion-exchange resin by the ions of the salts, present in that raw material.
Table 3.2.1.6 summarizes the best results of this work in comparison with results of other publications (Lu et al., 1991, Wu et al., 2007, Hou et al., 1998, Ding et al., 2003, Liu et al., 2007). It can be concluded that the yield of triacetin prepared by our methods is close to the results of other authors. However, we have to remark that the comparability of the results is limited as the reaction conditions (reaction temperature, molar ratio of acetic acid/glycerol etc.) were different in most cases.
Table 3.2.1.6 Comparison of the best results in case of triacetin synthesis
Authors Quality of glycerol Catalyst YTA %
Lu et al. Pure Acidic ion exchange
resin and MgSO4 87.0
Wu et al. Pure SO42- /ZrO2-TiO2 93.6
Hou et al. Pure Aminosulfonic acid >90.0
Ding et al. Pure H3PW12O40 >98.0
Liu et al. Pure p-toluenesulfonic acid/C 92.0
This work
Pure Amberlyst 15 92.0
Pure Amberlyst 36 92.0
Pure H2SO4 100.0
Partly purified H2SO4 99.0
Composition of salts filtered off after triacetin synthesis
Figure 3.2.1.1 and Table 3.2.1.7 show the typical composition of a salt which was filtered off after the triacetin synthesis from the reaction mixture. In this case triacetin was prepared from partly purified glycerol which contained inorganic salts.
Table 3.2.1.7 Composition of salts filtered off after triacetin synthesis
Component Content (%(m/m))
Na2SO4 x 10 H20 75.4
Na2SO4 (thenardite) 19.3
Na3H(SO4)2 3.9
Na5P3O10 0.9
Na2HPO4 0.5
Sum 100.0
The obtained salt consists of materials originated partly from the neutralization step of crude glycerol, but also such quantities derived from the neutralization of sulfuric acid by NaOH after the triacetin synthesis.
Thermal stability of triacetin
Figure 3.2.1.2 shows the DSC (differential scanning calorimetry) thermogram of triacetin purchased from Sigma-Aldrich and Figure 3.2.1.3 shows the DSC thermogram of triacetin prepared in this work from partly-purified glycerol.
According to the heat-flow vs. temperature curves the behavior of triacetin prepared by our process from partly purified glycerol is similar to the behavior of purchased material, seen from the point of view of thermal stability.
However, the thermal stability of our product is higher. The decomposition of pure commercial triacetin started at 198.7 °C, while our product started to decompose at 208.5
°C. Moreover, in case of purchased triacetin the rate of decomposition achieved its highest level at 207.3 °C whereas the maximal decomposition rate of our product was obtained at 220.9 °C.
Figure 3.2.1.2 DSC curve of pure commercial triacetin purchased from Sigma-Aldrich
Figure 3.2.1.3 DSC curve of triacetin prepared in this work by our process starting from
Distillation of triacetin
Although, it is known that separation of mono- di- and triacetin is not possible by simple distillation. Therefore, a final product purification of triacetin by distillation was investigated from the point of view of color-removal.
When a mixture containing MA, DA and TA was distilled, the composition of distillate and residue changed slightly. According to the material balance of mono- di- and triacetin, it can be concluded that during distillation conversion of monoacetins and triacetin to diacetins takes place.
However, the color value of material measured on Platinum-Cobalt color scale improved significantly (see Table 3.2.1.8).
Table 3.2.1.8 Compositions of triacetin and color values of triacetin measured on Platinum-Cobalt color scale before and after distillation under 0.6 kPa vacuum
Mass (g) Monoacetin (%)
Diacetin (%)
Triacetin (%)
Color (Pt-Co color values)
Raw material 200.0 2 44 51 382
Distillate 150.4 1 38 56 78
Residue 49.6 0 73 23 nm
Conclusions
A sulfuric acid catalyzed reaction, starting from partly purified glycerol, seems to be an appropriate way of triacetin synthesis. When the reaction water was removed by an azeotropically distillation with toluene or n-hexane triacetin yields of > 98% were obtained.
However, toluene offers some advantages compared with n-hexane such as flammability, boiling point, toxicity etc.
The most difficult problem concerning crude glycerol refining is the removal of salt which is formed during neutralization of the catalyst (KOH, NaOH).
Due to the fact that the presence of inorganic salts is not disturbing the formation of triacetin using homogenous catalysts like sulfuric acid or phosphoric acid, all formed salts can be removed by simple filtration. Such an operation is supported by the facts, that these salts are insoluble in the reaction media, but also easy filterable.
If applications of triacetin do not require colorless material, an expensive purification step by distillation can be ignored. As long as the yellow color of the product is undesirable, further investigation on decolorization (for example adsorption by activated carbon or decolorization by oxidant) is recommended to find the most convenient way of color removal.
Based on the experimental results the following material balance was calculated (see Figure 3.2.1.4 and Table 3.2.1.9).
75
Figure 3.2.1.4 Material balance of triacetin synthesis from crude glycerol Biodiesel
Table 3.2.1.9 Mass balance of triacetin synthesis from crude glycerol
Stream Input (kg) Output (kg)
Crude glycerol 100.00 -
Water 26.70 56.64
Phosphoric acid 12.30 -
Sulfuric acid 4.14 -
Acetic acid 96.14 -
Solvent (toluene, n-hexane) 150.00 150.00
Free fatty acids - 44.95
Inorganic salt - 21.15
Triacetin - 116.54
The manufacture of triacetin from crude glycerol using entraining agents allows to remove beside the reaction water also water which is remaining from the biodiesel process.
Therefore, any earlier dewatering step became obsolete.
Any activated-carbon treatment of crude glycerol seems to be unnecessary because triacetin needs to be decolorized at the end of the process, if the application requires low colored material.
3.2.2 Synthesis of tripropionin
The synthesis of triproprionin was performed on the analogy of the triacetin preparation.
The comparison of the results on reaction selectivity and tripropionin yield starting from different raw materials is shown in Table 3.2.2.1.
An excellent conversion of 100% for glycerol and a selectivity and yield of 98 % for tripropionin was obtained in case of pure glycerol, as it was predicted before the experiments.
When the starting glycerol was containing salt the yield of tripropionin dropped down to 86.8% (see Table 3.2.2.1)
These yields for tripropionin were much higher than in case of triacetin synthesis under similar conditions. This could be explained by the higher reaction temperature (tripropionin: 120-160 °C instead of triacetin: 115-145 °C) because in case of triproprionin synthesis reaction mixture contained components having higher boiling points (BpPA=141
°C > BpAA=118 °C; BTP= 290.7°C > BpTA= 258-260 °C ).
Table 3.2.2.1 Effect of quality of raw materials on the reaction selectivities and tripropionin yields of esterification of glycerol by propionic acid in presence of sulfuric acid (propionic acid/glycerol molar ratio: 6:1, 2-fold molar excess quantity, reaction temperature: 120-160 °C)
Raw
material Catalyst Entraining
solvent MP Selectivity % DP TP XG % YTP % Pure
glycerol H2SO4 - 0.0 2.0 98.0 100.0 98.0
Glycerol
cont. salt H2SO4 - 0.0 13.2 86.8 100.0 86.8
It was concluded that the straight-forward removal of water requires a distillation separation column with rectifying section. Namely, during the non-azeotropic distillation water removal without a rectification column a significant loss of propionic acid was observed.
Role of entraining solvents and ion exchange resins
Catalytic activity of Amberlyst A15 and Amberlyst A36 resins on the esterification of glycerol by propionic acid was also investigated. On the analogy of the triacetin
preparation, an entraining solvent was needed not to exceed the maximal operation temperature of ion-exchange resins in esterification reactions. Together with ion-exchange resins as esterification catalyst MIBK and n-hexane were tested as entraining agents.
It can be seen from Table 3.2.2.2 the highest yield was reached when raw material was pure glycerol and MIBK was used as entraining solvent in presence of Amberlyst A36 catalyst. Due to the higher boiling point of MIBK-water azeotrope, the temperature of the reaction media was higher (110-120 °C) than in case of n-hexane (68-74° C) and higher temperature may be favored by Amberlyst A36.
Moreover the same phenomenon was observed as in case of triacetin: when raw material was pure glycerol in presence of n-hexane Amberlyst A15 gave better result than Amberlyst A36 which may mean that the optimal working temperature of Amberlyst A15 is lower than that of the A36.
When partly purified glycerol is used as raw material ion-exchange resins are not as effective as in case of pure glycerol. Reaction product contained mainly dipropionin.
This could be explained by the deactivation of active sites of ion-exchange resin by the ions of the salts, present in that raw material.
Table 3.2.2.2 Comparison of the effects of catalysts, entraining solvents and quality of raw materials on reaction selectivity and tripropionin yield of esterification of glycerol (propionic acid/glycerol molar ratio: 6:1 with 2-fold molar excess quantity, reaction temperature: 110-120 °C in case of MIBK, 68-74° C in case of n-hexane and 100-124 °C in case of toluene)
When partly purified glycerol was used as raw material the best result, 96% yield of tripropionin was achieved in presence of H2SO4 catalyst and toluene as entraining solvent.
Distillation of tripropionin
Although, it was known that separation of mono-, di- and tripropionin is not possible by simple distillation. Therefore, a final product purification of tripropionin by distillation was investigated from the point of view of decolorization.
In the color-removal experiment a relatively pure bottom product containing 95%
tripropionin was distilled. The distillate contained the components by almost the same ratio as they were present in the raw material which confirms that the boiling points of dipropionins and tripropionin are close to each other. However, the color value of material measured on Platinum-Cobalt color scale improved significantly (see Table 3.2.2.3).
Considering the material balance of distillation, tripropionin is thermally stable at high temperature (up to ~200 °C).
Table 3.2.2.3 Compositions of tripropionin and color values of tripropionin measured on Platinum-Cobalt color scale before and after distillation under 0.6 kPa vacuum
Mass (g)
Monopropionin (%)
Dipropionin (%)
Tripropionin (%)
Color (Pt-Co color
value)
Raw material 200.0 0 3.6 95.0 423
Distillate 155.4 0 2.7 96.2 62
Residue 44.6 0 4.3 93.1 nm
Conclusions
A sulfuric acid catalyzed reaction, starting from partly purified glycerol, seems to be an appropriate way of tripropionin synthesis. When the reaction water was removed by an azeotropically distillation with toluene, tripropionin yield of > 96% was obtained. Considering a technical manufacturing toluene would provide technical and commercial benefit like low solubility in water and lower price.
The most difficult problem concerning crude glycerol refining is the removal of salt which is formed during neutralization of the catalyst (KOH, NaOH).
Due to the fact that the presence of salts is not disturbing the formation of tripropionin using homogenous catalysts like sulfuric acid, all formed salts can be removed by simple filtration. Such an operation is supported by the facts, that these salts are insoluble in the reaction media, but also easy filterable.
If applications of tripropionin do not require colorless material, an expensive purification step by distillation can be ignored. As long as the yellow color of the product is undesirable, further investigation on decolorization (for example adsorption by activated carbon or decolorization by oxidant) is recommended to find the most convenient way of color removal.
Based on the experimental results the following material balance was calculated for the tripropion synthesis from crude glycerol (see Figure 3.2.2.1 and Table 3.2.2.4).
81
Figure 3.2.2.1 Material balance of tripropionin synthesis from crude glycerol Biodiesel Vegetable oils, animal fats,
methanol, NaOH
Table 3.2.2.4 Mass balance of tripropionin synthesis from crude glycerol.
Stream Input (kg) Output (kg)
Crude glycerol 100.00 -
Water 26.70 56.64
Phosphoric acid 12.30 -
Sulfuric acid 4.14 -
Propionic acid 118.70 -
Solvent (toluene, n-hexane) 150.00 150.00
Free fatty acids - 44.95
Inorganic salt - 21.15
Tripropionin - 139.10
The manufacture of tripropionin from crude glycerol using entraining agents allows to remove beside the reaction water also water which is remaining from the biodiesel process. Therefore, any earlier dewatering step became obsolete.
Any activated-carbon treatment of crude glycerol seems to be unnecessary because tripropionin needs to be decolorized at the end of the process anyway, if the application requires colorless material.
Engine performance
The effects of tripropionin blending on engine performance characteristics and environmental repercussions are given in Table 3.2.2.5.
A small increase of total unburned hydrocarbons (THC) in blended fuel was observed which must be in connection with the inadequate burnout of tripropionin.
However, due to the high oxygen content of tripropionin a decrease of CO, smoke and exhaust temperature was observed.
Furthermore, the specific fuel consumption slightly increased which can be explained by the measured lower maximal motor torque.
Table 3.2.2.5 Comparison of engine performance characteristics and exhaust emissions from combustion of reference commercial diesel fuel and tripropionin-blended diesel fuel
Reference represents a promising material, which can be used as fuel additive.
3.2.3 Synthesis of glycerol carbonate Results of glycerol carbonate synthesis
Between two types of hydroxyl group in glycerol, two primary alcohols are presumably more reactive than a secondary hydroxyl group. This statement is supported by the experimental data reported by Dallos et al. (2000), who measured the effect of steric hindrance on molecular interactions of hydroxyl groups and the differences in pair-wise interactions between molecules and hydroxyl groups in primary and secondary positions.
Therefore, during the synthesis of glycerol garbonate intermediate C might be formed at
Therefore, during the synthesis of glycerol garbonate intermediate C might be formed at