PRODUCTION OF VALUE-ADDED CHEMICALS FROM GLYCEROL
VIA A SIX-MEMBERED CYCLIC ACETAL
By
Xi Hong

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirement
for the degree of
DOCTOR OF PHILOSOPHY
Chemical Engineering
2011

ABSTRACT
PRODUCTION OF VALUE-ADDED CHEMICALS FROM GLYCEROL
VIA A SIX-MEMBERED CYCLIC ACETAL
By
Xi Hong
The global growth in biodiesel production over the past decade has led to the availability
of significant quantities of glycerol as a byproduct of the triglyceride transesterification process.
Glycerol is a versatile feedstock for producing chemicals such as propylene glycol and
epichlorohydrin. Another attractive application of glycerol is to convert it into cyclic acetals via
direct condensation with aldehydes and ketones or transacetalization with another acetal
compound. The product, glycerol acetal, can be used as fuel additives to improve viscosity and
cold properties of biodiesel, and alleviate the pollution problems by reducing particulate
emissions from fuel combustion.
In this work, glycerol acetal was synthesized via the acid-catalyzed reaction of glycerol
with acetaldehyde or 1,1-diethoxyethane. Results show that at 25°C, the reaction gives nearly
theoretical yield in 2 h if 1,1-diethoxyethane is used as the acetal forming reagent. The kinetics
of this reaction was then investigated in batch reactors. The effect of solvent, reactant molar ratio,
catalyst loading, and temperature were examined. A second order kinetic model is proposed to
describe this reaction, and rate constants, activation energies, and equilibrium constants have
been calculated based on the experimental results.
The product from glycerol transacetalization is a mixture of four glycerol acetal isomers.
In the presence of acid catalyst, the isomers interconvert to form an equilibrated mixture with the
same composition. Vacuum distillation was carried out to separate the four acetals into their
purified

forms.

Result

show

that

the

cis-5-hydroxy-2-methyl-1,3-

dioxane (1) is the most volatile isomer among the four. When the distillation is carried out in the
presence of an acid catalyst to promote interconversion, isomer (1) is obtained as a nearly pure
distillate stream. This is further adapted to a continuous reactive distillation which produced 9096% isomer (1) in a lab scale column.
With the availability of pure isomer (1), new synthesis pathways are developed to obtain
valuable chemicals such as 1,3-propanediol and 1,3-dihydroxyacetone. For instance, by
oxidizing the remaining free hydroxyl group in isomer (1) and hydrolyzing the corresponding
ketone product, the tanning ingredient 1,3-dihydroxyacetone can be produced. The novelty and
advantage of this method is that the cyclic acetal structure in isomer (1) blocks the primary
hydroxyl groups of glycerol, thus can ensure a good selectivity of secondary alcohol oxidation
and eliminate the byproduct generated from primary alcohol reaction.
To assess the feasibility of this concept, oxidation of isomer (1) to the corresponding
ketone product was investigated using a variety of oxidants. Results show that a ruthenium
tetroxide-based catalyst has the highest efficiency.
Based on the experimental results obtained in this study, a novel process to convert
glycerol to the desired acetal (1) and then to 1,3-dihydroxyacetone is presented.

To My Family

iv

ACKNOWLEDGEMENTS

I would like to express my sincere gratitude to my advisors, Dr. Dennis Miller and Dr.
Carl Lira, for their continuous support, guidance and encouragement. I would also like to thank
Dr. Martin Hawley, Dr. Ramani Narayan, and Dr. Chris Saffron for their guidance and serving
on my graduate committee.
My sincere thanks also goes to Dr. James Jackson, Dr. Lars Peereboom and other fellow
labmates for their help on my research work and stimulating discussions.
Finally, I would like to thank my parents for supporting me spiritually throughout my life.

v

TABLE OF CONTENTS
LIST OF TABLES....................................................................................................................... viii
LIST OF FIGURES ........................................................................................................................ x
LIST OF SCHEMES..................................................................................................................... xii
Chapter I. Background and Introduction ........................................................................................ 1
1.1. Introduction.......................................................................................................................... 1
1.2. Scope of this dissertation ..................................................................................................... 5
References................................................................................................................................... 8
Chapter II. Formation of Cyclic Glycerol Acetals from Acetaldehyde or 1,1-Diethoxyethane ... 11
2.1. Introduction and background ............................................................................................. 11
2.2. Experimental section.......................................................................................................... 13
2.2.1. Materials ..................................................................................................................... 13
2.2.2. Experiments ................................................................................................................ 14
2.2.3. Analysis....................................................................................................................... 15
2.3. Results and discussion ....................................................................................................... 16
2.3.1. Direct acetalization versus transacetalization ............................................................. 16
2.3.2. Identification of products ............................................................................................ 16
2.3.2. Reaction conditions..................................................................................................... 21
2.3.3. Mass transfer effects ................................................................................................... 25
2.3.4. Kinetic model.............................................................................................................. 26
2.4. Conclusions........................................................................................................................ 35
References................................................................................................................................. 37
Chapter III. Separation of Cyclic Glycerol Acetal Isomers & Continuous Production of the 6Membered Isomer...................................................................................................... 39
3.1. Introduction and literature review...................................................................................... 39
3.2. Material and methods......................................................................................................... 42
3.2.1. Materials ..................................................................................................................... 42
3.2.2. Synthesis of GA isomers............................................................................................. 42
3.3.3. Distillation of GA isomers .......................................................................................... 43
3.2.4. Analytical methods ..................................................................................................... 44
3.3. Results and discussion ....................................................................................................... 45
3.3.1. GA synthesis and analysis .......................................................................................... 45
3.3.2. GA isomerization equilibrium .................................................................................... 45
3.3.3. Separation of GA isomers ........................................................................................... 49
3.3.4. Batch and continuous production and recovery of cis-6 isomer................................. 53
3.4. Conclusions........................................................................................................................ 60
References................................................................................................................................. 62
Chapter IV. Oxidation of 6-Membered Glycerol Acetal .............................................................. 64
4.1. Introduction and literature review...................................................................................... 64
vi

4.1.1. Direct oxidation of glycerol to DHA .......................................................................... 65
4.1.2. Alcohol oxidation to carbonyl compounds ................................................................. 67
4.1.3. Alcohol oxidation using green oxidants ..................................................................... 75
4.1.4. Oxidation of dioxanol to the corresponding ketone product....................................... 84
4.2. Experimental...................................................................................................................... 85
4.2.1. Materials ..................................................................................................................... 85
4.2.2. GC and GC-MS analysis............................................................................................. 86
4.3. Results and discussion ....................................................................................................... 86
4.3.1 Oxidation of cis-6 with pyridinium chlorochromate (PCC) ........................................ 86
4.3.2. Oppenauer oxidation ................................................................................................... 87
4.3.3. Ruthenium tetroxide catalyzed oxidation ................................................................... 90
4.3.4. Aerobic oxidation of cis-6 with molecular oxygen..................................................... 97
4.3.5. Oxidation with hydrogen peroxide ........................................................................... 102
4.3.6. Ozone oxidation ........................................................................................................ 103
4.4. Conclusions...................................................................................................................... 106
References............................................................................................................................... 109
Chapter V. Design and Construction of Reactive Distillation Column ...................................... 117
5.1. Overview.......................................................................................................................... 117
5.2. Introduction...................................................................................................................... 117
5.3. Column construction........................................................................................................ 118
5.4. Design and calculations ................................................................................................... 121
5.4.1. Temperature and pressure limit................................................................................. 121
5.4.2. Pipe thickness............................................................................................................ 122
5.4.3. Reboiler..................................................................................................................... 123
5.4.4. Condenser ................................................................................................................. 124
5.4.5. Sensor wiring ............................................................................................................ 124
References............................................................................................................................... 129
Chapter VI. A Novel Process Concept to Make 1,3-Dihydroxyacetone from Glycerol............. 130
Reference ................................................................................................................................ 136
Chapter VII. Summary and Conclusions .................................................................................... 137
Chapter VIII. Suggestions for Future Work ............................................................................... 139
References............................................................................................................................... 142
Appendix I. Mass Spectra of the Glycerol Transacetalization Products ................................... 144
Reference ................................................................................................................................ 159
Appendix II. Calculation of Weisz-Prater Criterion for Ascertaining the Influence of Mass
Transport on Transacetalization Kinetics ................................................................ 160
Reference ................................................................................................................................ 164
Appendix III. Calculation of High-Valent Ruthenium Species in Aqueous Solution ............... 165
References............................................................................................................................... 172

vii

LIST OF TABLES
Table 2-1. cis- and trans- conformational structures of glycerol acetal........................................ 12
Table 2-2. Key to compounds detected by gas chromatography (Figure 2-1).............................. 18
Table 2-3. Effect of ethanol as solvent for glycerol transacetalization reactions. Amberlyst-15
loading = 1 wt% based on glycerol, reaction temperature = 50°C. ........................... 22
Table 2-4. Comparison of equilibrium conversion for reactions with various reactant molar ratios
................................................................................................................................... 23
Table 2-5. Evaluation of internal mass transport resistances........................................................ 26
Table 2-6. Equilibrium constant of glycerol transacetalization reactions (Amberlyst-15 loading =
1 wt% based on glycerol) .......................................................................................... 27
Table 2-7. Rate constant and Weisz-Prater value at different reaction conditions....................... 34
Table 3-1. Summary of GA isomers recovered in the initial distillation...................................... 50
Table 3-2. Summary of GA isomers recovered in the second distillation .................................... 52
Table 3-3. Summary of GA isomers recovered in reactive-distillation ........................................ 55
Table 3-4. Thermodynamic data of acetaldehyde acetalization reaction with alcohols [21, 23]
(Equilibrium constants are based on mol fraction).................................................... 60
Table 4-1. Ozone oxidation experiments of secondary alcohols at 0°C. [51] .............................. 81
Table 4-2. Standard reduction potentials of selected oxidation agents [85]. ................................ 83
Table 4-3. Oppenauer oxidation experiments of cis-6.................................................................. 88
Table 4-4. Ketone compounds and their oxidation potentials [96]............................................... 89
Table 4-5. Ruthenium tetroxide catalyzed oxidation of cis-6 at room temperature, atmospheric
pressure. Catalyst was generated in situ with ruthenium dioxide.............................. 92
Table 4-6. Solubility of the inorganic salts used as the secondary oxidant. ................................. 93
Table 4-7. Vapor phase aerobic oxidation of cis-6 at atmospheric pressure. ............................... 98
Table 4-8. Aerobic oxidation of cis-6 using noble metal catalysts. Atmospheric pressure, air flow
rate = 25-30 mL/min................................................................................................ 100
Table 4-9. Aerobic oxidation of cis-6 using noble metal catalysts at elevated pressure. Oxygen
was used as oxidant. ................................................................................................ 101
viii

Table 4-10. Ozone oxidation experiments. Substrate: 20 g of 20% cis-6 solution; ozone (8-9 wt%
in oxygen): 5-8 L/min; room temperature; reaction time: 20 min........................... 104
Table 5-1. Antoine coefficients of some chemicals and temperature allowed at Psat = 300 psi. 122
Table 5-2. Parameters used for minimum wall thickness calculation. ....................................... 123

ix

LIST OF FIGURES
Figure 1-1. Production of biodiesel and glycerol acetal via reactive distillation [29].................... 5
Figure 2-1. GC analysis of glycerol acetalization products.......................................................... 17
Figure 2-2. Glycerol transacetalization products after 24h. Reaction temperature, 1) 40 °C; 2) 50
°C; 3) 70 °C. Molar ratio of reactants, DEE:glycerol= 1:2. Amberlyst-15 cationic
exchange resin loading = 1 wt% of glycerol. For interpretation of the references to
color in this and all other figures, the reader is referred to the electronic version of
this thesis (or dissertation. ......................................................................................... 21
Figure 2-3. GC analysis of glycerol acetalization products. Reactions were carried out at 40°C
with 1 wt% Amberlyst-15 loading (based on glycerol). Molar ratio,
glycerol:DEE:ethanol = A) 1.5:1:4; B) 1:1:4; C) 1:2:4. Reaction time = 24 h. ........ 24
Figure 2-4. Effect of stirring speed. Glycerol:DEE:ethanol = 1:1:12, catalyst loading = 1wt%
(based on glycerol), temperature = 40°C................................................................... 25
Figure 2-5. Second order reaction kinetic plot for glycerol transacetalization with DEE at 40°C,
glycerol:DEE:ethanol = 1.5:1:15, 1 wt% catalyst loading based on glycerol ........... 30
Figure 2-6. Second order reaction kinetic plot for glycerol transacetalization with DEE at 40°C,
glycerol:DEE:ethanol = 1:1:12, catalyst loading based on glycerol; 0.5%, 1%, 2%,
3% respectively.......................................................................................................... 31
Figure 2-7. Parity plot between the reaction rates observed in experiments and the ones predicted
by the model. ............................................................................................................. 32
Figure 2-8. Kinetics plot for glycerol transacetalization reactions at different temperatures.
Conditions: initial molar ratio glycerol:DEE:ethanol = 1:1:12; catalyst loading 1%
(based on the weight of glycerol) .............................................................................. 33
Figure 2-9. Arrhenius plot of forward rate constant of transacetalization of glycerol with DEE. 35
Figure 3-1 Mechanism of the interconversion between glycerol acetal isomers [10]. ................. 40
Figure 3-2. Vacuum distillation apparatus.................................................................................... 44
Figure 3-3. Isomer composition of three GA samples before and after equilibration at 25°C. a) –
Before; b) – After equilibration. ................................................................................ 47
Figure 3-4. Isomer equilibrium at different temperatures............................................................. 49
Figure 3-5. Initial distillation of GA isomer. Column pressure was 20.5±1.0 mmHg (except for
the first sample, where the pressure was 9.0 mmHg). ............................................... 51
x

Figure 3-6. Second distillation of GA isomers. Column pressure: 7.7±0.3 mmHg.................... 53
Figure 3-7. Reactive-distillation of GA isomers. Column pressure was 1.2±0.2 mmHg. .......... 55
Figure 3-8. Continuous reactive distillation of GA isomers. Column pressure = 2.9±0.2 mm Hg.
................................................................................................................................... 57
Figure 3-9. Process concept for continuous production of 2-methyl-5-hydroxy-1,3-dioxane (cis6 isomer) from glycerol. ............................................................................................ 58
Figure 4-1. DHA as the building block to other chemicals [3]..................................................... 65
Figure 4-2. Standard reduction potential of Ru species at various oxidation states [34].............. 72
Figure 4-3. Phase-transfer oxidation with ruthenium oxide [33].................................................. 74
Figure 4-4. Oxidative dehydrogenation mechanism of alcohol oxidation over metal catalysts.
[55]............................................................................................................................. 77
Figure 4-5. Oxometal vs. peroxometal pathways in metal-catalyzed alcohol oxidations with
hydrogen peroxide [67].............................................................................................. 79
Figure 4-7. Calculated population (%) of various ruthenium species at different pHs in an
2-4
aqueous solution with [Ru] = 3.8×10 M. RuO4 (●), RuO4 (○) and RuO4 (△).
................................................................................................................................... 95
Figure 4-8. Population (%) of various ruthenium species at different pHs of an aqueous solution
2-4
with [Ru] = 3.8×10 M. Experimental results. RuO4 (●), RuO4 (○) and RuO4
(△). [99] .................................................................................................................... 95
Figure 4-6. Ozone oxidation of cis-6 solution. ........................................................................... 104
Figure 5-1 Reactive distillation column constructed for this project.......................................... 120
Figure 5-2. Reboiler of reactive distillation column................................................................... 124
Figure 5-3. Condenser used for the reactive distillation column. ............................................... 124
Figure 5-4. Wiring diagram of the pressure control system of the column. ............................... 126
Figure 5-5. Wiring diagram of the liquid level control system of the column ........................... 127
Figure 6-1. Design of new DHA production process. ................................................................ 131

xi

LIST OF SCHEMES
Scheme 1-1. Processes of catalytic conversion of glycerol into useful chemicals [19]. ................ 3
Scheme 1-2. Dioxane and dioxolane isomers of glycerol acetal .................................................... 4
Scheme 1-3. 2-methyl-1,3-dioxan-5-one and 1,3-dihydroxyacetone (DHA). ................................ 6
Scheme 2-1. Acetalization and transacetalization reaction........................................................... 11
Scheme 2-2. Formation of GAA from glycerol acetals and DEE................................................. 19
Scheme 2-3. Decomposition of 1,1-diethoxyethane ..................................................................... 19
Scheme 2-4. Reaction between glycerol and ethyl vinyl ether..................................................... 20
Scheme 2-5. Reaction mechanism of the 6-membered glycerol acetal formation. ...................... 28
Scheme 3-1. Equilibrium constants of isomer interconversion. ................................................... 46
Scheme 4-1. From 6-membered GA isomer to its ketone counterpart. ........................................ 67
Scheme 4-2. Mechanism of Jones reactions [16].......................................................................... 67
Scheme 4-3. Mechanism of Swern oxidation [19, 23].................................................................. 69
Scheme 4-4: Alcohol oxidation by DMP [24]. ............................................................................. 70
Scheme 4-5. Mechanism of Oppenauer oxidation [28]. ............................................................... 71
Scheme 4-6. Mechanism of the RuO4-catalyzed oxidation of alcohols to aldehydes or ketones
[46]............................................................................................................................. 73
Scheme 4-7. Direct oxidation of alkene by ozone [84]. ............................................................... 82
Scheme 4-8. Direct oxidation of alcohols by ozone [76].............................................................. 82
Scheme 4-9. Structure of 2-tert-butyl-5-hydroxyl-1,3-dixane...................................................... 85
Scheme 4-10. RuO4-catalyzed oxidation under slightly basic condition. .................................... 96
Scheme 6-1. Processes of catalytic conversion of glycerol into useful chemicals. .................... 130
Scheme 6-2. Interconversion between DHA and glyceraldehydes [1]. ...................................... 133
Scheme 6-3. Overall reaction of the novel process to make DHA from glycerol. ..................... 134

xii

Chapter I. Background and Introduction
1.1. Introduction
The production of biodiesel from plant triglycerides leads to formation of by-product
glycerol, a versatile feedstock for producing chemicals such as propylene glycol and
epichlorohydrin. Glycerol obtained from biodiesel production usually contains a mixture of
methanol, water, inorganic salts (catalyst residue), free fatty acids, unreacted and partly reacted
triglycerides, methyl esters, and other compounds in varying quantities, depending on the
biodiesel process [1].
The price of high purity glycerol in the U.S. dropped from $1.00/lb to $0.40/lb between
1995 and 2005 [2], but has since rebounded as chemical manufacturers developed new uses for
the inexpensive feedstock instead of treating it as an environmental liability [3, 4].
Great effort has been made by scientists to search for new uses of glycerol. Suppes et al.
developed a green process to convert glycerol into 1,2-propylene glycol [5, 6]. This new
approach couples a copper-chromite catalyst with reactive distillation and uses hydrogen as coreagent to perform hydrogenolysis of glycerol. It first removes a water molecule from glycerol
acetol, and then adds a hydrogen molecule. This process uses a lower temperature (220 °C) and
lower pressure (<145 psi) than previous systems. In addition, the process converts glycerol to
propylene glycol more efficiently (yield >90%), and produces less by-product (ethylene glycol <
1.5%) than do similar routes or catalysts. Propylene glycol made from this process is
significantly cheaper than from petroleum.
Glycerol can also be used as a feedstock for the production of gaseous fuels [7].
Generally, the reforming of glycerol into H2 or a syn-gas (H2+CO) stream can be carried out by

1

1) aqueous-phase reforming [7-10] over various metal catalysts, 2) reforming in supercritical
water [7, 11, 12], or 3) steam reforming [13-15]. However, high temperature and high pressure
are required in these processes to achieve a reasonable yield and selectivity.
Other uses of glycerol include converting it into alkanes, acetaldehyde, short chain
alcohols, and acetone, etc. [11, 16-18]. Zhou and coworkers [19] reviewed the pathways for
obtaining various useful chemicals from glycerol, summarized in Scheme 1-1.
In addition to chemicals, glycerol can be converted to effective fuel additives that
improve combustion and emission properties and provide additional energy content. Fernando et
al. [3] discussed the possibility of blending glycerol derivatives such as propanediol and
propanol with gasoline as an automotive fuel. They obtained fuel blends with high energy value
and with octane numbers exceeding 100. However, they did not investigate the cold weather
properties of those fuel mixtures.
Noureddini developed a process to derive ether products from glycerol and isobutylene
using acid catalyst [20]. He found that mixing these ethers with biodiesel decreased its viscosity
and cloud point. Others [21] have reported that glycerol-ethers also to reduce particulate
emission from diesel and biodiesel combustion. However, the drawback of this approach is that
isobutylene is an expensive feedstock, which will increase the overall production cost.
Glycerol triacetate has been synthesized and tested as fuel additive [22]. The
experimental results show that this acetate product could improve cold flow and viscosity
properties of biodiesel. Delgado [23] described a procedure to make glycerol acetal (GA) and
glycerol acetate mixtures, and blended them with biodiesel in concentrations of approximately
0.5-20 wt%. He found the mixture could reduce the pour point and viscosity of biodiesel at low
temperatures.

2

Scheme 1-1. Processes of catalytic conversion of glycerol into useful chemicals [19].
HO

O

HO

OH

OH

O

O

OH

O

OH

HO

O
OH OH
dihydroxyacetone glyceraldehyde tartronic acid hydroxyethanoic acid
OH
oxidation
O
OH
OH
HO
OH
HO
OH HO
HO
OH
O HO O
catalyst
formic acid
O
O O
O
O
glyceric acid hydroxypyruvic acid mesoxalic acid oxalic acid

hydrogenolysis
catalyst

O

OH

1,3-propanediol

HO

OH

ROH

ethylene glycol

R=C1-C3

O

OH

acetol

catalyst

OH

HO

1,2-propanediol

dehydration

pyrolysis,
gasification

OH

HO

acrolein

CnH2n+2

CnH2n

ROH

(CO+H2)

(C+H2)

alkane

olefin

alcohol

syngas

carbon+hydrogen

catalyst

OH
OH

transesterification,HO
esterification
catalyst

OH

catalyst

OH

catalyst

O

OH

catalyst

O

OH

O

O

OH

OH

Ci

O

OH
j

O

O

OH

OH

glycerol carbonate

3

O

O

O

OH
diethers

glycerol 1-monoethers

O

O
OH

O

OH

O
carboxylation

O

monoclycides α -monobenzoyl glycerol glycerol dimethacrylate

monoethers
oligomerization,
polymerization

O

O
O

O

O
etherification

R HO

O

triether

polyglycerol methacrylates

It is well known that glycerol can be converted into cyclic acetals or ketals via reaction
with aldehydes, ketones or another acetal compound [24-26]; for example, 2-methyl-4hydroxymethyl-1,3-dioxolane and 2-methyl-5-hydroxy-1,3-dioxane (Scheme 1-2) are formed via
reaction of glycerol with acetaldehyde. This class of compounds has been proven to have
desirable properties as fuel additives via improving pour point and viscosity of biodiesel, and
reducing particulate emissions from fuel combustion [23, 27, 28].
Scheme 1-2. Dioxane and dioxolane isomers of glycerol acetal
O

O

HO

O

O

OH
2-methyl-4-hydroxymethyl-1,3-dioxolane

2-methyl-5-hydroxy-1,3-dioxane

In 2008, a patent [29] filed by Dr. Miller’s group at MSU proposed the concept of
making glycerol acetals (GA) and biodiesel in large scale using a reactive-distillation (RD)
technique. In the route they designed (Figure 1-1), the reaction mixture from a pre-reactor, which
contains mostly transesterified oil and crude glycerol, is introduced to the top of the RD column.
At the same time, a mixture of alcohol and GA forming reagent (aldehyde or acetal) is
introduced to the column of the bottom. The reactive zone in this column is packed with solid
catalyst packing. Residual oil and alcohol further react in this zone to achieve complete
transesterification. Simultaneously, GA is synthesized in the reactive zone. During the operation,
unreacted reagents are separated for recycling and the fuel mixture which consists of biodiesel
and GA products are obtained at the bottom of the RD column.

4

Figure 1-1. Production of biodiesel and glycerol acetal via reactive distillation [29]

1.2. Scope of this dissertation
In this study, effort was initially dedicated to assess and demonstrate the viability of the
process proposed for continuous production of glycerol acetal with biodiesel. To achieve this
goal, two synthetic routes were examined, i.e. glycerol acetalization with acetaldehyde and
transacetalization with 1,1-diehtoxyethane. The latter approach, which is water-free and has a
nearly 100% equilibrium yield, was selected for kinetic study and a kinetic model was then
proposed to describe the reaction. At the same time, an elevated pressure reactive distillation
column was designed and constructed for pilot scale demonstration of glycerol acetal production.
Under ordinary conditions, GA product obtained from either glycerol acetalization or
transacetalization is a mixture of dioxane and dioxolane isomers. A continuous reactive

5

distillation process was then developed to obtain the 6-membered dioxane isomer as the sole
product of GA formation reaction. In the process of investigating the interconversion and
separation of these isomers, we noticed that oxidizing the 6-membered dioxane GA isomer to its
ketone product (2-methyl-1,3-dioxan-5-one) gives the precursor to 1,3-dihydroxyacetone (DHA)
(Scheme 1-3), a high value specialty chemical that is the active ingredient in sunless tanning
lotions.
Scheme 1-3. 2-methyl-1,3-dioxan-5-one and 1,3-dihydroxyacetone (DHA).

O
O
O

O
OH

OH

1,3-dihydroxyacetone (DHA)

2-methyl-1,3-dioxan-5-one

Because conventional DHA production from glycerol fermentation has low productivity
and high production cost [32], we perceived an opportunity to develop a new, economical
process and thus changed our focus from production of GA to DHA. Therefore, a variety of
oxidation experiments were carried out to determine an efficient and green pathway to obtain the
ketone precursor from cis-6. This work forms the basis for the development of a novel process to
produce DHA from glycerol in yields of 90+%.

6

References

7

References
1. Chiu, C, Goff, M. and Suppes, G. (2005). Distribution of methanol and catalysts between
biodiesel and glycerin phases. AICHE Journal. 51, 1274-1278.
2. Tang, S, Boehme, H, Lam, H. and Zhang, Z. (2009). Pichia pastoris fermentation for phytase
production using crude glycerol from biodiesel production as the sole carbon source.
Biochemical Engineering Journal. 43, 157-162.
3. Fernando S, Adhikari, S, Kota, K. and Bandi, Z. (2007). Glycerol based automotive fuels
from future biorefineries. Fuel. 86, 2806-2809.
4. Silva, G, Mack, M. and Contiero, J. (2009). Glycerol: A promising and abundant carbon
source for industrial microbiology. Biotechnology Advances. 27, 30-39.
5. http://www.epa.gov/greenchemistry/pubs/pgcc/winners/aa06.html
6. Suppes, J, Sutterlin, W. and Dasari, M. (2005). Method of producing lower alcohols from
glycerol, US Patent Application. 20050244312.
7. Fernandez, Y, Arenillas, A, Diez, M, Pis, J. and Menendez, J. (2009). Pyrolysis of glycerol
over activated carbons for syngas production. Journal of Analytical and Applied Pyrolysis.
84, 145-150.
8. Huber, G. and Dumesic, J. (2006). An overview of aqueous-phase catalytic processes for
production of hydrogen and alkanes in a biorefinery. Applied Catalysis A. 111, 119-132.
9. Cortright R, Davda R. and Dumesic J. (2002). Hydrogen from catalytic reforming of
biomass-derived hydrocarbons in liquid water. Nature. 418, 964-967.
10. Huber, G, Shabaker, H. and Dumesic, J. (2003). Raney Ni-Sn catalyst for H2 production
from biomass-derived hydrocarbons. Science. 30, 2075-2077.
11. Buhler, W, Dinjus, H, Ederer, A, Kruse, A, Mas, C. and Supercrit, J. (2002). Ionic reactions
and pyrolysis of glycerol as competing reaction pathways in near- and supercritical water.
The Journal of Supercritical Fluids. 22, 37-53.
12. Xu, X, Matsumura, Y, Stenberg, J. and Antal, M. (1996). Carbon-catalyzed gasification of
organic feedstocks in supercritical water. Industrial and Engineering Chemistry Research. 35,
2522-2530.
13. Douette, A, Turn, S, Wang W. and Keffer, V. (2007). Experimental investigation of
hydrogen production from glycerin reforming. Energy & Fuels. 21, 3499-3504.
14. Hirai, T, Ikenaga, N, Miyake, T. and Suzuki, T. (2005). Production of hydrogen by steam
reforming of glycerin on ruthenium catalyst. Energy & Fuels. 19, 1761-1762.
8

15. Franco, C, Pinto, F, Gulyurtlu I. and Cabrita I. (2003). The study of reactions influencing the
biomass steam gasification process. Fuel. 82, 835-842.
16. Pagliaro, M, Ciriminna, R, Kimura, H, Rossi, M. and Pina, C. (2007). From Glycerol to
Value-Added Products. Angewandte Chemie - International Edition. 46, 4434-4440.
17. Garciaa, R, Besson, M. and Gallezota, P. (1995). Chemoselective catalytic oxidation of
glycerol with air on platinum metals. Applied Catalysis A. 127, 165-176.
18. Paine, J, Pithawallaa, Y, Naworala J. and Thomas, C. (2007). Carbohydrate pyrolysis
mechanisms from isotopic labeling: Part 1: The pyrolysis of glycerin: Discovery of
competing fragmentation mechanisms affording acetaldehyde and formaldehyde and the
implications for carbohydrate pyrolysis. Journal of Analytical and Applied Pyrolysis. 80,
297-311.
19. Zhou, C, Beltramini, J, Fan, Y. and Lu, G. (2008). Chemoselective catalytic conversion of
glycerol as a biorenewable source to valuable commodity chemicals. Chemical Society
Reviews. 37, 527-549.
20. Noureddini, H. (2001). Process for producing biodiesel fuel with reduced viscosity and a
cloud point below thirty-two (32). degrees Fahrenheit. US Patent. 6015440.
21. Olah, G. (1995). Cleaner burning and cetane enhancing diesel fuel supplements. US Patent.
5520710.
22. Melero, J, Grieken, R, Morales, G. and Paniagua M. (2007). Acidic mesoporous silica for the
acetylation of glycerol: Synthesis of bioadditives to petrol fuel. Energy and Fuels. 21, 17821791.
23. Delgado J. (2003). Procedure to obtain biodiesel fuel with improved properties at low
temperature. US Patent. 20030167681.
24. Deutsch, J, Martin, A. and Lieske, H. (2007). Investigations on heterogeneously catalysed
condensations of glycerol to cyclic acetals. Journal of Catalysis. 245, 428-435.
25. Garcia, E, Laca, M, Perez, E, Garrido, A. and Peinado, J. (2008). New class of acetal derived
from glycerin as a biodiesel fuel component. Energy & Fuels. 22, 4274-4280.
26. Silva, C, Gonçalves, V. and Mota C. (2009). Water-tolerant zeolite catalyst for the
acetalisation of glycerol. Green Chemistry. 11, 38-41.
27. Delfort, B, Durand, I, Jaecker, A, Lacome, T, Motagne, X. and Paille, F. (2005). Diesel fuel
compounds containing glycerol acetals. US Patent. 6890364.
28. Delfort, B, Durand, I, Jaecker, A, Lacome, T, Motagne, X. and Paille, F. (2006). Diesel fuel
compounds contain glycerol acetal carbonates. US Patent. 7097674.
29. Miller, D, Peereboom, L, Kolah, A. and Asthana, N. (2008). Process for production of a
9

composition useful as a fuel. US Patent. 7321052.
30. Ferreira, A, Barbe, J. and Bertrand, A. (2002). Heterocyclic acetals from glycerol and
acetaldehyde in port wines: Evolution with aging. Journal of agricultural and food chemistry.
50, 2560-2564.
31. Aksnes, G, Albriktsen, P. and Juvvik, P. (1965). Studies of cyclic acetal and ketal isomers of
glycerol. Acta Chemica Scandinavica. 19, 920-930.
32. Garcia, R, Besson, M. and Gallezot, P. (1995). Chemoselective catalytic oxidation of
glycerol with air on platinum metals. Applied Catalysis A: General. 127, 165-176.

10

Chapter II. Formation of Cyclic Glycerol Acetals from
Acetaldehyde or 1,1-Diethoxyethane
2.1. Introduction and background
Glycerol acetals can be synthesized by condensation with aldehydes or ketones, or by
–OH exchange with other acetal compounds. The direct condensation of glycerol with aldehydes
or ketones is known as acetalization [1-4], and the exchange of alcohols from another acetal is
called transacetalization [5]. These two reactions are demonstrated in Scheme 2-1.

Scheme 2-1. Acetalization and transacetalization reaction

O

R
R'

O
O
OH

HO
OH

+

R

OH

+

+

R'

O

R

HO
O

H2O

R'

O
R'
O
R'
OH

HO
OH

+

R

O

OH
O

+

R

+
O

R'

HO
O

11

2 ROH

Early investigations indicated that the acetal product of glycerol formed via either process
are composed of two cyclic compounds, a 5-membered 1,3-dioxolane and a 6-membered 1,3dioxane [4, 6-8].
In this work, we present results of glycerol acetal synthesis using acetaldehyde or 1,1diethoxyethane (DEE). With each of the reactants, the synthesized 4-hydroxymethyl-2-methyl1,3-dioxolane and 5-hydroxy-2-methyl-1,3-dioxane products contain cis- and trans isomers,
giving a total of four conformational isomers as shown in Table 2-1 [4, 9].

Table 2-1. cis- and trans- conformational structures of glycerol acetal
Structure

IUPAC name

Abbreviation

cis-4-hydroxymethyl-2-methyl-1,3-dioxolane

cis-5

trans-4-hydroxymethyl-2-methyl-1,3-dioxolane

trans-5

cis-5-hydroxy-2-methyl-1,3-dioxolane

cis-6

trans-5-hydroxy-2-methyl-1,3-dioxolane

trans-6

OH
O

O

OH
O

O

OH
O

HO

O

O

O

In contrast to direct acetalization, the transacetalization reaction of glycerol with another
acetal does not produce water (Scheme 2-1) and has a higher equilibrium constant than the direct
12

route. Because biodiesel production and subsequent purification are best accomplished in a
water-free environment [10], transacetalization provides the opportunity to combine glycerol
acetal formation with biodiesel production in an integrated operation without water formation.
In order to assess the viability of converting glycerol to its acetal products and provide a
theoretical basis for designing the reactive distillation process proposed in Dr. Miller’s patent
[11], it is necessary to better understand the transacetalization reaction. Since very few reports
have been published regarding the kinetics of this reaction [4, 6, 8], we have undertaken here a
kinetic study of glycerol transacetalization with DEE using Amberlyst-15 cationic exchange
resin as the catalyst. Effects of various reaction conditions, including reaction temperature, molar
ratio of reactants, and catalyst loading have been examined along with reactant miscibility and
measurement of the reaction equilibrium constant. A kinetic model describing the reaction is
presented; results are compared with acetal formation via direct condensation with acetaldehyde.
The results presented in this work provide information to understand similar
transacetalization processes between glycerol and other acetals, and establish the theoretic basis
for development of large-scale glycerol acetal production process.

2.2. Experimental section
2.2.1. Materials
Anhydrous glycerol (99.5%), 1-propanol (99.9%) and 1-butanol (99.9%) were obtained
from J.T. Baker. Ethanol (99.5%) and 1,1-diethoxyethane (99%) were purchased from Pharmco
and Aldrich, respectively. Acetaldehyde ( >99.5%) and amberlyst-15 cationic exchange resin
3

(16-50 mesh, 1.2g/cm ) was purchased from Sigma-Aldrich. Amberlyst-15 cationic exchange
resin was investigated as the catalyst. Before being used in reactions, it was pretreated at 90°C in

13

a vacuum oven for 2 hours. Hydrogen (99.999%) and helium (99.999%) were used for GC
analysis; both were purchased from Linde Gas, Inc.
2.2.2. Experiments
Glycerol acetalization with acetaldehyde was carried out in a 500 ml glass reactor cooled
with ice water bath. To initiate an experiment, 200 ml of reactants (acetaldehyde and glycerol)
plus 5% (weight to glycerol) Amberlyst-15 cationic exchange resin were added. The mixture was
then stirred for 4 h for glycerol conversion.
Glycerol transacetlization with DEE was carried out using a Parr 5000 Multireactor
System (Parr Instrument Company, Moline, IL), which includes six 75 ml stainless steel
autoclaves. To initiate an experiment, 60 mL of reactants (DEE, glycerol and solvent) plus
catalyst were placed in an autoclave and the vessel was sealed. The autoclave was heated at
2°C/min to the desired temperature and held for 5 min before the stirrer was activated to begin
the reaction. Essentially no reaction took place during this heating period.
One ml liquid aliquots were taken from a sample port for analysis during reaction using a
disposable syringe. Catalyst taken up in samples was eliminated by filtration to prevent further
reaction in the sample solution. The samples were then diluted and analyzed by gas
chromatography using the method described below.
At the end of reaction, the material in the reactor was collected and the catalyst was
removed by filtration. The solution was then evaporated in a rotary evaporator at 70°C under 25
mmHg to remove volatile compounds such as ethanol and DEE. The final product was then
weighed and analyzed by GC.

14

2.2.3. Analysis
Gas Chromatography
1-Butanol and 1-propanol were used as the internal standard and solvent, respectively, for
this analysis. To prepare samples for analysis, 0.16 ml 1-butanol and 0.2 ml of reaction sample
were added to a volumetric flask and the mixture was diluted with 1-propanol to 10 ml of
solution. A Perkin-Elmer Sigma 2000 gas chromatograph with FID and a DB-Wax capillary
column (15m × 0.53mm × 1.0 µm) was used for analysis. Ultra high purity helium (30 mL/min)
was used as the carrier gas; the column temperature program involved holding at 40°C for 3min,
heating at a rate of 30°C/min to 220°C, and holding there for 1 min. The total program time was
10 min. On-column injection was applied to analyze the samples. Injector and detector chamber
temperatures were set at 60°C and 300°C, respectively.
Gas chromatography – mass spectroscopy
The structure of the primary reaction products was identified using GC-MS analysis (GCAgilent 6890, MS-Agilent 5973) by injecting a 1µL sample onto a DB-wax capillary column
(30m × 0.25mm × 0.25 µm). The column temperature for GC-MS analysis included holding at
40°C for 5 min, heating at 20°C/min to 250°C and then holding for 10 min. Total program time
was 25.5 min. Injector temperature was set at 240°C. Electron impact (EI) mass spectra were
monitored from 20 to 300 m/z. The MS source was set at 230°C.

15

2.3. Results and discussion
2.3.1. Direct acetalization versus transacetalization
Direct glycerol acetalization with acetaldehyde was carried out in an ice water bath at
0 °C because acetaldehyde has a boiling point of 21°C. Two reactions were carried out with
Amberlyst-15

cationic

exchange

resin

as

catalyst;

the

reactant

molar

ratio

of

glycerol:acetaldehyde was 1:1 and 1:4, respectively. After 4 h of reaction, the yields of glycerol
acetal were 54% and 84%, respectively.
On the other hand, glycerol transacetalization with 1,1-diethoxyethane (DEE) using
Amberlyst-15 cationic exchange resin as catalyst is a much faster reaction. At 25°C, a reaction
with molar ratio of glycerol:DEE = 1:1 gives nearly theoretical yield of glycerol acetal after 2 h.
More importantly, DEE is much easier to use than acetaldehyde because it is much less volatile
and no special handling is required. Further, no water is formed in the reaction of DEE with
glycerol. Thus, it can be concluded, as the two main approaches to synthesize glycerol acetals,
glycerol transacetalization with DEE is a much better choice for reaction than acetalization with
acetaldehyde. Therefore, DEE was chosen as the glycerol acetal forming reagent for the
following kinetic study.
2.3.2. Identification of products
Glycerol transacetalization with DEE produces ethanol and glycerol acetals. However,
some by-products were also obtained in this reaction. Figure 2-1 presents a GC analysis result of
a typical reaction mixture. Structures of these compounds were identified using mass
spectroscopy and are listed in Table 2-2. The detail of structure determination is presented in
Appendix I.

16

The desired products include 2-methyl-4-hydroxymethy-1,3-dioxolane (peaks E and J)
and 2-methyl-5-hydroxy-1,3-dioxane (peaks H and I). The two peaks corresponding to each of
them are the cis- and trans- conformational structure of the isomers. This result is in agreement
with Ferreira [4] and Camara’s work [9]. The authors obtained the same glycerol acetals in their
study of glycerol acetalization with acetaldehyde, and also detected the four GC peaks of the
conformational isomers.

Figure 2-1. GC analysis of glycerol acetalization products

17

Table 2-2. Key to compounds detected by gas chromatography (Figure 2-1)
Peaks
A

B

1,1-diethoxyethane
(DEE)

C

D
1-butanol

1-propanol

ethanol

E, J
cis- (E, top) and trans- (J,
bottom) 5-hydroxy-2methyl-1,3-dioxane

(internal
standard)

(solvent)

OH
O
O
OH

O

O

OH

OH

HO

O
O

F, G

H, I

K, L

M

byproducts GAA

cis- (H, top) and
trans- (I, bottom) 4hydroxymethy-2methyl-1,3-dioxolane

hemi-acetals of
glycerol

Glycerol

O
O

O

O

O

O

HO

O

OH

O

and

O

OH

OH

O

O

OH

HO

O

O

OH

HO

and

O

O

OH

O

Byproducts obtained in this reaction include 1) the hemi-acetals (peaks K and L) of
glycerol and DEE, which are generally accepted as the intermediates in acetal synthesis [12], and
2) two complex acetals of glycerol (peaks F and G, together denoted GAA). Formation of GAA

18

arises from the reaction of the remaining free hydroxyl group in glycerol acetals (E, H, I, J) with
DEE to form a second hemi-acetal and ethanol. This reaction is illustrated in Scheme 2-2.

Scheme 2-2. Formation of GAA from glycerol acetals and DEE.

O

O

O

O

HO

O
+

O

O

O

+

O
O

O

OH

O
O

OH

O

The existence of these by-products reveals that glycerol transacetalization with DEE
involves stepwise conversion of glycerol to hemi-acetal to glycerol acetal and then to GAA. In
general, the amount of GAA formed is dependent on the initial reactant composition. The more
DEE used, the more GAA obtained. This will be further discussed in the section discussing the
of molar ratio section.
Another important side reaction is related to DEE disproportionation to ethyl vinyl ether
(EVE). It has been reported [13] that at high temperature or in the presence of acid catalyst, vinyl
ether can be formed by decomposition of acetal. As for DEE, the decomposition brings one
molecule of ethanol and one molecule of EVE as shown in Scheme 2-3.
Scheme 2-3. Decomposition of 1,1-diethoxyethane

Acid or
O

O

DEE

O

+

Ethyl vinyl ether (EVE)
19

OH
Ethanol

Temperature has a great influence on this reaction. Generally, with a certain amount of
Amberlyst-15 loading, the higher the reaction temperature, the more vinyl ether is produced. In
this study, no vinyl ether was found if the reaction was performed below 45 °C.
EVE is reactive and volatile; Shostakovsky and coworkers [14] report that in the presence
of HCl, EVE could also react with glycerol and stepwisely form the same glycerol acetals as
DEE and glycerol (Scheme 2-4).
Scheme 2-4. Reaction between glycerol and ethyl vinyl ether.

O
O

HO
OH +

HO
OH

OH
+

acid
O

HO

O

O

acid

+
O

OH

HO
O

O
O

OH

If acetal formation is the only consequence of EVE formation, no loss of reactant would
be realized. Unfortunately, in the presence of Amberlyst-15 cationic exchange resin, EVE readily
oligomerizes and forms a viscous compound with deep color. In a control experiment in which
pure EVE was stirred with 1 wt% of Amberlyst-15 at room temperature, the colorless liquid
quickly turned green and then dark brown in color rapidly, along with an accompanying sharp
increase in viscosity.
The production of this viscous compound has been noticed in the transacetalization
experiments. Figure 2-2 displayed the transacetalization reactions after 24 h performed at
different temperatures. The color of the reaction mixture changed from transparent to dark as the

20

reaction temperature increased from 40 °C to 70 °C. Therefore, to minimize loss of DEE via this
route and limit the by-product formation, the glycerol transactetalization reaction with DEE is
best carried out at low temperature (< 45°C).

1

2

3

Figure 2-2. Glycerol transacetalization products after 24h. Reaction temperature, 1) 40 °C;
2) 50 °C; 3) 70 °C. Molar ratio of reactants, DEE:glycerol= 1:2. Amberlyst-15 cationic
exchange resin loading = 1 wt% of glycerol. For interpretation of the references to color in
this and all other figures, the reader is referred to the electronic version of this thesis (or
dissertation.

2.3.2. Reaction conditions
2.3.2.1. Reagent miscibility and use of ethanol as the reaction solvent
When mixed, pure glycerol and DEE forms two liquid phases. In order to examine the
influence of this partial miscibility on reaction rate, experiments were carried out with and
without ethanol as a solvent at conditions. Table 2-3 reports the number of phases present and
the apparent reaction rate constant at several times for several reaction mixtures.
Without ethanol present, there is an initial period of low rate as a result of the presence of
two liquid phases. As reaction proceeds, formation of acetals and liberation of ethanol solubilizes
21

the reactants, greatly accelerating the reaction such that equilibrium conversion is nearly
achieved in 40 min. This is consistent with a previous report [6] in which reaction products were
observed to homogenize the two reactant phases.

Table 2-3. Effect of ethanol as solvent for glycerol transacetalization reactions. Amberlyst15 loading = 1 wt% based on glycerol, reaction temperature = 50°C.
Reactions

Molar Ratio

Reaction time (min)

glycerol:DEE:EtOH

0

2

5

10

A

1:1:0

2

2

2

1

B

1:2:0

2

2

2

1

C

1:2:0.78

2

2

1

1

D

1:2:3

1*

1

1

1

A

1:1:0

-

-

-

1.9

B

1:2:0

-

-

-

0.5

C

1:2:0.78

-

-

4.5

3.4

D

1:2:3

-

9.0

5.2

4.3

Number of phases

Apparent reaction rate
5
constant min (×10
2

L /mol•g.cat•s)

* The mixture is in two phases at room temperature, but forms single phase at the reaction
temperature.

From the apparent reaction rate constant calculated from a differential analysis of the
concentration versus time profile, it is seen that the 1:1 glycerol:DEE mixture (no ethanol added)
reacts more rapidly than the 1:2 glycerol:DEE mixture, because the reactants form a single phase
earlier in reaction.
Solubility measurements at room temperature showed that for a mixture with 1 mole
glycerol and 1 mole DEE, it requires 2.5 moles of ethanol to form a single liquid phase; while for

22

a mixture with 1 mole glycerol and 2 moles DEE, the amount of ethanol required to form a
single phase is 3.7 moles.
Adding ethanol to the reactant mixture increases rate substantially because reaction takes
place in a single phase from the beginning of reaction. Even with dilution of reactants because
of the ethanol added, glycerol conversion exceeded 80% in both Reactions C and D after 10
minutes (versus ~35% in Reaction B without ethanol addition). These results are reflected in the
apparent rate constants in Table 2-3.
It is worth noting that addition of ethanol, a reaction product, lowers the equilibrium
glycerol conversion slightly, from 99.6% to 98.9% and 98.3% for Reactions B, C and D,
respectively. Despite this, further experiments in this study were carried in ethanol as a cosolvent to obtain intrinsic transacetalization kinetics in a single reaction phase.
2.3.2.2. Molar ratios of reactants/GAA production
One of the most important variables affecting yield and selectivity of glycerol
transacetalization is the glycerol:DEE molar ratio. Table 2-4 shows the equilibrium conversions
of glycerol and DEE for glycerol:DEE reactant molar ratios from 1.5:1 to 1:2.
Table 2-4. Comparison of equilibrium conversion for reactions with various reactant molar
ratios
Reactant molar ratio
Glycerol

DEE Ethanol

1.5

1
1

4

1

2

Time
(h)

Temperature
(°C)

Equilibrium
conversion

24

23

40

DEE
89.2%

78.6%

81.5%

92.7%

4

1%

Glycerol
64.3%

4

1

Amberlyst-15
loading (wt to
glycerol)

50.1%

Although the presence of excess DEE improves glycerol conversion, it also leads to
production of more by-product GAA. Figure 2-3 displays the GC analysis of the final products
from the reactions listed in Table 2-4. As can be seen, the production of GAA is low if DEE is
the limiting reactant. When DEE is in excess, however, the yield of GAA increased substantially,
such that GAA concentration in the final product increases by a factor of ten when the initial
DEE:glycerol molar ratio is increased from 1:1 to 2:1.

Figure 2-3. GC analysis of glycerol acetalization products. Reactions were carried out at
40°C with 1 wt% Amberlyst-15 loading (based on glycerol). Molar ratio,
glycerol:DEE:ethanol = A) 1.5:1:4; B) 1:1:4; C) 1:2:4. Reaction time = 24 h.

24

2.3.3. Mass transfer effects
2.3.3.1. Effect of agitation speed
In order to minimize external mass transport resistances during reaction, several reactions
were carried out at different stirring speeds. Figure 2-4 shows that reaction rate is essentially
constant when stirring speed exceeds 800 rpm, indicating that external mass transfer resistances
can be neglected. All kinetics experiments were thus performed at a stirring speed of 850 rpm.
500 rpm

800 rpm

1000 rpm

45%
40%
DEE conversion

35%
30%
25%
20%
15%
10%
5%
0%
0

10

20

30

40

50

60

Time (min)

Figure 2-4. Effect of stirring speed. Glycerol:DEE:ethanol = 1:1:12, catalyst loading =
1wt% (based on glycerol), temperature = 40°C.
2.3.3.2. Effect of catalyst loading & Internal Mass Transfer Resistances
Acid catalyst is essential for the transacetalization reaction to proceed. Control
experiments showed negligible formation of cyclic acetals after 24 h without catalyst present.
Experiments were carried out with several loadings of Amberlyst-15 cationic exchange resin (0.5
25

wt% to 3 wt% of the initial mass of glycerol present) using a 1:1 glycerol:DEE feed ratio in
ethanol (glycerol:DEE:ethanol = 1:1:12) to avoid GAA formation and at 40°C to avoid DEE
decomposition.
As shown in Table 2-5, the observed reaction rates per unit volume of catalyst are almost
identical (~0.06 mol/L.cat·s) in each experiment, signifying that all catalyst participates in the
reaction and that external mass transport resistances do not affect rate.
Table 2-5. Evaluation of internal mass transport resistances
Reactant molar ratio temperature
catalyst loading
observed reaction rate
(mol/L cat/s)
(°C)
weight (g) wt to glycerol
glycerol:DEE: EtOH

Φ wp

0.1835
40

0.06

0.30

0.1249

2%

0.06

0.32

0.0608

1%

0.06

0.34

0.0314

1:1:12

3%

0.5%

0.07

0.35

Analysis of mass transport resistances within the Amberlyst resin particles was carried
out by estimating the Weisz-Prater observable modulus (Φ wp) at the reaction conditions of
highest reaction rate (Table 2-5). Values of Φwp are in the range of 0.30-0.35, indicating that
only mild mass transfer resistances are observed. Details of this calculation are presented in
Appendix II.
2.3.4. Kinetic model
As discussed above, DEE decomposes and forms EVE at reaction temperatures above
50°C. To minimize DEE decomposition, all reactions performed for the kinetic model
development were carried out at temperatures at or below 40°C. In addition, DEE was used as

26

the limiting reactant to minimize byproduct GAA formation. Finally, because the reaction is
rapid, all experiments were carried out with large excess of ethanol (at least 12:1 EtOH:glycerol)
to reduce intraparticle mass transport resistances.
2.4.4.1. Estimation of equilibrium constant
Experiments were carried out at 25°C and at 40°C for 24 h to determine the
concentration-based equilibrium constant of the acetal formation reaction. In the relationship
used for the equilibrium constant (Eq. 1), the 5-membered and 6-membered GA isomers are
treated together as one compound (subscripted ‘Acetal’). The decline in equilibrium constant
with increasing temperature (Table 2-6) shows that acetal formation is exothermic, as increasing
temperature has an unfavorable effect on reaction equilibrium.
KC =

2
C Acetal , Eq ⋅ CEtOH , Eq

(1)

CGlycerol , Eq ⋅ CDEE , Eq

Table 2-6. Equilibrium constant of glycerol transacetalization reactions (Amberlyst-15
loading = 1 wt% based on glycerol)
Reactant molar
ratio
glycerol:DEE:EtOH

Equilibrium conversion
glycerol
25°C

40°C

DEE
25°C

lnKC

25°C

40°C

1.5:1:4

62.3% 64.3% 93.0% 89.2% 1350

830

7.2

6.7

1:1:4

81.1% 78.6% 83.8% 81.5% 1390

890

7.2

6.8

1:1:8

77.1% 75.2% 77.2% 73.4% 1410

1030

7.2

6.9

1:1:12

67.9% 65.1% 69.5% 62.4% 1250

960

7.1

6.9

Avg.

7.2

6.8

27

40°C

Observed KC

25°C 40°C

2.3.4.2. Formulation of model
It is generally believed [15, 16] that the first step of the reaction is initiated with the
protonation of an oxygen in DEE followed by the substitution of the ethoxy group by glycerol.
After that, the second protonation-substitution step cyclizes the product as a compound with ring
structure (Scheme 2-5).
Scheme 2-5. Reaction mechanism of the 6-membered glycerol acetal formation.

H+
O

O

O

+
O
H

+ HO

OH
OH

- EtOH

O

OH

O
OH

+
O
H

O

- H+

OH
OH

H+

+
O
H

OH

O

- EtOH

H +
O

O

- H+

O

O

OH
OH

OH

According to the reaction stoichiometry, the elementary rate expression for the reaction
of glycerol with DEE (Scheme 2-1) can be written as follows:

−

⎛
⎞
dCDEE
1 2
= η kWC ⎜ CDEE CGlycerol −
CEtOH C Acetal ⎟
dt
KC
⎝
⎠
28

(2)

where η is the effectiveness factor, k is the reaction rate constant per mass of catalyst, WC is the
weight of catalyst per volume of solution.
The large values of the reaction equilibrium constant (Table 2-6) indicate that the forward
reaction dominates over much of the reaction such that the rate expression can be expressed as

−

dCDEE
= η kWC ⋅ CDEE CGlycerol
dt

(3)

2.3.4.3. Model fit over course of reaction
When initial concentrations of glycerol and DEE are different, the concentration of
glycerol at any time can be expressed as
CGlycerol = CGlycerol ,0 − (CDEE ,0 − CDEE )

(4)

Combining Eq. (3) with Eq. (4) and integrating gives the following equation
⎛ CGlycerol ⋅ CDEE ,0 ⎞
⋅ ln ⎜
⎟ = η kWC ⋅ t
⎜ CGlycerol ,0 ⋅ CDEE ⎟
CGlycerol ,0 − CDEE ,0
⎝
⎠
1

(5)

As shown in Figure 2-5, a plot of the left-hand side of Eq. (6) vs. time gives a straight
line, indicating that the model reasonably describes the experimental data.

29

ln(C glycerol*CDEE,0/Cglycero,0*CDEE)/(Cglycerol,0-CDEE,0)

0.6
0.5
0.4
y = 0.0089x + 0.0059

0.3

2

R = 0.9951
0.2
0.1
0
0

10

20

30

40

50

60

70

Time (min)
Figure 2-5. Second order reaction kinetic plot for glycerol transacetalization with DEE at
40°C, glycerol:DEE:ethanol = 1.5:1:15, 1 wt% catalyst loading based on glycerol
For the case where initial glycerol and DEE concentrations are equal, the molar balance
for DEE is integrated to give

1
CDEE

−

1
CDEE ,0

= η kWC ⋅ t

(6)

A plot of 1/CDEE vs. time for reactions with molar ratio of glycerol:DEE:ethanol =
1:1:12 at 40°C at several catalyst loadings is given in Figure 2-6, again illustrating that the
kinetic model adequately describes the experimental data. This is confirmed by a parity plot in

30

Figure 2-7 of experimental vs. predicted reaction rates for all experimental data; most predicted
rates fall within ±20% of the experimental data at the same conditions.

0.5%

1%

2%

3%

1.8
y = 0.0144x + 0.8803

1.7

2

R = 0.9989

1.6
y = 0.0105x + 0.8831

1.5

2

1/C DEE (L/mol)

R = 0.9934
1.4
1.3
y = 0.0054x + 0.8847
2

1.2

R = 0.9839

1.1
y = 0.0029x + 0.8635

1.0

2

R = 0.9899

0.9
0.8
0

20

40

60

80

100

Time (min)
Figure 2-6. Second order reaction kinetic plot for glycerol transacetalization with DEE at
40°C, glycerol:DEE:ethanol = 1:1:12, catalyst loading based on glycerol; 0.5%, 1%, 2%,
3% respectively.

31

y=0.8x

y=1.25x

y=x

15
Model prediction

5

Reaction rate (×10 mol/g.cat•s)

20

10

5

0
0

5

10

15

20

5

Reaction rate (×10 mol/g.cat•s)
Experimental data
Figure 2-7. Parity plot between the reaction rates observed in experiments and the ones
predicted by the model.
2.3.4.4. Temperature dependence and activation energy

The kinetic model was applied to the transacetalization reaction for temperatures in the
range of 25°C to 40°C with a feed ratio of glycerol:DEE:ethanol = 1:1:12 (Figure 2-8). From the
slope of the lines in Figure 2-8, values of ηkWC were determined. Values of effectiveness factor
values (η) were calculated using Eq. (7) and the effectiveness factor Thiele modulus diagram
provided by Smith [17]. In this equation, Φwp is the Weisz-Prater value, and Φ s is the Thiele
modulus for a spherical pellet. The results are presented in Table 2-7.
32

Φ wp = ηΦ 2
s

(7)

25°C

30°C

35°C

40°C
y = 0.0054x + 0.8847

1.2

2

R = 0.9838
y = 0.0041x + 0.8828
2

R = 0.9907

1/C DEE (L/mol)

1.1

y = 0.0031x + 0.8834
2

R = 0.9737
1

y = 0.0017x + 0.8827
2

R = 0.9093
0.9

0.8
0

10

20

30

40

50

60

Time (min)
Figure 2-8. Kinetics plot for glycerol transacetalization reactions at different temperatures.
Conditions: initial molar ratio glycerol:DEE:ethanol = 1:1:12; catalyst loading 1% (based
on the weight of glycerol)

33

Table 2-7. Rate constant and Weisz-Prater value at different reaction conditions

Initial
Reaction rate
reaction rate WeiszCatalyst constant (k)
per unit mass Prater
Reactant Molar ratio Reaction loading
Effectiveness
5
of catalyst
temp.
value
( ×10
Factor (η)
(glycerol:DEE:EtOH) (°C)
(wt to
( Φ wp )
5
glycerol ) 2
(×10
L /mol•g.cat•s)
mol/g.cat•s)
1.5:1:15

40

15

0.39

0.74

9.7

9.5

0.30

0.81

2%

10.6

10.2

0.32

0.79

1%

11.4

10.7

0.34

0.78

0.5%

1:1:12

17.3

3%
40

1%

12.0

11.2

0.35

0.77

8.1

8.2

0.26

0.84

5.6

6.1

0.19

0.90

3.1

3.4

0.10

0.91

35
30

1%

25

According to Arrhenius equation
E ⎛1⎞
ln k = ln A − a ⎜ ⎟
R ⎝T ⎠

(8)

A plot of of ln(k) versus 1/T in Figure 2-9 gives an activation energy of Ea = 66.50
7 2

kJ/mol. And a pre-exponential factor of A = 1.5×10 L /mol•g.cat•s.

34

3
2.5

ln(k)

2
1.5
y = -7.9989x + 28.037
2
R = 0.9832

1
0.5
0
3.15

3.2

3.25

3.3
3

3.35

3.4

-1

1/T (×10 K )
Figure 2-9. Arrhenius plot of forward rate constant of transacetalization of glycerol with
DEE.

2.4. Conclusions
Glycerol transacetalization with DEE is a complete reaction that at 25°C, equilibrium
conversions of reactant can be achieved in 2 h. This reaction is described by a simple second
order kinetic model. The major products formed in the reaction are the cis- and trans- forms of 5membered and 6-membered cyclic acetals. The reaction is best carried out at low temperature to
avoid formation of ethyl vinyl ether from DEE and with DEE as the limiting reactant to avoid
over-acetalization of the primary acetal products. High selectivity of reactants to desired
products are achieved at these conditions.

35

References

36

References
1. Silva, C, Gonçalves, V. and Mota C. (2009). Water-tolerant zeolite catalyst for the
acetalisation of glycerol. Green Chemistry. 11, 38-41.
2. Hill, H. and Hibbert, H. (1922). Studies on reactions relating to carbohydrates and
polysaccharides. VI. Relative ease of formation of five- and six-membered heterocyclic
carbon-oxygen configurations. Journal of American chemical society. 45, 3117-3124.
3. Hill, H, Whelen, M. and Hibbert, H. (1928). Studies on the reactions relating to
carbohydrates and polysaccharides. XV. The isomeric benzylidene glycerols. Journal of
American chemical society. 50, 2235-2242.
4. Ferreira, A, Barbe, J. and Bertrand, A. (2002). Heterocyclic acetals from glycerol and
acetaldehyde in port wines: Evolution with aging. Journal of agricultural and food chemistry.
50, 2560-2564.
5. Piantadosi, C, Anderson, C, Brecht, E. and Yarbro, C. (1958). The preparation of cyclic
acetals by transacetalation. Journal of American chemical society. 80, 6613-6617.
6. Deutsch, J, Martin, A. and Lieske, H. (2007). Investigations on heterogeneously catalysed
condensations of glycerol to cyclic acetals. Journal of Catalysis. 245, 428-435.
7. Shevchuk, A, Podgornova, V. and Piskaikina, E. (1999). Features of synthesis of 4hydroxymethyl-1,3-dioxolanes from glycerol and aldehydes. Russian Journal of Applied
Chemistry. 72, 853-856.
8. Aksnes, G, Albriktsen, P. and Juvvik, P. (1965). Studies of cyclic acetal and ketal isomers of
glycerol. Acta Chemica Scandinavica. 19, 920-930.
9. Camara J, Marques J, Alves A. and Silva Ferreira, A. (2003). Heterocyclic acetals in Madeira
wines. Analytical and Bioanalytical Chemistry. 375, 1221-1224.
10. Freedman, B., Buttfield, R. and Pryde, E. (1986). Transesterification of soybean oil. Journal
of The American Oil Chemistry Society. 63, 1375-1380.
11. Miller, D, Peereboom, L, Kolah, A, Asthana, N. and Lira, C. (2008). Process for production
of a composition useful as a fuel. US Patent. 7321052.
12. Solomons, G. (1994). Fundamentals of organic chemistry. New York : J. Wiley & Sons.
13. Dakka, J. and Goris, H. (2006). Clean process for propanal oxidation to lactic acid. Catalysis
Today. 117, 265-270.
14. Shostakovsky, M, Zhebrovsky, V. and Medelyanovskaya, M. (1954). Reactions between
vinyl compounds and multifunctional compounds, Communication 1. Reactions of alkyl
vinyl ethers with glycerol. Russian Chemical Bulletin. 3, 137-141.
37

15. Du, J, Peng, Y. and Ding, Y. (2003). Synthesis and characterization of polymer networks
based on two kinds of macromonomer. Colloid and Polymer Science. 281, 90-95.
16. Cintrat, J, Leat-Crest, V, Parrain, J, Grognec, E, Beaudet, I. and Quintard, J. (2004).
Preparation of chiral 2-stannyloxazolidines and first considerations on the transacetalization
reaction mechanism. European Journal of Organic Chemistry. 4251-4267.
17. Smith, J. (1981). Chemical engineering kinetics, third edition. NY: McGraw-Hill.

38

Chapter III. Separation of Cyclic Glycerol Acetal Isomers &
Continuous Production of the 6-Membered Isomer
3.1. Introduction and literature review
Cyclic acetal formation from glycerol has been demonstrated in Chapter II. This class of
compounds has been shown to be good fuel additive as it can improve fuel combustion and
reduce particle emissions [1-4]. In a 2008 patent, Miller et al. [5] proposed a continuous process
to synthesize these compounds along with the formation of biodiesel to obtain a composition
useful as a fuel. In addition, glycerol acetal can be used as flavour [6] and the building-block for
surfactants [7]. Moreover, the six-membered dioxane isomer can serve as an important precursor
for the production of value-added chemicals from glycerol, such as 1,3-dihydroxyacetone [8] and
1,3-propanediol [9].
As mentioned in Chapter II, under ordinary conditions, the acetal products formed via
either acetalization or transacetalization pathway with glycerol are 2-substituted-5-hydroxy-1,3dioxane and 2-substituted-4-hydroxymethyl-1,3-dioxolane compounds, which when formed from
asymmetric ketones or aldehydes are produced as both cis- and trans- conformational isomers for
a total of four products.
If desired, a 5-membered ring dioxolane compound (2,2-dimethyl – 4-hydroxymethyl1,3-dioxolane) can be synthesized with high selectivity (> 99%) by reacting acetone with
glycerol [10, 11]. The selectivity to the five-membered ring dioxolane is unique to acetone; with
other acetal forming reagents, the 6-member ring dioxane compounds are always co-produced
along with their 5-membered dioxolane counterparts in quantities determined by the kinetics and
equilibrium constants of the interconversion isomerization reactions.

39

Aksnes et al. [12] calculated the isomerization equilibrium constant and pointed out that
the exchange between the two isomers only happens in the presence of acid catalyst. Deutsch et
al. [10] investigated the influence of reaction parameters on the selectivity of cyclic glycerol
acetal formation. They carried out acetalization and transacetalization reactions using various
solid catalysts, temperature, solvents with reagents including aldehydes, ketones and linear
acetals. In those reactions, the isomer ratio dioxane:dioxolane is usually 60:40. In addition, the
equilibrium ratio can be achieved via isomer interconversion even after the initial conversion of
reactant is over. For instance, in the experiments of glycerol acetalization with benzaldehyde
using Nafion-H NR-50 or Montmorillonite K-10, the authors received 50:50 mixtures of dioxane
and dioxolane isomers at reaction completion. However, when the mixtures were left at room
temperature for a few days, the ratio shifted to 60:40. They further proposed a mechanism
(Figure 3-1) for this interconversion as that in the presence of acid catalyst, the activated oxygen
in the benzylic position is the intermediate for this ring transformation.

benzyl cation (chain)
OH

OH

OH
O

O (+) OH

O

C6H5

H

+ H+

OH

- H+
O
C6H5

O

C6H5

H

- H+
OH

(+)
H

O

H

C6H5

(+)
O
H
H

O
C6H5

O
H

+ H+

cyclic oxonium cation
(5-membered cyclus)

cyclic oxonium cation
(6-membered cyclus)

Figure 3-1 Mechanism of the interconversion between glycerol acetal isomers [10].

40

Lab-scale direct separation of the dioxane and dioxolane isomers was carried out by
Aksnes and coworkers [12] in the 1960s using gas chromatography; large scale production of
pure GA isomers, however, is still a challenge.
It was reported by Showler and Darley [13] in the review of condensation of glycerol
with aldehydes and ketones that the most reliable method of obtaining the acetal isomers in their
pure form was by organic synthesis, although the cis isomer of 6-membered glycerol acetal could
be separated by crystallization.
The solubility of the cis-5-hydroxyl-2-phenyl-1,3-dioxane (cis-HPD) is significantly
different from the other three isomers in a 1:1 benzene−ligroin solution, and the isolation of cisHPD was accomplished by crystallizing it from this mixture [9, 14]. However, the yield of cisHPD in those work was only 25%. Moreover, the product obtained was not 100% pure, it
contained small amount of trans-HPD and very little the 5-membered isomers.
In another study, Carlsen et al. [8] isolated pure cis-HPD by crystallization in diethyl
ether. They used acid catalyst to re-equilibrate the mother liquor to obtain more cis-PHD. After
multi-step equilibration and crystallization, they achieved a yield of 46% of cis-HPD.
Recently, Deutsch et al. [10] discussed the feasibility of separating these isomers by
converting them into their derivatives with sufficient difference in physical properties, such as
boiling point, polarity, and solubility. However, these procedures require extra expensive
purification steps and are always accompanied by significant loss of target product.
It has been reported by Hill et al. [15] that as a mixture, GA product boils in the range of
176-187°C at atmospheric pressure. However, under reduced pressure, the isomers demonstrate
different volatilities. For instance, the authors collected two distinct products in a distillation

41

experiment with fraction I boiling at 86-88°C (21 mmHg) and fraction II boiling at 101-103.5°C
(20 mm Hg). Therefore, it is possible to separate GA isomers via vacuum distillation.
In this chapter, the synthesis of GA isomers via glycerol transacetalization with DEE
using Amberlyst-15 cationic exchange resin is described. The equilibrium compositions of the
GA isomer mixtures formed in transacetalization are characterized, and the use of reactive
distillation to produce the pure cis-6 dioxane isomer is presented. Based on these findings, a
continuous process was developed to make cis-6 dioxane isomer as the sole product of glycerol
condensation with acetaldehyde or DEE.

3.2. Material and methods
3.2.1. Materials
Anhydrous glycerol (99.5%), 1-propanol (99.9%), and 1-butanol (99.9%) were obtained
from J.T. Baker. Ethanol (Pharmco, 99.5%), 1,1-diethoxyethane (DEE, Sigma-Aldrich, 99%),
and sodium carbonate (Sigma-Aldrich, >99%) were used without further purification.
Amberlyst-15 cationic exchange resin (16-50 mesh, Sigma-Aldrich), used as the catalyst, was
pretreated at 90°C in a vacuum oven for 2 hours prior to use. Hydrogen (99.999%) and helium
(99.999%) were used for GC analysis; both were purchased from Linde Gas, Inc.

3.2.2. Synthesis of GA isomers
Cyclic GA isomers were synthesized at 45°C in a 20 L round bottom flask by adding
glycerol (3.1 kg, 33.9 mol), DEE (4.0 kg, 33.9 mol), and Amberlyst-15 cationic exchange resin
(120 g) and stirring for 4 hours. After reaction, the mixture was filtered to eliminate catalyst and
then subjected to evaporation in a rotary evaporator at 70°C under an absolute pressure of 25
mmHg to remove unreacted DEE, product ethanol, and any volatile byproducts or impurities.

42

The final mixture contained the four GA isomers and only trace amounts (less than 1 wt% total)
of unreacted glycerol and byproducts.

3.3.3. Distillation of GA isomers
A glass batch vacuum distillation column was constructed to separate the four GA
isomers (Figure 3-2). The column (height = 3.6 m, i.d. = 50 mm) was packed with 0.6 cm 316
Stainless Steel PRO-PAK protruded metal distillation packing (Scientific Development Co.). The
column reboiler, consisting of a two-liter round-bottom flask, was heated with a 750 W electric
heating mantle. Column pressure was adjusted using a Model RV8 vacuum pump (Edwards
Vacuum, Ltd.). The column was equipped with a condenser and a dry ice-cooled cold trap to
prevent loss of distillate through the vacuum pump. Vapor temperature was measured at the top
of the column using a thermometer.
To initiate a distillation, a measured quantity of GA isomers and boiling chips were
loaded in the reboiler. The column was then evacuated to a desired pressure, and the reboiler
contents were heated slowly (2°C to 3°C/min initially) to generate GA vapor. When steady state
was achieved throughout the column at total reflux (as indicated by steady temperatures in the
reboiler and throughout the column), the top product was withdrawn slowly, at a rate of 1 to 2
ml/min, by controlling the power to the reboiler. In all experiments performed in the column,
liquid reflux took place via condensation along the column wall; no external reflux was returned
to the column from the distillate product stream.

43

Figure 3-2. Vacuum distillation apparatus.

3.2.4. Analytical methods
GC analysis: 1-Butanol and 1-propanol were used as the internal standard and the
solvent, respectively, for the analysis. To prepare samples, 0.16 ml of n-butanol and 0.2 ml of
reaction sample were added to a 10 ml volumetric flask and the flask was filled with 1-propanol.
The analysis was carried out on a Perkin-Elmer Sigma 2000 gas chromatograph equipped with a
44

flame ionization detector. A DB-Wax capillary column (15m × 0.53mm × 1.0 µm) with oncolumn injection was used for analyses with the following temperature program: hold at 40°C
for 3min, ramp to 220°C at 30°C/min and hold for 1 min. Ultra high purity helium was used as
the carrier gas at a flow rate of 30 mL/min. Injector and detector chamber temperatures were set
at 60°C and 300°C, respectively.

Gas Chromatographic – Mass Spectroscopic Analysis: Structures of the reaction products
were identified using gas chromatography – mass spectroscopy (GC-Agilent 6890, MS-Agilent
5973). A 1µL sample was separated in a DB-wax capillary column with a temperature program
similar to that described above; electron impact (EI) mass spectra were monitored from 20 to 300
m/z. The MS source was set at 230°C.

3.3. Results and discussion
3.3.1. GA synthesis and analysis
The transacetalization of glycerol with DEE to produce the GA required for this study
was carried out at a near-stoichiometric molar ratio and proceeded nearly to completion. In the
final reaction mixture, only 3.8% of unreacted DEE and 1.5% of glycerol remained as unreacted
compounds. The combined yield of GA isomers was greater than 95%, and no products other
than GA isomers and ethanol were detected in the reaction mixture.
Following reaction and removal of volatile species (DEE and ethanol) by rotary
evaporation, GC-MS analysis of the product mixture showed the presence of four different
structures corresponding [16, 17] to the four GA isomers displayed in Table 2-1 of Chapter II.

3.3.2. GA isomerization equilibrium
In the presence of an acid catalyst, the four GA isomers interconvert and reach an
equilibrated composition [10, 12, 16, 17]. Figure 3-3a presents three GA samples, obtained from
45

the distillation experiments below, with different isomer compositions: Sample 1 is rich in cis-6
isomer, Sample 2 is rich in cis-5 and trans-5 isomers, and Sample 3 is rich in trans-5 and trans-6
isomers.
After stirring each sample for 5 hours at 298 K in the presence of 5 wt% (relative to GA)
Amberlyst-15 cationic exchange resin, identical isomer compositions were established for all
three samples (Figure 3-3b). At this temperature, the equilibrated concentrations of cis-6, cis-5,
trans-5 and trans-6 are 37±1%, 17±1%, 13±1%, and 33±1%, respectively. The molar
percentages of 6-membered isomers and 5-membered isomers are approximately 70% and 30%,
respectively, at equilibrium.
The concentration-based equilibrium constants of isomerization were calculated based on
Scheme 3-1 and Equations (1) to (3). The values are K1,2 = 0.46; K2,3 = 0.76; K3,4 = 2.54. With
these three values the equilibrium constants between any two isomers can be calculated.
OH

OH
O

O

K1,2

cis-6

OH
K2,3

O

O

O

O

K3,4
HO

trans-5

cis-5

O

O

trans-6

Scheme 3-1. Equilibrium constants of isomer interconversion.

K1,2 =

[cis -5]
[cis -6]

(1)

K 2,3 =

[trans -5]
[cis -5]

(2)

K3,4 =

[trans -6]
[trans -5]

(3)

46

cis-6

a)

cis-5

trans-5

trans-6

Concentration (wt%)

100
80
60
40
20
0
1

2

3

Sample number

cis-6

b)

cis-5

trans-5

trans-6

100

Concentration (wt%)

60
40
20

32

32

34

13

14

13

17

17

16

38

37

37

1

80

2

3

0
Sample number

Figure 3-3. Isomer composition of three GA samples before and after equilibration at 25°C.
a) – Before; b) – After equilibration.
47

Although equilibration of GA isomers was observed within a few hours in the presence of
5 wt% Amberlyst-15 catalyst, it is reported in the literature that isomerization can take many
years to reach equilibrium without acid present [16, 17] and that isomerization rate remains very
slow below a certain threshold concentration of acid [12]. Sufficient acid was therefore used in
all experiments to drive the isomerization reactions rapidly to equilibrium.
In addition to requiring a sufficient quantity of catalyst present, rapid isomerization also
requires vigorous agitation (to ensure mass transfer between solid catalyst and solution) and
elevated temperature. For example, Sample 1 took 12 hr to reach equilibrium at 0°C, but only
one hour at 80°C. Increasing temperature also shifts the equilibrium composition (Figure 3-4):
for Sample 1, increasing temperature from 0°C to 80°C decreased the sum of (cis- + trans-) 6membered isomers from 78% to 54%. However, the total quantity of cis- (cis-5 + cis-6) and
trans- (trans-5 + trans-6) isomers remained essentially constant over all temperatures at 55±2%
and 45±2%, respectively.

48

cis-6

cis-5

trans-5

trans-6

Concentration (wt%)

100
80
60

34

9
13

40
20

45

22

34

30

28

13

16

18

16

20

22

26

37

35

32

32

25

40

60

80

21

0
0

Temperature (°C)

Figure 3-4. Isomer equilibrium at different temperatures.
The higher concentration of cis- isomers at equilibrium has been attributed [18] to
intramolecular hydrogen bond formation. The cis-6 isomer is stabilized by a relatively strong
intramolecular hydrogen bond, with the corresponding trans-6 bond weaker and cis-5 and trans-5
bonds weakest and nearly identical. Therefore, the cis- isomers are slightly favored.

3.3.3. Separation of GA isomers
Two vacuum distillation experiments were carried out to obtain each of the GA isomers
in their pure forms. In the first distillation, 896 g (800 ml) of a GA mixture containing 35% cis-6,
18% cis-5, 15% trans-5, and 32% trans-6 isomers was placed in the reboiler. To avoid isomer
interconversion, the solution was maintained at a weakly basic state by adding 50 g of sodium
carbonate to the mixture. This initial distillation was carried out at a pressure of 20.5 ± 1.0 mm
Hg (except for Sample 1, where the pressure was temporarily reduced to 9.0 mmHg).

49

Twelve 60-mL distillate samples were collected over the course of the distillation and
analyzed by GC. Ninety percent of the initial charge to the still pot was recovered as distillate in
these samples (Table 3-1). The material remaining in the still pot consisted primarily of the trans6 isomer and glycerol; inclusion of this material gave essentially complete (100±2%) closure of
the material balance for the experiment.

Table 3-1. Summary of GA isomers recovered in the initial distillation
Initial GA charge
Mass

Product recovery in distillate

Percent
recovered in
distillate

(g)

Composition
(wt%)

Mass (g)

Cis-6

311

35

288

36

93

Cis-5

166

18

152

19

92

Trans-5

132

15

120

15

91

Trans-6

287

32

246

30

86

Total

896

100

806

100

90

Percent of total

(%)

Figure 3-5 gives the reboiler liquid temperature, vapor temperature at the top of the
column, and composition (via GC) of the distillate samples taken during the distillation. The
results clearly show the order of decreasing volatility of the GA isomers as cis-6, cis-5, trans-5,
and finally trans-6. As the more volatile isomers are removed from the still pot, liquid and vapor
temperature gradually increase.

50

Cis-6

Cis-5

Trans-5

Trans-6

Distillate vapor temp.

Reboiler temp.

90

180

80

160

70

140

60

120

50

100

40

80

30

60

20

40

10

20
0
0

50

100

150

200
Time (min)

250

300

350

Tem perature (°C)

200

0

Isom er concentration (w t% )

100

400

Figure 3-5. Initial distillation of GA isomer. Column pressure was 20.5±1.0 mmHg (except
for the first sample, where the pressure was 9.0 mmHg).
The first three samples collected during the distillation contained high concentrations
(>90%) of the cis-6 isomer, with the first one composed of 99% cis-6. Likewise, the last two
samples contained >90% of the trans-6 isomer, with the final sample composed of 97% trans-6.
Clearly, these species separate well from the other isomers present in the mixture.
In Figure 3-5, it can be seen that pure cis-5 isomer and trans-5 isomer were not obtained
in the distillation experiment, as no sample contained more than 75 wt% cis-5 isomer or more
than 60 wt% trans-5 isomer. A second distillation was therefore performed at a lower pressure
(7.7±0.3 mmHg) in order to separate and obtain pure samples of the 5-membered GA isomers.
In this run, a charge of 914 g (816 ml) of a mixture of 17 wt% cis-6, 38 wt% cis-5, 43 wt% trans5, and 2 wt% trans-6 isomers was fed into the reboiler along with 50 g of sodium carbonate. At
the lower pressure, the liquid temperature was about 20°C lower than in the first distillation.
51

From this distillation, a total of 744 g of distillate product (81% of that fed) was collected
in 21 total samples with an average sample size of 31 mL (Table 3-2). Recovery of unseparated
GA isomers from the reboiler again gave essentially 100% closure of the material balance.

Table 3-2. Summary of GA isomers recovered in the second distillation
Initial GA charge

Product
distillate

Mass (g)

Composition (wt%)

Mass (g)

recovered

in Percent
recovered in
distillate
Percent of total (%)

Cis-6

154

17

152

20

99

Cis-5

351

38

336

45

96

Trans-5

391

43

252

34

65

Trans-6

18

2

4

1

21

Total

914

100

744

100

81

The lower pressure of this distillation significantly improved the quality of separation of
the cis-5 and trans-5 isomers (Figure 3-6). Several samples contained more than 90% of a 5membered isomer; among them, the sample at 201 min contained 93 wt% cis-5 isomer, and the
sample at 559 min contained 97 wt% trans-5 isomer.

52

Cis-6

Cis-5

Trans-5

Trans-6

Distillate vapor temp.

Reboiler temp.

90

80

80

70

70

60

60

50

50

40

40

30

30

20

20

10

10

0

0
0

50

100

150

200

250 300 350
Time (min)

400

450

500

550

Temperature (°C)

100

90
Isomer composition (w t%)

100

600

Figure 3-6. Second distillation of GA isomers. Column pressure: 7.7±0.3 mmHg.

The above distillation experiments show that it is possible to achieve a good separation of
GA isomers by optimizing operating conditions such as pressure in a packed distillation column.

3.3.4. Batch and continuous production and recovery of cis-6 isomer
As mentioned earlier, the 6-membered dioxane GA isomers are potential precursors for
several valuable chemical products. In the above distillation of GA isomers, the maximum
quantity of 6-membered isomers achievable is determined by the feed composition, since
reactions are intentionally suppressed by base. However, if interconversion of GA isomers is
catalyzed during distillation, it is possible to obtain high purity 6-member cis isomer at higher
purity than in an equilibrium batch reaction. Removing the most volatile cis-6 isomer by

53

distillation from an isomer mixture containing an acid catalyst provides a driving force toward
re-equilibration in the reboiler, thus producing more cis-6 isomer.

As long as the isomer

interconversion rate is faster than the distillation rate, the charge in the still pot will approach its
equilibrium composition and provide a stable cis-6 source such that cis-6 is the sole product
recovered in the distillate.
To demonstration of this concept, a batch reactive distillation was carried out by charging
1344 g (1200 ml) of a mixture of 32 wt% cis-6, 26 wt% cis-5, 20 wt% trans-5, and 22 wt%
trans-6 GA isomers along with 100 g of Amberlyst-15 cationic exchange resin to the reboiler pot
of the glass column. For this experiment, the column pressure was maintained at 1.2±0.2 mmHg.
During the distillation, 16 distillate samples were collected with an average volume of 50 ml; the
average distillate rate was maintained at approximately 2 ml/min.
As shown in Figure 3-7 and Table 3-3, the reactive distillation experiment produced a
high concentration of cis-6 isomer (from 89% to 96%) in all samples; the total quantity of
distillate produced (895 g) contained 93 wt% cis-6 isomer. The other species in the distillate was
the cis-5 isomer; no trans- isomers were collected. The low pressure of this experiment (1.2 mm
Hg) resulted in low reaction temperatures of 42–52°C. Some fluctuation in temperature, due to
pressure fluctuation, was observed, but overall the reboiler temperature remained relatively
constant over the course of the experiment. This is evidence that the interconversion rate is rapid
relative to the distillation rate, as the reboiler composition and thus the liquid bubble point was
essentially constant over the course of distillation.

54

cis-6

cis-5

trans-5

trans-6

Distillate vapor temp.

Reboiler temp.

90

80

80

70

70

60

60

50

50

40

40

30

30

20

20

10

10

0

0
0

100

200

300
Time (min)

400

500

600

Figure 3-7. Reactive-distillation of GA isomers. Column pressure was 1.2±0.2 mmHg.

Table 3-3. Summary of GA isomers recovered in reactive-distillation
Product recovered in
distillate

Initial GA charge
Mass (g)

Composition
(wt%)

Mass (g)

Percent of total

Percent
recovered in
distillate
(%)

Cis-6

430

32

836

93

194

Cis-5

349

26

59

7

17

Trans-5

269

20

0

0

0

Trans-6

296

22

0

0

0

Total

1344

100

895

100

67

55

Tem peratu re (°C)

100

90
Isom er com position (w t% )

100

Although only about two-thirds of the initial charge to the still pot was distilled, results to
that point indicate that if the reboiler pot is continuously charged with GA isomers to make up
the consumption, continuous interconversion of GA isomers and recovery of cis-6 product is
possible.
To demonstrate this, a continuous distillation experiment was therefore carried out. In
this run, 504 g (450 mL) of a mixture of 36 wt% cis-6, 16 wt% cis-5, 14 wt% trans-5, and 34
wt% trans-6 GA isomers was charged to the reboiler pot of the glass column along with 50 g of
Amberlyst-15 cationic exchange resin. For this experiment, the column pressure was maintained
at 2.9±0.2 mmHg. Once the distillate product started to condense at the top of the column, a
continuous feed of the same GA isomer mixture was introduced to the column reboiler to
maintain its liquid level constant. The feed rate was adjusted to match the distillate rate from the
column, which was approximately 1.8 mL/min. During the distillation, 16 distillate samples were
collected with an average volume of 59 mL.
As shown in Figure 3-8, this experiment produced samples with high concentration of
cis-6 isomer (from 90% to 96%) continuously during operation of the column. The reboiler
temperature and distillate vapor temperature were essentially constant during operation,
indicating that the reboiler composition was constant and that the column was operating at steady
state. The total quantity of distillate obtained was 1028 g, more than twice the amount of the
initial charge to the reboiler, demonstrating that continuous production of the cis-6 isomer from
the GA acetal mixture is feasible. Based on these results, it is fully expected that an appropriately
designed commercial-scale column with sufficient separation capacity would provide essentially
pure cis-6 isomer.

56

Cis-6

Cis-5

Trans-5

Trans-6

Distillate vapor temp.

Reboiler temp.

90

80

80

70

70

60

60

50

50

40

40

30

30

20

20

10

10

0

0
0

100

200

300
Time (min)

400

T em p eratu re (°C )

100

90
Iso m er co m p o sitio n (wt% )

100

500

Figure 3-8. Continuous reactive distillation of GA isomers. Column pressure = 2.9±0.2
mm Hg.

Discussion
A process concept to produce cis-6 isomer from glycerol is proposed in Figure 3-9. In
this process, glycerol and aldehyde or dialkyl acetal are introduced to Reactor 1 to produce
mixed GA isomers. Following Reactor 1, reaction mixture are heated to remove product ethanol,
water, and acetaldehyde. The mixture of GA isomers is then introduced to the reboiler of a
continuous reactive distillation column containing acid catalyst. The cis-6 isomer is produced as
a continuous distillate stream from the column.

57

Figure 3-9. Process concept for continuous production of 2-methyl-5-hydroxy-1,3-dioxane
(cis-6 isomer) from glycerol.
In this process, glycerol and acetaldehyde are inexpensive chemical substances that are
readily available. However, the challenge of using acetaldehyde is the formation of water along
with the acetals and the reaction is reversible. Shevchuk et al. [19] investigated the
thermodynamic equilibrium of glycerol acetalization with paraldehyde. They found that the
formation of GA is favored at high temperature, however, the highest equilibrium constant they
observed was 1.3, corresponding to a yield of 54%. Therefore, this reaction is most
advantageously conducted in a reactive distillation column with acetaldehyde in large amount
excess as suggested by Miller et al. [6]. In the reactive distillation process they proposed,
glycerol and acetaldehyde are introduced to the top and bottom of a reactive distillation column,
respectively. The reactive zone in this column is packed with solid catalyst packing. GA isomers
are synthesized in the reactive zone. During the operation, unreacted reagents and water are
58

separated for recycling from the top, and the product which consists of GA isomers can be
obtained at the bottom of RD column.
On the other hand, if a dialkyl acetal substance is used as surrogate for acetaldehyde, it
must be synthesized prior to the GA formation reaction. And then, the reaction of glycerol with
the acetal compound can proceed to completion and thus a simple stirred tank or packed bed
reactor suffices as Reactor 1.
It has been reported that such dialkyl acetals can be produced by the acid-catalyzed
acetalization reaction with 2 mol of monohydric alcohol and 1 mol of aldehydes [20]. In case of
formation of dialkyl acetal of acetaldehyde, recent research has been focused on using
heterogeneous catalysts and the readily available alcohols, such as methanol [21, 22], ethanol
[23-25] or butanol [26]. Thermodynamic and kinetic investigation indicates that with these
alcohols, the acetalization reaction is slightly exothermic and equilibrium-limited (Table 3-4).
Various solutions have been proposed to overcome this limit and improve the acetal productivity,
e.g. using organic solvents to modify the equilibrium in a continuous flow reactor [24];
application of fixed bed adsorptive reactor [22], moving bed reactor (MBR) [22, 26] or simulated
moving bed reactor (SMBR) [26] to integrate reaction and separation in a single unit apparatus,
etc. With these technologies, the equilibrium can be shifted toward the desired product, and thus
commercial scale production of dialkyl acetal from green natural resources is feasible.
Once the cyclic acetal forming reagent (acetaldehyde or its dialkyl acetal) is provided,
the process proposed in Reactor 1, Figure 3-9 can be operated in a semi-batch mode or a
continuous mode to match the distillation rate of the cis-6 isomer.

59

Table 3-4. Thermodynamic data of acetaldehyde acetalization reaction with alcohols [21, 23]
(Equilibrium constants are based on mol fraction).
Reaction Temperature (K)
293.15

Methanol /
acetaldehyde =
2:1 [21]

Equilibrium
conversion of
acetaldehyde

313.15

323.15

333.15

71.3

69.1

69.0

66.5

65.3

5.35

4.03

4.01

2.92

2.55

59.0

57.3

55.7

--

52.3

2.20

1.92

1.66

--

1.21

(%)
Equilibrium
constant

Ethanol /
acetaldehyde =
2.2:1 [23]

303.15

Equilibrium
conversion of
acetaldehyde
(%)
Equilibrium
constant

3.4. Conclusions
Four isomers are formed in the transacetalization of glycerol with 1,1-diethoxyethane. In
the presence of acid catalyst, the isomers interconvert to form an equilibrated mixture that
contains similar quantities of the four products. Distillation of the isomers shows that the cis-6
dioxane isomer (2-methyl-5-hydroxy-1,3-dioxane) is the most volatile, and is distilled
preferentially such that nearly pure cis-6 isomer is initially produced. When an acid catalyst is
added to the mixture of isomers such that interconversion can occur during distillation, it is
possible to produce 90+% cis-6 isomer in quantitative yield from glycerol. A larger column or
controlled reflux could lead to higher purities. This work demonstrates that a continuous process
is possible.

60

References

61

References
1. Delgado, J. (2003). Procedure to obtain biodiesel fuel with improved properties at low
temperature. US Patent Application. 20030167681.
2. Delfort, B, Durand, I, Jaecker, A, Lacome, T, Motagne, X. and Paille, F. (2005). Diesel fuel
compounds containing glycerol acetals. US Patent. 6890364.
3. Delfort, B, Durand, I, Jaecker, A, Lacome, T, Motagne, X. and Paille, F. (2006). Diesel fuel
compounds contain glycerol acetal carbonates. US Patent. 7097674.
4. Silva, P, Goncalves, V. and Mota, C. (2010) Glycerol acetals as anti-freezing additives for
biodiesel. Bioresource Technology. 101, 6225-6229.
5. Miller, D, Peereboom, L, Kolah, A, Asthana, N. and Lira, C. (2008). Process for production
of a composition useful as a fuel. US Patent. 7321052.
6. Climent, M, Velty, A. and Corma, A. (2002). Design of a solid catalyst for the synthesis of a
molecule with blossom orange scent. Green Chemistry. 4, 565-569.
7. Piasecki, A, Sokolowski, A, Burczyk, B. and Kotlewska, U. (1997). Chemical structure and
surface activity .30. Synthesis and surface properties of chemodegradable anionic surfactants:
Sodium (2-n-alkyl-1,3-dioxan-5-yl)sulfates. Journal of the American Oil Chemists' Society.
74, 33-37.
8. Carlsen, P, Sorbye, K, Ulven, T. and Aasbo, K. (1996). Synthesis of benzylidene-protected
dihydroxyacetone. Acta Chemica Scandinavica. 50, 185-187.
9. Wang, K, Hawley, M. and DeAthos, S. (2003). Conversion of glycerol to 1,3-propanediol via
selective dehydroxylation. Industrial & Engineering Chemistry Research. 42, 2913-2923.
10. Deutsch, J, Martin, A. and Lieske, H. (2007). Investigations on heterogeneously catalysed
condensations of glycerol to cyclic acetal. Journal of Catalysis. 245, 428-435.
11. Da Silva, C, Gonçalves, V. and Mota C. (2009). Water-tolerant zeolite catalyst for the
acetalisation of glycerol. Green Chemistry. 11, 38-41.
12. Aksnes, G, Albriktsen, P. and Juvvik, P. (1965). Studies of cyclic acetal and ketal isomers of
glycerol. Acta Chemica Scandinavica. 19, 920-930.
13. Showler, A. and Darley, P. (1967). Condensation Products of Glycerol with Aldehydes and
Ketones: 2-Substituted m-Dioxan-5-ols and 1,3-Dioxolane-4-methanols. Chemical Reviews.
67, 427-440.
14. Whelen, M. and Hibbert, H. (1928). Studies on the Reactions Relating to Carbonydrates and
Polysaccharides. XV. The Isomeric Benzylidene Glycerols. Journal of the American Oil
Chemists' Society. 50, 2235-2242.
62

15. Hill. H, Hill, A. and Hibbert, H. (1928). Studies on the reactions relating to carbohydrates
and polysaccharides. XVI. Separation and identification of the isomeric ethylidene glycerols.
Journal of the American Chemical Society. 50, 2242-2249.
16. Da Silva Ferreira, A, Barbe, J. and Bertrand, A. (2002). Heterocyclic acetals f from glycerol
and acetaldehyde in port wines: Evolution with aging. Journal of Agricultural and Food
Chemistry. 50, 2560-2564.
17. Câmara, J, Marques, J, Alves, A. and Da Silva Ferreira, A. (2003). Heterocyclic acetals in
Madeira wines. Analytical and Bioanalytical Chemistry. 375, 1221-1224.
18. Gelas, J. and Rambaud, R. (1969). Cyclic ketal glycerine derivatives 1. Synthesis and
spectroscopic study of hydroxylated dioxolanes and dioxanes. Bulletin de la Société
Chimique de France. 4, 1300-1304.
19. Schevchuk, A, Podgornova, V. and Piskaikina, E. (1999). Features of synthesis of 4hydroxymethyl-1,3-dioxolanes from glycerol and aldehydes. Russian Journal of Applied
Chemistry. 72, 853-856.
20. Distilleries, D. (1932). Process for the Manufacture of Acetal. US Patent. 1,850,836.
21. Gandi, G, Silva, V. and Rodrigues, A. (2005). Process development for dimethylacetal
synthesis: Thermodynamics and reaction kinetics. Industrial & Engineering Chemistry
Research. 44, 7287-7297.
22. Gandi, G, Silva, V. and Rodrigues, A. (2006). Synthesis of 1,1-dimethoxyethane in a fixed
bed adsorptive reactor. Industrial & Engineering Chemistry Research. 45, 2032-2039.
23. Silva, V. and Rodrigues, A. (2001). Synthesis of diethylacetal: Thermodynamic and kinetic
studies. Chemical Engineering Science. 56, 1255-1263.
24. Gomez, M, Arrua, L. and Abello, M. (2001). Synthesis of 1,1-diethoxyethane from
bioethanol. Influence of catalyst acidity. Reaction Kinetics and Catalysis Letters. 73, 143-149.
25. Gomez, M, Arrua, L. and Abello, M. (2004). Synthesis of 1,1-diethoxyethane using a
continuous flow reactor: catalyst deactivation and effect of feed purity and of solvent
addition. Journal of Chemical Technology and Biotechnology. 79, 391-396.
26. Graca, N, Pais, L, Silva, V. and Rodrigues, A. (2010). Oxygentated biofuels from butanol for
diesel blends: Synthesis of the acetal 1,1-dibutoxyethane catalyzed by Amberlyst-15 ionexchange resin. Industrial & Engineering Chemistry Research. 49, 6763-6771.

63

Chapter IV. Oxidation of 6-Membered Glycerol Acetal
4.1. Introduction and literature review
Chapter II demonstrated the feasibility of continuous production of pure cis-6 GA isomer
from glycerol. With the availability of this substance, new synthesis pathways can be developed
to obtain valuable chemicals. For instance, by oxidizing the remaining free hydroxyl group in
cis-6 and hydrolyzing the corresponding ketone product, 1,3-dihydroxyacetone (DHA) can be
produced [1]. Likewise, by dehydroxylation and hydrolysis, 1,3-propanediol can be synthesized
from cis-6 [2].
DHA is a very important chemical product. Because of its non-toxic tanning effect, it has
been extensively used in the cosmetics industry. DHA, as the active ingredient of sunless tanning
lotions, has a growing market demand. In addition, DHA can be used as a versatile building
block for the organic synthesis of a variety of fine chemicals (Figure 4-1) [3].
Currently, commercial synthesis of DHA is performed via glycerol fermentation
approaches utilizing microorganisms [3, 4]. Although the microbial processes can provide high
selectivity to DHA, they have some notable disadvantages such as low productivity and high
production cost [5]. For example, in the glycerol fermentation with Gluconobacter oxidans, it
takes ~32 h to obtain 87-94% yield of DHA [6]. Moreover, glycerol and DHA produced in these
processes can cause inhibition of microorganism growth [7]. Thus, although biochemical
technologies have increased the DHA production to a prominent level, the production rate has
not reached the level yield to meet the desired commercial demand. A process for direct
oxidation of glycerol is highly desirable.

64

H

O

(H)
HO

N

(OH)

HS

OH

OH

N

HO

OH

N

OH
R

N

O

O

H
H

O

HO

R

O

O
O

OR
R
O
OEt
HOCH2
O O
CH2OH
OEt
OH
HO
HO
MeO OMe

OH
CH3

H3C

Cosmetics
Pharmaceuticals

NH2
N

N
HOOC

O
N
H

Methotrexate

O

O

OR

O

HOOC

R

DHA

O
H3C

OH

HO

OH

HO

O

O

O

N

N

HO
NH2

NH2

OH

HO
O

OH
O
Ph

O
O

O-OP(OH)2

O
O

HO

DHA
N
CH3

R

O

Ph

OH

Figure 4-1. DHA as the building block to other chemicals [3].
4.1.1. Direct oxidation of glycerol to DHA
Scientists have attempted conventional chemical oxidation processes to make DHA from
glycerol using metal catalysts with air or molecular oxygen [8-12]. The goal is to achieve direct
oxidization of the secondary hydroxyl group in glycerol to a carbonyl bond. Unfortunately, this
reaction is always accompanied with primary alcohol oxidation and a variety of products (i.e.
glyceroaldehyde, glyceric acid, etc.) are produced. For example, in a liquid phase batch reactor
with Au/C or Au-Pd/C catalyst, glycerol was completely oxidized to a complex mixture after 5 h.
65

The main product was glyceric acid (45%); the selectivity to DHA, however, was only ~10% [810]. Platinum catalysts have also been tested in this application. Results show that they are more
selective for oxidation of primary alcohols [11, 12]. Kimura and co-workers reported that by
using 5 wt% Pt on carbon, it is possible to obtain DHA from glycerol, but the yield was only 4%
for DHA at 37% conversion [13].
Recently, it was found that introducing Bi to a Pt/C catalyst could enhance the selectivity
toward the oxidation of secondary hydroxyl groups [13]. With such catalysts, 50% selectivity to
DHA was achieved at ~70% conversion of glycerol [5, 13, 14]. Hu et al. [6] examined the effect
of various reaction parameters on Bi-Pt/C catalyzed DHA production from glycerol. They found
the optimum reaction conditions for maximum DHA yield were using a catalyst of 3% Pt0.6%Bi/C, reacting at 80°C, 30 psig O2, and initial pH ≈ 2. Under these conditions, a maximum
yield of DHA of 48.1% was achieved.
Overall, direct conversion of glycerol to DHA is faster than fermentation approaches.
However, such reactions always require expensive catalysts and harsh reaction conditions. In
addition, the yield of DHA is not very high because of competition for primary alcohol oxidation
product, and is thus not suitable to meet the requirements for large scale production.
Compared with glycerol, cis-5-hydroxy-2methyl-1,3-dioxane (cis-6) appears to be a
much better substrate for DHA synthesis. Because in the molecule of cis-6, the primary hydroxyl
groups are blocked by the cyclic acetal bonds, high selectivity of secondary alcohol oxidation
can be expected. Therefore, if a suitable oxidation reagent is found for this purpose, a more
productive process can be established to make DHA. The next section provides an overview of
alcohol oxidation methods.

66

4.1.2. Alcohol oxidation to carbonyl compounds
The desired product from oxidation of the 6-membered GA isomer, as shown in Scheme
4-1, is a ketone compound, 2-methyl-1,3-dioxan-5-one (G-tone).

Scheme 4-1. From 6-membered GA isomer to its ketone counterpart.
O

O

HO

O
O

O

6-membered GA isomer

G-tone

In the last two centuries, various approaches have been developed to oxidize alcohols to
their carbonyl products. Among them, the most commonly used reagents are the chromiumbased chemicals. In 1946, Jones et al. [15] discovered the “Jones reaction” (Scheme 4-2). With
the Jones reagent, which is chromium trioxide and diluted aqueous sulfuric acid in acetone
solution, primary and secondary alcohols can be oxidized to carboxylic acids and ketones,
respectively. In this reaction, alcohol is first converted into chromic acid ester, and then the ester
evolves to the final product [16, 17]. The drawback of the Jones reaction is the use of sulfuric
acid, which may cause side reactions in some acid-sensitive functional groups.

Scheme 4-2. Mechanism of Jones reactions [16].

67

Scheme 4-2. cont’d

In 1968, Collins reagent [18] was developed for alcohol oxidation. Collins reagent is the
complex of chromium trioxide with pyridine in dichloromethane. Chromium trioxide derivatives
are very strong oxidizing agents, however, the preparation procedure is quite dangerous in that it
has the potential to explode in the presence of some organics [19].
Other chromium-based reagents include pyridinium chlorochromate (PCC) [20] and
pyridinium dichromate (PDC) [21]. PCC is a yellow-orange solid, which is prepared from the
reaction of pyridinium, chromic acid and concentrated hydrochloric acid. This oxidant is very
effective in transforming alcohols to their carbonyl products in high yield at room temperature
[19]. For primary alcohols, PCC selectively oxidizes the substrates to aldehydes and rarely does
the over-oxidation to carboxylic acids.
PDC is produced by adding pyridinium into aqueous chromium trioxide solution. This
bright-orange solid is a very efficient alcohol oxidant. Over-oxidation of primary alcohols into
carboxylic acids can be prevented by controlling the amount of water in the reaction media [19].
In general, chromium-based reagents are cheap and easy to prepare. They can normally
guarantee a good yield of oxidation product. However, one noticeable disadvantage of these
reagents is that chromium compounds are highly toxic, corrosive, and carcinogenic. Large-scale
application may cause serious environmental problems. In addition, they are strong oxidants that

68

may lead to explosion under some conditions. Therefore, application of these reagents is usually
limited to the lab-scale.
The Swern oxidation [22] is a widely used chromium free oxidation procedure. It uses
oxalyl chloride-activated dimethyl sulfoxide (DMSO) and organic bases (i.e. triethylamine) as
the oxidation reagent. This reaction can be performed under mild conditions and has a wide
tolerance of functional groups [23]. The mechanism of this reaction involves the formation of a
five-membered ring transition state and then the decomposition of the ring to dimethyl sulfide
and the desired aldehyde or ketone products (Scheme 4-3) [19, 23]. Generally, Swern reactions
must be carried out at very low temperatures because of the poor thermal stability of activated
DMSO. Its application in organic synthesis is also somewhat limited by the stinky odor of the
byproduct dimethyl sulfide [19].

Scheme 4-3. Mechanism of Swern oxidation [19, 23].

In 1983, Dess and Martin [24] developed the hypervalent iodine compound Dess-Martin
periodinane (DMP, Scheme 4-4) reagent as the oxidant for alcohols. DMP has many advantages
over chromium- and DMSO-based reagents. It offers rapid, complete, selective and mild
oxidation of alcohols to aldehydes or ketones at room temperature [19, 25]. However, DMP has
some disadvantages in application including its high price, the risk of explosion if decomposed,
and the requirement of tedious purification steps after reaction [19].

69

Scheme 4-4: Alcohol oxidation by DMP [24].

TEMPO, 2,2,6,6-Tetramethylpiperidin-1-yl)oxyl, or (2,2,6,6-tetramethylpiperidin-1-yl)
oxidanyl, is broadly employed in organic synthesis as a catalyst for the oxidation of alcohols [19,
26]. This compound is prepared by oxidation of 2,2,6,6-tetramethylpiperidine. It is very efficient
in alcohol oxidation reactions such that they can reach equilibrium in a few minutes [19]. A
routine procedure of using TEMPO [19, 27] was established as mixing 1 mol% of a TEMPO
related stable nitroxide radical with excess of bleach (NaClO), KBr, and NaHCO3 in a biphasic
medium of water and dichloromethane or tetrahydrofuran. In this process, NaOCl is the
secondary oxidant that re-oxidizes the reduced TEMPO product. NaHCO3 is used to adjust the
pH value. The limitation of TEMPO related oxidation is the competing reactions by HClO
produced from NaClO [19]. In addition, separation of the reaction mixture, which contains
inorganic salts and organic substances, is quite a challenge for large-scale application.
Another classic and useful method for alcohol oxidation is Oppenauer oxidation [28, 29].
Oppenauer oxidation is the reverse process of Meerwein-Ponndorf-Verley (MPV) reduction. The
mechanism of Oppenauer oxidation, as shown in Scheme 4-5, is believed to involve a sixmembered cyclic transition state in which both the carbonyl compound and the reducing alcohol
are coordinated to the metal ion [28]. Generally, the oxidation reaction is performed with excess
of a hydride acceptors (ketone or aldehydes) and catalytic amounts of Al(III) alkoxide such as
aluminum t-butoxide, aluminum isopropoxide, or trimethyl aluminum. The advantage of
70

Oppenauer oxidation over the traditional oxidation processes is that this reaction is highly
selective for secondary alcohols and does not oxidize other sensitive functional groups such as
amines and sulfides [30].

Scheme 4-5. Mechanism of Oppenauer oxidation [28].

Oppenauer oxidation has been proved to be extremely successful in the oxidation of
sterols. However, its application in synthesis of ketones outside the field of steroids met a more
limited success because of less favorable thermodynamics [19] and side reactions such as
dehydration of alcohols and aldol condensations [28] which are induced by the basic character of
the aluminum alkoxides. In addition, trimethyl aluminum is very sensitive to moisture, so if it is
used as a catalyst, the reaction must be carried out under strictly anhydrous conditions [31]. Such
drawbacks greatly limited the application of Oppenauer oxidation.
Ruthenium-based catalysts have great potential for catalytic oxidation of various organic
compounds because of their wide scope of oxidation states (from -2, divalent to -8, octavalent)

71

and various coordination of geometries in each electron configuration [19, 32, 33]. A Latimer
diagram of the reduction potentials of ruthenium species in base are given in Figure 4-2 below:

Figure 4-2. Standard reduction potential of Ru species at various oxidation states [34].
For alcohol oxidation, the Ru-based catalytic systems include RuCl2(PPh3)3 with
oxidants such as N-methylmorpholine N-oxide [35], iodosylbenzene [36], and peracetic acid [37],
RuCl3 with hydrogen peroxide [38], K2RuO4 with peroxodisulfate [39], and RuO4 with sodium
hypochlorite [19, 40], etc.
Ruthenium tetroxide (RuO4), in particular, is a very powerful oxidant that can easily
convert secondary alcohols, including very hindered ones, to ketone and primary alcohols to
carboxylic acids very rapidly at room temperature [19].
Ruthenium tetroxide is a golden-yellow solid that has a high solubility in non-polar
solvents such as carbon tetrachloride (CCl4) and dichloromethane (CH2Cl2). The partition
coefficient of RuO4 in CCl4-water system is in the range of 59-60 [19, 40]. Because it has high
volatility (melting point = 25°C) and always reacts violently with common organic substrates,
RuO4 is generally dissolved in CCl4 and handled as solution [19, 32, 40]. This solution can be
easily prepared by stirring an aqueous solution of sodium metaperiodate (NaIO4) [19], sodium
hypochlorite (NaClO) [1] or sodium bromate (NaBrO3) [41, 42] with a suspension of hydrated
ruthenium dioxide [19, 32] or ruthenium (III) chloride [41, 42] in CCl4.
72

In early experiments, a stoichiometric amount of RuO4 was employed for oxidation of
organic compounds [19, 43]. Later, the procedure was simplified to using catalytic amounts of
RuO4 in the presence of NaIO4 as secondary oxidant in a biphasic water-CCl4 system [44]. In
1965, Parikh and Jones further developed this method in that, rather than being prepared
independently, RuO4 is generated in situ by oxidation with excess of NaIO4 of catalytic hydrated
RuO2 [45].
As shown in Scheme 4-6, the mechanism of RuO4-catalyzed oxidation of secondary
alcohols to aldehydes or ketones involves a two-electron transfer process [46-48]. It is generally
accepted that the reaction rate is determined by the hydride transfer process from alcohol to the
oxidant [49].

Scheme 4-6. Mechanism of the RuO4-catalyzed oxidation of alcohols to aldehydes or
ketones [46].

For secondary alcohol oxidations, one molecule of RuO4 can oxidize two molecules of
the substrates to the corresponding ketone product, while RuO4 is reduced to RuO2 [19]. The

73

reaction is usually performed in a biphasic system (i.e. water-CCl4) [33]. As shown in Figure 4-3,
in such a system, the substrate is oxidized by RuO4 in the organic phase. The co-product is a
black precipitate of RuO2, which can be re-oxidized to RuO4 by the secondary oxidant (i.e.
NaIO4, NaBrO3, NaClO) in the aqueous solution. The regenerated RuO4 then enters the organic
phase again for another cycle of oxidation. In this way, the alcohol substrate can be continuously
oxidized to the desired product with catalytic amounts of RuO4 while the secondary oxidant is
consumed. When the reaction is finished, the residue RuO4 species in the organic phase can be
easily recovered by adding isopropanol to make it RuO2 precipitate [19].

Figure 4-3. Phase-transfer oxidation with ruthenium oxide [33].
Phase-transfer reaction with ruthenium-based catalysts is very efficient in oxidizing many
organic substrates. However, this process is much less effective for oxidizing alcohols with high
polarity. Such alcohols have low solubility in organic solvent, thus cause a poor mixing with the
oxidant. Another drawback of this reaction is that separation of the product from the solution
with ruthenium oxides and oxidizing salts is always a problematic issue.

74

4.1.3. Alcohol oxidation using green oxidants
In the last a few decades, there was a growing demand for developing environmentally
benign processes to eliminate the pollution problems encountered in conventional chemical
industry. The requirement for “green chemistry” has led to extensive research in catalytic alcohol
oxidation using clean oxidants such as air, molecular oxygen (O2) [32], hydrogen peroxide
(H2O2) [32], and ozone (O3) [50, 51], etc. Such substances have many advantages over classical
oxidation reagents that they are easy to get, inexpensive, and the sole byproduct generated in
alcohol oxidation reactions is water. Thus the use of clean oxidants can greatly diminish the
pollution problems caused by using conventional methods.
Homogeneous catalysts, in many circumstances, exhibited good performances in alcohol
oxidation with these clean oxidants [33, 52, 53] However, industrial scale application of these
catalysts may cause many corrosion problems and increase the difficulty of post-reaction
separation and purification. Furthermore, the recycle of homogeneous catalysts is always a
challenge.
Heterogeneous catalysts, on the other hand, are of significant interest to chemical
engineers because of their easy handling, simple separation, and reusability. Therefore, in this
study, the investigation was focused on using heterogeneous solid catalysts with clean oxidants.

Molecular oxygen and air
Heterogeneous catalysts for gas or liquid phase oxidation of alcohols using molecular
oxygen or air as the oxidant has been broadly studied [54, 55-57]. Numerous types of materials
containing transition metals such as Pt, Pd, Au, Ag, Ru, V and Mo etc. exhibited good catalytic
activity in this application [55]. Among them, the most frequently used heterogeneous catalysts

75

usually consist of platinum group-metals, e.g. Pt, Pd and Ru, as the active components [12, 55].
It has been reported that these metals can activate the alcohol substrates and molecular oxygen
under close to ambient conditions [55]. For instance, with Pt- or Pd-based catalysts, alcohol
oxidation reactions can achieve close to 100% yields of carbonyl compounds or carboxylic acids
at mild reaction conditions (e.g. 293K-363K, atmosphere pressure of O2) [14, 55, 56, 58].
Mallat and Baiker [55] reviewed the oxidation of alcohols with molecular oxygen on
various solid catalysts in aqueous solutions. They concluded that the most accepted mechanism
model of this type of reaction is the oxidative dehydrogenation procedure (Figure 4-4) [12, 55].
In this approach, alcohols adsorb and dehydrogenate on the metal surface in two steps, i.e.
dehydrogenation of the O-H bond to a carbocation intermediate, and then dehydrogenation of the
β-C-H bond to the carbonyl product. At the same time, the adsorbed oxygen molecule oxidizes
the produced hydrogen atoms to shift the equilibrium toward carbonyl compound and
accomplish a complete redox cycle. The rate-determining step of this reaction, which can be
accelerated by bases, is generally believed to be the release of the β-C-H hydrogen [55, 59].
Recently, this model has been further developed. Experiments [55, 60] indicate that the primary
effect of oxygen in this reaction is not oxidation of the produced hydrogen atoms, but
regeneration of the deactivated catalyst sites via oxidative removal of the strongly adsorbed
poisoning byproducts.

76

Subscript ‘ad’ refers to adsorbed species
Figure 4-4. Oxidative dehydrogenation mechanism of alcohol oxidation over metal catalysts.
[55].
In general, heterogeneous catalysts containing noble metals are very efficient in aerobic
oxidation of a wide range of alcohol substrates, especially for activated alcohols such as aromatic,
allylic or acetylenic species. However, the same catalysts were found to be much less effective
for oxidation of non-activated alcohols such as aliphatic and alicyclic alcohols [53, 61,
62].Yamaguchi and Mizuno [63] investigated the Ru/γ-Al2O3 catalyzed oxidation of benzyl
alcohol and p-substituted benzyl alcohols with molecular oxygen. They found that the order of
reactivity for such benzyl alcohols was in the sequence of p-CH3O > p-CH3 > p-H > p-Cl > pNO2, indicating that an electron-donating group can weaken the β-C-H bond and thus boost the
reaction. On the other hand, an electron-withdrawing substituent plays a negative role in this
reaction.
Recently, a few new catalysts have been developed to overcome obstacles for oxidation
of non-activated alcohols with molecular oxygen. Such catalysts can be divided into two
categories: homogeneous complex compounds [64, 65] and nano particles of noble metals such
as Pt, Pd, Ru and Ir [62, 66]. However, these catalysts have some significant drawbacks. For
example, oxidation with homogeneous catalysts is typically carried out in toluene solvent, which
can cause many environmental problems [53, 64, 65]. The nanoparticle-catalyzed process
77

requires an extended reaction time (20-40 h) to achieve a good yield of aliphatic ketone or
aldehyde (78% to 99%) [62, 66]. In addition, the non-activated alcohols examined in these
studies are aliphatic alcohols, the activity of those catalysts on alcohols with electronwithdrawing groups is unknown.
Although aerobic oxidation with noble metal catalysts is an extensively used synthetic
approach, it is generally uneconomical because of the high price of the catalysts. In addition, the
deactivation of the catalyst is always a problematic issue that decreases the efficiency and
increases the production cost [12, 55].

Hydrogen peroxide
Metal-catalyzed oxidations with hydrogen peroxide (H2O2) can be divided into two
categories including peroxometal and oxometal species, respectively, as the active oxidants
(Figure 4-5) [67]. In the peroxometal pathway, the metal ion remains in the same oxidation state
during the catalytic cycle. In contrast, oxometal pathways involve a two-electron change in
oxidation state of the metal ion.
Peroxometal pathways are typically observed with early transition metal ions such as
0

Mo(VI), W(VI), Ti(VI) and Re(VII). These metals have a d configuration and are relatively
weak oxidants. On the other hand, oxometal pathways are characteristic of late transition
elements and first row transition elements, e.g. Cr(VI), Mn(V), Os(VIII), Ru(VI) and Ru(VIII),
which are strong oxidants in high oxidation states [32]. In addition, depending on the substrates,
some metals can operate via both pathways. For instance, vanadium(V) operates via a
peroxometal pathway in olefin epoxidations and via an oxometal pathway in alcohol oxidations
[68].

78

Figure 4-5. Oxometal vs. peroxometal pathways in metal-catalyzed alcohol oxidations with
hydrogen peroxide [67].
Alcohol oxidations using metal catalysts and hydrogen peroxide as the oxidant are
usually carried out in a biphasic system for phase-transfer operation [69-71]. Recently, a
microwave-assisted oxidation process using aqueous hydrogen peroxide was reported [72]. The
experimental procedure involves a simple mixing of an alcohol, sodium tungstate (Na2WO4),
and tetrabutylammonium hydrogen sulfate (TBAHS) followed by the addition of 30% aqueous
H2O2 in a 25:1:1:125 molar ratio for primary and a 25:1:1:40 molar ratio for secondary alcohols
in a monomode microwave reactor. Good yields (60%- 97%) of carboxylic acid or ketone
products from primary or secondary alcohols can be obtained within 20 min.

79

Commercial application of H2O2 in alcohol oxidation is always carried out using
titanium silicate (TS-1) as the catalysts [73]. TS-1 is an isomorphously substituted hydrophobic
molecular sieve and the progenitor of the so-called redox molecular sieves [74]. A major reason
for the superior activity of TS-1 is that the hydrophobicity of the framework stimulates the
adsorption of the relatively hydrophobic substrates and hydrogen peroxide from an aqueous
medium [75]. However, the scope of alcoholic substrates that can be used in this process is
limited by the pore size of TS-1. In addition, in the presence of allylic functionality, it gives a
high selectivity for epoxidation products [67].

Ozone oxidation
Ozonation of organics in the absence of any catalyst is attracting more and more attention
[76]. Ozone is well known as a powerful oxidant agent capable of oxidizing many organic
compounds [50] such as alcohols [77, 78], olefins and aromatic substrates [79-81]. When ozone
is used in organic synthesis, it can be simply decomposed back to oxygen and no special
neutralization or separation is required at the end of the reaction. Moreover, since ozone is being
produced in situ, complex logistics in transport, storage and disposal are not required.
For ozonation of secondary alcohols, it was found that the substrates could be rapidly and
completely converted into the corresponding ketone products (Table 4-1) under mild conditions
[51, 82].

80

Table 4-1. Ozone oxidation experiments of secondary alcohols at 0°C. [51]
Alcohol

Ketone Yield (%)

Alcohol

Ketone Yield (%)

2-propanol

83

2-nonanol

57

2-butanol

72

4-nonanol

68

2-pentanol

69

4-decanol

71

3-pentanol

81

5-decanol

65

2-hexanol

66

2,4-dimethyl-3pentanol

83

3-hexanol

74

cyclopentanol

53

2-heptanol

62

cyclohexanol

65

3-heptanol

82

cycloheptanol

74

4-heptanol

70

cyclooctanol

69

2-octanol

71

cyclododecanol

61

4-octanol

66

In general, it is believed that ozone oxidation of organics in aqueous solution proceeds
via two mechanisms: direct oxidation or radical decomposition [83]. The first mechanism
involves the direct attack of ozone molecules on the substrate [76, 84]. This reaction usually
occurs to some extent at any pH values for oxidation of alkene, alkyne compounds (Scheme 4-7)
or the –CH bond of alcohols and ethers (Scheme 4-8).

81

Scheme 4-7. Direct oxidation of alkene by ozone [84].

Scheme 4-8. Direct oxidation of alcohols by ozone [76].

O

R
HO

H

R
HO

O

R'

OOH

aldehydes and ketones

R'

O
+

The second mechanism involves the decomposition of ozone into radical species (mainly
hydroxyl radicals, ·OH). The hydroxyl radical, as shown in Table 4-2, is one of the strongest
oxidants which reacts in a nonselective way with organic substrates at high reaction rates [83,
85].
In a low pH (0-4) solution, the formation of OH radicals is ascribed to the occurrence of
the following reactions [86]:
O + O2

O3

O + H2O

2 ·OH

At alkaline pHs, the radical decomposition mechanism can be accelerated by the
following cycle [87]:
O3 + OH
-

-

·O3 + H2O

-

·O3 + ·OH
-

·OH + OH + O2
82

Table 4-2. Standard reduction potentials of selected oxidation agents [85].
O

Redox reaction

E (Volt) at 25°C

F2 + e− ⇔ 2 F −

2.87

⋅OH + H + + e− ⇔ H 2O

2.33

O3 + 2 H + + 2e− ⇔ O2 + H 2O

2.07

H 2O2 + 2 H + + 2e− ⇔ 2 H 2O

1.76

Cl2 + e− ⇔ 2Cl −

1.36

O3 + H 2O + 2e− ⇔ O2 + 2OH −

1.24

In the research of wastewater treatment, some new procedures have been discovered to
improve the oxidative power of ozone and H2O2. Experiments show that the combination of
ozone (O3) with hydrogen peroxide (H2O2) [88, 89], O3 with UV radiation [90], or H2O2 with
UV radiation gives rise to the formation of OH radicals and other reactive species. The related
reactions [90, 91] are:
-

+

2 O3 + H2O2 (HO2 + H )
O3 + hν + H2O
H2O2 + hν

2 ·OH + 3 O2

·OH + other active species

2 ·OH

With the ability to generate more hydroxyl radicals, the reactivity of O3/H2O2, UV/O3
and UV/H2O2 system significantly exceeds that of using ozone or hydrogen peroxide alone [90].

83

Such technologies, termed as “advanced oxidation processes” (AOP), are extensively applied in
waste water treatment to degrade some tough pollutants.
Recently, some metal oxides, such as γ-alumina and titanium oxide, were found to be
effective catalysts in ozonation of organics [92]. Ni and Chen [93] investigated γ-Al2O3
catalyzed ozonation of 2-chlorophenol. They found that at an acidic pH value, the addition of γAl2O3 to the reaction medium increased the efficiency of the ozonation process by 83.7%
compared to using ozonation alone. At neutral and basic pH, the increments of ozonation
efficiency were 100% and 17%, respectively.
Generally, it is believed that the presence of catalyst increases ozone dissolution [92].
The proposed mechanism of catalytic ozonation on metal oxides [92, 93] assumes that the
adsorption of both organic molecules and ozone takes place simultaneously on the surface of the
catalyst. As the result of ozone adsorption and its conversion, OH and other radicals [92] are
generated. Free radicals can initiate an oxidation reaction both on the surface of the catalyst and
in the bulk of the aqueous phase. In this way, the catalysts improve the efficiency of ozonation.
Although ozonation is a very effective process for oxidizing various organic substrates,
its application is usually limited to water treatment. It is rarely used for large scale organic
synthesis, because its strong oxidative ability can over-oxidize the substrate and result in many
undesired byproducts. In addition, the corrosion and damage caused by ozone is an important
consideration.

4.1.4. Oxidation of dioxanol to the corresponding ketone product
In 1995, Majewski et al. [94] investigated the oxidation of the 6-membered cyclic acetal
synthesized from glycerol and tert-butyraldehyde to the corresponding carbonyl product. The
substrate, as shown in Scheme 4-9, has similar molecular structure as 5-hydroxyl-2-methyl84

dixoane (cis-6). The authors attempted a variety of oxidation methods including Swern, Collins,
Fetizon’s, PDC, and Jones reactions. They found that the reaction either failed or gave very low
yield. The possible reason for this, they presumed, was due to the instability of dioxanone
product.

Scheme 4-9. Structure of 2-tert-butyl-5-hydroxyl-1,3-dixane.

In this chapter, the oxidation of cis-6 to its ketone products is carried out with various
oxidation methods. The efficiency of these methods was discussed and compared with literature.
The experimental results provide a basis for further oxidant selection.

4.2. Experimental
4.2.1. Materials
Anhydrous diethyl ether (99.5%, ACS grade), acetone (99.7%), sodium hydroxide (>
99%), 1-propanol (99.9%), 1-butanol (99.9%), hydrogen peroxide (30% aqueous solution),
toluene (99.7%, ACS grade) were obtained from J.T. Baker.
Anhydrous dichloromethane (99.8%), anhydrous o-dichlorobenzne (99%), sodium
carbonate (> 99%), sodium bromate (> 99%), potassium bromate (> 99.8%), calcium
hypochlorite (> 98%), sodium periodate (> 99%), sodium iodide (> 99%) were purchased from
Sigma-Aldrich. Amberlyst-15 cationic exchange resin (16-50 mesh, 1.2g/cm3), purchased from

85

Sigma-Aldrich, was investigated as the catalyst. Before being used in reactions, it was pretreated
at 90°C in a vacuum oven for 2 hours.
Pyridinium chlorochromate (98%), 3-nitrobenzaldehyde (> 99%), trimethyl aluminum
(2.0 M in chlorobenzene), aluminum isopropoxide (> 99.99%), aluminum tert-butoxide (> 98%),
ruthenium dioxide hydrate (99.9% trace metals basis, 28% water content), ruthenium trichloride
hydrate (reagent plus, 46% Ru content), platinum on carbon (5% Pt absis), platinum on alumina
(5% Pt basis), palladium on activated carbon (5% Pd basis), ruthenium on carbon (5% Ru basis),
titanium silicate (99.8% trace metals basis) were obtained from Aldrich.
Bismuth modified platinum on carbon (5% Pt and 1% Bi) was obtained from Johnson
Matthey. Sodium hypochlorite (15% aqueous solution) was purchased from Pharmco and
Aldrich. Aluminum oxide (99.9%, γ-Al2O3) was obtained from Alfa Aesar. Cyclohexanone (>
99%) and 1,4-Benzoquinone (> 98%), were obtained from Fluka. Potassium bromide (ACS
reagent, > 99.5%) was purchased from Spectrum.
Hydrogen (99.999%) and helium (99.999%) used for GC analysis, and oxygen (99.999%)
used for oxidation experiment were purchased from Linde Gas, Inc.

4.2.2. GC and GC-MS analysis
The GC and GC-MS methods used in this investigation are the same as the ones
described in Chapters I and II.

4.3. Results and discussion
4.3.1 Oxidation of cis-6 with pyridinium chlorochromate (PCC)
In a preliminary experiment, PCC was used to oxidize cis-6 to its ketone product. This
experiment was carried out according to the procedure described by Piancatelli, et al [95]. In a
500 ml round bottomed flask fitted with a reflux condenser and magnetic stirrer, 32.5 g of
86

pyridinium chlorochromate (PCC) was suspended in 200 ml of anhydrous dichloromethane. Cis6 isomer (11.8 g) in dichloromethane (20 ml) was added to the magnetically stirred suspension.
After reaction at room temperature for 1.5 h, 200 ml of dry diethyl ether was added and the
organic solution was decanted. Following the reaction, GC analysis showed the presence of Gtone in the final mixture. The conversion of cis-6 was 30%, and the selectivity to G-tone was
nearly 100%.
This experiment illustrates the feasibility of producing the ketone precursor of DHA from
cis-6. However, the production method must be improved because PCC and dichloromethane are
toxic. Hence, further study was carried out with more environmentally benign catalysts and
oxidants.

4.3.2. Oppenauer oxidation
Oppenauer oxidation proceeds via exchanging two hydrogen atoms between a ketone and
an alcohol. In this study, the Oppenauer oxidation experiment of cis-6 was carried out in a dried
500 mL round bottom flask equipped with a condenser. Into this flask was introduced a
measured quantity of cis-6 GA isomer, oxidizing reagent such benzoquinone and 3nitrobenzaldehyde, and anhydrous organic solvent, followed by the aluminum-base catalyst, i.e.
aluminum tert-butoxide, aluminum isopropoxide, and trimethyl aluminum. If trimethyl
aluminum (AlMe3) is used, the reaction flask must be capped with septa and purged with
nitrogen to maintain an anhydrous atmosphere. The chlorobenzene solution of trimethyl
aluminum was transferred to the reactor with an oven dried syringe.
After loading of chemicals, the mixture was stirred and heated to the desired temperature
(from 25°C to 113°C) using a heating mantle for reaction. Samples were taken after the reaction
for GC analysis. The reaction conditions and results are presented in Table 4-3.

87

Table 4-3. Oppenauer oxidation experiments of cis-6.
Catalyst
Reaction loading (wt Oxidant
to cis-6)

Solvent

Molar ratio
(cis-6 :
oxidant :
solvent)

Temp. Reaction
(°C)
time (h)

G-tone
yield
(%)

1

P (50%)

AT

Toluene

1:8:9

25

46

2

2

M (7.5%)

3-NB

Toluene

1:3:19

25

20

0

3

M (7.2%)

BQ

Benzene

1:2:15

25

24

0

4

B (22%)

CH

Toluene

1:5:30

113

24

0

5

B (25%)

BQ

Toluene

1:5:30

113

24

0

6

B (30%)

3-NB

Toluene

1:5:28

113

24

0

7

P (290%)

CH

-

1:100

100

24

0

Catalysts: B = aluminum tert-butoxide, Al(t-BtO)3, M = trimethyl aluminum, AlMe3; P =
aluminum isopropoxide, Al(i-PrO)3.
Oxidants: AT= acetone; BQ = benzoquinone; CH = cyclohexanone; 3-NB = 3-nitrobenzaldehyde.

As shown, negligible product was produced via this pathway even after a long reaction
period. Recalling the reaction mechanism (Scheme 4-5) that Oppenauer oxidation is the reverse
process of Meerwein-Ponndorf-Verley (MPV) reduction, the poor yield indicates that the
equilibrium is not favored towards the formation of G-tone.
The result can be explained by comparing the oxidation potentials of the reactants and
product. Adkins and coworkers [96] calculated the relative redox potentials for a wide rage of
carbonyl-alcohol systems via polarographic experiments. They found that under ordinary
conditions, dialkyl ketones have the lowest oxidation potentials and, therefore, their
corresponding secondary alcohols have the best reducing capacities. However, if the carbonyl
bond of the ketone is on a ring structure or the α-hydrogen of the ketone is replaced by an
alkoxyl group, its oxidation potential will increase remarkably. Accordingly, the alcohol
88

counterparts of such ketones have poor activity in Oppenauer reaction. The relation of ketone
structure and its oxidation potential is well illustrated in Table 4-4 by comparing two series of
ketone compounds.

Table 4-4. Ketone compounds and their oxidation potentials [96].
IUPAC name of ketone

Structure

Oxidation potential (mV)

O

4-Heptanone

101
O

110

3-Pentanone

O
162

Cyclohexanone

O
O

2-Methoxycyclohexanone

218

O

Acetone

129
O

Methoxyacetone

189

O
O

1,3-Dimethoxyacetone

O

O

350

From the theory developed by Adkins and et al., it can be anticipated that the oxidation
potential of G-tone would be even higher than 1,3-dimethoxyacetone because it contains a ring

89

structure and two α-carbon connected methoxy groups. Hence, oxidation of its alcohol
counterpart, cis-6, is expected to be extremely difficult via Oppenauer oxidation with other
carbonyl compounds.

4.3.3. Ruthenium tetroxide catalyzed oxidation
A typical operation procedure for ruthenium tetroxide-catalyzed oxidation of cis-6 started
with charging 10 ml of cis-6 solution (4 to 10 wt% in water) to a glass reactor with hydrated
ruthenium dioxide (RuO2·xH2O) or ruthenium trichloride (RuCl3·xH2O) powder. The mixture
was then agitated for 5 min with a magnetic stirring bar. After that, 10 to 20 mL of saturated
solution of a secondary oxidant, such as NaClO, NaBrO3, KBrO3, NaIO4, etc., was added to the
well stirred reaction mixture at a rate of 5-10 mL/min. Ruthenium tetroxide (RuO4) is then
generated in situ by oxidation of RuO2 or RuCl3. A small amount of base, such as sodium
carbonate (Na2CO3) or sodium hydroxide (NaOH), was added to maintain a slightly basic pH
condition.
Table 4-5 presents the results of cis-6 oxidation experiments catalyzed by rutheniumbased substances. It is clear that this type of reaction is highly exothermic and very efficient. In
some reactions over 90% yield of G-tone was achieved within 30 min.
An important feature of this type of reaction is that it must be performed with a secondary
oxidant (consuming oxidant), such as NaClO, NaBrO3, NaIO4, and KBrO3. Such inorganic salts
have strong oxidative ability in aqueous solution, however, they cannot oxidize cis-6 without the
Ru catalyst present. As shown, negligible G-tone was produced in Reactions 1 and 2, indicating
that the real oxidant which converts cis-6 to G-tone is a high-valent ruthenium species generated

90

in situ from RuO2·xH2O, or RuCl3·xH2O with the oxidizing salts. This ruthenium species has
been identified as RuO4 [19, 45, 97].
Conventional RuO4-catalyzed alcohol oxidation often employs a biphasic system, in
which RuO2·xH2O or RuCl3·xH2O is reacted with the secondary oxidant in water and the RuO4
generated is extracted into an organic solvent for alcohol oxidation [19, 97]. The most commonly
used organic solvents are CH2Cl2, CCl4, etc [19, 33, 44]. However, due to the poor cis-6
solubility in non-polar solvents, phase-transfer reaction cannot be applied. In this study, water
was used as the only solvent.
The effect of water on this reaction is clearly illustrated by Reactions 10 and 11. Without
water, no product was obtained in 24 h. However, if a small amount of water (0.4 g) is used, the
yield of G-tone increased rapidly from 0 to 20% in 30 min.

91

Table 4-5. Ruthenium tetroxide catalyzed oxidation of cis-6 at room temperature,
atmospheric pressure. Catalyst was generated in situ with ruthenium dioxide.
Reaction

Cis-6
(g)

Water
(g)

Catalyst
(RuO2·xH2O)
loading (g)

Secondary oxidant
solution/mass
(g)/concentration

Time
(min)

G-tone
yield
(%)

1

0.41

10

0

NaClO/4/10%

30

0

2

0.42

10

0

KBrO3/10/saturated

1440

0

3

4.0

50

0.40

NaClO/100/10%

30

80

0.39

11

0.31

NaClO/4/10%

30

76

NaClO/5/10%

60

95

Ca(ClO)2/10/
saturated

20

87

NaIO4/10/saturated

75

88

KBrO3/10/ saturated

45

> 99

KBrO3/10/saturated

75

95

4

5

0.45

10

0.05

6

0.40

10

0.30

7

0.42

10

0.26

8

0.43

10

0.32

9

4.1

10

0.40

KBrO3/100/saturated

20

97

10

0.31

0

0.10

KBrO3/0.31/powder

1440

0

11

0.31

0.4

0.10

KBrO3/0.31/powder

30

20

12

0.30

10

0.41

NaBrO3/10/10%

20

96

13

10.0

10

0.51

NaBrO3/100/20%

40

90

14

3.0

10

0.10

NaBrO3/12.5/20%

20

95

a

a

a. RuCl3·xH2O was used as catalyst.

Selection of an appropriate secondary oxidant is of significant importance for RuO4 catalyzed alcohol oxidation. In general, the five inorganic salts tested in this study are all good
92

options for lab scale experiments. However, for large scale production, fully consideration must
be taken with respect to the stability, solubility and economic factors of the oxidizing salts. For
instance, calcium hypochlorite is not stable in air, but reacts with carbon dioxide to form calcium
carbonate and release chlorine.
2 Ca(ClO)2 + 2 CO2 → 2 CaCO3 + 2 Cl2 + O2
Sodium hypochlorite loses its oxidation ability via decomposition to NaCl at elevated
temperature. Since a large amount of heat is generated in the process of cis-6 oxidation to G-tone,
if NaClO is used as the secondary oxidant it must be fed extremely slowly to avoid thermal
decomposition.
Sodium periodate and potassium bromate have relatively better thermal stability than
Ca(ClO)2 and NaClO. However, they are much less soluble in water (Table 4-6). If they are
chosen to be the secondary oxidant, more water is required to dissolve the salts, which will
increase the separation cost after reaction.

Table 4-6. Solubility of the inorganic salts used as the secondary oxidant.
Salt

Solubility in water at 20°C
(g/100 mL)

Concentration of saturated
solution (%)

NaClO

29.3 (0°C)

23

NaBrO3

36.4

27

NaIO4

8.0

7

KBrO3

6.9

6

Ca(ClO)2

21.0

17

Last, sodium bromate, which has good thermal stability and the highest solubility in
water, is the best option for this reaction.

93

Another parameter, pH value of the reaction medium, is of paramount importance for this
reaction, because it can influence the stability and catalytic activity of RuO4 [47]. In this study, it
was noticed that during the course of some reactions, especially the ones with NaClO, the
reaction mixture turned into dark orange or green in color. This is because RuO4 generated in the
-

-

aqueous phase is unstable. It can be reduced by OH ions to perruthenate (RuO4 ), a green
-

species, which in turn, is reduced more slowly by OH ions to an orange species, ruthenate
2-

(RuO4 ). [98, 99]. Thus, in sodium hypochlorite solution, which typically has a pH higher than
11, RuO4 is readily reduced to the coloring ionic species.
The formation of perruthenate and ruthenate is highly dependent on the pH value. Figures
2-

-

4-7 display the distribution of RuO4 , RuO4 , and RuO4 as a function of pH in aqueous
solutions with Ru concentration = 0.00038 M. The plots are obtained via electrochemical
calculations for soluble ruthenium species (see in Appendix III for details). As can be seen,
-

RuO4 is only stable up to pH = 8, RuO4 regime is dominant between pH of 8 and 11.5, and
2-

above 11.5, RuO4

is the dominant species. The trends predicted by electrochemical calculation

are in good agreement with the experimental data reported by Mills and Holland [99] (Figure 48).

94

100

Population (%)

80
60
40
20
0
0

2

4

6

8

10

12

14

pH

Figure 4-7. Calculated population (%) of various ruthenium species at different pHs in an
2-4
aqueous solution with [Ru] = 3.8×10 M. RuO4 (●), RuO4 (○) and RuO4 (△).

Figure 4-8. Population (%) of various ruthenium species at different pHs of an aqueous
2-4
solution with [Ru] = 3.8×10 M. Experimental results. RuO4 (●), RuO4 (○) and RuO4
(△). [99]

95

Ruthenate and perruthenate species are much weaker oxidants than ruthenium tetroxide
[98]. To ensure a good performance, most RuO4 catalyzed oxidations generally require acidic
conditions. However, it is still possible to conduct the oxidation in slightly basic condition via
the disproportion equilibrium of perruthenate as shown in Scheme 4-10 [47, 100]. Because the
reaction between RuO4 and organic substrates is very fast, the more reactive RuO4 species is
-

-

consumed much faster than the more stable RuO4 species. Thus, the equilibrium of RuO4

disproportion is shifted to the right side. In the meanwhile, the secondary oxidant is able to
reoxidize the ruthenate species to perruthenate, which can again disproportionate. Hence, even in
-

a basic solution, the high reactivity of RuO4 and the ability of RuO4 to undergo a disproportion
reaction can provide an efficient oxidation cycle for an organic substrate.

Scheme 4-10. RuO4-catalyzed oxidation under slightly basic condition.

For cis-6 oxidation with RuO4 catalyst, acidic conditions must be avoided, because in the
presence of acid and large amount of water cis-6 readily hydrolyzes to glycerol and acetaldehyde.
Furthermore, acetaldehyde produced can be oxidized by RuO4 to acetic acid, which will further
96

increase the acidity of the solution and cause more cis-6 to decompose. Thus, to avoid loss of
reactant, basic pH must be maintained in the reaction mixture.
High-valent ruthenium species can be reduced by isopropanol to RuO2 hydrate, which is
insoluble in water and precipitates as black solid [19]. This method can be used for catalyst
recycling. In Reaction 14, 13 g of isopropanol was added to the final mixture; after stirring for 2
h and filtering, 97% of the Ru species was recovered as RuO2 hydrate.

4.3.4. Aerobic oxidation of cis-6 with molecular oxygen
Aerobic oxidation of cis-6 was performed in vapor phase or liquid solution.
Vapor phase experiments were carried out under atmosphere pressure in a fixed bed
reactor equipped with a downstream condenser. In a typical experiment, a measured quantity of
catalyst, such as 5% Pt/Al2O3, 5% Pt/C and RuO2, supported by glass wool was loaded into the
reactor. The catalyst was purged with helium gas at 200°C for 1 h before reaction. After that, cis6 vapor was generated in a round bottom flask heated by heating mantel and carried to the
reactor by air (air flow rate = 25 mL/min). During reaction, the reactor temperature was
maintained in the range of 210 to 214°C. The reactor effluent was collected in a tube connected
to the downstream condenser and immersed in an ice water bath.
The experimental results of vapor phase aerobic oxidation of cis-6 are listed in Table 4-7.
GC analysis of cis-6 vapor showed that no reaction occurred up stream of the reactor, indicating
that the reactant is stable under the vapor conditions.

97

Table 4-7. Vapor phase aerobic oxidation of cis-6 at atmospheric pressure.
Catalyst

Loading
(g)

Reaction
temp. (°C)

Air flow rate

5%
Pt/Al2O3

0.61

212

25

Negligible yield of G-tone,
other GA isomers produced.

5% Pt/C

0.83

210

25

Negligible yield of G-tone,
other GA isomers produced.

RuO2·xH2O

0.49

214

30

No reaction occurred

Reaction product

(mL/min)

Oxidation of cis-6 to G-tone was insignificant at the conditions listed in Table 4-7. At the
same time, some undesired products, trans-6 and 5-membered GA isomers, were detected in the
final product of Pt-catalyzed reactions, indicating that interconversion of GA isomers was
catalyzed by Pt-based catalyst.
Table 4-7 also shows that ruthenium dioxide has no effect on cis-6 oxidation or GA
isomer re-equilibration, indicating that RuO2 was not oxidized to high valent RuO4 at this
reaction conditions.
Liquid phase aerobic oxidation of cis-6 at atmospheric or elevated pressure is
summarized in Tables 4-8 and 4-9, respectively.
For atmospheric pressure oxidation experiments, a 500 mL three-necked round bottom
flask was used. A typical run started with charging of 5 g of cis-6, 25 g of solvent, and a
measured quantity of catalyst such as 5% Pd/C to the flask. Thereafter, air flow was introduced
into the mixture at a rate of 20-30 mL/min via stainless steel tubing. The mixture was then stirred
with a magnetic bar and heated to the desired temperature for reaction from 1.5 h to 24 h.
Samples were taken after the reaction and subjected to GC analysis.

98

High pressure oxidation liquid-phase experiments were carried out in Parr reactors
connected to an oxygen cylinder. A typical reaction occurred with an initial charge of 30 g of 1.8
M cis-6 solution and a measured quantity of catalyst, such as 5% Ru/C and 5% Pt/C etc., in the
reaction vessel. The system was stirred and purged with oxygen for 10 min before heating started.
After reaching the desired temperature (from 60°C to 95°C), the oxygen pressure in the reactor
was increased to the set point (from 30 psi to 250 psi) by adjusting the valve installed between
the reactor and cylinder. When reaction is finished, the mixture was filtered and subjected to GC
analysis.
In order to prevent isomer interconversion and hydrolysis of cis-6 in aqueous solution,
small amount of bases, such as sodium hydroxide (NaOH), was added to the reaction mixture to
maintain a slightly basic pH in the solution.

99

Table 4-8. Aerobic oxidation of cis-6 using noble metal catalysts. Atmospheric pressure, air
flow rate = 25-30 mL/min.
Cis-6
(g)

Catalyst loading (g)

Solvent (g)

10

5% Pd/C (0.40)

-

50

24

No oxidation occurred

1.5

5% Pd/C (0.48)

Water (9)

25

24

No oxidation occurred

5.9

5% Pt/Al2O3 (0.59)

Toluene (40)

60

24

No oxidation occurred

6.0

RuO2·xH2O (0.61)

Water (30)

60

22

No oxidation occurred

5.1

RuO2·xH2O (0.28)

o-DCB (26)

177

1.5

No oxidation occurred

4.1

5% Ru/C (1.25)

o-DCB (26)

175

23

No oxidation occurred

5.9

5% Pt/Al2O3 (0.51)

Water (30)

65

18

4% G-tone yield, glycerol,
acetic acid and other
byproducts detected

20

Over 90% cis-6
conversion, 30% G-tone
yield, glycerol, acetic acid
and other byproducts
detected

4.0

1%Bi-5%Pt/C (0.25)

Water (36)

o-DCB: o-dichlorobenzene.

100

Temp. Time
(°C)
(h)

60

Result

Table 4-9. Aerobic oxidation of cis-6 using noble metal catalysts at elevated pressure.
Oxygen was used as oxidant.
Cis-6
(g)

Catalyst loading
(g)

Temp.
(°C)

5.0

RuO2·xH2O
(0.51)

60

30

24

No reaction occurred

5.0

5% Ru/C (0.53)

60

30

24

No reaction occurred

5.0

5% Pt/C (1.06)

85

100

24

No G-tone produced, cis-6
decomposed to glycerol and
converted to other GA isomers

5.0

5% Pt/Al2O3
(0.75)

85

150

24

Cis-6 decomposed to glycerol
and acetaldehyde which was
further oxidized to acetic acid

5.0

RuO2·xH2O
(0.65)

75

150

22

Cis-6 decomposed to glycerol
and acetaldehyde which was
further oxidized to acetic acid

24

No G-tone produced, isomer
equilibrium shifted and acetic
acid produced

5.0

5% Pt/C (0.87)

95

Pressure Time
(psi)
(h)

250

Result

Under atmospheric pressure, cis-6 was inert in the Pd and Ru catalyzed reactions even at
a temperature of 177°C. It indicates that cis-6 is more difficult to oxidize than other alcohol
substrates. In aqueous solution, Pt-based catalysts exhibited low catalytic activity. In one
reaction, a small amount of cis-6 (4%) was oxidized to G-tone in the presence of 5% Pt/Al2O3
catalyst in 18 h. The yield of G-tone increased markedly to 30% when Bi promoted Pt/C catalyst
was used.
It was reported that addition of Bi to Pt catalyst could improve its performance for
secondary alcohol oxidation [6, 13]. However, in the Bi-Pt/C catalyzed reaction presented in
Table 4-8, the improved yield of G-tone could be the result from the oxidation of the secondary

101

hydroxyl group of glycerol instead of cis-6 dehydrogenation. Because a large amounts of cis-6
was decomposed to glycerol and acetaldehyde in this reaction. Acetaldehyde was then oxidized
to acetic acid which further promoted cis-6 decomposition.
Side reactions were more prevalent at elevated pressure and temperature (Table 4-9). In
some reactions, especially those catalyzed by Pt-based compounds, although base was initially
added to the solution to obtain a slightly basic pH, the reactant was completely decomposed
before oxidation occurred.
The results from the aerobic oxidation experiments indicate that under special conditions,
Pt catalyst has low activity for cis-6 oxidation. However, it also catalyzes isomer interconversion
and decomposition of cis-6. Recalling the reaction mechanism [55] and the order of alcohol
reactivity discovered by Yamaguchi and Mizuno [63], it can be assumed that the presence of
oxygen in the heterocyclic structure of cis-6 may have the same effect as electron-withdrawing
groups and thus strongly suppress the reactivity of alcohol.

4.3.5. Oxidation with hydrogen peroxide
The initial oxidation experiments using hydrogen peroxide solution as oxidant was
carried out at room temperature with RuO2·xH2O and RuCl3·xH2O catalysts. In these
experiments, hydrogen peroxide decomposed immediately upon addition to the reaction mixture.
It was reported [101] that RuCl3 can facilitate the decomposition of H2O2 and the concentration
of H2O2 in aqueous RuCl3 solution can only be sustained at the mmol/L level. Thus, more than
90% of H2O2 was consumed via unproductive decomposition.

102

Further experiments were carried out in Parr reactors. In one of them, 5 g of cis-6 was
mixed with a 30 g of 30% aqueous H2O2 solution and stirred at 60°C for 22 h using a magnetic
bar. However, no reaction occurred under this condition.
The other reaction was conducted with H2O2 and 0.5 g of titanium silicate (TS-1) as
catalyst. The vessels were then well sealed and heated to 60°C. Alcohol oxidation with TS-1 and
H2O2 is a commercialized technology [73]. However, the reagents did not exhibit any oxidation
effect on cis-6. In this experiment, no oxidation product was detected in 22 h. Moreover, cis-6
was slightly decomposed to glycerol and acetaldehyde under this reaction condition.

4.3.6. Ozone oxidation
Ozonation of cis-6 was carried out in a self-designed reactor (Figure 4-6) constituting of a
glass tube with a fritted disc at the bottom. To initiate an experiment, a measured quantity of cis6 and solvent (water or 30% aqueous H2O2 solution) were added to the reactor and maintained
room temperature in a water bath. Ozone-oxygen mixture was produced in situ by passing dry
oxygen (5-8 L/min) through a Trailigaz ozone generator (model number OZC-1001; Praxair,
Cicinnati, OH); the mixture contains of 8-9 wt% of O3. During reaction, O3 was introduced into
the reaction medium as small gas bubbles by passing the ozone-oxygen stream through the fritted
disc. The formation of these small gas bubbles improved the dispersion of ozone within the
reaction mixture. A dry ice-cooled trap was used to collect the volatile product in the reactor
effluent. Any unreacted ozone was decomposed by venting the exhaust gas through aqueous KI
solution.

103

Figure 4-6. Ozone oxidation of cis-6 solution.
Table 4-10. displays the result of a series of ozone oxidation experiments. In Reaction 1,
ozone gas was used as the only oxidizing reagent. The yield of G-tone in this reaction is less than
1%, indicating that ozone alone has poor activity for cis-6 oxidation.

Table 4-10. Ozone oxidation experiments. Substrate: 20 g of 20% cis-6 solution; ozone (8-9
wt% in oxygen): 5-8 L/min; room temperature; reaction time: 20 min.
Reaction

Reagent (g)

Solvent

pH

G-tone yield (%)

1

--

Water

7

<1

2

--

Water

10

<1

3

--

30% H2O2

7

1

4

--

30% H2O2

10

1

5

γ-Al2O3 (1.0)

Water

7

<1

6

RuO2·xH2O (0.2)

Water

7

<1

7

RuCl3·xH2O (0.2)

Water

10

1

8

RuCl3·xH2O (0.2) + KBr (0.5)

Water

7

13

104

Reactions 2 to 5 examine the activity of hydroxyl radicals on cis-6 oxidation. The
hydroxyl radical is a very strong oxidant with a redox potential of 2.33 V [85]. It was reported
that increasing pH [87] or adding hydrogen peroxide [88, 89] to aqueous ozone solution can
promote the formation of OH radicals. However, in Reactions 2 to 4, no improvement of G-tone
yield was observed when base or hydrogen peroxide was added to the reaction medium. The
result indicates that formation of OH radical in the solution is not essential for cis-6 oxidation.
The mechanism of G-tone formation is more likely to be from the direct attack of ozone on the
substrate, as described by Canterino et al [76].
In Reaction 5, γ-Al2O3 was used as a catalyst for ozonation. It is believed that on the
surface of metal oxides, ozone can be adsorbed and decomposed to OH and other radicals to
initiate oxidation reactions on both the catalyst surface and in the bulk of the aqueous phase.
However, in Reaction 5, the ozonation efficiency was not increased by addition of γ-Al2O3. This
result further reinforces the conclusion that oxidizing radicals have no effect on cis-6 oxidation,
but that formation of G-tone is the result of ozone direct attack.
Early in this chapter, ruthenium tetroxide was proven to be an excellent reagent for cis-6
oxidation. In this section, two experiments (Reactions 6 and 7) were designed to test whether
ozone or OH radical can be used as the secondary oxidant for RuO4 formation. Unfortunately,
this goal was not accomplished. As shown, at pH = 7 or 10, addition of RuO2 and RuCl3 to the
reaction mixture made no difference to the yield of G-tone, indicating that RuO4 was not
generated in the solution.
In Reaction 8, potassium bromide and ruthenium trichloride were added to the reaction
mixture. As a result, yield of G-tone increased significantly to 13%. This is because ozone is able
105

to oxidize bromide to bromate in aqueous solution [102, 103]. Consequently, in this solution,
RuO4 was able to be generated from RuCl3 and bromate produced in situ. Therefore, cis-6 was
oxidized to G-tone by RuO4. The rate of this reaction is likely to be controlled by the oxidation
of bromide to bromate, because formation of RuO4 and oxidation of cis-6 have been proven to
be very fast reactions.
It is reasonable to presume that bromide can be recycled and continuously oxidized to
bromate in this solution. Thus, if the reaction time is extended to long enough, it can be expected
that conversion of cis-6 to G-tone will proceed to completion.
When ozone and bromide salt are used in RuO4-based oxidation, the oxidizing salt
required can be reduced from stoichiometric ratio to catalytic amount. This improvement will
greatly help the separation operation after reaction.

4.4. Conclusions
Oxidation of cis-6 GA isomer to the corresponding ketone product provides a new
synthesis route for DHA production. A broad range of catalysts and oxidizing reagents including
green oxidants such as molecular oxygen, hydrogen peroxide and ozone have been examined for
this purpose. However, due to the heterocyclic ring structure and high oxidation potential of cis-6,
most of the reagents exhibited poor activity.
Ruthenium tetroxide based oxidation appears to be the most effective method for this
reaction. With this reagent, complete conversion of cis-6 to G-tone can be achieved very rapidly.
However, separation of the desired product from the reaction mixture which contains water,
soluble ruthenium species and oxidizing salts such as bromate, periodate and hypochlorite is a
big challenge. An improvement of this method was found to be using bromide salt and ozone as
106

the oxidant. In this way, bromate salt can be generated in situ, and only catalytic amount of salt
is required for reaction.

107

References

108

References
1. Carlsen, P, Sorbye, K, Ulven, T. and Aasbo, K. (1996). Synthesis of benzylidene-protected
dihydroxyacetone. Acta Chemica Scandinavica. 50, 185-187.
2. Wang, K, Hawley, M. and DeAthos, S. (2003). Conversion of glycerol to 1,3-propanediol via
selective dehydroxylation. Industrial & Engineering Chemistry Research. 42, 2913-2923.
3. Hekmat, D, Bauer, R. and Fricke, J. (2003). Optimization of the microbial synthesis of
dihydroxyacetone from glycerol with Gluconobacter oxydans. Bioprocess and Biosystems
Engineering. 26, 109-116.
4. Hu, Z, Zheng, Y. and Shen, Y. (2010). Dissolved-oxygen-stat fed-batch fermentation of 1,3dihydroxyacetone from glycerol by Gluconobacter oxydans ZJB09112. Bioprocess and
Biosystems Engineering.15, 651-656.
5. Garcia, R, Besson, M. and Gallezot, P. (1995). Chemoselective catalytic oxidation of
glycerol with air on platinum metals. Applied Catalysis A: General. 127, 165-176.
6. Hu, W, Knight, D, Lowry, B. and Varma, A. (2010). Selective Oxidation of Glycerol to
Dihydroxyacetone over Pt-Bi/C Catalyst: Optimization of Catalyst and Reaction Conditions.
Industrial & Engineering Chemistry Research. 2010, 49, 10876–10882.
7. Mishra, R, Jain, S. and Kumar, A. (2008). Microbial production of dihydroxyacetone.
Biotechnology Advances. 26, 293-303.
8. Demirel, S, Lehnert, K, Lucas, M. and Claus, P. (2007). Use of renewables for the production
of chemicals: Glycerol oxidation over carbon supported gold catalysts. Applied Catalysis: B.
70, 637-643.
9. Demirel, S, Kern, P, Lucas, M. and Claus, P. (2007). Oxidation of mono- and polyalcohols
with gold: Comparison of carbon and ceria supported catalysts. Catalysis Today. 122, 292300.
10. Villa, A, Campione, C. and Prati, L. (2007). Bimetallic gold/palladium catalysts for the
selective liquid phase oxidation of glycerol. Catalysis Letters. 115, 133-136.
11. Vinke, P, De Wit, D, De Goede, A. and Van Bekkum, H. (1992). Noble metal catalyzed
oxidation of carbohydrates and carbohydrate derivatives. Studies in Surface Science and
Catalysis. 72, l-20.
12. Mallat, T. and Baiker, A. (1994). Oxidation of alcohols with molecular oxygen on platinum
metal catalysts in aqueous solutions. Catalysis Today. 19, 247-283.
13. Kimura, H. (1993). Selective oxidation of glycerol on a platinum-bismuth catalyst by using a
fixed bed reactor. Applied Catalysis A: General. 105, 147-158.

109

14. Brandner, A, Lehnert, K, Bienholz, A, Lucas, M. and Claus, P. (2009). Production of
biomass-derived chemicals and energy: chemocatalytic conversions of glycerol. Topics in
Catalysis. 52, 278– 287.
15. Bowden, K, Heilbron, I. and Jones, E. (1946). Researches on acetylenic compounds. Part I.
The preparation of acetylenic ketones by oxidation of acetylenic carbinols and glycols.
Journal of the Chemical Society. 39-45.
16. Lanes, R. and Lee, D. (1968). Chromic acid oxidation of alcohols: A simple experiment on
reaction rates. Journal of Chemical Education. 1968, 45, 269-274.
17. Westheimer, F. and Nicolaides, N. (1949). The kinetics of the oxidation of 2deuteropropanol-2 by chromic acid. Journal of American Chemical Society. 71, 25-28.
18. Collins, J, Hess, W. and Frank F. (1968). Dipyridine-chromium (VI) oxide oxidation of
alcohols in dichloromethane. Tetrahedron Letters. 9, 3363-3366.
19. Tojo, G. and Fernandez, M. (2006). Oxidation of alcohols to aldehydes and ketones. NY:
Springer.
20. Corey, E. and Suggs, W. (1975). Pyridinium chlorochromate. An efficient reagent for
oxidation of primary and secondary alcohols to carbonyl compounds. Tetrahedron Letters. 16,
2647–2650.
21. Hudlicky, M. (1990). Oxidations in Organic Chemistry. Washington D.C.: American
Chemical Society.
22. Omura, K.; Swern, D. (1978). Oxidation of alcohols by "activated" dimethyl sulfoxide. A
preparative, steric and mechanistic study. Tetrahedron. 34, 1651-1660.
23. http://en.wikipedia.org/wiki/Swern_oxidation
24. Dess, D. and Martin, J. (1983). Readily accessible 12-I-5 oxidant for the conversion of
primary and secondary alcohols to aldehydes and ketones. Journal of Organic Chemistry. 48,
4155-4156.
25. http://en.wikipedia.org/wiki/Dess–Martin_periodinane
26. http://en.wikipedia.org/wiki/TEMPO
27. Anelli, P, Biffi, C.; Montanari, F. and Quici, S. (1987). Fast and selective oxidation of
primary alcohols to aldehydes or to carboxylic acids and of secondary alcohols to ketones
mediated by oxoammonium salts under two-phase conditions. Journal of Organic Chemistry.
52, 2559-2562.
28. Graauw, C, Peters, J, Van Bekkum, H. and Husken, J. (1994). Meerwein-Ponndorf-Verley
reductions and Oppenauer oxidations: An integrated approach. Synthesis. 1007-1017.

110

29. Klomp, D, Maschmeyer, T, Hanefeld, U. and Peters, J. (2004). Mechanism of
homogeneously and heterogeneously catalysed Meerwein–Ponndorf–Verley–Oppenauer
reactions for the racemisation of secondary alcohols. Chemistry-A European Journal. 10,
2088-2093.
30. Otvos, L, Gruber, L. and Meisel-Agoston, J. (1965). The Meerwein-Ponndorf-VerleyOppenauer reaction. Investigation of the reaction mechanism with radiocarbon. Racemization
of secondary alcohols. Acta Chimica Academiae Scientarium Hungaricae. 43, 149–153.
31. Graves, C, Zeng, B. and Nguyen, S. (2006). Efficient and selective Al-catalyzed alcohol
oxidation via Oppenauer chemistry. Journal of American Chemical Society. 128, 1259612597.
32. Backvall, J. (2004). Modern oxidation methods. Wiley-VCH.
33. Naota, T, Takaya, H. and Murahashi S. (1998). Ruthenium-catalyzed reactions for organic
synthesis. Chemical Reviews. 98, 2599-2660.
34. Silverman, M. and Levy, H. (1954). Polarographic Studies of Ruthenium in Oxidation States
IV, VI, VII and VIII1. Journal of the American Chemical Society. 76, 3319-3321.
35. Sharpless, K, Akashi, K. and Oshima, K. (1976). Ruthenium catalyzed oxidation of alcohols
to aldehydes and ketones by amine-n-oxides. Tetrahedron Letters. 17, 2503-2506.
36. Muller, P. and Godoy, J. (1981). Catalyzed oxidation of alcohols and aldehydes with
iodosylbenzene. Tetrahedron Letters. 22, 2361-2364.
37. Murahashi, S, Naota, T, Oda, Y. and Hirai, N. (1995). Ruthenium-catalyzed oxidation of
alcohols with peracids. Synlett. 733-734.
38. Rothenberg, G, Barak, G. and Sasson Y. (1999). Unusual phase transfer mechanism of the
ruthenium-catalyzed oxidation of alcohols with hydrogen peroxide. Tetrahedron. 55, 63016310.
39. Kim, K, Kim, S, Song, Y. and Hahn, C. (1987). Selective oxidation of allylic and benzylic
alcohols using potassium ruthenate (K2RuO4) under phase-transfer catalysis conditions.
Synthesis. 11, 1017-1018.
40. Fernandez, M. (2007). Oxidation of alcohols to carboxylic acids. NY: Springer.
41. Singh, B. and Srivastava, S. (1991). Kinetics and mechanism of ruthenium tetroxide
catalysed oxidation of cyclic alcohols by bromate in base. Transition Metal Chemistry. 16,
466-468.
42. Desai, S, Halligudi, N. and Nandibewoor, S. (2002). Kinetics and mechanism of ruthenium
(III)-catalysed oxidation of allyl alcohol by acid bromate-autocatalysis in catalysis.
Transition Metal Chemistry. 27, 207-212.

111

43. Djerassi, C. and Engle, R. (1953). Oxidations with ruthenium tetroxide. Journal of the
American Chemical Society. 75, 3838-3840.
44. Nakata, H. (1963). Oxidation reaction of steroid alcohols by ruthenium tetroxide.
Tetrahedron. 19, 1959-1963.
45. Parikh, V. and Jones, J. (1965). Oxidation of sugars with ruthenium dioxide-sodium
periodate: a simple method for the preparation of substituted keto sugars. Canadian Journal
of Chemistry. 43, 3452-3453
46. Caputo, J. and Fuchs, R. (1967). The oxidation of cyclobutanols and aromatic rings with
ruthenium tetroxide. Tetrahedron Letters. 1967, 4729-4731.
47. Plietker, B. (2005). Selectivity versus reactivity – recent advances in RuO4-catalyzed
oxidations. Synthesis. 15, 2453-2472.
48. Deng, L.; Ziegler, T. (1997). Theoretical study of the oxidation of alcohol to aldehyde by d(0)
transition-metal-oxo complexes: Combined approach based on density functional theory and
the intrinsic reaction coordinate method Organometallics. 16, 716-724.
49. Lee, D. and Vandenen, M. (1972). Oxidation of 2-propanol by ruthenium tetroxide. Canadian
Journal of Chemistry. 50, 2000-2009.
50. Camel, V. and Bermond, A. (1998). The use of ozone and associated oxidation processes in
drinking water treatment. Water Research. 32, 3208-3222.
51. Waters, W, Rollin, A, Bardwell, B, Schneider, J. and Aanerud, T. (1976). Oxidation of
secondary alcohols with ozone. Journal of Organic Chemistry. 41, 889-891.
52. Nishimura, T, Maeda, Y, Onoue, T, Ohe K. and Uemura, S. (2000). Palladium(II)-catalysed
oxidation of alcohols under an oxygen atmosphere in a fluorous biphase system. Journal of
Chemical Society-Perkin Transactions. 24, 4301-4305.
53. Brink, G, Arends, I. and Sheldon, R. (2000). Green, catalytic oxidation of alcohols in water.
Science. 287, 1636-1639.
54. Besson, M. and Gallezot, P. (2000). Selective oxidation of alcohols and aldehydes on metal
catalysts. Catalysis Today. 57, 127-141.
55. Mallat, T. and Baiker, A. (2004). Oxidation of alcohols with molecular oxygen on solid
catalysts. Chemical Reviews. 104, 3037-3058.
56. Knozinger, H. (1997). Handbook of heterogeneous catalysis. Weinheim: VCH.
57. Hoelderich, W. (2000). Environmentally benign manufacturing of fine and intermediate
chemicals. Catalysis Today.62, 115-130.
58. Donze, C, Korovchenko, P, Gallezot, P. and Besson, M. (2000). Aerobic selective oxidation
112

of (hetero) aromatic primary alcohols to aldehydes or carboxylic acids over carbon supported
platinum. Applied Catalysis B: Environmental. 70, 621-629.
59. Vinke, P, Van Dam, H. and Van Bekkum, H. (1990). Platinum catalyzed oxidation of 5hydroxymethylfurfural. Studies in Surface Science and Catalysis. 55, 147-158.
60. Keresszegi, C, Burgi, T, Mallat, T. and Baiker, A. (2002). On the role of oxygen in the
liquid-phase aerobic oxidation of alcohols on palladium. Journal of Catalysis. 211, 244-251.
61. Yamada Y, Arakawa, T, Hocke, H. and Uozumi, Y. (2007). A Nanoplatinum Catalyst for
Aerobic Oxidation of Alcohols in Water. Angewandte Chemie-International Edition. 46,
704-706.
62. Wang, T, Shou, H, Kou, Y. and Liu, H. (2009).Base-free aqueous-phase oxidation of nonactivated alcohols with molecular oxygen on soluble Pt nanoparticles. Green Chemistry. 11,
562-568.
63. Yamaguchi, K. and Mizuno, N. (2003). Scope, kinetics, and mechanistic aspects of aerobic
oxidations catalyzed by ruthenium supported on alumina. Chemistry-A European Journal. 9,
4353-4361.
64. Nishimura, T, Onoue, T, Ohe, K. and Uemura S. (1998). Pd(OAc)(2)-catalyzed oxidation of
alcohols to aldehydes and ketones by molecular oxygen. Tetrahedron Letters. 39, 6011-6014.
65. Marko, I, Giles, P, Tsukazaki, M, Chelle-Regnaut, I, Urch, C. and Brown, S. (1997).
Efficient, aerobic, ruthenium-catalyzed oxidation of alcohols into aldehydes and ketones.
Journal of the American Chemical Society. 119, 12661–12662.
66. Uozumi, Y. and Nakao, R. (2003). Catalytic oxidation of alcohols in water under
atmospheric oxygen by use of an amphiphilic resin-dispersion of a nanopalladium catalyst.
Angewandte Chemie-International Edition. 42, 194-197.
67. Sheldon, R, Arends, I. and Dijksman, A. (2000). New developments in catalytic alcohol
oxidations for fine chemical synthesis. Catalysis Today. 57, 157-166.
68. Sheldon, R. (1993). Homogeneous and heterogeneous catalytic oxidations with peroxide
reagents. Topics in Current Chemistry. 1993, 164, 21-43.
69. Martin, S. and Garrone, A. (2003). Efficient solvent-free iron(III) catalyzed oxidation of
alcohols by hydrogen peroxide. Tetrahedron Letters. 44, 549-552.
70. Sato, K, Aoki, M, Tagaki, J. and Noyori, R. (1997). Organic solvent- and halide-free
oxidation of alcohols with aqueous hydrogen peroxide. Journal of the American Chemical
Society. 119, 12386–12387.
71. Barak, G, Dakka, J. and Sasson, Y. (1988). Selective oxidation of alcohols by a H2O2-RuCl3
system under phase-transfer conditions. Journal of Organic Chemistry. 53, 3553-3555.

113

72. Bogdal, D. and Łukasiewicz, M. (2000). Microwave-assisted oxidation of alcohols using
aqueous hydrogen peroxide. Synlett. 1, 143–145.
73. Notari, B. (1996). Microporous crystalline titanium silicates. Advances in Catalysis. 41, 253334.
74. Arends, I, Sheldon, R, Wallau, M. and Schuchardt, U. (1997). Oxidative transformations of
organic compounds mediated by redox molecular sieves. Angewandte Chemie-International
Edition. 36, 1144-1163.
75. Adam, W, Corma, A. and Reddy, T. (1997). Diastereoselective epoxidation of allylic
alcohols with hydrogen peroxide catalyzed by titanium-containing zeolites or
methyltrioxorhenium versus stoichiometric oxidation with dimethyldioxirane: Clues on the
active species in the zeolite lattice. Journal of Organic Chemistry. 62, 1997, 3631-3637.
76. Canterino, M, Marotta, R, Di Somma, I. and Andreozzi, R. (2009). A kinetic investigation on
the ozonation of glycerol and its oxygenated derivatives. Ozone: Science & Engineering. 31,
445-453.
77. Hoigne, J. and Bader, H. (1983). Rate constants of reactions of ozone with organic and
inorganic compounds in water. I. Non-dissociating organic compounds. Water Research. 17,
173-183.
78. Zimin, Y. Trukhanova, N. Strel’tsova, I. and Komissarov V. (2000). Kinetics of polyol
oxidation by ozone in aqueous solutions. Kinetics and Catalysis. 41, 749–752.
79. Hoigne, J. and Bader, H. (1983). Rate constants of reactions of ozone with organic and
inorganic compounds in water. II. Dissociating organic compounds. Water Research. 17,
185-194.
80. Bailey, P. (1982). Ozonation in Chemistry. New York : Academic Press.
81. Von Gunten, U. (2003). Ozonation of drinking water: Part I. Oxidation kinetics and product
formation. Water Research. 37, 1443-1467.
82. Whiting, M, Bolt, A. and Parish, J. (1968). Reaction between ozone and saturated
compounds. Advances in Chemistry Series. 77, 4-14.
83. Staehelin, J. and Hoigne, J. (1985). Decomposition of ozone in water in the presence of
organic solutes acting as promoters and inhibitors of radical chain reactions. Environmental
Science & Technology. 19, 1206-1213.
84. Criegee, R. (1975). Mechanism of ozonolysis. Angewandte Chemie International Edition in
English. 14, 745-752.
85. Huang, C, Dong, C. and Tang, Z. (1993). Advanced chemical oxidation: Its present role and
potential future in hazardous waste treatment. Wastewater Management. 13, 361-377.

114

86. Sehested, K, Corfitzen, H, Holcman, J, Fischer, C. and Hart, E. (1991). The primary reaction
in the decomposition of ozone in acidic aqueous solutions. Environmental Science &
Technology. 25, 1589-1596.
87. Staehelin, J. and Hoigne, J. (1985). Decomposition of ozone in water: rate of initiation by
hydroxide ions and hydrogen peroxide. Environmental Science & Technology. 16, 678-681.
88. Acero, J. (2000). Influence of carbonate on the ozone/hydrogen peroxide based advanced
oxidation process for drinking water treatment. Ozone Science & Engineering. 22, 305-328.
89. Andreozzi, R, Caprio, V, Insola, A, Marotta, R. and Sanchirico, R. (2000). Advanced
oxidation processes for the treatment of mineral oil-contaminated wastewaters. Water
Research. 34, 620-628.
90. Reisz, E, Schmidt, W, Schuchmann, H. and Von Sonnta, C. (2003). Photolysis of ozone in
aqueous solutions in the presence of tertiary butanol. Environmental Science & Technology.
37, 1941-1948.
91. Popov, E, Mametkuliyev, M, Santoro, D, Liberti, L. and Eloranta, J. (2010). Kinetics of UVH2O2 advanced oxidation in the presence of alcohols: The role of carbon centered radicals.
Environmental Science & Technology. 44, 7827-7832.
92. Kasprzyk-Hordern, B, Ziole, M. and Nawrocki, J. (2003). Catalytic ozonation and methods
of enhancing molecular ozone reaction in water treatment. Applied Catalysis B:
Environmental. 46, 639-669.
93. Ni, C. and Chen, J. (2001). Heterogeneous catalytic ozonation of 2-chlorophenol aqueous
solution with alumina as a catalyst. Water Science and Technology. 43, 213-220.
94. Majewski, M, Gleave, M. and Nowak, P. (1995). 1,3-Dioxan-5-ones: Synthesis,
deprotonation, and reactions of the lithium enolates. Canadian Journal of Chemistry. 73,
1616-1626.
95. Piancatelle, G, Scettri, A. and D’Auria, M. (1982). Pyridinium chlorochromate: A versatile
oxidant in organic synthesis. Synthesis. 245-258.
96. Adkins, H, Elofson, R, Rossow, A. and Robinson C. (1949). The oxidation potentials of
aldehydes and ketones. Journal of the American Chemical Society. 71, 3622-3629.
97. Griffith, W. and Kwong, E. (2003). Alkene and alkyne oxidative cleavage catalyzed by RuO4
in environmentally acceptable solvents. Synthetic Communications. 33, 2945-2951.
98. Martin, V, Palazon, J. and Rodriguez, C. (2001). Ruthenium (VIII) oxide. Encyclopedia of
reagents for organic synthesis. NY: Wiley.
99. Mills, A. and Holland, C. (1997). Investigation into the nature of the oxoruthenate species
used to mediate the oxidation of an organic substrate by hypochlorite in a biphasic system.
115

Journal of Chemical Research. 368-369.
100. Beattie, J. (1990). Ruthenium-catalysed homogeneous oxidation processes. Pure Applied
Chemistry. 62, 1145-1146.
101. Rothenberg, G, Barak, G. and Sasson, Y. (1999). Unusual phase transfer mechanism of
the ruthenium-catalyzed oxidation of alcohols with hydrogen peroxide. Tetrahedron. 55,
6301-6310.
102. Haag, W. and Hoigen, J. (1993). Ozonation of bromide-containing waters: kinetics of
formation of hypobromous acid and bromate. Environmental Science & Technology. 17,
261-267.
103. Von Gunten, U. (2003). Ozonation of drinking water: Part II. Part II. Disinfection andbyproduct formation in presence of bromide, iodide or chlorine. Water Research. 37, 1469-1487.

116

Chapter V. Design and Construction of Reactive Distillation
Column
5.1. Overview
In 2008, a patent [1] filed by colleagues in our research group at MSU proposed the
concept of making glycerol acetals and biodiesel in large scale using a reactive-distillation
technique. To demonstrate the viability of this process, an elevated pressure reactive distillation
column was constructed for pilot scale test. In this chapter, the advantage of reactive distillation
operation over conventional biodiesel process is discussed, and the parameter design and
construction of this column is presented.

5.2. Introduction
Continuous oil transesterification to make biodiesel was proposed as early as the 1940s [2,
3] and has been studied continuously since [4-8]. Current biodiesel production primarily uses
batch reactors with alkali-catalysts and downstream separation units. In such a process, raw
materials and catalysts react in the batch reactor and produce a two-phase mixture. The upper
phase consists of esters, alcohol and catalysts; the lower phase is rich in glycerol and residual
transesterified triglycerides. The mixture is further separated in a flash chamber or distillation
column to obtain qualified biodiesel product. This process is continuous and fast, but needs labor
and intense energy. Because, in this process, excess alcohol must be used to improve oil
conversion, thus a large amount of energy is required to recover the excess alcohol and
byproduct glycerol, and catalyst waste in the solution requires treatment.
A prospective alternative to the classic process is using reactive distillation (RD). RD is a
unit operation that integrates reaction and separation in a single vessel. It is increasingly being
117

applied to various chemical systems, i.e. etherification, esterification, ester hydrolysis and
acetalization, etc [9].
Recently, in our lab, colleagues have successfully demonstrated the use of reactive
distillation to recover propylene glycol and ethylene glycol from aqueous solution [10], and to
form ethyl lactate from aqueous lactic acid solution and ethanol [11].
RD holds a promising potential to reduce investment and operating costs by overcoming
chemical equilibrium, achieving energy integration and performing difficult separations. These
features are appropriate solutions for the problems encountered in biodiesel industry today.
However, very few studies on RD in biodiesel production have been reported, especially for
acid-catalyzed processes. In one of these studies, He developed a RD system for biodiesel
synthesis using an alkali-catalyst [12]. He found that 1) short reaction time and high unit
productivity can be achieved in this unit, and 2) much lower excess alcohol is required for the
reaction. Yet, this process cannot avoid saponification problems caused by alkali-catalysts. Thus,
as described in Chapter I, a more comprehensive RD process is proposed with the application of
heterogeneous acid catalysts.
To demonstrate the viability of this process, an elevated pressure reactive distillation
column was designed and constructed for pilot scale test.

5.3. Column construction
The pilot-scale reactive distillation column at MSU was designed with a maximum
capacity of producing 3 L of glycerol acetal per hour. The temperature and pressure allowed are
o

190 C and 2 MPa (300 psi), respectively.
As shown in Figure 5-1, the main body of the column consists of six sections. Five of
them are 3 ft (91.5 cm) in length; the top section is 1 ft (30.5 cm). On the top, there is an elbow
118

connecting the column and the condenser. The material of the column is 304 stainless steel (2
inch, schedule 40).
At the bottom of the column, there is a reboiler with a 2.4 kW heating system. The
reboiler working volume is around 1.5 L. It is equipped with a level sensor and solenoid valve to
maintain a constant level and allow product withdrawal.
On the top of the column, there is a total condenser with a chiller capable of achieving a
o

condenser temperature of -20 C. A stainless steel reservoir is used to collect the condensed
chemical and allow product withdrawal. Between the condenser and reservoir, there is a solenoid
valve which adjusts the column pressure by releasing or accumulating condensed chemical in the
condenser. In addition, the reservoir has an overflow outlet to maintain the liquid level inside.
The column is equipped with three pumps, one reflux pump which returns condensed
chemical to the column, and two feed pumps which dispense feed solutions to the column at a
controlled rate.
The column has several ports along its length allowing internal temperature measurement,
introduction of feed, and sample withdrawal. The column is wrapped with electric heating tapes
that are controlled by surface thermocouples and Omega controllers to a temperature just below
the internal column temperature to minimize the heat loss. Heat loss is further minimized by
insulating the column with bands of glass wool.

119

Figure 5-1 Reactive distillation column constructed for this project
120

Inside the column, there are three sections: top non-reactive stripping section (45.7 cm), a
reactive section, and a non-reactive enriching section (73.7 cm). The reactive section, 360 cm in
height, is packed with Katapak-S structured packing elements (Sulzer chemtech Ltd.) filled with
0.3 mm Aberlyst-70 cationic exchange resin as the catalyst. Empty Katamax structured packing
(Koch-Glitsch, Ltd.) is used to fill non-reactive column sections.
In addition, safety relief valve and rupture disk are installed on the top to protect the
column against over pressure.

5.4. Design and calculations
5.4.1. Temperature and pressure limit
o

The operation temperature of the column cannot exceed 190 C which is the maximum
operation temperature of Amberlyst-70 cationic exchange resin. In addition, saturated pressure of
the distillation compound at the operation temperature must be less than the designed column
pressure (300 psi).
The relation of saturated pressure and operation temperature can be estimated using
Antoine’s Equation:

logP = A-

B
B
−C
or T =
C+T
A − log P

(1)

Table 5-1. presents the values of Antoine coefficients [13] and the calculation of
operation temperature allowed at 300 psi. For different systems, the smaller number between
o

190 C and Tmax must be used as the temperature limit of the column.

121

Table 5-1. Antoine coefficients of some chemicals and temperature allowed at Psat = 300 psi.
o

Chemical

A

B

C

Tmax ( C) @ Psat =
300 psi

Acetaldehyde

7.31377

1175.27

244.272

132

Ethanol

8.13484

1662.48

238.131

183

Methanol

8.09126

1582.91

239.096

166

Water

8.05573

1723.64

233.076

212

5.4.2. Pipe thickness
The formula [14] used to calculate minimum pipe thickness tm required for a designed
pressure (t < D/6) is

tm =

PD0
+C
2( SE + PY )

(2)

Values of the parameters used in Eq. (2) are listed in Table 5-2.
For 2 inch SS 304 schedule 40 pipe, the calculated value of tm at 300 psi is 0.132 inches.
It is less than the standard wall thickness of this pipe which is 0.154 inches. Thus, this pipe was
selected for the column construction.

122

Table 5-2. Parameters used for minimum wall thickness calculation.
”

Nomenclature

Unit

Values for 2 SS 304 schedule
40 pipe

P

design pressure

Psi

300

D0

outside diameter of pipe

Inch

2.375

C

Sum of allowance
(corrosion, thread depth)

Inch

0.085

S

Basic allowable stress at
maximum temperature

Psi

10,000 (at 315 C)

E

Longitudinal weld joint
factor

--

0.8

Y

Coefficient

--

0.4

o

5.4.3. Reboiler
The reboiler (Figure 5-2) of the reactive distillation column is purchased from Pope
Scientific Inc., Saukville, WI. The working volume is around 1.5 L. It is equipped with a level
sensor and solenoid valve to maintain a constant level and allow product withdrawal.

123

Figure 5-2. Reboiler of reactive distillation column.
Initially, two 1.5 kW cartridge heaters (14 inches in length, 1.25 inches diameter) were
equipped with the reboiler to generate chemical vapor. The heaters, mounted on a blind flange,
were kept in the lower half of the reboiler to ensure constant immersion in liquid when the
reboiler is half full. However, during operation, the blind flange was heated to red hot due to its
direct contact with the heaters. To solve this problem, the cartridge heaters were replaced by two
304 stainless steel coil heaters (1.2 kW, 12 feet total length, 0.125 inches diameter). The coil
heaters have 5 inches of transition piece (cold zone) at the end to avoid overheating of the flange.

5.4.4. Condenser
The condenser equipped with the column was purchased from Exergy, LLC. It is a 316
stainless steel tube-in-tube heat exchanger (Figure 5-3). The diameters of inside and outside
tubes are 0.25 and 0.5 inches, respectively. The total heat exchange surface area is 44.5 square
inches. During operation, vapor condensed in the inside tube, while cooling water flows in the
outside tube. The designed capacity of this condenser is 3000 W.

Figure 5-3. Condenser used for the reactive distillation column.
5.4.5. Sensor wiring

124

Temperature profile along the column is monitored manually by thermocouple controllers
installed on a control box. Other sensors including pressure transducer, level sensor, relay and
solenoid valves are connected to a data acquisition system (purchased from National Instruments
Corporation, model number: USB 6259) and controlled by Labview software.

Pressure control
The major components of the pressure control system include pressure transducer, relay,
24V DC power and solenoid valve. The pressure transducer (model number: PX41T0-500AI)
and relay (model number: SSR330DC10) are purchased from Omega.
Figure 5-4 displays the wiring diagram of the pressure control system. During operation,
the data acquisition unit measures the amplitude of the signal generated from the pressure
transducer. A Labview program computes the amplitude measured by the data acquisition unit to
determine a reflectivity. When the signal is higher than the setpoint, the program triggers the DC
power channel of the data acquisition unit, and then the relay turns on the solenoid valve to
release pressure from the column.

125

Figure 5-4. Wiring diagram of the pressure control system of the column.
Liquid level control
The magnetic sensor used to control the liquid level in the distillate collector is purchased
from National magnetic sensors, Inc (model number: BB212OO43OO510-11S2S). As shown in
Figure 5-5, the level sensor is an on-off type float switch that when the liquid level reaches the
setpoint, it closes the circuit and turns on the solenoid valve to drain liquid from the distillate
collector.

126

Figure 5-5. Wiring diagram of the liquid level control system of the column

127

References

128

References
1. Miller, D, Peereboom, L, Kolah, A. and Asthana, N. (2008). Process for production of a
composition useful as a fuel. US Patent. 7321052.
2. Trent, W. (1945). Process of treating fatty glycerides. US Patent. 2383632.
3. Allen, H, Rock, G. and. Kline, A. (1945). Process for treating fats and fatty oils. US Patent.
2383579.
4. Pinto, A, Guarieiro, L, Rezende, M, Ribeiro, N, Torres, E, Lopes, W, Pereira, P. and Andrade,
J. (2005). Biodiesel: An overiew. Journal of the Brazilian Chemical Society. 16 1313-1330.
5. Noureddini, H, Harkey, D. and Medikonduru, V. (1998). Continuous process for the
conversion of vegetable oils into methyl esters of fatty acids. Journal of the American Oil
Chemists' Society. 75, 1775-1783.
6. Peterson, C, Cook, J, Thompson, J. and Taberski, J. (1999). Continuous Flow Biodiesel
Production. Applied Engineering in Agriculture. 18, 5-11.
7. Darnoko, D. and Cheryan, M. (2000). Continuous production of palm methyl esters. Journal
of the American Oil Chemists' Society. 77, 1269-1272.
8. Antolin, G, Tinaut, F, Briceno, Y, Castano, V, Perez, C. and Ramirez, A. (2002).
Optimization of biodiesel production by sunflower oil transesterification. Bioresource
Technology. 83, 111–114.
9. Sundmacher, K. and Kienle, A. (2003). Reactive Distillation-Status and future directions.
Weinheim: Wiley-Vch.
10. Dhale, A, Myrant, L, Chopade, S, Jackson, J. and Miller, D. (2004). Propylene glycol and
ethylene glycol recovery from aqueous solution via reactive distillation. Chemical
Engineering Science. 59, 2881-2890.
11. Navinchandra, A, Kolah, A, Dung, V, Lira, C. and Miller, D. (2005). A continuous reactive
separation process for ethyl lactate formation. Organic Process Research & Development. 9,
599-607.
12. He, B, Singh, A. and Thompson, J. (2006). A novel continuous-flow reactor using reactive
distillation for biodiesel production, Transactions of the ASABE. 49, 107-112.
13. Yaws, C. (2009). Yaws' Handbook of Antoine Coefficients for Vapor Pressure (2nd
Electronic Edition). NY: Knovel.
14. Perry, R. and Green, D. (1997). Perry's chemical engineers' handbook. NY: McGraw-Hill.

129

Chapter VI. A Novel Process Concept to Make 1,3Dihydroxyacetone from Glycerol
Using 5-hydroxyl-2-methyl-1,3-dioxane as intermediate provides relatively quick,
environmentally safe and economical routes for production of 1,3-dihydroxyacetone (DHA)
from glycerol. Based on the experimental results presented in the previous chapters, a novel
process of making DHA from glycerol is developed and described here.
Figure 6-1 presents the overall scheme of a novel DHA production process based on
acetaldehyde as the glycerol acetal forming reagent. In fact, this reagent can be any aldehyde,
ketone and acetal (eg. 1,1-diethoxyethane, DEE). For practical operation, selection of this
reagent is dependent on its price, availability and reproducibility. As shown in Figure 6-1, the
new concept of DHA production involves three major steps: 1) formation of glycerol acetal (GA)
and reactive distillation to get the 6-membered GA isomer 5-hydroxyl-2-methyl-1,3-dioxane
(cis-6); 2) oxidation of cis-6 GA isomer to the corresponding ketone product 2-methyl-1,3dioxan-5-one (G-tone); 3) conversion of G-tone to DHA via hydrolysis. Acid-catalyzed
hydrolysis of G-tone (Scheme 6-1) is a very fast and complete reaction. In this reaction, a
molecule of acetaldehyde is produced with DHA. In one experiment, 5 wt% (weight to substrate)
of Amberlyst-15 cationic exchange resin was added to the final mixture of Reaction 14 in Table
4-5. When the mixture was heated to 30°C, complete hydrolysis of G-tone was achieved in 10
min.

Scheme 6-1. Processes of catalytic conversion of glycerol into useful chemicals.
O

O

O
O

+

H 2O

Acid

+
OH

130

OH

CH 3 CHO

Figure 6-1. Design of new DHA production process.

131

Figure 6-1. Cont’d

In the first stage of this process, glycerol and acetaldehyde are introduced to Reactor 1 to
form cyclic glycerol acetals. The reaction mixture, which contains four isomers of glycerol acetal,
is then evaporated to get rid of water and unreacted acetaldehyde for recycling. After that, the
mixture of glycerol acetal isomers is introduced to a vacuum distillation column, in which the
cis-6 GA isomer can be separated and obtained from the top of this column. At the same time,
other isomers, which have higher boiling points, react in the reboiler to re-establish isomerization
equilibrium. The continuous recycling of the less-volatile glycerol acetal isomers through the

132

equilibration-distillation procedure can transform all of the products into the desired 6-membered
structure.
In the second stage of this design, cis-6 GA isomer is introduced to the oxidation Reactor
2 to form 2-methyl-1,3-dioxan-5-one (G-tone). A preferred operation for this stage is to use
ruthenium tetroxide-related oxidation.
In the last stage, G-tone and an excess amount of water are reacted, perhaps in a reactivedistillation column, to hydrolytically decompose G-tone to the desired product DHA and
acetaldehyde, which can be recycled or reformed into DEE. The possible application of reactive
distillation along with a large amount of water will overcome any chemical equilibrium
limitations of this reaction, such that a high yield of DHA can be achieved from this process.
Storage and handling of DHA have some special requirements [1]. DHA presents as
monomer in aqueous solution, but can gradually tautomerize into glyceraldehydes in alkaline
environment (Scheme 6-2). Under alkaline conditions, various isomerization and condensation
reactions occur and ultimately lead to the formation of brown colored oligomers [1].

Scheme 6-2. Interconversion between DHA and glyceraldehydes [1].

Alkaline envrionment

Acidic environment
O

OH

OH

OH

OH

O
o

DHA content can be dramatically reduced if stored at temperature higher than 40 C. Via
decomposition or rearrangement reactions, aldehydes and ketones e.g. methyl-glyoxal can be
o

formed [1]. The recommended storage temperature for DHA is 4 C.

133

Overall, the novel process developed in this study is a green and environmentally friendly
process that no polluted gas is generated. After reaction, catalysts and organics, which are
acetaldehyde or ethanol, can be easily recovered and recycled from the aqueous solution. From
this point of view, the entire process only consumes glycerol and oxygen (Scheme 6-3). These
features make it a more economical approach than current DHA production technologies; and it
holds great promise of achieving commercialization.

Scheme 6-3. Overall reaction of the novel process to make DHA from glycerol.
O

OH

+
O

+

1/2 O2
OH

O

134

OH

H 2O

Reference

135

Reference
1. http://www.dihydroxyacetone-dha.com/

136

Chapter VII. Summary and Conclusions
Glycerol transacetalization with 1,1-diethoxyethane (DEE) in the presence of Amberlyst15 cationic exchange resin gives the cis- and trans- forms of 5-membered and 6-membered cyclic
acetals. At 25°C, equilibrium conversion of reactants can be achieved in 2 h. The reaction is best
carried out at low temperature to avoid formation of ethyl vinyl ether from DEE, and with DEE
as the limiting reactant to avoid over-acetalization of the primary acetal products. A kinetic study
was carried out to understand the transacetalization process and establish the theoretical basis for
large scale glycerol acetal production. The result shows that the reaction can be described by a
second order kinetic model.
Further investigation of glycerol acetal product from the reaction shows that in the
presence of acid catalyst, the four isomers interconvert to form an equilibrated mixture that
contains similar quantities of the four products. Vacuum distillation can be used to obtain each
individual isomer with a purity in excess of 90%. The distillation experiments also indicate that
the cis-6 dioxane isomer (2-methyl-5-hydroxy-1,3-dioxane) is the most volatile, and is distilled
preferentially such that nearly pure cis-6 isomer is initially produced. When an acid catalyst is
added to the mixture of isomers such that interconversion can occur during distillation, it is
possible to produce the cis-6 isomer in nearly quantitative yield from glycerol. Based on these
results, a reactive distillation process is proposed to produce pure cis-6 isomer as the sole product
from glycerol continuously.
Oxidation of cis-6 GA isomer to the corresponding ketone product (2-methyl-1,3-dioxan5-one) provides a new synthesis route for 1,3-dihydroxyacetone (DHA) production. A broad
range of catalysts and oxidizing reagents including green oxidants such as molecular oxygen,

137

hydrogen peroxide and ozone were then tested for this purpose. Due to the heterocyclic ring
structure and high oxidation potential of cis-6, most of the reagents exhibited poor activity.
However, ruthenium tetroxide-based oxidation demonstrated high activity toward oxidation of
cis-6. With this reagent, complete conversion of cis-6 to its ketone product was achieved in 1 h at
room temperature.
Based on this discovery, a novel process is designed for DHA production. In this route,
glycerol is first converted to cis-6 glycerol acetal isomer. After that, the cis-6 isomer is oxidized
to its ketone product and followed by a hydrolysis process to the final DHA product.
This new method has many advantages over the current DHA production technologies.
Comparing to the conventional microbial approaches, this method does not require expensive
microorganisms for fermentation; and has much shorter reaction time. Comparing to direct
oxidation of glycerol, oxidation of the cis-6 substrate, with the primary alcohols blocked by a
cyclic acetal structure, gives a much higher oxidation selectivity towards the secondary hydroxyl
group. We estimate that a DHA yield in excess of 90% can be achieved from glycerol using this
technology. Thus, the new method holds a good potential for commercial success.

138

Chapter VIII. Suggestions for Future Work
In this study, the experimental results indicate that preparation of the 6-membered GA
isomer (cis-6) and hydrolysis of 2-methyl-1,3-dixoan-5-one (G-tone) to the final product 1,3dihydroxyacetone (DHA) can be carried out continuously with high yield (over 90%). The
biggest challenge to accomplish DHA production via the process proposed in the previous
chapter is oxidation of cis-6 and separation of G-tone from the reaction mixture.
Cis-6, with heterocyclic structure and high oxidation potential, is beyond the capacity of
many conventional oxidation reagents. The most effective methods we discovered so far is using
ruthenium based catalyst and oxidizing salts as consuming oxidant. Although complete oxidation
of cis-6 can be achieved using this method, separation of G-tone product from the reaction
mixture which contains soluble ruthenium species and inorganic salts is quite a challenge.
Thus, a more convenient oxidation method is highly desirable to make this process
eligible for commercialization purpose. As for future work recommendation, I would suggest the
following research directions for consideration.
First, development of new catalysts for vapor phase oxidation of cis-6. Comparing with
liquid phase reaction, vapor phase oxidation has a big advantage that it employs only pure cis-6
and air or molecular oxygen as reactant. Hence, the tedious separation operations encountered in
conventional oxidation processes can be eliminated. Copper [1] and nickel catalysts [2] that have
been used for aerobic oxidation of methanol and other alcohol substrates to their carbonyl
products could be good candidates for this purpose.
Second, application of electro-catalytic method for cis-6 oxidation. In 2005, Ciriminna et
al. [3] reported an electro-catalytic oxidation method to make DHA from glycerol. They obtained

139

an oxidation mixture with 25% DHA yield in 20 h. If an electro-catalytic process can be
developed for direct cis-6 oxidation or ruthenium tetroxide catalyzed cis-6 oxidation, the
following separation process can also be considerably simplified.
Third, dehydroxylation of cis-6 to a 1,3-propanediol precursor. As discussed in the
Chapter III, cis-6 has a high oxidation potential and is less likely to be oxidized. From the
opposite point of view, it may be a good substrate for reduction reaction such as dehydroxylation.
Thus, if an appropriate method is found to reduce cis-6 to 2-methyl-1,3-dioxane, cis-6 can serve
as a good intermediate for 1,3-propanediol synthesis.

140

References

141

References
1. Davies, P. and Mariotti, G. (1996). Oxidation of methanol at Cu(110) surfaces: New TPD
studies. Journal of Physical Chemistry. 100, 19975-19980.
2. Ma, Z. and Zaera, F. (2006). Organic chemistry on solid surfaces. Surface Science Reports.
61, 229-281.
3. Ciriminna, R, Palmisano, G, Pina C, Rossi, M. and Pagliaro, M. (2005). One-pot
electrocatalytic oxidation of glycerol to DHA. Tetrahedron Letters. 47, 6993-6995.

142

Appendices

143

Appendix I. Mass Spectra of the Glycerol Transacetalization
Products
GC and GC-MS analysis of the reaction mixture from glycerol transacetalization with
DEE is presented in Figures A1-1 to A1-9.

Figure A1-1. Gas chromatogram of transacetalization product mixture.

144

Figure A1-2. Mass spectrum of Peak E

145

Figure A1-3. Mass spectrum of Peak F

146

Figure A1-4. Mass spectrum of Peak G

147

Figure A1-5. Mass spectrum of Peak H

148

Figure A1-6. Mass spectrum of Peak I

149

Figure A1-7. Mass spectrum of Peak J

150

Figure A1-8. Mass spectrum of Peak K

151

Figure A1-9. Mass spectrum of Peak L

152

Glycerol acetal isomers were identified by their retention index and mass spectrum as
compared with Ferreira’s report [1]. Comparing the experimental result and the reported mass
spectrum [1], we can conclude that the isomers of the glycerol acetal obtained in this study are
cis-5-hydroxy-2-methyl-1,3-dioxane (Peak E), cis-4-hydroxymethyl-2-methyl-1,3-dioxolane
(Peak H), trans-4-hydroxymethyl-2-methyl-1,3-dioxolane (Peak I) and trans-5-hydroxy-2methyl-1,3- dioxane (Peak J).

153

Structure analysis of species F, G, K, and L is presented in Table A1-1 and A1-2.

Table A1-1. Molecule fracture identification for species F and G.
m/z
189

175

Fracture
M-1

M-15

O

O CH
2

O

O

O

CH2

O

O

or

145

O

M-45
O

M-73

M-89

O

CH

O

O

O

CH3

CH
O

or
CH2

CH
M-105

O

or

O

85

O

O

CH3

O

101

O

or

O
117

CH

O CH

CH CH3
O C
H2

CH

or

CH CH3
C O
H2

73

M-117

-CHCH3(OCH2CH3)

58

M-132

-CHO CH2CH3

45

M-145

-OCH2CH3

29

M-161

-CH2CH3

So, the structures of species F and G are

154

O

O

O
O

O

O

O

O
O

and

.

They are 4-(1-ethoxy-ethoxymethyl)-2-methyl-1,3-dioxalane and 5-(1-ethoxy-ethoxy)-2methyl-1,3-dioxane, M.W. = 190.

155

Table A1-2 molecule fracture identification for species K and L.

m/z
163

Fracture
M-1
OH

HO

149

M-15

O

HO

O

O

OH

133

M-31

CH

or

O

CH
O

O

OH

119

M-15

O

CH2

CH2

O

or OH
CH

O

HO

OH

HO
O

OH

or

CH

CH2

HO

75

M-89

73

M-91

-CHCH3(OCH2CH3)

58

M-106

-CHO CH2CH3

45

M-119

-OCH2CH3

29

M-135

-CH2CH3

or HO

OH

So, the structures of species K and L are

156

CH

OH

OH

HO
O
O

O

HO

and

O

OH

M.W. 164.

157

Reference

158

Reference
1. Ferreira, A, Barbe, J. and Bertrand, A. (2002). Heterocyclic acetals from glycerol and
acetaldehyde in port wines: Evolution with aging. Journal of agricultural and food chemistry.
50, 2560-2564.

159

Appendix II. Calculation of Weisz-Prater Criterion for
Ascertaining the Influence of Mass Transport on
Transacetalization Kinetics
The equation used to calculate Weisz-prater criterion is

RG ⋅ L2
Φ wp =
CS ⋅ DAB, eff
The values of the parameters used in this equation are listed in Table A2-1.

Table A2-1. Parameters used to calculate Weisz-prater criterion

Parameter

Denotation

Unit

Φ wp

Wiesz-Prater criterion

RG

observed reaction rate per unit
catalyst

L

Catalyst size, L=R/3

CS

bulk concentration of reactant

mol/L

DAB

molecular diffusion coefficient

m /s

ε

porosity of catalyst

τ

Tortuosity ( τ ≅ 1/ ε )

mol/L.cat•s
m

2

o

For reaction performed at 40 C with 1 wt% of Amberlyst-15 cationic exchange resin
3

catalyst (0.0608 g, density = 1.2 g/cm ) and reactant molar ratio glycerol:DEE:ethanol = 1:1:12,
the observed reaction rate constant is 9.0 ×10−5 L/mol ⋅ s . The initial concentration of glycerol
and DEE is 1.1 mol/L (1:1:12 glycerol:DEE:EtOH); volume of solution is 60 ml. Catalyst
160

3

volume is 0.051 cm when it is saturated with ethanol, the catalyst swells to twice its original
3

size, so the volume to be used is 0.102 cm . Then, the observed reaction rate per unit catalyst is

RG = (9.0 ×10−5 L / mol ⋅ s ) × (1.1mol / L)2 × 60ml ÷ 0.102cm3 = 0.06mol / L.cat ⋅ s
The characteristic length L is related to particle size as L = R/3. For dry Amberlyst-15
catalyst, the average radius is 0.15 mm. Considering swelling by a factor of two in ethanol,
1
L = 0.15 × 2 3

/ 3 = 0.063mm

CS is the bulk concentration of reactant,

CS = 1.1mol / L
DAB,eff is the effective molecular diffusion coefficient, DAB,eff = ε ⋅ DAB / τ . Catalyst
porosity ( ε ) for the dry catalyst is taken as 0.5; tortuosity, ( τ ) is taken as 1/ ε according to the
classical random pore model. Considering catalyst swelling,

ε = 1 − 0.5 / 2 = 0.75 , τ = 1/ 0.75 = 1.33
The molecular diffusion coefficient DAB can be evaluated using Wilke-Chang equation,

DAB = 7.4 ×10−12

T ⋅ ϕB ⋅ M B
μ B ⋅ν 0..6
A

The values of the parameters used in Wilke-Chang equation are listed in Table A2-2.

161

Table A2-2. Parameters used in Wilke-Chang equation [1]

Parameter

Denotation

Value

Unit

ϕB

association factor of solvent

1.5

MB

molecular weight of solvent

46

T

Temperature

313

K

μB

viscosity of solvent

1.2

mPa•s

vA

molar volume of solute

73

cm

DAB = 7.4 ×10−12 ⋅

3

T ⋅ ϕB ⋅ M B
313 ⋅ 1.5 ⋅ 46
= 7.4 ×10−12 ⋅
= 1.22 ×10−9 m2 / s
0..6
0.6
1.2 ⋅ 73
μ B ⋅ν A

DAB, eff = ε ⋅ DAB / τ = 0.75 × (1.22 ×10−9 m2 / s ) /1.33 = 6.86 ×10−10 m2 / s
So, for this reaction,
Φ wp =

RG ⋅ L2
0.06mol / L ⋅ s × (0.063 × 10−3 m)2
=
= 0.34
CS ⋅ D AB,eff
1.1mol / L × 6.86 × 10−10 m 2 / s
In the same manner, the Weisz-Prater criterion for other reactions is calculated. The

value of ~0.3 indicates mild mass transport resistances in the kinetic experiments as noted.

162

Reference

163

Reference
1. Pryde, A. and Gilbert, M. (1979). Applications of high performance liquid chromatography.
New York: Wiley.

164

Appendix III. Calculation of High-Valent Ruthenium Species
in Aqueous Solution
The following reactions and equations are taken from Pourbaix’s book ‘Atlas of
electrochemical equilibria in aqueous solutions’ [1]. In these reactions, H2RuO5 is the dissolved
form of RuO4 species. E0,x represents the potential of the ionic solution.
−
H 2 RuO5 + e− ⇔ RuO4 + H 2O

E0， = 1.001 − 0.0591⋅ log
1

(1)

[ H 2 RuO5 ]
−
[ RuO4 ]

−
RuO4 + xH 2O + 4 H + + 3e− ⇔ RuO2 ⋅ xH 2O( s) + 2 H 2O

(2)

−
E0， = 1.533 + 0.0788 pH − 0.0197 ⋅ log[ RuO4 ]
2

H 2 RuO5 + 4 H + + 4e− ⇔ RuO2 + 3H 2O

(3)

E 0， = 1.400 + 0.0591 pH − 0.0148 ⋅ log[ H 2 RuO5 ]
3
−
2
RuO4 + e− ⇔ RuO4 −

E0， = 0.590 − 0.0591⋅ log
4

(4)
−
[ RuO4 ]

2
[ RuO4 − ]

2
RuO4 − + xH 2O + 4 H + + 2e− ⇔ RuO2 ⋅ xH 2O( s) + 2 H 2O

(5)

2
E0， = 2.005 + 0.1182 pH − 0.0295 ⋅ log[ RuO4 − ]
5

The standard potential E o for each reaction is the first item on the right side, i.e., for
reaction 1, the standard potential is 1.001 V.
165

−
To calculate the equilibrium constant for a system containing RuO2 ⋅ xH 2O , RuO4 and

H 2 RuO5 , we need to combine Reactions (1) and (2), or (2) and (3). For example,
(1) – 3 × (2)
−
RuO2 ⋅ xH 2O( s) + 3H 2 RuO5 ⇔ 4 RuO4 + ( x + 1) H 2O + 4 H +

(6)

The standard potential of Reaction (6) is
o
3 ⋅ E o =3 ⋅ E1 -3 ⋅ E o =-1.596
6
2
−
According to Nernst equation, at equilibrium, the correlation between [ RuO4 ] and

[ H 2 RuO5 ] must satisfy Eq. (6),
Eo =
6

−
⎞
[ RuO4 ]4
RT
0.0591 ⎛
log K =
log
− 4 pH ⎟
⎜
⎟
nF
3 ⎜
[ H 2 RuO5 ]3
⎝
⎠

(6)

Eq (6) can be simplified to
[ H 2 RuO5 ]0.75
−
[ RuO4 ]

=10(6.751− pH )

(7)

Eq (7) can also be obtained by relating reactions (1) and (3). From Eq (7), it can be seen
−
that RuO4 is favored in a basic pH, while H 2 RuO5 is favored in acidic pH.

−
2
In the same way, the correlation between [ RuO4 ] and [ RuO4 − ] can be calculated using

Eq (2), Eq (4) and Eq (5), and the result is
−
[ RuO4 ]0.6667
2
[ RuO4 − ]

=101.33⋅(11.9 − pH )

(8)

2
Eq (8) indicates that RuO4 − is favored at even higher pH value.

166

With Eq (7) and Eq (8), the distribution of the three Ru species can be calculated for a
given solution. For example, Figure A3-1 gives the distribution of Ru species between pH = 0 to
14 in an aqueous solution with total amount of [Ru] = 0.00038 mol/L. This result is in agreement
with the experimental data reported by Mills (Figure A3-2) [2].
100

Population (%)

80
60
40
20
0
0

2

4

6

8

10

12

14

pH

Figure A3-1. Calculated population (%) of various ruthenium species at different pHs in an
2-4
aqueous solution with [Ru] = 3.8×10 M. RuO4 (●), RuO4 (○) and RuO4 (△).

167

Figure A3-2. Population (%) of various ruthenium species at different pHs of an aqueous
2-4
solution with [Ru] = 3.8×10 M. Experimental results. RuO4 (●), RuO4 (○) and RuO4
(△). [2]

-

Electro-chemical calculation of RuO2 and BrO3 system
−
−
In a system containing RuO2 and BrO3 , RuO2 is oxidized to RuO4; while BrO3 may be

reduced to BrO − or Br − . To understand the interactions between the species, the possible half
reactions are listed below (taken from ref [1]).

H 2 RuO5 + 4 H + + 4e− ⇔ RuO2 ( s) + 3H 2O

(3)

E0， = 1.400 + 0.0591 pH − 0.0148 ⋅ log[ H 2 RuO5 ]
3
−
BrO3 + 4 H + + 4e− ⇔ BrO − + 2 H 2O

(9)

168

E0， = 1.341 − 0.0591 pH + 0.0148 ⋅ log
9

−
[ BrO3 ]

[ BrO − ]

−
BrO3 + 6 H + + 6e− ⇔ Br − + 3H 2O

(10)

E0， = 1.423 − 0.0591 pH + 0.0098 ⋅ log
10

−
[ BrO3 ]

[ Br − ]

−
If BrO3 was reduced to BrO − , the complete reaction can be expressed by combining

Reactions (3) and (9).
(9) – (3)
−
BrO3 + RuO2 ( s) + H 2O ⇔ H 2 RuO5 + BrO −

(11)

o
o
o
4 ⋅ E11 =4 ⋅ E9 -4 ⋅ E3 =4 × 1.341-4 ×1.400= -0.236
o
E11 = -0.059

According to Nernst equation, at equilibrium, the equilibrium constant can be calculated
from Eq. (12),
o
E11 =

[ H RuO5 ] ⋅ [ BrO − ]
RT
0.0591
log K =
= − 0.059
⋅ log 2
−
nF
4
[ BrO3 ]

(12)

K = 10−4
It indicates that this reaction is not thermodynamically favored. Therefore, BrO − is not
the reduction product in the system.
−
Likewise, if BrO3 was reduced to Br − , the complete reaction can be expressed by

combining Reactions (3) and (10).
2 × (10) – 3 × (3)

169

−
2 BrO3 + 3RuO2 ( s ) + 3H 2O ⇔ 3H 2 RuO5 + 2 Br −

(13)

o
o
o
12 ⋅ E13 =12 ⋅ E10 -12 ⋅ E3 =12 × 1.423-12 ×1.400= 0.276
o
E13 = 0.023

According to Nernst equation, at equilibrium, the equilibrium constant can be calculated
from Eq. (14),
o
E13 =

[ H 2 RuO5 ]3 ⋅ [ Br − ]2
0.0591
RT
log K =
=0.023
⋅ log
−
12
nF
[ BrO3 ]2

(14)

K = 104.67
The large value of the equilibrium constant indicates that this reaction is
thermodynamically favored. Therefore, Br − is the reduction product in the system. Also, note
that water is consumed in the reaction, which helps to explain why the reaction did not run
without water present.

170

References

171

References
1. Pourbaix, M. (1966). Atlas of electrochemical equilibria in aqueous solutions. NY: Pergamon
Press
2. Mills, A. and Holland, C. (1997). Investigation into the nature of the oxoruthenate species
used to mediate the oxidation of an organic substrate by hypochlorite in a biphasic system.
Journal of Chemical Research, Synopses. 368-369.

172