YAN Et Al, 2022

Download as pdf or txt
Download as pdf or txt
You are on page 1of 9

Chemical Engineering Journal 444 (2022) 136498

Contents lists available at ScienceDirect

Chemical Engineering Journal


journal homepage: www.elsevier.com/locate/cej

Dehydrochlorination of β-chlorohydrin in continuous microflow system:


Reaction kinetics and process intensification
Zifei Yan, Chencan Du, Yubin Wang, Jian Deng, Guangsheng Luo *
The State Key Lab of Chemical Engineering, Department of Chemical Engineering, Tsinghua University, Beijing 100084, China

A R T I C L E I N F O A B S T R A C T

Keywords: Glycidyl esters, as a class of terminal epoxides, are important raw materials for the synthesis and modification of
Dehydrohalogenation epoxy resins. Their synthetic routes via the dehydrochlorination of β-chlorohydrin esters are typically conducted
Microreactor under heterogeneous condition. Under the control of multiphase mass transfer, the inherently fast reaction can
Mass transfer
exhibit a very low observed reaction rate. Here we developed a continuous microflow system to determine the
Glycidyl ester
kinetic parameters of dehydrochlorination in the presence of aqueous NaOH solution. The same device was
applied for the continuous production of glycidyl esters, and the conditions including flow velocity, alkali
concentration, phase ratio and temperature were fully optimized. Comparing the obtained reaction performances
in the microreactor with the conventional batch reactors, the required reaction time was shortened by 1/20 and
the space–time yield was increased by 15 times, which is attributed to the mass transfer enhancement and the
continuous operating mode. Finally, a mathematical model was established to calculate the volumetric mass
transfer coefficients, whose value were in the range of 0.005–0.03 s− 1. The established microflow system and
research methods could have high merit for guiding the kinetic studies and process intensification of other
dehydrohalogenation reactions.

widely used one is still the aqueous solution of alkali metal hydroxide
1. Introduction [16] and calcium hydroxide [17]. Many efforts have been made to study
the reaction kinetics of dehydrochlorination process in the presence of
Terminal epoxides are a class of important materials in chemical alkali solution. Ma et al. [18] established a stirred reactor assembled
industry and are crucial building blocks in organic synthesis [1–3]. The with a conductivity measurement device to determine the variation of
small molecule terminal epoxides such as ethylene epoxide, propylene OH– concentration in situ. In this way, the dehydrochlorination kinetics
epoxide and epichlorohydrin, act as monomers in the ring-opening of dichloropropanol was determined in the timescale of 0–10 s, and the
polymerization to prepare polyether rubber [4]. In addition, the termi­ intrinsic kinetics was demonstrated to be extremely fast under homo­
nal epoxides with long aliphatic or aromatic chains can play versatile geneous condition. Zhang et al. [19] used a microreactor to investigate
roles in pharmaceutical community [5], coating industry [6] and envi­ the kinetic parameters of dehydrochlorination of 2,3-dichloro-1-propa­
ronmental engineering [7]. Glycidyl products are a large class of this nol, and presented the dehydrochlorination mechanism. Then, consid­
kind of terminal epoxides, mainly including glycidyl esters (glycidyl ering the effect of water concentration on the reaction rate, Filho et al.
methacrylate [8], glycidyl neodecanoate [9]) and glycidyl ethers [20] further obtained the intrinsic dehydrochlorination kinetics in a
(diglycidyl ethers of bisphenol-A [10]), they are vital synthetic blocks to continuous milliscale tubular reactor. The studies on the dehydrochlo­
produce and modify epoxy resins and waterborne coatings. rination kinetics of dichloropropanol are relatively sufficient, but for
Dehydrochlorination of β-chlorohydrin is the final reaction step in β-chlorohydrins with long aliphatic and aromatic carbon chains, the
the chlorohydrin method to prepare terminal epoxides. There are many related research is lacking. It is relatively difficult to obtain the homo­
types of alkaline that can be used in the dehydrochlorination processes, geneous dehydrochlorination kinetics because the reaction is very fast
including basic ion-exchange resin [11], metal oxide [12,13], solid and the side reactions including hydrolysis and ring-opening could occur
caustic soda [14] and halohydrin dehalogenase [15], whereas the most at high temperature [21]. Thus, it is meaningful to establish an effective

Abbreviations: CHPDA, 3-chloro-2-hydroxypropyl neodecanoate; ECH, epichlorohydrin; EPDA, 2, 3-epoxypropyl neodecanoate; GC, gas chromatography; MIBK,
methyl isobutyl ketone; NDA, neodecanoic acid.
* Corresponding author.
E-mail address: gsluo@tsinghua.edu.cn (G. Luo).

https://doi.org/10.1016/j.cej.2022.136498
Received 28 December 2021; Received in revised form 11 April 2022; Accepted 16 April 2022
Available online 21 April 2022
1385-8947/© 2022 Elsevier B.V. All rights reserved.
Z. Yan et al. Chemical Engineering Journal 444 (2022) 136498

Conventionally, the reaction time of several hours is unavoidable.


Nomenclature Moreover, the high temperature and high alkali concentration are
beneficial to improve reaction rate, but in post-reaction, a prolonged
A specific surface area, m2⋅m− 3 phase separation under these harsh conditions can lead to the hydrolysis
A pre-exponential factor, L⋅mol− 1⋅s− 1 of ester products.
Ea activation energy, kJ⋅mol− 1 In summary, it is of great importance to explore the intrinsic dehy­
FC mass flow rate of CHPDA, g⋅s− 1 drochlorination kinetics of chlorohydrin esters and enhance the liquid­
k reaction rate constant, L⋅mol− 1⋅s− 1 –liquid mass transfer to increase the observed reaction rate. And
koa volumetric mass transfer coefficient, s− 1 developing a continuous synthetic process is also important to further
M molar flow rate of CHPDA, mol⋅s− 1 improve the productivity of glycidyl esters. A conventional batch reactor
QO volumetric flow rate of organic phase, ml⋅s− 1 is challenging to employ for the determination of the kinetic parameters
QW volumetric flow rate of aqueous phase, ml⋅s− 1 in very fast reactions because the precise control of reaction time in the
rs interface reaction rate, mol⋅m− 2⋅s− 1 timescale of seconds or even sub-seconds is difficult [19]. In addition, in
VR reactor volume, ml the industrial stirred tank reactors to perform liquid–liquid heteroge­
X conversion of CHPDA neous process, the size of the dispersed droplets is very large (generally
τ residence time, s in the range of 1–3 mm) and the mass transfer coefficient is only in the
η space–time yield, g⋅cm− 3⋅h− 1 order of 10-6 m/s, leading to the fact that the heterogeneous mass
φ volume fraction of oil phase transfer could restrict the reactions.
Microfluidic technology holds the advantages of fast mixing, rapid
mass and heat transfer, accurate control of residence time in second or
sub-second scale, and very low consumption of materials, so it has been
platform to determine the kinetic parameters accurately, so as to guide widely acknowledged to be a strong tool for process intensification [26]
the production of glycidyl esters and glycidyl ethers. and kinetics determination of fast reactions including aromatic nitration
Unlike the kinetic research, the industrial dehydrochlorination pro­ [27], oxidation [28], radical polymerization [29] and so on [30]. Some
cesses to produce terminal epoxides are typically carried out under researchers established microfluidic systems to study the reaction ki­
heterogeneous condition because the reactant concentrations used in netics and heterogeneous dehydrochlorination process of dichlor­
production are much higher than those in kinetic study, and the solu­ opropanol [19,20]. However, there is still no report on the
bility of β-chlorohydrins in alkali/water solution are always very low. dehydrochlorination of β-chlorohydrins with long carbon chains in a
Makhin and Klebanov [22] studied the dehydrochlorination process of microreactor. In this work, we focus on the dehydrochlorination process
chlorohydrin ether of bisphenol A using methyl isobutyl ketone (MIBK) of β-chlorohydrin esters to prepare glycidyl esters in the presence of
and toluene as solvents respectively. The observed reaction rate in MIBK aqueous NaOH solution. The specific reaction we choose to systemati­
was twice as fast as that in toluene because MIBK and water could form a cally study is the dehydrochlorination of 3-chloro-2-hydroxypropyl
homogeneous reaction system, but toluene and water are mutually neodecanoate (CHPDA) (Fig. 1(c)), which is the last reaction step to
insoluble, so the reaction only proceeded at the interface. Under het­ produce 2,3-epoxypropyl neodecanoate (EPDA) (it is an important
erogeneous condition, the dehydrochlorination may be controlled by synthon of environmentally friendly coatings, the application status and
mass transfer, leading to the fact that the observed reaction rate was market prospects of EPDA could be found in the following literatures
much slower than the intrinsic kinetics. [31–33].) A microtube reaction system was established to determine the
The preparation of glycidyl esters via dehydrochlorination of intrinsic reaction kinetics under homogeneous condition, then this
β-chlorohydrins is also faced with this dilemma. Fig. 1 shows the het­ platform was also successfully applied to intensify the liquid–liquid
erogeneous dehydrochlorination process to produce glycidyl esters and heterogeneous mass transfer process. Finally, the continuous dehydro­
the corresponding operating conditions in batch reactors [21,23–25]. chlorination of CHPDA was realized based on microfluidic technology,

Fig. 1. Typical reaction conditions of dehydrochlorination of β-chlorohydrins to produce glycidyl esters, including (a) glycidyl methacrylate [25], (b) 3-Chloro-2-
hydroxypropyl 4-(chloromethyl)benzoate [24] and (c) glycidyl neodecanoate [21,23].

2
Z. Yan et al. Chemical Engineering Journal 444 (2022) 136498

greatly shortening the reaction time and improving the space–time yield analyzed after extraction. The set-up shown in Fig. 2(b) was also used to
in EPDA production. produce EPDA via the heterogeneous dehydrochlorination process. At
this time, CHPDA was no longer dissolved in ethanol but delivered as a
2. Materials and methods neat reagent. Unlike the operating procedures in kinetic determination,
the quench agent (phosphate buffer solution) was pre-added to the
2.1. Chemicals sample vessel. After collecting samples within 1–5 s, the vessel was
shaken to terminate the reaction.
The β-chlorohydrin, 3-chloro-2-hydroxypropyl neodecanoate
(CHPDA), was prepared via the acidolysis ring-opening reaction of
2.3. Sample analysis
neodecanoic acid (NDA) and epichlorohydrin (ECH) according to the
related patents [23] and our previous works [33]. Vacuum distillation
Gas chromatography (GC) can be applied to analyze the concentra­
(75 ◦ C, vacuum degree > 0.095 MPa, 2 h) was applied to remove the
tion of CHPDA, EDPA and NDA, however, the samples containing a large
trace amount of water and ECH in the product. After purification, the
amount of brine cannot be directly analyzed by GC. Thus, octane was
prepared product contained CHPDA (92 wt%), 2,3-epoxypropyl neo­
used as the extractant to extract the dissolved and dispersed organic
decanoate (EPDA, 4 wt%), diester (3 wt%), 1,3-dichloro-2-propanol (0.5
substances into the oil phase. The distribution coefficients of CHPDA,
wt%) and NDA (0.5 wt%). NDA was purchased from Shanghai Ethyl
EPDA and NDA in octane/water system were determined in shaker tests,
Chemical Co., LTD, with the purity of 99 wt% tested by titration. EPDA
and the values were all larger than 50 at room temperature and 7.0 pH.
was purchased from Guangdong Weng Jiang Reagent, and its epoxy
This means more than 98% of organic substances could be extracted into
group content was 4195 mmol/kg (corresponding to 95.8 wt% purity).
oil phase by using 1:1 volumetric ratio of octane to water. After
All other reagents were purchased from Aladdin Chemistry Co., LTD and
extraction and 30 min for phase separation in static state, the oil phase
classified as analytical grade. The purities of these reagents could meet
was analyzed by GC. The testing conditions are listed in Table 1. The
the experimental requirements, so the purification steps were not
relative correction factor of CHPDA, EPDA and NDA in the internal
required.
standard method were 1.264, 1.127, and 1.020 respectively, and the
correlation coefficients of the standard curves were all larger than 0.999.
2.2. Experimental procedures The hydrolysis of the terminal epoxy group of EPDA could lead to the
formation of glycerin neodecanoate, but no such hydrolysis product was
The target reaction of this work, dehydrochlorination of 3-chloro-2- detected by GC. In addition, the hydrolysis of the ester group in CHPDA
hydroxypropyl neodecanoate (CHPDA), is displayed in Fig. 2(a), and the and EPDA generated sodium neodecanoate, which further converted to
set-up is shown in Fig. 2(b). In the experiments to determine reaction NDA and transported to oil phase during the extraction process. In this
kinetics, CHPDA was dissolved in ethanol and NaOH was dissolved in way, the degree of hydrolysis can be obtained via GC analysis. The re­
water, both with the concentration of 0.05–0.2 mol/L. The two reactants sults showed that the hydrolysis rates were lower than 1% in all of our
were delivered by constant-flux pumps (also called HPLC pumps, Beijing experiments. Thus, the side reactions caused by hydrolysis were
Satellite Co. Ltd.) and preheated to the reaction temperature in stainless neglected.
steel capillaries with inner diameter of 0.5 mm and external diameter of Conversion of CHPDA (X) is the key parameter to analyze the reac­
1.6 mm. Later, the two streams mixed in a T-micromixer (aperture of the tion kinetics and performances, it is calculated by Eq. (1), where ωC (%)
mixed position is 0.25 mm) and constituted a homogeneous system, then is the mass fraction of CHPDA in the organic phase after extraction,
flowed into a stainless-steel capillary coil with inner diameter of 2 mm
Mo (g) is the mass of organic phase after extraction, FC0 (g/s) is the initial
and external diameter of 3 mm to proceed the reaction. The reaction coil
mass flow rate of CHPDA, ts (s) is the sampling time, which is calculated
with different lengths (1, 2, 3, 5 and 7 m) were used to adjust the re­
by dividing the mass of the sample Ms (g) by the total mass flow rate
action time. The preheating coil, the mixer and the reaction coil were
Ft (g/s).
immersed into an oil bath to control temperature. When leaving the oil-
bath zone, an excessive amount of phosphate buffer solution (NaH2PO4 ωC Mo ωC Mo
X = 1− = 1− (1)
and Na2HPO4 dissolved in water, pH = 7.0) was pumped into the FC0 ts FC0 MFts
reacting system to quench the dehydrochlorination by consuming the
remaining hydroxide ions. The samples were collected at the outlet and Space-time yield (η, g⋅cm− 3⋅h− 1) is another important parameter to
evaluate the reactor performances. It is calculated by Eq (2), where
FC0 (g/s) is the initial mass flow rate of CHPDA, X is the conversion of
CHPDA, VR(cm3) is the reactor volume, MEPDA and MCHPDA are molec­
ular weights of EPDA and CHPDA, whose values are 228.3 and 264.8 g/
mol respectively.

MEPDA F 0C X
η = 3600 × (2)
MCHPDA VR

Table 1
GC testing conditions.
Device: Agilent GC-7890A Detector: FID
Injection temperature: 300 οC
Detection temperature: 300
ο
C
Column: HP-5 / 30 m × 0.32 mm × 0.25 μm Sample: 1 μL
Carrier gas: N2, 1 ml/min; Split ratio: 30:1 Quantitative method: internal
standard
Column temperature: start at 50 οC, maintain for 2 Internal standard substance:
min; 10 οC/min to 160 οC; 5 οC/min to 200 οC; 10 methyl benzoate
ο
C/min to 300 οC, maintain for 3 min
Fig. 2. Experimental procedures and set-up.

3
Z. Yan et al. Chemical Engineering Journal 444 (2022) 136498

3. Results and discussions

3.1. Reaction kinetics of dehydrochlorination of CHPDA

In the studies on the reaction kinetics of dehydrochlorination of


dichloropropanol [19,20] in microreactors, it was demonstrated that the
plug flow model was applicable when the Reynolds number was higher
than 150. In this work, the physical properties and reactor dimension are
close to the previous studies, and the Reynolds number varies from 200
to 400. Therefore, it is reasonable to utilize the plug flow model to solve
the mass balance in the microtube reactor (Eq. (3)), where τ(s) is the
residence time of the reactants. The reaction heat of CHPDA with NaOH
in ethanol/H2O solution under 50 ◦ C was measured to be 31.1 ± 2.9 kJ/
mol (experimental device, conditions and results of calorimetric tests are
listed in Supplementary Material). The heat absorbed by the reaction is
relatively small, so the isothermal condition is easy to realize in the
microreaction system. In order to avoid the influence of mixing in the
inlet section, the sample obtained without reaction coil is assumed to be
“time zero” and the corresponding concentration of CHPDA is assumed Fig. 4. Plot of f(X) versus residence time to verify the hypothesis of first order
to be “initial concentration”. The concentration of NaOH can be calcu­ to NaOH. Reaction conditions: the initial concentration of NaOH and CHPDA
lated according to the conversion of CHPDA and mass balance (4). after mixing were 0.0676 mol/L and 0.0758 mol/L respectively; the total flow
rate was 19.8 ml/min; temperature was 30 ◦ C.
dCi
= − ri (i = CHPDA, OH− ) (3)

⎛ ⎞
0
CNaOH = CNaOH 0
− (CCHPDA − CCHPDA ) (4) X
1 ⎜1 − a ⎟
f (X) = 0 ln⎝ ⎠ = kτ (a ∕
= 1) (6)
Under the condition that the molar ratio of NaOH to CHPDA was CCHPDA (a − 1) 1− X
larger than 10, the reaction order of CHPDA was determined. As shown
in Fig. 3, the conversion (X) of CHPDA was calculated according to the The slope (k) of the curve, f(X) versus τ, represents the reaction rate
time profile of its concentration. Then, the accurate linear relation be­ constant. Fig. 5(a) shows the values of k at different reaction tempera­
tween ln(1-X) versus τ indicates that the order of CHPDA is first. tures, and the plot of lnk versus 1/RT is displayed in Fig. 5(b). According
The concentration of NaOH cannot be directly analyzed because to Arrhenius equation (Eq (7)), the activation energy (Ea) and pre-
excessive phosphate buffer was added for quenching. If the experiment exponential factor (A) were calculated to be 75.9 ± 5.0 kJ/mol and
is performed in excessive CHPDA, and NaOH concentration is calculated 1.47 × 1012 L⋅mol− 1⋅s− 1 respectively.
by the decrease of CHPDA concentration (Eq (4)), the experimental error
Ea
could be magnified. Thus, hypothesis-retesting method is used to lnk = lnA − (7)
RT
determine the reaction order of NaOH. If the order of NaOH is first, just
as the previous studies [18–20] on the dehydrochlorination of dichlor­ The obtained reaction kinetics of dehydrochlorination of CHPDA is
opropanol signified, the rate law can be written as Eq (5). Substituting compared with the dehydrochlorination of 1,3-dichloro-2-propanol
Eq (4) to Eq (5) and integrating the equation, the relation between the [20], and the results are listed in Table 2. Using the similar micro­
conversion of CHPDA and residence time can be deduced as given by Eq reactor setup, the obtained kinetic equations and reaction orders are the
(6), where a = C0NaOH /C0CHPDA is the ratio of the initial concentrations of same, but there are significant differences in activation energy and pre-
NaOH to CHPDA. Fig. 4 shows the plot of f(X) versus τ, the stringent exponential factor. According to the recognized reaction mechanism of
linear relation indicates that the hypothesis of first order to NaOH is dehydrochlorination (see Fig. 6) [19], activation energy represents the
feasible. energy difference between the transition state and the initial state. The
transition state of CHPDA is more stable due to the presence of side
r= −
dCCHPDA
= kCNaOH CCHPDA (5) chains, thus exhibiting a relatively lower activation energy. The pre-
dτ exponential factor of CHPDA is far smaller than 1,3-dichloro-2-propa­
nol, which is because the large steric hindrance reduces the frequency

Fig. 3. Kinetic results to obtain the reaction order of CHPDA. (a) The time profile of CHPDA concentration. (b) The conversion of CHPDA versus residence time.
Reaction conditions: the initial concentration of NaOH and CHPDA after mixing were 0.680 mol/L and 0.0614 mol/L respectively; the total flow rate was 19.5 ml/
min; temperature was 30 ◦ C.

4
Z. Yan et al. Chemical Engineering Journal 444 (2022) 136498

Fig. 5. (a) Plot of f(X) versus residence time at temperature range of 30–50 ◦ C. (b) Plot of lnk versus 1/RT. Reaction conditions: the initial concentration of NaOH and
CHPDA after mixing were 0.0676 mol/L and 0.0758 mol/L respectively; the total flow rate was 19.8 ml/min.

(3) The slug flow pattern was observed at the outlet of microtube
Table 2
reactor. It was demonstrated that the residence time distribution
Comparison of homogeneous dehydrochlorination kinetics of CHPDA and 1,3-
of slug flow in microtubes is narrow [34], thus the plug flow
dichloro-2-propanol in NaOH solution.
model was applied to calculate the average residence time (Eq
Entry CHPDA 1,3-dichloro-2-propanol (8)), where VR is the volume of reactor, Qw and Qo are volumetric
[20]
flow rate of aqueous and organic phases respectively.
Setup for kinetic research Microreactor Microreactor
Determination method Gas Gas chromatography VR
τ= (8)
chromatography Qw + Qo
Timescale 0–60 s 0–25 s
Kinetic equation r = kCNaOH CCHPDA r = kCNaOH CDCP The time profiles of CHPDA conversion obtained in the heteroge­
Activation energy (kJ/mol) 75.9 ± 5.0 82.6 ± 1.2 neous and homogeneous dehydrochlorination are shown in Fig. 7, in
Pre-exponential factor 1.47 × 1012 4.84 × 1014 which the heterogeneous results were measured by experiments and
(L⋅mol− 1⋅s− 1)
homogeneous results was calculated according to reaction kinetics by
k at 50 ◦ C (L⋅mol− 1⋅s− 1) 1.45 21.2
assuming that the two phases were miscible. The results indicate that the
observed reaction rate under heterogeneous condition is significantly
lower than that in homogeneous condition, and the most prominent
reason is that the liquid–liquid heterogeneous mass transfer restricts the
dehydrochlorination reaction.

Fig. 6. The reaction mechanism of dehydrochlorination of β-chlorohydrin with 3.2.1. Effect of flow velocity
hydroxide ions. Fig. 8 shows the observed reaction rates of dehydrochlorination at
different flow velocity in the microtube reactor. There is no significant
of effective collisions between reactant molecules. Considering both difference in the conversion-time curves of these three experiments. This
effects, the dehydrochlorination of CHPDA is slower than 1,3-dichloro- phenomenon indicates that as the velocity reached 2.12 cm/s, simply
2-propanol (e.g., the reaction rate constant of CHPDA dehydrochlori­ increasing the flow rate cannot further facilitate the mass transfer and
nation at 50 ◦ C is only one fifteenth of 1,3-dichloro-2-propanol), but it is improve the observed reaction rate.
still a fast reaction.

3.2. Heterogeneous dehydrochlorination in microtube reactor

In the above discussion about reaction kinetics, CHPDA/ethanol


solution and NaOH/water solution, both with relatively low concen­
tration, formed a homogeneous reaction system. This is beneficial to
eliminating the effect of mass transfer and obtaining the intrinsic reac­
tion kinetics. The following discussion focuses on the production of
EPDA via the heterogeneous dehydrochlorination of CHPDA. The main
differences include two aspects: CHPDA was no longer dissolved in
ethanol but delivered as a neat reagent, the mass fraction of NaOH in
water increased to 20 wt%~40 wt%. Both of these two measures are
aimed at reducing the amount of inert solvent, which can bring many
difficulties and increases the cost in the post separation and purification
processes. Some characteristics about the heterogeneous reaction sys­
tem should be stressed here:

(1) The mutual solubility of the organic and aqueous phases is Fig. 7. CHPDA conversion in the heterogeneous experiment and the results
extremely low because of the high hydrophobicity of CHPDA. calculated according to reaction kinetics using pseudo-homogeneous model.
(2) The change of volumetric flow rate of organic and aqueous phases Reaction conditions in heterogeneous experiment: mass fraction of NaOH was
before and after the dehydrochlorination is measured to be less 20 wt%; flow rate of NaOH/H2O = 6 ml/min; flow rate of CHPDA = 2 ml/min;
than 3%. It is reasonable to ignore this small variation. temperature was 50 ◦ C.

5
Z. Yan et al. Chemical Engineering Journal 444 (2022) 136498

When the concentration of NaOH is lower than 20 wt%, the variation of


interfacial tension is not obvious. However, as the concentration
increased from 20 wt% to 30 wt%, the interfacial tension declined
significantly, accompanied by a substantial increase in reaction rate (see
Fig. 9(a)).

3.2.3. Effect of phase ratio


The effect of phase ratio on dehydrochlorination reaction was stud­
ied using the optimal 40 wt% NaOH solution, and the results are dis­
played in Fig. 10. When the water-to-oil phase ratio is greater than 1:1
(the alkali is more than three-times excessive), the phase ratio has little
effect on the observed reaction rate. However, as the phase ratio further
reduces from 1:1 to 1:2 and 1:3, corresponding to the molar ratio (n
(NaOH):n(CHPDA)) of 2:1 and 1.4:1, the reaction rate is severely
restricted. The time profiles of NaOH concentration are shown in Fig. 10
(b). Combing Fig. 10 (a) and (b), it can be figured out that when NaOH
concentration drops below 28.7 wt% for the experiment in phase ratio =
1:2 and below 26.2 wt% for phase ratio = 1:3, the dehydrochlorination
Fig. 8. Time profiles of CHPDA conversion at different flow velocities in reaction becomes very slow. In a way, the effect of phase ratio can also
microtube reactor. Reaction conditions: concentration of NaOH solution = 20
be attributed to the change of NaOH concentration. In order to keep the
wt%; phase ratio of water to oil = 3:1; temperature = 50 ◦ C.
dehydrochlorination reaction in a relatively high rate, it is essential to
maintain the alkali concentration above 30 wt%. Therefore, the optimal
3.2.2. Effect of NaOH concentration molar ratio of NaOH to CHPDA is 4:1, corresponding to water-to-oil
Measured at 50 ◦ C, the distribution coefficient of NaOH in 2-ethyl­ phase ratio of 1:1 when 40 wt% NaOH solution is used.
hexanol/water system is 9 × 10-5 and the solubility of CHPDA in
water is lower than 0.05 wt%. This indicates that the reactants are hardly 3.2.4. Effect of temperature
soluble in the other phase, the reaction mainly occurs at the oil–water Fig. 11 shows the curves of CHPDA conversion versus residence time
interface. In this way, raising the reactant concentration in bulk phase to at different reaction temperatures. As the temperature rises from 35 ◦ C
increase the concentration gradient could be an efficient method for to 90 ◦ C, the observed reaction rate at the initial stage shows a clear
mass transfer enhancement. CHPDA was used as a neat reagent in the upward trend because the mutual solubility of the two reactants in­
heterogeneous dehydrochlorination reaction, so there is basically no creases and the interfacial tension declines. However, as the reaction
room to further increase its concentration. However, the concentration proceeds, it could be found that the highest temperature, 90 ◦ C, is not
of NaOH is easy to adjust by changing its initial mass fraction in water, conducive to the complete conversion of CHPDA to EPDA in the late
and the effect is shown in Fig. 9(a). When 20 wt% NaOH solution was reaction stage. On the contrary, the time profile of conversion at 50 ◦ C is
used, the conversion of CHPDA was only 35% with a residence time of the highest among these five experiments when the residence time ex­
2.7 min, while when 40 wt% NaOH solution was used, the conversion ceeds 80 s. This phenomenon may be due to the fact that the high
could reach 98%. The significant improvement in observed reaction temperature can promote the coalescence of droplets by reducing the
rate, more specifically, in mass transfer rate, is attributed to the increase mechanical strength of the interfacial film and the viscosity of the sys­
of concentration gradient of alkali between the bulk phase and the tem. When using 65 ◦ C or higher temperature, the coalescence of
oil–water interface. droplets could be faster, which is detrimental to the heterogeneous re­
The interfacial tension is another factor that can indirectly affect the action because it reduces the specific surface area. Therefore, 50 ◦ C is a
liquid–liquid mass transfer by changing the characteristic dispersion relatively better temperature condition to promote the conversion rate
size. Generally, in microreactors, the liquid–liquid two phases with above 90%.
lower interfacial tension are more inclined to form a more stable
dispersion system with a larger specific surface area [35], which is
beneficial to mass transfer. Fig. 9(b) shows the interfacial tension be­
tween EPDA and NaOH solution with concentration range of 0–40 wt%.

Fig. 9. The effects of NaOH concentration on the dehydrochlorination of CHPDA. (a) Time profiles of CHPDA conversion at different initial mass fractions of NaOH.
Reaction conditions: flow rate of NaOH/H2O = 6 ml/min; flow rate of CHPDA = 2 ml/min; temperature was 50 ◦ C (b) Interfacial tension between EPDA and NaOH/
H2O solution with concentration range of 0–40 wt%, measured by pendant drop method (DataPhysics Instruments, OCAH220) at 20 ◦ C, 101 kPa.

6
Z. Yan et al. Chemical Engineering Journal 444 (2022) 136498

3.3. Comparisons of CHPDA dehydrochlorination in microreactor and


conventional batch reactor

Table 3 lists the operating conditions and the obtained reaction


performances of CHPDA dehydrochlorination in the conventional batch
reactors and in the microreactor established in this work. Generally, the
reaction time longer than 1 h is required in stirred batch reactors.
Because of the higher surface area and concentration gradient in a
microreactor, the mass tansfer rate is typically 1–2 orders of magnitude
higher than that in conventional stirred equipment [36]. Therefore,
under the same conditions of temperature and molar ratio (comparing
Experiment No.3 with No.2), the reaction time in the microreactor is less
than 1/4 of that in the stirred reactor, and the space–time yield was
increased by 4 times. Experiment No. 4 lists the optimal operating
conditions and the obtaiend reaction performances in the established
microflow system. When the molar ratio n(NaOH):n(CHPDA) = 4:1 was
employed, the NaOH concentration could be maintained higher than 30
wt% during the dehydrochlorination process, thus making the reaction
time further reduced from 20 min to 2.7 min (comparing Experiment
No.4 with No.3).
The condition with the high molar ratio of NaOH to CHPDA is
challenging to employ in conventional batch processes, in which dehy­
drochlorination and phase separation proceed at the same stirred
equipment. Under the conditions of high alkali concentration (>25 wt%)
and high temperature (>50 ◦ C), the product EPDA is peculiarly prone to
hydrolysis [21]. Therefore, the feeding amount of NaOH solution must
be strictly limited. On the contrary, in the continuous microflow sys­
tems, the reaction time is great shortened and the phase separation can
also be conducted in a continuous way with the temperature set much
lower than that in the reaction. This is why the hydrolysis rate is still
lower than 1% even under the conditions of 90 ◦ C and molar ratio of 4:1.
It should be stressed that the excessive NaOH solution can be recy­
cled by phase separation and re-used for dehydrochlorination by
Fig. 10. The effect of phase ratio on the performances of heterogeneous replenishing solid NaOH to 40 wt% concentration again. The precipi­
dehydrochlorination reaction. (a) Time profiles of CHPDA conversion at tated NaCl could be recovered by chlor-alkali industry [17].
different phase ratios. (b) Time profiles of NaOH concentration at different
phase ratios. Reaction conditions: total flow rate of the two phases (QO + QW)
= 8 ml/min, phase ratio of water to oil varied from 3:1 to 1:3, NaOH concen­ 3.4. Volumetric mass transfer coefficient in microtube reactor
tration = 40 wt%, temperature = 50 ◦ C.
In order to establish a mathematical model to calculate the volu­
metric mass transfer coefficient, the mass balance of CHPDA in a dif­
ferential volume of microtube reactor is analyzed as shown in Fig. 12,
where M (mol/s) is the molar flow rate of CHPDA, rs (mol⋅m− 2⋅s− 1) is the
interface reaction rate, ΔV(m3) and ΔS (m2) are volume and interface
area of the differential section respectively, a = ΔV
ΔS
represents the spe­
cific surface area. Thus, mass balance can be expressed as Eq (9).
MV+ΔV − MV = − rS ΔS = − rS aΔV (9)
It was demonstrated that the dehydrochlorination takes place at the
interface and it is controlled by mass transfer. When NaOH is largely
excessive to CHPDA (molar ratio is higher than 10) and its concentration
is very high (>35 wt%), the mass transfer resistance in aqueous phase
can be ignored and the concentration of CHPDA at the interface can be
regarded as 0. Therefore, the interface reaction rate, rS , can be expressed
as the product of the mass transfer coefficient (ko (m⋅s− 1)) and CHPDA
concentration in the oil phase (Eq (10)).
I
rS = ko (CCHPDA − CCHPDA ) = ko CCHPDA (10)

Fig. 11. The effect of temperature on the performances of heterogeneous Combining Eqs. (9) and (10), the mass balance equation in differ­
dehydrochlorination reaction. Reaction conditions: NaOH concentration = 40 ential form is expressed as given by Eq (11).
wt%, flow rate of NaOH/H2O = 4 ml/min, flow rate of CHPDA = 4 ml/min.
dM
= − ko aCCHPDA (11)
dV
Substituting the expression of M (Eq (12), φ = QOQ+Q
O
W
is the volume
fraction of oil phase) in Eq (11), the relation between CHPDA concen­
tration and residence time can be deduced as given by Eq (13).

7
Z. Yan et al. Chemical Engineering Journal 444 (2022) 136498

Table 3
Reaction conditions and performances of dehydrochlorination of CHPDA in the microreactor and conventional batch reactors.
No. Reactor types NaOH concentration (wt Molar ratio n(NaOH): n Temperature Reaction time Conversion Space-time
%) (CHPDA) (◦ C) (min) (%) yield
(g⋅cm− 3⋅h− 1)

1 Batch [23] 40 1.65:1 70 60 95.0 0.53


2 Batch [21] 40 2:1 50 90 95.6 0.34
3 Microreactor (this work) 40 2:1 50 20 94.1 1.48
4 Microreactor (this work, 40 4:1 50 2.7 93.8 8.2
optimal)

φ 1
ko a = ln( ) (15)
τ 1− X
Fig. 14 shows the calculated values of ko a at different reaction
temperature, which is distributed in the range of 0.005–0.03 s− 1.
Comparing with the obtained intrinsic reaction rate, mass transfer pro­
cess still controls the dehydrochlorination of CHPDA although enhanced
by microreactor. The process could be further intensified by increasing
Fig. 12. Mass balance of CHPDA in a differential volume of microtube reactor. the specific surface area, for example, using micropacked bed reactors
[37].
M = QO CCHPDA = φ(QO + QW )CCHPDA (12)
4. Conclusion
dM d[φ(QO + QW )CCHPDA ] dCCHPDA
= =φ = − ko aCCHPDA (13) The dehydrochlorination of CHPDA with aqueous NaOH solution is
dV dV dτ
systematically studied in a microtube reactor. The reaction kinetics was
Through integration, the equation of conversion versus residence
determined under homogeneous condition by using ethanol as solvent.
time is obtained:
The orders of CHPDA and NaOH are both first, the activation energy and
1 pre-exponential factor are 75.9 ± 5.0 kJ⋅mol− 1 and 1.47 × 1012
φln( ) = ko aτ (14)
1− X L⋅mol− 1⋅s− 1 respectively. Furthermore, in the same reaction set-up, the
Fig. 13 shows the excellent linear relation between φln(1−1 X) and τ continuous production of EPDA via the dehydrochlorination of CHPDA
is conducted without any organic solvent, constituting a liquid–liquid
when the molar ratio of NaOH to CHPDA is larger than 10, which proves
heterogeneous reaction system. It was demonstrated that the heteroge­
the feasibility of the model. The volumetric mass transfer coefficient
neous reaction is definitely controlled by mass transfer because the
(ko a), which is the slope of the curve, is calculated to be 5.84 ± 0.36 ×
observed reaction rate is much lower than that under homogeneous
10-3 s− 1.
condition. It was demonstrated that increasing the concentration of
Under the obtained optimal dehydrochlorination conditions, in
NaOH solution is the most direct and effective way to enhance mass
which the phase ratio is 1:1 and molar ratio of NaOH to CHPDA is 4:1,
transfer because it can improve the concentration gradient and reduce
the volumetric mass transfer coefficient can be calculated according to
the interfacial tension. In order to keep dehydrochlorination at a high
Eq (15), and the point with CHPDA conversion less than 20% is selected
rate, the NaOH concentration should be maintained higher than 30 wt%.
for calculation because the change of NaOH concentration is so small
Under the optimal dehydrochlorination conditions (CNaOH = 40 wt%,
that the promise of ignoring the mass transfer resistance in the aqueous
phase ratio (water to oil) = 1:1, temperature = 50 ◦ C), 93.8% conversion
phase is still valid.
of CHPDA is obtained within 2.7 min. Compared with the conventional
batch processes to produce EPDA, the space–time yield is enlarged by 15

Fig. 13. The curve of φln(1−1 X) versus residence time in microtube reactor. Fig. 14. Volumetric mass transfer coefficients in the organic phase at different
Reaction conditions: NaOH concentration = 40 wt%, temperature = 50 ◦ C. QO reaction temperatures. Reaction conditions: NaOH concentration = 40 wt%, QO
= 2 ml/min, QW = 6 ml/min, molar ratio of NaOH to CHPDA was 12.4. = 4 ml/min, QW = 4 ml/min, molar ratio of NaOH to CHPDA was 4.0.

8
Z. Yan et al. Chemical Engineering Journal 444 (2022) 136498

times. In addition, the mathematical model was established to determine [11] M. Dargelos, M.E. Borredon, A. Gaset, Synthesis of epihalohydrins by
dehydrohalogenation of glycerol dihalohydrins-utilization of a basic ion-exchange
the volumetric mass transfer coefficient in the microtube reactor. This
resin, Aust. J. Chem. 45 (1992) 1327–1330.
work may provide a valuable reference for the industrial transition of [12] G.M. Lari, G. Pastore, C. Mondelli, J. Perez-Ramirez, Towards sustainable
dehydrochlorination of β-chlorohydrins from batch to continuous pro­ manufacture of epichlorohydrin from glycerol using hydrotalcite-derived basic
duction mode. Further investigation on the process simulation and in­ oxides, Green Chem. 20 (2018) 148–159.
[13] P. Yang, X. Xue, Z. Meng, R. Zhou, Enhanced catalytic activity and stability of Ce
dustrial scale-up based on the microfluidic technology are very doping on Cr supported Hzsm-5 catalysts for deep oxidation of chlorinated volatile
meaningful. organic compounds, Chem. Eng. J. 234 (2013) 203–210.
[14] CN 106632150 A, 2016.
[15] J.H. Schrittwieser, I. Lavandera, B. Seisser, B. Mautner, W. Kroutil, Biocatalytic
cascade for the synthesis of enantiopure β-azidoalcohols and β-hydroxynitriles,
Declaration of Competing Interest Eur. J. Org. Chem. 2009 (2009) 2293–2298.
[16] R. Morodo, R. Gérardy, G. Petit, J.M. Monbaliu, Continuous flow upgrading of
glycerol toward oxiranes and active pharmaceutical ingredients thereof, Green
The authors declare that they have no known competing financial
Chem. 21 (2019) 4422–4433.
interests or personal relationships that could have appeared to influence [17] A.M. Krzyżanowska, E. Milchert, W.M. Paździoch, Technological parameters of
the work reported in this paper. dehydrochlorination of 1,3-dichloropropan-2-Ol to epichlorohydrin, Ind. Eng.
Chem. Res. 52 (32) (2013) 10890–10895.
[18] L. Ma, J.W. Zhu, X.Q. Yuan, Q. Yue, Synthesis of epichlorohydrin from
Acknowledgement dichloropropanols: kinetic aspects of the process, Chem. Eng. Res. Des. 85 (2007)
1580–1585.
[19] J.S. Zhang, Y.C. Lu, Q.R. Jin, K. Wang, G.S. Luo, Determination of kinetic
We gratefully acknowledge financial support from National Natural parameters of dehydrochlorination of dichloropropanol in a microreactor, Chem.
Science Foundation of China (No. 21991104). Eng. J. 203 (2012) 142–147.
[20] C.A. de Araujo Filho, S. Heredia, K. Eränen, T. Salmi, Advanced millireactor
technology for the kinetic investigation of very rapid reactions:
Appendix A. Supplementary data dehydrochlorination of 1,3-dichloro-2-propanol to epichlorohydrin, Chem. Eng.
Sci. 149 (2016) 35–41.
[21] Z. Yan, J. Deng, Y. Chen, G. Luo, Preparation of 2,3-epoxypropyl neodecanoate:
Supplementary data to this article can be found online at https://doi.
process optimization and mechanism discussion, Ind. Eng. Chem. Res. 59 (2020)
org/10.1016/j.cej.2022.136498. 19168–19176.
[22] M.N. Makhin, M.S. Klebanov, Reaction kinetics of dehydrochlorination of the
chlorohydrin esters of bisphenol a, Kinet. Catal.+ 61 (2020) 238–241.
References [23] US 8802872 B2, 2014.
[24] JP 2014210738 A, 2014.
[1] A. Berkessel, C. Rollmann, F. Chamouleau, S. Labs, O. May, H. Gröger, Practical [25] WO 2016047277 A1, 2016.
two-step synthesis of an enantiopure aliphatic terminal (S)-epoxide based on [26] Z. Yan, J. Tian, K. Wang, K.D.P. Nigam, G. Luo, Microreaction processes for
reduction of haloalkanones with “designer cells”, Adv. Synth. Catal. 349 (2007) synthesis and utilization of epoxides: a review, Chem. Eng. Sci. 229 (2021) 116071.
2697–2704. [27] D. Russo, G. Tomaiuolo, R. Andreozzi, S. Guido, A.A. Lapkin, I. Di Somma,
[2] M. Tokunaga, J.F. Larrow, F. Kakiuchi, E.N. Jacobsen, Asymmetric catalysis with Heterogeneous benzaldehyde nitration in batch and continuous flow microreactor,
water: efficient kinetic resolution of terminal epoxides by means of catalytic Chem. Eng. J. 377 (2019), 120346.
hydrolysis, Science 277 (1997) 936–938. [28] J. Tan, L. Du, Y.C. Lu, J.H. Xu, G.S. Luo, Development of a gas-liquid
[3] A.N. Rainsberry, J.G. Sage, M.L. Scheuermann, Iridium-promoted conversion of microstructured system for oxidation of hydrogenated 2-
terminal epoxides to primary alcohols under acidic conditions using hydrogen, ethyltetrahydroanthraquinone, Chem. Eng. J. 171 (2011) 1406–1414.
Catal. Sci. Technol. 9 (2019) 3020–3022. [29] L. Qiu, K. Wang, S. Zhu, Y. Lu, G. Luo, Kinetics study of acrylic acid polymerization
[4] G.W. Yang, Y.Y. Zhang, R. Xie, G.P. Wu, High-activity organocatalysts for polyether with a microreactor platform, Chem. Eng. J. 284 (2016) 233–239.
synthesis via intramolecular ammonium cation assisted S(N)2 ring-opening [30] Z. Yan, J. Tian, C. Du, J. Deng, G. Luo, Reaction kinetics determination based on
polymerization, Angew. Chem. Int. Edit. 59 (2020) 16910–16917. microfluidic technology, Chinese J. Chem. Eng. 41 (2022) 49–72.
[5] M.W. Bedore, N. Zaborenko, K.F. Jensen, T.F. Jamison, Aminolysis of epoxides in a [31] L. Sun, Application of Vinyl Esters and Glycidyl Esters of Tertiary Carboxylic Acids
microreactor system: a continuous flow approach to В-amino alcohols, Org. Process in Waterborne Industrial Coatings. Modern Paint and Finishing 20 (2017) 12-17,
Res. Dev. 14 (2010) 432–440. 52.
[6] D.C. Webster, Glycidyl carbamate functional resins and their applications: a [32] Y. Xiao, C. Huang, Y. Li, Application of Glycidyl Neodecanoate, Guangzhou Chem.
review, Polym. Int. 70 (6) (2021) 710–719. Ind. 45 (2017) 13–16.
[7] X. Sun, L. Yang, H. Xing, J. Zhao, X. Li, Y. Huang, H. Liu, Synthesis of [33] Z. Yan, C. Du, G. Luo, J. Deng, Remarkable improvement of epoxide ring-opening
polyethylenimine-functionalized poly(glycidyl methacrylate) magnetic reaction efficiency and selectivity with water as a green regulator, Reaction Chem.
microspheres and their excellent Cr(Vi) ion removal properties, Chem. Eng. J. 234 Eng. 6 (2021) 2159–2169.
(2013) 338–345. [34] K. Wang, L. Li, P. Xie, G. Luo, Liquid-liquid microflow reaction engineering,
[8] E. Bryuzgin, V. Klimov, M.D. Le, A. Navrotskiy, I. Novakov, The superhydrophobic Reaction Chem. Eng. 2 (5) (2017) 611–627.
state stability of coatings based on copolymers of glycidyl methacrylate and alkyl [35] J.H. Xu, P.F. Dong, H. Zhao, C.P. Tostado, G.S. Luo, The dynamic effects of
methacrylates on cotton fabric surface, Fiber. Polym. 21 (2020) 1032–1038. surfactants on droplet formation in coaxial microfluidic devices, Langmuir 28
[9] Z. Yan, Y. Hu, C. Du, J. Deng, K. Wang, G. Luo, Tetramethylammonium (2012) 9250–9258.
neodecanoate as a recyclable catalyst for acidolysis reaction of epichlorohydrin [36] K. Wang, G. Luo, Microflow extraction: a review of recent development, Chem.
with neodecanoic acid, J. Catal. 385 (2020) 44–51. Eng. Sci. 169 (2017) 18–33.
[10] V. Kumar, Y.K. Bhardwaj, S. Sabharwal, Coating characteristics of electron beam [37] L. Sang, X. Feng, J. Tu, B. Xie, G. Luo, J. Zhang, Investigation of external mass
cured bisphenol a diglycidyl ether diacrylate resin containing 1,6-hexanediol transfer in micropacked bed reactors, Chem. Eng. J. 393 (2020), 124793.
diacrylate on wood surface, Prog. Org. Coat. 55 (2006) 316–323.

You might also like