You are on page 1of 7

964 Ind. Eng. Chem. Res.

2010, 49, 964–970

New Process for Producing Epichlorohydrin via Glycerol Chlorination


E. Santacesaria,* R. Tesser, M. Di Serio, L. Casale, and D. Verde
Dipartimento di Chimica dell’UniVersità “Federico II” di Napoli, Italy

The strong growth of biodiesel production occurring in the last years has determined the availability of a
great amount of the byproduct glycerol. Many researches in the world are therefore oriented to find new
possible uses for glycerol also with the aim of reducing the cost of biodiesel. In this paper the chlorination
of glycerol with gaseous hydrochloric acid to obtain 1,3-dichlorohydrin and then epichlorohydrin will be
described. All the advantages of this process will be examined and discussed. The behavior of the different
proposed catalysts (normally compounds containing carboxylic acid groups), the reaction kinetics, the effect
of the catalyst concentration, the effect of HCl pressure, the vapor-liquid phase equilibria of the reaction
products in the reaction environment, and the most convenient operative conditions have been studied,
concluding with useful suggestions for the design of the industrial plants.

1. Introduction or minimized. All the occurring reactions have been studied in


In biodiesel production, about 10% by weight of glycerol is a previous work11 and there a kinetic approach, based on a
obtained as byproduct. Therefore, by increasing the amount of reliable reaction mechanism, has been developed. For this
biodiesel produced the availability of glycerol increases, too. It purpose, a reaction scheme of the following type can be
will be imperative, therefore, to find new uses for glycerol considered:
suitable to consume the large amounts of this substance derived
from the biodiesel production also with the aim of reducing
the overall costs. There are only two possible strategies to solve
this problem: to use glycerol as raw material for obtaining
commodities or to use glycerol for producing fuels additives.
A comparison of the kinetic behavior of different catalysts
For this purpose, the use of glycerol for producing acrolein and
have already been made in the mentioned work,11 but in the
acrylic acid1-4 is an example of the first type, while, the glycerol
present work this aspect will be further deepened with the aim
etherification (5,6) is an example of the second one. In this paper
to find a correlation between the structure of the catalyst and
we will study, in particular, the glycerol chlorination by using
the observed activities and selectivities. Moreover, the possibility
gaseous HCl to obtain as main product 1,3-dichlorohydrin,
of a continuous operation will also be demonstrated on
which is an important intermediate for synthesizing epichloro-
experimental bases.
hydrin used in the production of epoxy resins. Epichlorohydrin
At last, kinetic runs performed in total refluxing conditions,
is normally produced in industry starting from 1,2- (70%) and
that have been used for determining the kinetic laws and related
1,3-dichlorohydrins (30%) that are obtained in a mixture using
parameters11 have been compared with other runs performed
propylene as raw material. The predominance of 1,2-dichloro-
in HCl stripping conditions, by condensing, collecting, and
hydrin is a drawback of this process, because, the reaction rate
analyzing the stripped compounds along the run. The most
for obtaining epichlorohydrin from this isomer is about 20 times
abundant stripped compound was 1,3-dichlorohydrin because
slower than from 1,3-dichlorohydrin, and this has negative
of its relatively high volatility, demonstrating in this way the
consequences on the reactor sizing (a reactive distillation
possibility of a continuous operation with product recovery at
column) and on the formation of undesired byproduct.5,6 The
a high degree of purity. By using the kinetic approach developed
reaction of glycerol with hydrochloric acid, in the presence of
in a previous work,11 and with the vapor-liquid equilibrium
carboxylic acids as catalyst, occurs in two steps giving mainly
data described by UNIFAC model, in this paper we simulated
1-monochlorohydrin with small amounts of 2-monochlorohy-
the kinetic runs performed in stripping conditions by using a
drin. The entrance of a second chlorine in the last compound is
commercial process simulator (ChemCad 5.2), obtaining a good
strongly inhibited, and 1,3 dichlorohydrin is obtained almost
agreement between calculated and experimental measurements
exclusively in the second chlorination step. In the first papers
for both the reactor and the distillate reservoir.
and patents published on the subject, acetic acid was used as
Finally, by using the kinetic and modeling approach, also
catalyst and HCl in aqueous phase as reagent. In this condition
developed in previous literature work,5,6 we quantitatively show
the reaction rates were very slow. Different patents7-10 and
and discuss the advantage of using a feed stream of 1,3-
papers11,12 have recently been published on glycerol chlorination
dichlorohydrin instead of the mixture 1,2- (70%) and 1,3-
and the construction of some industrial plants has recently been
dichlorohydrins (30%), normally used in the process via propene
announced. All these papers and patents, published more
for producing epichlorohydrin.
recently, suggest the use of gaseous HCl and carboxylic acids
as catalyst, other than acetic acid, because they have boiling
points higher than that of acetic acid at 117 °C. In the presence 2. Experimental Section
of gaseous HCl the reaction is faster and by using the new 2.1. Reactants and Apparatus. All the reactants have been
proposed catalysts the loss of catalyst by evaporation is avoided purchased by Aldrich Co. at the maximum available degree of
* To whom correspondence should be addressed. E-mail: purity with the exception of gaseous hydrochloric acid that has
elio.santacesaria@unina.it. been purchased in a bomb from Air Liquide Italia Co.
10.1021/ie900650x  2010 American Chemical Society
Published on Web 08/05/2009
Ind. Eng. Chem. Res., Vol. 49, No. 3, 2010 965
products’ molar concentration are reported in Figures 2 and 3,
corresponding to the runs in the presence of respectively tartaric
and malonic acid. The first catalyst is one of the more selective
in the production of monochlorohydrin,13 while, the second is
active and selective in giving 1,3-dichlorohydrin.7
For all the catalysts reported in Table 1 similar experimental
data have been collected. Moreover, runs at different temper-
atures have been made only for malonic and tartaric acid,
developing a kinetic approach based on a reasonable reaction
mechanism. At last, some kinetic runs have been made with
acetic acid with different catalyst concentrations and different
HCl pressures (these last runs were performed in semibatch
conditions).
Runs in Stripping Conditions. Two runs have been per-
Figure 1. Experimental apparatus: (1) jacketed reactor made of Pyrex glass
formed by using the HCl stream as stripping agent through the
or in hastelloy steel, with a capacity of 700 mL; (2) thermostat for the condenser (Nr. 7) in Figure 1 and using malonic acid as catalyst.
circulation of thermal fluid (glass reactor) of electrical heating system (steel In the run at 100 °C only the final composition of the distillate
reactor); (3) peristaltic pump; (4) valve for performing samples withdrawal reservoir have been determined, while, for the run at 110 °C,
of the reaction mix; (5) cylinder of hydrochloric acid; (6) vertical condenser
for total reflux operation; (7) horizontal condenser for stripping operation;
the distillate has been analyzed at different intervals of time
(8) bottle for collecting the condensed reaction product; (9) bubbler provided during the run. The aim of such runs was the evaluation of the
with a porous disk for introducing gaseous HCl, which ensures effective possibility to selectively recover the desired reaction product
contact between the gas and liquid phases; and (10) system for suppressing that is 1,3-dichlorohydrin. The experimental results related to
the excess HCl (two or three Drechsel bottles, arranged in series, each
containing a solution of sodium hydroxide).
the stripping runs are reported in Table 2 together with the other
adopted operating conditions. As an example, in Figure 4 and
The scheme of the used experimental apparatus is reported 5, the composition-time profiles are reported for, respectively,
in Figure 1. As reactor body, both a jacketed glass reactor and the liquid phase in the reactor and the products recovered in
a hastelloy steel reactor have been used, this last in the case of the distillate after condensing.
runs in the condition of elevated pressure. The reactor is operated Runs under Pressure in Semibatch Conditions. A recent
in continuous mode for what concerns the flow of HCl and in paper published by Bell et al.,12 has reported a detailed behavior
batch modality regarding glycerol and catalyst. After flowing of the reaction of glycerol with HCl at different pressures from
through the reactor and the distillate accumulator, the unreacted 15 to 110 psi in semibatch conditions. It is very interesting to
HCl flows into a series of alkaline trap solutions for its complete observe that the pressure of HCl has a great effect on both the
neutralization. In this device both total reflux and stripping runs glycerol consumption rate and the products distribution. By
can be performed. More details about the reactor and the simply increasing the HCl pressure from 20 to 50 psi they show
operative procedure for experimental runs can be found else- a dramatic change of behavior because in the first case a
where.11 negligible amount of 1,3-dichlorohydrin is formed with a
2.2. Analysis. On the collected samples a first neutralization conversion of glycerol of about 60% reaching a plateau, while,
operation is performed in order to eliminate the dissolved HCl at the higher pressure 1,3-dichlorohydrin becomes predominant
and the acid used as catalyst. On about 3 cm3 of sample, 0.5 g and glycerol is totally converted in about 30 min. It must be
of calcium carbonate is added in a vial and the mixture is kept pointed out that the authors used wet glycerol containing 9%
at 100 °C for 30 min in order to remove also the water formed by weight of water. The presence of water very probably affects
by the reaction. Subsequently, the solid is removed by centrifu- the equilibrium of the reaction at low HCl pressure. We
gation and the composition is determined by GC analysis. The confirmed this finding, and in Figure 6 there is a comparison
GC analytical method is the following: column, CHROMPACK
between an experimental run performed at atmospheric pressure
CP Wax; stationary phase, 100% polyethyleneglycol; length,
with flowing HCl and a run performed under constant pressure
30 m; i.d., 0.25 mm; film thickness, 0.25 µm; FID detector;
of 5.5 bara. In the first case water is slowly removed from the
helium as gas carrier; injector temperature, 250 °C; detector
system considering that the condenser was cooled with water
temperature, 280 °C; temperature ramp, 1 min at 40 °C; heating
at 15-20 °C, and this explain the fast glycerol conversion,
rate, 20 °C/min to 100 °C, then 40 °C/min up to 200 °C, then
while, on the other hand, the formation of 1,3-dichlorohydrin
hold for 4 min. The sample of the reaction mixture is first diluted
is strongly promoted by the HCl pressure. In Figure 7, the
with methanol in a volumetric ratio of 1:20 and 1 µL of solution
is injected into the GC. instantaneous feed of hydrochloric acid along the time for this
2.3. Experimental Runs. Catalytic Screening: Runs in run is reported. The flow of HCl is necessary to maintain the
Total Reflux Conditions. To extend the catalytic screening total pressure inside the reactor at the prefixed value of 5.5 bara
started in our previous work,11 in the present paper we have and represents, indirectly, a measurement of the hydrochloric
tested other compounds containing carboxylic groups as catalysts acid consumption rate. It is interesting to observe that in the
in the glycerol chlorination reaction. The tested catalysts and first part of the run the decrease in the HCl consumption rate is
the experimental conditions adopted are summarized in Table almost linear while, subsequently, the ration rate is slower and
1 together with the obtained results in terms of products a tenuous decrease of HCl consumption proceeds through the
distribution (1; 2; 1,3; and 1,2 isomers) after 3 h of reaction. In rest of the run.
each run related to the reported catalysts, different samples of All these observations suggest that HCl concentration in the
liquid phase have been taken during the reaction and analyzed liquid phase is not linear with the pressure and that gas-liquid
accordingly to the GC method previously described. Two mass transfer is very probably operative especially in the initial
examples of the evolution with time of the experimental part of the run. Both these aspects require further investigations.
966 Ind. Eng. Chem. Res., Vol. 49, No. 3, 2010

Table 1. Experimental Runs for Catalyst Screening: Products Distribution after 3 h of Reactiona
products distribution after 3 h of reaction (mol %)
catalyst pKa amount of catalyst (g) Gly 1 2 1,3 1,2
acetic acidb 4.75 10.00 0.86 63.01 5.98 29.82 0.33
malonic acidb 2.84 18.00 0.57 55.85 7.24 35.87 0.46
propionic acidb 4.87 13.00 0.00 49.88 8.77 41.00 0.35
adipic acid 4.43 24.85 0.00 21.93 4.94 72.23 0.89
succinic acidb 4.20 20.00 1.86 60.88 6.84 30.05 0.37
citric acidb 3.13 34.00 5.15 70.45 6.49 17.61 0.26
levulinic acidb 4.59 20.60 0.39 60.04 6.98 32.20 0.37
pivalic acidb 5.10 18.00 88.05 9.19 1.65 1.11 0.00
benzoic acidb 4.19 13.90 86.34 12.38 1.27 0.00 0.00
trichloroacetic acidb 0.70 13.60 82.91 15.08 2.01 0.00 0.00
tartaric acid 3.03 25.50 14.14 78.43 7.42 0.00 0.00
fumaric acid 3.05 20.30 5.91 85.08 7.99 1.02 0.00
oxalic acid 1.25 15.40 34.15 60.06 5.53 0.23 0.01
maleic acid 1.88 20.30 5.50 76.08 9.38 8.78 0.30
formic acid 3.74 7.00 24.15 66.63 7.86 1.32 0.04
chloro-succinic acid 2.67 25.60 5.82 82.16 7.94 4.08 0.00
EDTAc 2.00 10.52 46.39 44.69 4.93 3.84 0.13
a
Other experimental common conditions: T ) 100°C, HCl flow rate ) 24 g/min, glycerol loaded ) 200 g, catalyst concentration ) 8% by moles,
reaction time ) 3 h, total reflux conditions except for HCl. b Results taken from Tesser et al. 2007. c Run under HCl pressure of 6 bara.

Figure 2. Evolution in time of the experimental product distribution for Figure 3. Evolution in time of the experimental product distribution for
the run with tartaric acid at T ) 100 °C: (9) glycerol; (2) 1-monochloro- the run with malonic acid at T ) 100 °C: (9) glycerol; (2) 1-monochlo-
hydrin; (b) 2-monochlorohydrin; (0) 1,3-dichlorohydrin; (1) 1,2-dichlo- rohydrin; (b) 2-monochlorohydrin; (0) 1,3-dichlorohydrin; (1) 1,2-
rohydrin. dichlorohydrin.

Pseudocontinuous Runs. Finally, with the aim to demon- and other catalysts, having higher boiling points, have been
strate the possibility to perform the 1,3-dichlorohydrin produc- tested and proposed in different patents and papers.7-11 In the
tion process in continuous modality, a specific run was already mentioned previous work11 malonic acid catalyst was
performed in which the continuous conditions were simulated considered for a detailed kinetic approach, and an hypothesis
by withdrawing portions of reactive mixture followed by a of reaction mechanism was presented there. In particular, the
corresponding addition of fresh glycerol (containing dissolved reaction mechanism, used for the kinetic model development,
catalyst). This operation was repeated at intervals of time of was a mechanism assumed valid for all the OH substitutions
1 h. The reaction was conducted with malonic acid as catalyst, with chlorine: a first acid-catalyzed OH esterification followed
at 100 °C and atmospheric pressure, under a continuous flow by an alkyl-oxygen bond scission favored by vicinal OH group
of HCl, and a clear distillate was collected. The overall that restores the catalytic species as in the reaction scheme
composition of the reaction products (reactor + distillate) is reported below. It must be pointed out that chlorination of
reported in Table 3 together with other experimental conditions catalyst molecules was not evidenced by using GC-MS
adopted for the run. analysis. However, from the results reported in Table 1, it is
evident that some catalysts, tested in the present work, have
shown a remarkable selectivity toward the production of
3. Results and Discussion
monochlorohydrins. Moreover, the concentration profiles re-
In a previous work11 catalysts screening was performed by ported in Figure 2 show a particular behavior according to which
comparing the activities of different carboxylic acids (see in dichlorohydrins concentration remains quite low in the initial
Table 1 catalysts marked with superscript b) in the formation part of the run. After an induction period, a sensible increase
of chlorohydrins from glycerol with respect to acetic acid that in 1,3-dicholorohydrin production has been observed while
was the first catalyst proposed in the literature.14-16 Acetic acid 1-monochlorohydrin begins to decrease. These experimental
has the drawback of a relatively low boiling point (117 °C) observations cannot be explained on the basis of the previously
Ind. Eng. Chem. Res., Vol. 49, No. 3, 2010 967
Table 2. Mole % Composition in the Reactor and in the Distillate Accumulator for Runs at 100 and at 110 °C. Catalyst: Malonic Acid
T ) 100 °C reactor T ) 110 °C reactor T ) 110 °C distillate tank
t (min) Gly 1 2 1,3 1,2 Gly 1 2 1,3 1,2 Gly 1 2 1,3 1,2
0 100 0 0 0 0 100 0 0 0 0 0 0 0 0 0
30 - - - - - 39.20 53.71 5.80 1.29 0.00 0 18.47 0.00 81.13 0.40
60 1.62 76.15 6.18 16.05 0.00 12.60 75.40 7.30 4.74 0.00 0 21.55 1.64 76.08 0.73
120 0.00 56.48 7.74 35.24 0.54 1.11 60.20 7.92 30.30 0.47 0 5.53 0.27 93.30 0.90
180 0.00 39.10 9.71 50.41 0.78 1.54 42.86 12.15 42.55 0.90 0 6.89 0.85 91.40 0.86
240 0.00 30.15 10.52 58.36 0.97 0.05 31.87 12.98 53.84 1.26 0 2.40 0.49 96.11 0.99
300 0.00 24.88 11.69 62.34 1.09 0.06 31.11 18.57 48.82 1.43 0 1.78 0.27 96.92 1.03
a
Note: Composition in the distillate tank for run at 100 °C after 300 min of reaction: Gly ) 0; 1 ) 3.62; 2 ) 0.52; 1,3 ) 94.90; 1,2 ) 0.96.

Table 3. Overall Mole % Composition (Reactor + Distillate) for the Pseudocontinuous Runa
reaction time (h) glycerol 1-monochlorohydrin 2-monochlorohydrin 1,3-dichlorohydrin 1,2-dichlorohydrin
0 100.00 0.00 0.00 0.00 0.00
1 5.18 60.79 6.31 27.33 0.38
2 0.19 46.20 6.27 46.77 0.57
3 1.08 34.87 7.44 55.83 0.78
4 3.89 22.33 6.85 66.07 0.86
5 1.88 16.70 8.40 71.94 1.09
6 0.60 14.35 8.33 75.60 1.12
7 0.89 15.60 9.40 72.87 1.24
8 3.18 17.03 10.06 68.58 1.15
9 1.06 15.13 8.82 73.76 1.22
a
Run operative conditions: initial charge of glycerol, 260.6 g; total glycerol added, 84,52 g; catalyst malonic acid, 8% molar with respect to glycerol;
temperature, 100 °C; atmospheric pressure, continuous flow of hydrogen chloride.

proposed reaction mechanism and a different mechanistic Tesser et al.11 On the contrary, when Keq1 . Keq2 the catalyst
hypothesis must be introduced considering diversifications gives mainly monochlorohydrin because the catalyst gives
between the first and the second chlorination reaction. mainly the ester1 as intermediate. Then, it is interesting to
observe that, with some exceptions, the catalysts having pKa
greater than 4 are selective to dichlorohydrin, while, the
catalysts with pKa between 1.2 and 3 are more selective to
monochlorohydrins. More acidic carboxylic acids, like for
example trichloroacetic acid, are not active in the reaction.
Very probably it is necessary to form esters for promoting
the reaction, but the ester formed must be not too stable to
allow the subsequent reaction steps. Clearly pKa is not the
only factor influencing the selectivity, because the presence
of more than one carboxylic group in the same molecule and/
or the presence of other functional groups are also important.
More deepened mechanistic studies are clearly necessary for
a fine-tuning of the interpretation of the selectivity-structure
relationship. The selectivity to monochlorohydrins open a
perspective to the industrial production and use of monochlo-
rohydrins and glycidol that could become building blocks
for many other syntheses on the basis of the most general
interest to find new uses for glycerol.
The two runs in which a stripping operation has been
In this case we have supposed that both monochlorohydrin performed by means of the HCl stream itself (Table 2 and
and dichlorohydrin production proceed through the formation
Figures 3 and 4) have been simulated by means of a commercial
of two ester species of different reactivity in which the catalyst
process simulation package (Chemcad 5.2). The reference
is distributed. For this purpose, a simplified scheme of the
scheme and the equation for the material balance related to the
following type can be written:
stripping system are reported in Figure 8. The vapor-liquid
Keq2 equilibria of the system have been described by means of the
1-monochlorohydrin + catalyst {\} ester2 + water UNIFAC model for liquid phase activity coefficients while the
kinetics parameters have been taken from our previous inves-
Keq1 tigation.11 The simultaneous description of liquid phase reaction
glycerol + catalyst {\} ester1 + water
and multicomponent vapor-liquid phase partition, allow a
satisfactory accuracy in the prediction of the experimental data
The values of the two equilibrium constants Keq1 and Keq2 collected for both the reactor and the flashed condensed
are of paramount importance for the selectivity of the catalyst composition. It is interesting to observe that the stripping
toward monochlorinated or dichlorinated products. If Keq1 ≈ operation allows the recovery of a product with a molar
Keq2 the formation of mono and dichlorohydrins follows a composition above 95% in 1,3-dichlorohydrin. It is obvious,
reaction-in-series mechanism according to that proposed by that a rectification will give better performances. Rectified 1,3-
968 Ind. Eng. Chem. Res., Vol. 49, No. 3, 2010

Figure 4. Composition profiles for the run at 110 °C in the reactor. (9) Figure 6. Comparison of runs at different pressure: (9) glycerol at 1 bar;
glycerol; (b) 1-monochlorohydrin; (1) 2-monochlorohydrin; (2) 1,3- (b) 1,3-dichlorohydrin at 1 bar; (2) glycerol at 5.5 bara; (1) 1,3-
dichlorohydrin; ([) 1,2-dichlorohydrin. Comparison between the experi- dichlorohydrin at 5.5 bar. Glycerol, 150 g; catalyst acetic acid, 7.82 g.
mental data and Chemcad prediction.

Figure 7. Instantaneous consumption rate of HCl in the semibatch run under


Figure 5. Composition profiles for the run at 110 °C in the distillate pressure. Total pressure kept at 5.5 bara; glycerol, 150 g; catalyst acetic
accumulator: (2) 1,3 dichlorohydrin, (b) 1-monochlorohydrin. Comparison acid, 8 g.
between the experimental data and Chemcad prediction.

dichlorohydrin could be advantageously employed in a epichlo-


rohydrin production process as will be shown in the following
paragraph.
Finally, a confirmation of the feasibility of the 1,3-dichloro-
hydrin production process in continuous modality has been
obtained from a pseudocontinuous run for which results are
reported in Figure 9. From this plot we can appreciate that, after
about 5 h of operation, the overall product concentrations
(reactor + distillate) are quite stationary and, in particular, the
desired 1,3-dichlorinated product is obtained at a concentration Figure 8. Reference scheme for the material balance used in the simulation
level of about 70%. of stripping runs: (R) reactor, (D) distillate accumulator, (V) liquid phase
volume.
Simulation of Epichlorohydrin Production Process. The
conventional process for epichlorohydrin production, via To better emphasize the advantages involved in the use of
propene, starts from a mixture of dichlorohydrin isomers of glycerol chlorination, promoted by catalysts that are selective
which the molar composition is around 70% of 1,2 isomer toward the production of 1,3-dichlorohydrin, a simulation has
and 30% of 1,3-dichlorohydrin.5,6 Starting from this mixture, been made for comparing the yields of the two process
by HCl elimination in alkaline conditions, the desired alternatives. The first one is the reactive distillation (pilot
epichlorohydrin product is formed and, simultaneously, scale) of the mixture with the traditional composition (1,3-/
monochlorohydrin and subsequently glycidol can be formed 1,2-, 30/70%) and the second one is the same unit operation
as by products. The overall reactions scheme is reported in performed on pure 1,3-dichlorohydrin. The simulation of
Figure 10. epichlorohydrin reactive column has been performed with a
Ind. Eng. Chem. Res., Vol. 49, No. 3, 2010 969
column. The yield is defined as the ratio between the distilled
epichlorohydrin and the fed dichlorohydrins. It is interesting
to observe that, from the plot reported in Figure 11, a higher
epichlorohydrin yield is obtainable also at low reflux ratios
with a feed that is constituted by pure 1,3-dichlorohydrin. A
confirmation of the superiority of a process based on this
kind of feed to the epichlorohydrin reactive column section
is evident also if we consider the trend of epichlorohydrin
yield for the case of 30/70% mixture: only at a very high
reflux ratio is a good product yield obtained, but it is not
comparable.

4. Conclusions
In this work the basis of a production process of epichlo-
rohydrin from glycerol was developed. The kinetics and
mechanism of the first stage of the process, glycerol
conversion to chlorohydrins, was deepened and different
Figure 9. Overall % molar concentration profiles (reactor + distillate) for carboxylic acids, not previously tested, were used as catalyst
the run in pseudocontinuous conditions. for this first reaction step. A rough correlation was found
between the value of pKa of the catalyst and its selectivity
toward mono- (pKa < 3) or dichlorinated (pKa > 4) com-
pounds. Some explorative runs were performed also under
HCl pressure. A pseudocontinuous run was performed in
order to verify the feasibility of industrial continuous
operation. Finally, a run in stripping conditions was reported
that satisfied the scope of a continuous recovery of the
products from the reaction mixture. This last approach was
particularly promising, allowing the recovery of a product
with a molar concentration of 1,3-dichlorohydrin higher than
95%. The possible industrial exploitation of these results were
investigated by comparing the traditional epichlorohydrin
production process starting from propene, with the proposed
Figure 10. Reactions scheme for epichlorohydrin production.
process starting from glycerol that allows the selective
recovery of 1,3-dichlorohydrin. In the last case, the reactive
column used for epichlorohydrin production, could be
operated with lower reflux ratio, maintaining high product
yield.

Literature Cited
(1) Ott, L.; Bicker, M.; Vogel, H. Catalytic dehydration of glycerol in
sub- and supercritical water: A new chemical process for acrolein production.
Green Chem. 2006, 8, 214–220.
(2) Watanabe, M.; Iida, T.; Aizawa, Y.; Aida, T. M.; Inomata, H.
Acrolein synthesis from glycerol in hot compressed water. Bioresour.
Technol. 2007, 98, 1285–1290.
(3) Chai, S. H.; Wang, H. P.; Liang, Y.; Xu, B. Q. Sustainable production
of acrolein: Gas-phase dehydration of glycerol over Nb2O5 catalyst. J. Catal.
2007, 250, 342–349.
(4) Tsukuda, E.; Sato, S.; Takahashi, R.; Sodesawa, T. Production of
acrolein from glycerol over silica-supported heteropoly acids. Catal.
Commun. 2007, 8, 1349–1353.
Figure 11. Comparison between reactive distillation for epichlorohydrin (5) Carra’, S.; Santacesaria, E.; Morbidelli, M. Synthesis of epichloro-
production: (2) feed mixture, 1,2-dichlorohydrin (70%)/1,3-dichlorohydrin hydrin by elimination of hydrogen chloride from chlorohydrins. 1. Kinetic
(30%); (b) feed only 1,3-dichlorohydrin. aspects of the process. Ind. Eng. Chem. Process Des. DeV. 1979, 18, 424–
427.
(6) Carra, S.; Santacesaria, E.; Morbidelli, M. Synthesis of epichloro-
commercial process simulation package (Chemcad 5.2) by hydrin by elimination of hydrogen chloride from chlorohydrins. 2. Simula-
tion of the reaction unit. Ind. Eng. Chem. Process Des. DeV. 1979, 18,
introducing the kinetic expression and parameters taken from 428–433.
Carrà et al.5,6 In both the process alternatives considered, (7) Siano, D.; Santacesaria, E.; Fiandra, V.; Tesser, R.; Di Nuzzi, G.;
the reactive column has been simulated with the following Di Serio, M.; Nastasi, M. (ASER srl). Process for the production of a,g-
specifications: total pressure, 1 bar; reboiler heat duty, 17000 dichlorohydrin from glycerin and hydrochloric acid. PCT Patent, WO
kcal/h; 15 theoretical plates; liquid holdup, 0.02 L/stage. The 111810 A2, 2006.
(8) Krafft, P.; Gilbeau, P.; Gosselin, B.; Claessens, S. (Solvay). Process
comparison between the two different feeds to the column, for producing dichloropropanol from glycerol, the glycerol coming eventu-
reported in Figure 11, is made by plotting the epichlorohydrin ally from the conversion of animal fats in the manufacture of biodiesel.
yields as a function of the reflux ratio adopted in the reactive PCT Patent, WO 054167 A1, 2005.
970 Ind. Eng. Chem. Res., Vol. 49, No. 3, 2010

(9) Schreck, D.; Kruper, W.; Varjian, R. D.; Jones, M. E.; Campbell, a Renewable Feedstock for Epichlorohydrin Production. The GTE Process.
R. M.; Kearns, K.; Hook, B. D.; Briggs, J. R.; Hippler, J. G. (Dow Clean: Soil, Air, Water. 2008, 36, 657–661.
Global Tech. Inc.). Conversion of multihydroxylated-aliphatic hydro- (13) Santacesaria, E.; Di Serio, M.; Tesser, R. (Eurochem Engineering
carbon or ester thereof to a chlorohydrin. PCT Patent, WO 020234 A1, srl). Process for monochlorohydrins production from glycerol and hydro-
2006. chloric acid. PCT Patent WO 132770 A1, 2008.
(10) Kubicek, P.; Sladek, P.; Buricova, I. (Spolek Pro Chemickou). (14) Britton, E. C.; Heindel, R. L. U.S. Patent 2.144.612, 1939.
Method of preparing dichloropropanols from glycerine. PCT Patent, WO (15) Britton, E. C.; Slagh, H. R. U.S. Patent 2.198.600, 1940.
020234 A1, 2006. (16) Thompson, W. P. Eur. Patent EP 931.211, 1963.
(11) Tesser, R.; Santacesaria, E.; Di Serio, M.; Di Nuzzi, G.; Fiandra,
V. Kinetics of glycerol chlorination with hydrochloric acid: A new route ReceiVed for reView April 23, 2009
to R,γ-dichlorohydrin. Ind. Eng. Chem. Res. 2007, 46, 6456–6465. ReVised manuscript receiVed July 15, 2009
(12) Bell, B. M.; Briggs, J. R.; Campbell, R. M.; Chambers, S. M.; Accepted July 16, 2009
Gaarenstroom, P. D.; Hippler, J. G.; Hook, B. D.; Kearns, K.; Kenney,
J. M.; Kruper, W.; Schreck, D.; Therlault, C. N.; Wolfe, C. P. Glycerin as IE900650X

You might also like