Rs

JPa

b

c

P

1

TgtitCmraogtf

M

0d

food and bioproducts processing 8 7 ( 2 0 0 9 ) 171–178

Contents lists available at ScienceDirect

Food and Bioproducts Processing

journa l homepage: www.e lsev ier .com/ locate / fbp

evalorization of glycerol: Comestible oil from biodieselynthesis

ordi Boneta,∗, Jose Costaa, Romain Sireb, Jean-Michel Reneaumeb, Alexandra Elenalesuc, Valentin Plesuc, Grigore Bozgac

Departament d’Enginyeria Química, Universitat de Barcelona, Martí i Franquès 1, 6th Floor, E-08028 Barcelona, SpainEcole Nationale Supérieure en Génie des Technologies Industrielles, UPPA/LATEP Rue Jules Ferry, BP 7511, 64 075 PAU Cedex, FranceDepartment of Chemical Engineering, University POLITEHNICA of Bucharest, Centre for Technology Transfer in the Process Industries, 1olizu St., Building A, Room A-056, RO-011061, Sector 1, Bucharest, Romania

a b s t r a c t

High dependence on fossil fuel has caused increase of carbon dioxide concentration in the atmosphere. The actual

political trends are towards an increased use of renewable fuels from agricultural origin. One of the main products of

the European biorefineries is biodiesel. The main reaction involved in biodiesel synthesis produces a large amount of

glycerol as by-product. Two aspects are arising in this respect: the glycerol obtained as residue and the food conversion

to fuel. This paper deals with the revalorization of the residual glycerol stream to obtain triacetin (glyceryl triacetate),

the lightest comestible oil. The application of glycerol as raw material to produce triacetin is not new. The goal of this

paper is to check the feasibility of this transformation in an efficient integrated continuous process which is suitable

for processing high quantities of glycerol. A kinetic model was determined experimentally for the production of

triacetin from glycerol and acetic acid in the absence of catalyst. The results showed that by process integration of

the reaction and distillation in the same unit (reactive distillation), a more sustainable process can be developed.

The proposed configuration output is checked by rigorous simulation.

© 2009 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

Keywords: Glycerine; Triacetin; Reactive distillation; Process simulation; Biodiesel

diesel are imported and consequently this leads to its price

. Introduction

he EU is supporting biofuels with the aim of reducingreenhouse gas emissions, boosting the decarbonization ofransport fuels, diversifying fuel supply sources, offering newncome opportunities in rural areas and developing long-erm replacement for non-renewable fossil fuels (Europeanommission, 2006). The fossil fuel represents 88% of the pri-ary energy consumed in the world. However, fossil fuel

esources are becoming more and more difficult to extractnd process. Nowadays, the energy supply strongly dependsn the events occurring in the main countries which providereat fossil fuel resources. Moreover, the Asian industrializa-

ion and growth requires more energy, leading to higher pricesor the oil (Fig. 1). In the current context of high oil prices, the

Abbreviations: EU, European Union; RD, reactive distillation; FID, flamt, million tons.∗ Corresponding author. Tel.: +34 934 021 310; fax: +34 934 021 291.

E-mail address: bonet@ub.edu (J. Bonet).Received 7 October 2008; Received in revised form 25 May 2009; Accep

960-3085/$ – see front matter © 2009 The Institution of Chemical Engioi:10.1016/j.fbp.2009.06.003

development of sustainable energetic sources for the transportbecomes a priority. In EU, the main focus is on diesel, whoseprice is higher than the gasoline for the first time (Fig. 2). At thebeginning the diesel was used mainly for slow engines suchas for agriculture, fishing ships and transport, for this reasonrecording lower taxes. Nowadays, the improvement of dieselengines leads to performances similar to gasoline enginesand lower fuel consumption. This caused a 75% use of dieselproduction in regular cars, while only 18% is used for the agri-culture. The diesel–gasoline consumption ratio in the 1990swas around 1:10 and today it is around 7:3. The higher demandof diesel in EU generates excess of gasoline, large amounts of

e ionization detector; GC, gas chromatography; USD, US dollars;

ted 10 June 2009

increase. This situation has a direct impact on food prices bythe increase of production and transport costs. EU established

neers. Published by Elsevier B.V. All rights reserved.

172 food and bioproducts processing 8 7 ( 2 0 0 9 ) 171–178

Nomenclature

kj kinetic constantAj pre-exponential factor—Arrhenius expressionEaj activation energy (kJ/kmol)T temperature (K)n number of molesx molar fraction—liquid phasey molar fraction—vapour phase

Fig. 3 – World population. (http://www.census.gov/ipc/

Fig. 1 – Crude oil prices (http://www.wtrg.com/prices.htm).

the aim to mix 5.75% of biofuels in gasoline and diesel until2010 (O’Driscoll, 2007), USA, Canada, Australia or Japan havetaken similar measures. Biodiesel production in the EU wasestimated to be about 6 Mt in 2006 and is forecasted to increaseto about 12 Mt in 2010 (Behr et al., 2008). A great expansion ofthe biodiesel production in the next decades in EU is expected.

On the other hand, a continuous increase of popula-tion (doubled in last 40 years (Fig. 3)), leads to a leakage ofresources. In 2005, the total production of fats and oils was inthe range of 144 Mt. Nowadays, only 14% of this material is pro-cessed within the chemical industry (Behr et al., 2008). The useof comestible oil to produce biodiesel can lead to a decrease ofthe oil available as food for people and farm animals. Anotherissue related to biodiesel production is the large surplus ofglycerol generated. Although glycerol has been a well-knownrenewable chemical for centuries, its commercial relevancehas increased considerably in the last few years due to itsrising unavoidable formation as a by-product of biodiesel pro-duction. A 25% molar of the output stream is the by-product

glycerol (10% in weight). The glycerol market is already satu-rated and more than half of the industries producing glycerol

Fig. 2 – Gasoline and diesel prices.

www/idb/worldpop.html).

have already closed (Fig. 4). The main strategy used has beento find new applications for the glycerol, apart of the ones thatalready exist (Leffingwell and Lesser, 1945). Fig. 5 shows someof the main applications (Pagliaro et al., 2007). Nowadays, alarge number of biodiesel producers are incinerating the glyc-erol (O’Driscoll, 2007). An efficient way to convert glycerol tofood grade triacetin would be a plausible way to revalorizeglycerol because it could be used at least in animal nutrition.Triacetin is one of the main chemical products obtained fromglycerol. There are several industrial synthesis alternatives,one of them being the reaction of glycerol and acetic acid. Thetriacetin is a transparent oil, combustible and of bitter taste.It occurs naturally as cod-liver oil, in butter, and other fats(Grant, 1972). Also known as glyceryl triacetate, it is reported tofunction as a cosmetic biocide, plasticizer, solvent in cosmeticformulations and also as biodiesel fuel component (García etal., 2008). It is a commonly used carrier for flavours and fra-grances. Triacetin was generally recognized as safe humanfood ingredient by the Food and Drug Administration. Moreinformation on its properties and toxicity is presented in theFinal Report on the Safety Assessment of Triacetin (2003).

The conversion of glycerol to triacetin is an existing proce-dure which should increase its performance to process highquantities of glycerol in the most environmental friendly way.Nakamori (1952) found that the quantity of triacetin generatedwas very low compared to the diacetin (glyceryl diacetate) andmonoacetin (glyceryl monoacetate) synthesis intermediates.There are several studies in literature searching for catalystsfor the reaction of glycerol and acetic acid. For instance Luand Ma (1991) obtain 87% triacetin in the product streamusing acidic ion exchange resin and MgSO4 at room temper-ature for 72 h. Yang and Lu (1996) and Wu et al. (2007) useSO4

2−/ZrO2−TiO2, Hou et al. (1998) use aminosulfonic acid,Zhang (1999) use SnCl4·5H2O/C, Zhang and Yuan (2001) usephosphotungstic acid, Ding et al. (2003) use H3PW12O40, Dongand Guo (2003) use solid sulfated Fe2O3/TiO2, Melero et al.(2007) obtain the best performances using sulfonic acids, Liuet al. (2007) use p-toluenesulfonic acid/C, Li et al. (2007) useionic liquids ([HSO3-pmim] [PTSA]). Several patents proposevarious approaches for triacetine synthesis in the presence ofcatalysts, such as Bremus et al. (1981), Gawrikow et al. (1982),Pechenev et al. (1995), Mitsuya and Ogawa (1996), Mhaskar andKulkarni (2002). The general strategies followed to increasethe conversion to triacetine involve the use of acetic acid inlarge excess, the use of acetic anhydride to eliminate water

generated, or the simple distillation for water removal.

The innovation proposed in this paper is not related tothe catalyst, but to the process. The kinetic parameters are

food and bioproducts processing 8 7 ( 2 0 0 9 ) 171–178 173

F and( emb

aaeuslmaRe

ig. 4 – European glycerol producer stocks in metric tons (a)1995–2005) (Oleoline Glycerine Market Report, num. 71, Dec

djusted to experimental data obtained in the absence ofny catalyst. As the system is highly non-ideal, the phasequilibrium is described with activity coefficients calculatedsing classical UNIFAC method. The feasibility of triacetinynthesis by reactive distillation is verified by rigorous simu-ation, performed with AspenPlus® implementing the kinetic

odel developed. The reactive distillation (RD) provides a new

pproach to the increase of conversion in triacetine synthesis.D combines the reaction and the distillation within the samequipment, overcoming chemical equilibrium limitations. The

Fig. 5 – The market for glycerol (volumes and i

price development of 80% crude glycerol in USD (b)er 2005).

products are continuously removed from the reactive mediaand a total conversion of glycerol can be achieved in a singlecolumn.

2. Materials and methods

The components used in the experiments are glycerol of

99% purity (GC) from Panreac (ref. 151339.1212), glacial aceticacid (Reag. Ph. Eur.) of 99.7% purity (GC) from Panreac (ref.131008.1611) and triacetin of 99% purity from Aldrich (ref.

ndustrial use) (Source: Novaol, May 2002).

cessing 8 7 ( 2 0 0 9 ) 171–178

Fig. 6 – Measured and calculated time evolutions of thethree reaction products concentrations at 120 ◦C (Themarked points represent experimentally measured valuesand the lines calculated values).

Table 1 – Estimated values of the reaction rate constantsat 120 ◦C (concentration basis in molar fractions andkinetic constants in s−1).

k1 = 8.87 × 10−2 k−1 = 1.32k2 = 8.74 × 10−2 k−2 = 3.32 × 10−1

k3 = 1.23 × 10−2 k−3 = 6.54 × 10−2

Fig. 7 – Measured and calculated time evolutions of thethree reaction products concentrations at 160 ◦C (Themarked points represent experimentally measured values

curve, is quite well represented by the proposed kinetic model.The values of the Arrhenius expression re-confirm the already

Table 2 – Estimated values of the reaction rate constantsat 160 ◦C (concentration basis in molar fractions andkinetic constants in s−1).

k1 = 1.26 × 10−3 k1 = 1.99 × 10−3

k2 = 1.24 × 10−3 k−2 = 5.75 × 10−3

k3 = 1.91 × 10−4 k−3 = 1.30 × 10−4

Table 3 – Estimated activation energies andpre-exponential factors.

J (reaction) Aj Eaj (kJ/kmol) A−j Ea−j (kJ/kmol)

174 food and bioproducts pro

525073). Monoacetin and diacetin are not available pure in themarket.

Gas chromatography has been used for the analysis since itis proved to provide good results in the triacetin analysis (Lu,1991; Ogawa et al., 1988; Uematsu et al., 1997). The gas chro-matograph used is a HP 7890A with a column HP-5 Part. no.19091J-413 of 30 m × 0.32 mm × 0.25 �m, FID injector at 310 ◦Cand detector at 300 ◦C, slope of temperature of 15 ◦C/min from100 ◦C to 200 ◦C resulting in an analysis time around 7 min.The injected volume is of 1 �L and the pressure is at 13.1 psi.The response factor (ratio of areas versus ratio of moles) is lin-ear without bias. The water generated by the reaction is notdetectable.

The final kinetic determinations were performed in astirred pressurized reactor PPI Industries Series LC of 0.300 Lcapacity, pressurized at 10.7 bar to avoid the vaporization ofthe mixture and with a stirring speed of 290 rpm.

The number of components which can be obtained con-sidering all the combinations between acetic and dimerisedacetic acid with glycerol rises up around 22 feasible products.When an excess of acetic acid is used, there are a few maincomponents detected. To simplify the kinetic adjustments, thereacting system is reduced to three consecutive main reac-tions:

Glycerol + aceticacid ↔ monoacetin + H2O

r1 = k1cglycerinecacetic acid − k−1cmonoacetincH2O (1)

Monoacetine + aceticacid ↔ diacetin + H2O

r2 = k2cmonoacetinecacetic acid − k−2cdiacetincH2O (2)

Diacetin + aceticacid ↔ triacetin + H2O

r3 = k3cdiacetincacetic acid − k−3ctriacetincH2O (3)

An Arrhenius-type model is used:

ki = Ai e−Eai/RT (4)

The reaction rate is negligible at room temperature. How-ever, at boiling point, the reaction rate is fast enough and thepresence of catalyst is not required. The kinetic determina-tions are performed in the absence of catalyst, at 393 and 433 K,with an excess of acetic acid four times the required quantityimposed by stoichiometry.

3. Results

3.1. Kinetic parameters determination

The kinetic studies available in literature related to triacetinproduction are oriented towards the identification of an opti-mal catalyst for this reaction. This paper presents a kineticmodelling study at higher temperatures able to be imple-mented in a reactive distillation system as well as in thetraditional system of reactor followed by a train of distilla-

tion columns operating at higher temperature conditions. Theexperimental results and the estimated values of rate con-stants at 120 ◦C are shown in Fig. 6 and Table 1 and at 160 ◦C in

and the lines calculated values).

Fig. 7 and Table 2. The regression of the Arrhenius parametersis identified in Table 3. The fitted lines are obtained by simul-taneous estimation of the 12 Arrhenius parameters, using theleast squares method, based on the experimental data givenin both figures (Figs. 6 and 7).

Both figures show the expected behaviour of the consec-utive reactions and provide the corresponding concentrationevolution in time. The triacetin kinetics, following a sigmoid

1 6.9 3.1 × 104 190 3.3 × 104

2 6.8 3.1 × 104 220 3.8 × 104

3 2.4 3.4 × 104 200 4.3 × 104

food and bioproducts processing 8 7 ( 2 0 0 9 ) 171–178 175

Fa

kt

waolagTbal

3p

Titaraqmdstiu

Fig. 9 – Traditional process scheme for the production of

ig. 8 – Vapour–liquid (liquid) equilibrium for the mixturecetic acid with water and glycerol with triacetin.

nown information available in literature: the inverse reac-ions are favoured compared to the direct ones.

The vapour–liquid equilibrium of the involved mixturesas calculated (Fig. 8). Water is the lowest boiling compound

nd it does not form any azeotrope with any componentf the mixture, therefore it could be collected in the distil-

ate of the reactive distillation column. It must be taken intoccount that there is a dimerisation of the acetic acid in theas phase when the vapour–liquid equilibrium is calculated.he heaviest component is glycerol followed by triacetin;oth are immiscible (liquid–liquid phase split) and form anzeotrope which divides the residue curve map in two distil-ation regions.

.2. Rigorous simulation of the triacetin synthesisrocess

he rigorous simulation of the triacetin synthesis processs performed in different process configurations: the tradi-ional system of a reactor followed by distillation column trainnd the reactive distillation system. The first alternative rep-esented by the traditional scheme is evaluated taking intoccount an excess of acetic acid three times the stoichiometricuantity, at a pressure of 5 bar. The reactor output contains aixture composed of acetic acid, water, glycerol, monoacetin,

iacetin and triacetin. A first distillation column is used toeparate the water and acetic acid from the rest of the mix-

ure. The unreacted glycerol due to equilibrium limitationss collected in a second distillation column. The final prod-ct is a mixture of monoacetin, diacetin and triacetin which

Table 4 – Traditional process streams—triacetin synthesis (mole

GLY HAC F1

Glycerol 1.00 0 0.02Acetic acid 0 1.00 0.83Water 0 0 0.11Triacetin 0 0 0.02Diacetin 0 0 0.02Monoacetin 0 0 0.02Mole flow (kmol/h) 5 75 80

triacetin.

is difficult to separate, containing the triacetin in a low con-centration. The scheme and details of the system streams areshown in Fig. 9 and Table 4.

Water and acetic acid are the lowest boiling point compo-nents of the mixture (373 and 391 K, respectively). From thevapour–liquid equilibrium data estimated taking into accountthe acetic acid dimerisation in vapour phase, a pinch zonenear pure water is detected. The water generated by reactionand the excess of acetic acid are recovered in the distillate ofthe first distillation column.

The heaviest components of the mixtures are the glyceroland the produced oils. Glycerol and triacetin form a minimumboiling azeotrope. According to this fact, at least two distilla-tion regions are expected: one with a bottom product enrichedin glycerol and another with a bottom product enriched intriacetin. Monoacetin and diacetin have similar volatilitiesthan triacetin but they are totally converted to triacetin dueto the continuous removal of water from the reaction media.Therefore, a total conversion of glycerol in a single reactive dis-tillation column is expected (Fig. 10), collecting pure triacetinin the bottoms and a mixture of water and acetic acid in thedistillate.

Some other process schemes are possible. Collecting a mix-ture of glycerol and triacetin in the bottoms can be followedby their separation in a decanter due to immiscibility. Glyc-erol can be further recycled to the column whereas triacetin ispurified in a second distillation column. However, this secondalternative would require a higher number of columns andrecycled streams.

The same operating conditions used for the traditionalsystem are also applied to the RD process simulation using

AspenPlus: a pressure of 5 bar and an excess of acetic acidthree times the stoichiometric quantity. The residence time

fraction).

D1 D2 B1 B2

0 0.25 0.25 0.240.89 0 0 00.11 0 0 00 0.60 0.33 7.16E−50 4.11E−3 0.31 0.670 0.14 0.12 0.09

75 2.71 5 2.29

176 food and bioproducts processing 8 7 ( 2 0 0 9 ) 171–178

Fig. 12 – Dependence of the reflux and the number ofstages.

Fig. 13 – Distillate flow rate analysis to determine thebottoms product purities.

Fig. 10 – Reactive distillation process scheme for theproduction of triacetin.

in the column stages (reaction time) is fixed to five minutes.This time could be decreased using a catalyst. The goal ofthe present paper is to check the feasibility of triacetin syn-thesis process by reactive distillation without optimization.Therefore, 70 stages are considered. A further optimizationcan provide a smaller column. The reflux ratio is adjusted toobtain 99% molar triacetin as the bottom product. The aceticacid and glycerol are fed on stages 3 and 2, respectively. Thefeed streams are at boiling point. The column profile com-position for 70 stages is presented in Fig. 11. The reflux ratioconsidered is 2.51 and the influence of the required number of

stages on the reflux is shown in Fig. 12. A sensitivity analysisof the distillate flow rate and its influence on the monoacetin,diacetin and triacetin concentrations at the column bottoms

Fig. 11 – Composition profile in the reactive distillationcolumn.

Fig. 14 – Temperature profile in the reactive distillation

column.

is shown in Fig. 13. A maximum value of diacetin concentra-tion is obtained at a distillate flow rate value of 26 kmol/h,but it is not obtained pure. Triacetine is the only compoundwhich can be collected pure in the bottom of the column. Inthis situation, water and the excess of acetic acid are obtainedin the distillate. The glycerol and monoacetin are depicted insmall concentration along the column. Most of the compo-sition changes occur near the column top and bottom. Thecolumn profile composition is almost constant and very rich

in acetic acid on most of the stages. Fig. 14 shows the columntemperature profile. A sharp increase in temperature takes

food and bioproducts processing 8 7 ( 2 0 0 9 ) 171–178 177

Table 5 – Reactive distillation process streams—triacetin synthesis (mole fractions).

GLY HAC D B

Glycerol 1.00 0 79 ppm 8 ppbAcetic acid 0 1.00 0.668 916 ppmWater 0 0 0.332 23 ppbTriacetin 0 0 823 ppb 0.990Diacetin 0 0 7 ppm 0.009Monoacetin 0 0 15 ppm 6 ppm

45

piTT

4

TtspbmTtunaitidctrtlRsatmivticcwtst

otda

A

FCn

Mole flow (kmol/h) 5

lace on a few stages interval as a consequence of the suddenncrease of triacetin composition as it is described in Fig. 11.he stream compositions and flow rate summary are given inable 5.

. Conclusions

he EU is promoting the use of biodiesel but a high quan-ity of glycerol is generated as by-product and its market isaturated. The conversion of the glycerol to triacetin couldrovide a plausible solution. The kinetic parameters able toe implemented in reactive distillation processes are deter-ined experimentally and fitted to an Arrhenius-type model.

he integration of the reaction in a distillation column leadso the advantages of reactive distillation (energy of reactionsed directly for the distillation, reaction rate accelerated,o danger of hot spots, total conversion of glycerol withoutny catalyst employed). A high purity of triacetin is obtainedn the reactive distillation column bottom due to the con-inuous removal of water generated by the reaction whichs collected finally in the distillate. According to the kineticsetermined and vapour–liquid equilibrium data estimated bylassical UNIFAC method, the reactive distillation is provedo be an appropriate alternative for this reaction system, theesult being confirmed by rigorous simulation. Compared tohe traditional system of reactor followed by a train of distil-ation columns, in which the glycerol conversion is not total,D provides greater advantages. Additionally, the traditionalcheme leads to a final product consisting of a difficult to sep-rate mixture of monoacetin, diacetin and triacetin, in whichhe triacetin is in a lower concentration than the diacetin or

onoacetin. By using RD separation scheme, this drawbacks avoided. Monoacetin, diacetin and triacetin have similarolatilities and are difficult to separate by a traditional dis-illation column. Therefore, the continuous removal of watern the distillate of the reactive column is able to force the totalonversion and provide pure triacetin at the bottom. Triacetinan become a low cost but high volume chemical producthich could be used in cattle nutrition or other purposes. Due

o the purpose of triacetin usage as a bulk product, it is con-idered that its production should be carried out integrated inhe biorefinery.

Future work will be oriented towards a more detailed studyf the process kinetics and of the suitability of reactive dis-illation alternative, complemented with the experimentaletermination of vapour–liquid equilibrium data not yet avail-ble and process optimization.

cknowledgements

inancial support from the Ministerio de Educación yiencia (Grant Nr. CTQ2005-05065/PPQ) and from the Roma-ian Government National Agency for Scientific Research

75 5

Programme PNII (Project number 22138/2008 TINOCIPand Project number 71053/2007 GLICEVAL) is gratefullyacknowledged.

References

Behr, A., Eilting, J., Irawadi, K., Leschinski, J. and Lindner, F., 2008,Improved utilisation of renewable resources: New importantderivatives of glycerol. Green Chem, 10(1): 13–30.

Bremus, N., Dieckelmann, G., Jeromin, L., Rupilius, W. andSchuett, H., 1981, Continuous production of triacetin, DE3004660, 34 p.

Ding, B., Jin, Z., Zhang, C. and Pan, G., 2003, Synthesis of glycerintriacetate catalyzed by H3PW12O40 and purified by activatedcarbon. Jingxi Shiyou Huagong, 2: 47–50.

Dong, Z. and Guo, J., 2003, Study on synthesis of glyceroltriacetate with solid superacid catalyst. Jingxi HuagongZhongjianti, 33(4): 20–21.

European Commission, 2006, Communication from theCommission of the European Communities. An EU Strategyfor Biofuels (SEC (2006) 142). Brussels, 8.2.2006 COM (2006) 34final.

Final Report on the Safety Assessment of Triacetin., 2003, Int JToxicol, 22(3): 1–10.

García, E., Laca, M., Pérez, E., Garrido, A. and Peinado, J., 2008,New class of acetal derived from glycerin as a biodiesel fuelcomponent. Energy Fuels, 22: 4274–4280.

Gawrikow, A., Hein, M., Reklat, D. and Thoms, K., 1982, Glyceroltriacetate, DD 156803, 8 p. Patent.

Grant, J. (ed) 1972, Hackh’s chemical dictionary (p. 688).Hou, J., Zhang, Q., Shi, W. and Li, Y., 1998, New process for

synthesis of triacetin. Henan Huagong, 15(6): 18–19.Leffingwell, G. and Lesser, M.A., (1945). Glycerin. Its industrial and

commercial applications. (Chemical Pub. Co, New York).Li, H., Yu, S., Liu, F., Li, H. and Xie, C., 2007, Synthesis of glycerol

triacetate using acidic functional ionic liquid as catalyst.Huagong Jinzhan, 26(8): 1120–1124.

Liu, H., Lu, Y. and Gong, S., 2007, Study on synthesis of glyceroltriacetate using p-toluenesulfonic acid supported by activatedcarbon as catalyst. Hebei Gongye Keji, 24(1): 21–23.

Lu, Y., 1991, Determination of triacetin by gas chromatography.Fenxi Ceshi Tongbao, 10(5): 77–79.

Lu, Z. and Ma, W., 1991, Studies on esterification reactions. (I).Esterification of glycerin with acetic acid using ion exchangeresin and drying agent as catalysts. Guangzhou ShiyuanXuebao, Ziran Kexueban, 2: 53–56.

Mhaskar, S.Y., Kulkarni, S.M., 2002, Esterification process forobtaining triacetin with a low acetic acid content, IN 188658,12 p. Patent.

Melero, J.A., Van Grieken, R., Morales, G. and Paniagua, M., 2007,Acidic mesoporous silica for the acetylation of glycerol:Synthesis of bioadditives to petrol fuel. Energy Fuels, 21(3):1782–1791.

Mitsuya, Y. and Ogawa, Y., 1996, Preparation of triacetin, JP08027061, 3 p. Patent.

O’Driscoll, C., 2007, Seeking a new role for glycerol. Biofpr:Biofuels Bioproducts and Biorefining, 1(1): 7.

Nakamori, I. and Glycerides, V., 1952, Triacetin produced by thereaction between acetic acid and glycerol. Nippon KagakuKaishi (1921–1947). Ind Chem Sect, 55: 225–227.

cessi

Zhang, M. and Yuan, X., 2001, Synthesis of glycerol triacetate

178 food and bioproducts pro

Ogawa, T., Fujii, R., Tanaka, K., et al., 1988, A method fordetermination of triacetin in foods by gas chromatography. JFood Hyg Soc Jpn, 29: 180–184.

Pagliaro, M., Ciriminna, R., Kimura, H., Rossi, M. and Della Pina,C., 2007, From glycerol to value-added products. Angew ChemInt Ed, 46: 4434–4440.

Pechenev, Y.G., Savinova, M.A., Ankudimov, V.A., Bastrakov, N.I.,Fedotov, P.I. and Dmitrieva, T.A., 1995, Process for preparingtriacetin, RU 2,047,596, Patent.

Uematsu, Y., Sadamasu, Y., Hirata, K., et al., 1997, Determinationof carrier solvents in flavor preparations by GC and GC/MS. JFood Hyg Soc Jpn, 38: 452–459.

ng 8 7 ( 2 0 0 9 ) 171–178

Wu, H., Yu, B. and Ge, S., 2007, Shengxiang, Complex solid superacid SO4

2−/ZrO2–TiO2 used in synthesis of glycerol triacetate.Huagong Jinzhan, 26(7): 1041–1043.

Yang, S. and Lu, B., 1996, Application of SO42−/ZrO2–TiO2 in

synthesis of glycerin triacetate. Xibei Shifan Daxue Xuebao,Ziran Kexueban, 32(2): 43–45.

Zhang, F., 1999, Synthesis of glycerol triacetate catalyzed bySnCl4·5H2O/C. Jingxi Huagong, 16(6): 30–32.

catalyzed by phosphotungstic acid. Hecheng Huaxue, 9(5):469–472.