Hybrid catalytic effects of K2CO3 on the synthesis of salicylic acid from carboxylation of phenol with CO2

Article abstract of DOI:10.1007/s11164-015-2025-2

As a base-promoted Kolbe-Schmitt carboxylation reaction, the mechanism of synthesis of salicylic acid derivatives from phenols with CO₂ in the industry is still unclear, even up to now. In this paper, synthesis of 3,6-dichloro salicylic acid (3,6-DCSA) from 2,5-dichloro phenoxide and CO₂ was investigated in the presence of K₂CO₃. We show the reaction can proceed by itself, but it goes at a slower rate as well as a lower yield, compared to the case with the addition of K₂CO₃. However, the yield of 3,6-DCSA is only minorly affected by the size of K₂CO₃, which cannot be explained from the view of catalytic effect. Therefore, K₂CO₃ may on one hand act as a catalyst for the activation of CO₂ so that the reaction can be accelerated, while on the other hand, it also acts as a co-reactant in deprotonating the phenol formed by the side reaction to phenoxide, which is further converted to salicylate.

Full text of DOI:10.1007/s11164-015-2025-2

Res Chem Intermed
DOI 10.1007/s11164-015-2025-2
Hybrid catalytic effects of K2CO3 on the synthesis
of salicylic acid from carboxylation of phenol with CO2
Miaofei Gu1 Xingxing Yan1 Zhenmin Cheng1
•
•
Received: 12 December 2014 / Accepted: 19 March 2015
Ó Springer Science+Business Media Dordrecht 2015
Abstract As
a base-promoted Kolbe–Schmitt carboxylation reaction, the
mechanism of synthesis of salicylic acid derivatives from phenols with CO2 in the
industry is still unclear, even up to now. In this paper, synthesis of 3,6-dichloro
salicylic acid (3,6-DCSA) from 2,5-dichloro phenoxide and CO2 was investigated in
the presence of K2CO3. We show the reaction can proceed by itself, but it goes at a
slower rate as well as a lower yield, compared to the case with the addition of
K2CO3. However, the yield of 3,6-DCSA is only minorly affected by the size of
K2CO3, which cannot be explained from the view of catalytic effect. Therefore,
K2CO3 may on one hand act as a catalyst for the activation of CO2 so that the
reaction can be accelerated, while on the other hand, it also acts as a co-reactant in
deprotonating the phenol formed by the side reaction to phenoxide, which is further
converted to salicylate.
Keywords 3,6-Dichrolo salicylic acid Á Kolbe–Schmitt reaction Á Carboxylation Á
CO2 utilization
Introduction
3,6-Dichloro salicylic acid (3,6-DCSA) is the main intermediate in producing 3,6-
dichloro-2-methoxy benzoic acid (Dicamba), a benzoic acid herbicide widely used
for the control of broadleaf weeds and several grassy weeds [1]. 3,6-DCSA is also
used for the synthesis of some fine chemicals, such as chemiluminescences,
pharmaceuticals, and pigments. Compared to some other synthesis routes from
1,2,4-trichloro benzene and 2,5-dichloro aniline, the Kolbe–Schmitt reaction is the
&
Zhenmin Cheng
zmcheng@ecust.edu.cn
1
State Key Laboratory of Chemical Engineering, East China University of Science
and Technology, Shanghai 200237, People’s Republic of China
123

M. Gu et al.
most common one to generate aromatic hydroxyl carboxylic acids, owing to high
purity of the product, as well as the wide availability of material source [2]. Usually,
3,6-DCSA is prepared by carboxylation of 2,5-dichloro phenoxide (2,5-Cl2PhOK),
the dehydrated product of 2,5-dichloro phenol (2,5-DCP), with carbon dioxide in
organic solvents at elevated pressures, as shown in Eq. (1):
ð1Þ
It is suggested that at high temperatures, the side reaction of Kolbe–Schmitt will
take place as presented in Eq. (2) [2, 3], in which mono-metallic salicylate can be
further converted to di-metallic salicylate, with equimolar phenol generated. In our
previous study on the synthesis of 3,6-DCSA by the Kolbe–Schmitt reaction [4], it
was verified that the side reaction can completely proceed through the Brønsted–
Lowry proton transfer mechanism, and therefore, the overall stoichiometric
equation of synthesizing 3,6-DCSA by the Kolbe–Schmitt reaction can be written
in the form of Eq. (3) by summing Eqs. (1) and (2), which means the theoretical
maximum yield of 3,6-DCSA is only 50 %. This low yield is believed to result from
the side reaction; hence, the key method to improve the yield of 3,6-DCSA is to
convert the 2,5-DCP produced via auxiliary reaction to 2,5-Cl2PhOK that can be
carboxylated [5–7].
ð2Þ
ð3Þ
Research interests in the past have emphasized the mechanism of the Kolbe–
Schmitt reaction both theoretically and experimentally [8–18], and the
‘‘3-intermediates and 3-transition states’’ mechanism proposed by Markovic et al.
[17] was well accepted; it involves three steps to generate 3,6-DCSA. In the first
step, a polarized O–Na bond of sodium phenoxide is attacked by the carbon dioxide
molecule, and the intermediate PhONa–CO2 is formed. In the next, the electrophilic
carbon atom attacks the benzene ring primarily at the ortho position, thus forming
123

Hybrid catalytic effects of K2CO3 on the synthesis of…
two new intermediates. In the final step, the product sodium salicylate is formed by
the 1, 3-proton shift.
However, as a promoter of CO2–carboxylation reactions in the synthesis of
organic carbonates [19–21], the catalytic mechanism of K2CO3 remains unclear. It
has been suggested by Iijima et al. [22] that in carboxylation of phenol with CO2
catalyzed by K2CO3 under supercritical conditions, the hydroxyl group of phenol
bonds with the basic O2- site on the surface of K2CO3 to produce a phenoxyl anion,
which then releases a hydrogen atom, to react with CO2 adsorbed on the adjacent
Lewis acid sites for generating a salicylate anion. Yamaguchi et al. [23] provided a
possible catalytic mechanism of K2CO3 as a base catalyst through studying some
new solid bases derived from supported metal nitrates or carbonates, and they
believe that K2CO3 supported on metal-oxide supports would decompose at
moderate temperatures to produce (K?)2-CO32-, which by occupying a basic active
center for the carboxylation, meanwhile releases one CO2 molecule. In addition, Cai
et al. [21] used the isotope tracer technique to study the direct synthesis of dimethyl
carbonate (DMC) promoted by K2CO3, and they found that the carbonate group of
DMC is partially derived from K2CO3, indicating that K2CO3 is the precursor of
carboxyl group. Their conclusion seems in accordance with the K2CO3 decompo-
sition mechanism proposed by Yamaguchi et al. [23].
In this work, the mechanism of carboxylating 2,5-DCP with CO2 to form 3,6-
DCSA was investigated with a series of experimental and theoretical methods,
which would be essential for the improvement and optimization of the Kolbe–
Schmitt reaction catalyzed by the solid base.
Experimental section
Synthesis of 3,6-DCSA
2,5-DCP was commercially obtained with the purity above 99.5 % in the mass
fraction. A typical procedure for preparing 3,6-DCSA can be found in our previous
report [4] Commercial potassium carbonates were ground and screened to different
diameter particles.
Catalyst characterization
Basic properties of catalysts were measured by Temperature Programmed
Desorption (TPD) using a CHEMBET3000 chemical absorber. Samples loaded in
a U-shaped quartz reactor were pretreated in situ in a N2 flow at 300 °C to remove
water and carbonates formed during storage, cooled down to room temperature, and
then exposed to a flowing mixture of 10 % of CO2 in He (100 ml/min) until surface
saturation was achieved (2 h). Weakly adsorbed CO2 was removed by flushing with
N2. Finally, the temperature was increased to 800 °C at a rate of 8 °C/min. CO2
production was monitored using a quadruple mass detector.
123

M. Gu et al.
K123CO3 isotope tracer experiment
Tracer isotope reagent K123CO3 was used to reveal the behavior of K2CO3 in the
Kolbe–Schmitt reaction. K123CO3 was provided by Sigma Aldrich, with an
abundance of 13C up to 98 %. After acidification, the mixture of product 3,6-
DCSA and unreacted 2,5-DCP was extracted by ethyl acetate and analyzed by
Liquid Chromatography–Electro Spray Ionization–Tandem Mass Spectrometry
(LC–ESI–MS). LC–ESI–MS was run on a Finnigan TSQ Quantum quadrupole mass
spectrometer (Thermo Electron Corporation, San Jose, CA, USA) equipped with an
ESI ion source and a Surveyor auto-sampler and an MS-pump (Thermo Electron
Corporation). Spray voltage was 3200 V. Sample Cone voltage was 39 V. The
temperature of Desolvation and Source were 120 and 80 °C, respectively. Ion
polarity was negative.
Computational method
To gain an insight into the activation of CO2 absorbed on K2CO3, a theoretical
calculation based on density functional theory (DFT) was performed using Material
Studio 5.0. The Perdew–Burke–Ernzerh (PBE) correlation functional form of the
generalized gradient approximation (GGA) for exchange and correlation energies
was used in the Dmol3 module [24–26]. Double numerical plus polarization (DNP)
function was utilized in all calculations. The threshold values of convergence
criteria for energy, gradient, and displacement convergence were 1 9 10-5 Ha,
˚
˚
0.002 Ha/A, and 0.005 A, respectively. A ‘‘fine’’ convergence criterion was adopted
for all calculations. For the self-consistent field cycle convergence, the threshold
was 1.0 9 10-5 eV. Zero point vibrational energy (ZPVE) corrections were
calculated using the harmonic vibrational modes of surface adsorbates.
The surface energy is defined as a function of the discrepancy between the total
energies of the system before and after the deformation, which quantifies the
disruption of intermolecular bonds formed during creation of a surface. The specific
surface energy was calculated by the following formula:
Esurf ¼ ½Eslab À EbulkŠ=2A
where Eslab refers to the total energy of the slab, Ebulk is the energy for bulk that has
ð4Þ
the same number of atoms with the slab, and A is the surface area. A vacuum region
˚
of 15 A is added above the surface to ensure negligible interaction between periodic
images normal to the surface. In order to ensure computational accuracy, all the
molecules are allowed to relax freely.
The adsorption energy is defined as:
DEads ¼ EðsurfþadsÞ À ðEsurf þ Eads
Þ
ð5Þ
where EðsurfþadsÞ is the total energy of the surface with adsorbate, and Eads is the
energy of the free adsorbate.
123

Hybrid catalytic effects of K2CO3 on the synthesis of…
Results and discussion
Optimization of the reacting conditions
Selection of combinations of raw materials and catalysts
Various combinations of raw materials and catalysts for synthesizing salicylic acid
have been attempted. According to published reports, phenol, alkali and alkaline earth
metal phenoxide are assumed to be the appropriate raw materials for the
carboxylation. Among these reports, Iijima et al. [22] prepared salicylic acid by
directly carboxylating phenol with CO2 under supercritical conditions. Rahim et al.
[27] investigated the effects of alkaline earth metal on the Kolbe–Schmitt reaction,
indicating that potassium and sodium phenoxides show good yields of ortho products.
As for the catalyst in the Kolbe–Schmitt reaction, carbonates and bicarbonates are
most commonly used. Since a complex gas–liquid–solid multiphase is involved, some
ionic liquids [28, 29] and phase transfer catalysts (PTC) [30] have been introduced to
improve the homogeneity of the system, expecting to enhance the carboxylation yield.
Furthermore, a heterogeneous photocatalytic system was reported to synthesize
salicylic acid from phenol and CO2 [31], and the supported catalysts were suggested
to improve the catalytic activity and recovery efficiency [32], All the above reports
inspired us to find the best combination of raw material and catalyst to obtain a high
yield. Typical results of various combinations are shown in Table 1.
Table 1 Carboxylation of various alkaline earth metal phenoxides with CO2 using different catalysts
Entry Raw
material
Catalyst
Ratio of
catalyst/
raw material
Reaction condition
Yieldd
(%)
CO2
MPa
Temp.
°C
Time
h
1
2,5-DCP
2,5-DCP
2,5-DCP
–
–
6
160
160
160
160
160
4
6
7
6
4
0
2
KHCO3
K2CO3
0.5
0.5
0.5
0.5
0.05
6
9.4
3
6
27.7
26.1
52.2
\23.9
\23.9
0
4
2,5-Cl2PhOK KHCO3
2,5-Cl2PhOK K2CO3
2,5-Cl2PhOK Ionic liquida
2,5-Cl2PhOK K2CO3 ? Ionic liquida 0.5 ? 0.05
2,5-Cl2PhOK PTCb
7
5
6
6
4–8
4–8
4–8
6
140–160 4–8
140–160 4–8
140–160 4–8
7
8
0.05
9
2,5-Cl2PhOK K2CO3 ? PTC
2,5-Cl2PhOK Na2CO3
0.5 ? 0.05
0.5
160
160
4
6
52.9
31.7
\27.5
\34.8
10
11
12
a
6
2,5-Cl2PhOK Na2CO3
2,5-Cl2PhOK Supported K2CO3c
0.5
4–8
4–8
140–160 4–8
140–160 4–8
0.5
Ionic liquid: [BMIM](1-butyl-3-methylimidazolium)-HCO3 was used
b
c
d
PTC: tetra-n-butylammonium bromide was used
Catalyst carrier used including c-Al2O3, zeolite, activated carbon, silica
The yield of the carboxylation based on the amount of 2,5-DCP used; For the 2,5-DCP ? CO2 system,
para-production is ignored due to its scarce yield; so does the following study
123

M. Gu et al.
It can be concluded from the experimental data that 2,5-DCP cannot react with CO2
unless alkaline catalyst is added (the role of alkaline catalyst will be discussed in detail
in the later Sects. of 3.4), and alkaline metal phenoxide exhibits good reactivity.
Markovic et al. [12] found that the yield of the para substituted product increases with
an increasing ionic radius of the alkali metal (Li, Na, K, Ru, Ce) used. Taking into
account of both reaction effect and economic cost, potassium 2,5-dichloro phenoxide
is selected as the most appropriate raw material for 3,6-DCSA synthesis. It seems that
the phase transfer catalysts, ionic liquids and supported catalysts do not achieve an
excellent effect as expected, but are even worse than the traditional carbonate catalyst.
Therefore, we suggest that the poor yield is not caused by the mass transfer in the
multiphase system. Actually, Massoth and Doraiswamy [33] proposed that chemical
reaction might be the controlling step rather than diffusion, by deriving the mass
transfer model of the Kolbe–Schmitt carbonation of 2-naphthol.
Influence of the reacting condition
Potassium 2,5-dichloro phenoxide and K2CO3 were selected as the optimal raw
material and catalyst, respectively. Carboxylation was conducted under the general
conditions of the Kolbe–Schmitt reaction, namely high temperature (100–160 °C)
and high pressure (6–8 MPa), both with and without catalyst. The starting feed of
material phenoxide was 0.5 mol. Stirring speed was required to be more than
600 rpm, which is high enough to eliminate the impact of external diffusion
between the solid catalyst and phenoxide. The content of the reaction product in
aqueous phase and organic phase was analyzed by high performance liquid
chromatography and computed as exhibited in Table 2.
Table 2 Effect of reaction pressure, temperature, time, and catalyst on yield of 3,6-DSCA
Temp/ °C CO2/MPa Time/h K2CO3/mol Aqueous phase
Organic
phase
3,6-DCSA/ 2,5-DCP/ 2,5-DCP/mol
Total
moles/mol
mol
mol
180
160
140
120
100
80
6
6
6
6
6
6
8
4
6
6
6
6
6
6
4
4
4
4
4
4
4
4
9
7
5
3
5
5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
–
Coking
0.261
0.229
0.143
0.06
0.058
0.083
0.188
0.411
0.498
0.127
0.186
0.114
0.12
0.172
0.178
0.16
0.491
0.489
0.491
0.49
0.019
0
0
0.498
160
160
160
160
160
160
160
160
0.183
0.185
0.196
0.198
0.175
0.182
0.267
0.295
Coking
0.127
0.179
0.18
0.498
0.489
0.499
0.482
0.475
0.497
0.478
–
–
0.143
0.123
0.058
0.049
0.164
0.171
0.172
0.134
–
0.5
0.75
123

Hybrid catalytic effects of K2CO3 on the synthesis of…
Experimental data indicated that the Kolbe–Schmitt reaction between phenoxide
and CO2 can proceed without catalyst, but both the yield and the reaction rate are much
lower than those using K2CO3. The reaction exhibits a better yield at higher
temperatures, and does not occur below 80 °C, which proves that a higher temperature
is favorable to form 3,6-DCSA. However, temperatures higher than 180 °C will cause
serious coke, leading to extremely low yields. The yield is enhanced with increasing
pressures; but it goes down if the pressure exceeds the supercritical point. This
phenomenon demonstrates that supercritical reactions do not always show advan-
tages, especially for a gas–liquid-solid reaction [34]. A possible explanation is that a
gas–liquid system tends to be homogeneous, which is more like a gas phase, with a
worse dispersion property compared to the liquid solution.
Role of K2CO3 on the reaction
The promoting effect of K2CO3 is carefully examined in this study because of its
abnormal performance. The influences of the particle size and the usage amount of
K2CO3 on the yield of 3,6-DCSA are shown in Figs. 1 and 2, respectively. It can be
concluded that the particle size does not have a significant impact on the yield, as
the yield exhibits a deviation\5 % with different particle sizes of K2CO3. However,
the usage amount of K2CO3 has a great impact on the salicylate yield, and the
increasing trend is apparent until the ratio of K2CO3 to 2,5-DCP is up to 1.5. These
results show that the reaction is solely influenced by the usage amount of K2CO3,
rather than its surface area, indicating that the Kolbe–Schmitt reaction is not a
surface-controlled reaction.
Detailed kinetic data with different usage amounts of K2CO3 are displayed in
Fig. 3. With the increasing usage amount of K2CO3, the reaction rate, which refers
60
55
50
45
40
80~100
60~80
40~60
20~40
10~20
1~10
Mesh number of K2CO3
Fig. 1 The effect of particle size of K2CO3 on the yield of 3,6-DCSA (the ratio of K2CO3 to 2,5-DCP is
1:1)
123

M. Gu et al.
60
50
40
30
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00
The mole ratio of K2CO3 to 2,5-DCP
Fig. 2 The effect of the usage amount of K2CO3 on the yield of 3,6-DCSA
55
50
45
40
35
30
K2CO3:2,5-DCP=1:1
K2CO3:2,5-DCP=0.5:1
no K2CO3
25
20
0
1
2
3
4
5
6
Reaction time/hr
Fig. 3 Influence of amount of catalyst on the yield of 3,6-DCSA changing along with reaction time
(reaction temperature: 160 °C; reaction pressure: 6 MPa)
to the slope of the curve plotted versus reaction time, is raised, and the time to reach
equilibrium is shortened. What is worth mentioning is that the equilibrium
conversion is also improved by K2CO3, which means K2CO3 acts not only as a
catalyst, but also somewhat as a co-reactant, since the conventional catalytic theory
declares that catalyst can only accelerate the reaction, but cannot improve the
equilibrium yield.
123

Hybrid catalytic effects of K2CO3 on the synthesis of…
Hybrid catalytic effect of K2CO3
K123CO3 Isotope tracing
To examine whether K2CO3 could provide a carboxyl source as expected in Cai and
Yamaguchi et al.’s [21, 23] conclusions, isotope reagent K123CO3 was used instead
of K2CO3 to trace the carbon atom in the carboxylation pathway. The aqueous
mixture of 3,6-DCSA and unreacted 2,5-DCP was analyzed by electrospray
ionization mass spectrometry, and the result is shown in Fig. 4. As a negative ion
detective mode, mass-to-charge (m/z) ratio on the x-axis is [M-1]-, while the
relative intensity (%) of each peak is exhibited on the y-axis. 3,6-DCSA exhibited
three obvious peaks: 205.95 (100.0 %), 207.95 (63.9 %), and 209.95 (10.2 %),
which are contributed by three kinds of natural isotopes of chlorine atoms in the 3,6-
DCSA molecule, respectively. Yet, the molecular ion peak involving 13C can be
neglected. Therefore, K2CO3 is in no way a carboxyl group source in the Kolbe–
Schmitt reaction.
TPD profile of K2CO3
Decomposition of K2CO3 was observed on TPD profile by monitoring CO2
(m/z = 44) as the decomposition product in this study. Figure 5 shows the change
of TPD profiles of K2CO3 with and without CO2 adsorption. K2CO3 itself gives a
single peak at around 875 °C, which closely corresponds to its decomposition
temperature. The CO2-adsorbed K2CO3 shows two extreme decompositions,
reaching the maximum around 100 and 140 °C in the range of the reaction
Fig. 4 Mass spectra of productions using K123CO3 as isotope labeling reagent
123

M. Gu et al.
K2CO3 with CO2 adsorption
K2CO3 without CO2 adsorption
200
400
600
800
1000
Temperature/°C
Fig. 5 CO2-temperature programmed desorption curves of K2CO3
temperatures. It can be inferred that two weak basic sites exist on the surface of
K2CO3, which could be the active centers for CO2.
The theoretical analysis
On the basis of experimental studies on the crystal structure of potassium carbonate
by Idemoto et al. [35] and Becht et al. [36], the geometry structure of K2CO3 body
phase was optimized, and the accuracy was verified by comparing with
experimental results (Table 3). The relative error is between -1.35 and 2.17 %,
which is \5 %, indicating that the DFT calculations under the above parameter
settings are reliable.
The optimized powder crystal structure is shown in Fig. 6. Bulk K2CO3 belongs
to the P121/c1 space group, with four formula units in the primitive cell. Its (001)
surface has the lowest surface energy, where the reaction is most likely to occur
[37]. CO2 is easily activated at the basic site due to its weak electroaffinity. The
analysis of electronic density distribution of K2CO3 (001) surface is illustrated in
Fig. 7, which also indicates a potential Lewis basic center: the red loop is the
electron-rich center. Therefore, the top site of the potassium atom, the bridge site
and hollow site of oxygen atoms are likely to be the active sites for CO2.
Table 3 Structural parameters of K2CO3 by DFT calculation
˚
a/A
˚
b/A
˚
c/A
˚
Bond length/A
Structure
parameters
a/°
b/°
c/°
O1–C O2–C O3–C
Experiments[131] 5.64
9.83
9.88
6.83
6.96
90.00
90.00
98.70
97.37
90.00
90.00
1.28
1.32
1.32
1.28
1.28
DFT
5.76
1.281
Relative error
2.17 % 0.52 % 1.88 %
0.00 % -1.35 %
0.00 % 0.27 % 0.25 % 0.21 %
123

Hybrid catalytic effects of K2CO3 on the synthesis of…
Fig. 6 The optimized crystal structure of bulk K2CO3
Fig. 7 Electronic density distribution of the (001) surface of K2CO3
The possible structure of CO2 adsorbed on the (001) surface was optimized.
According to Seiferth et al.’s [38] research on adsorption of CO2 on chromium
surface, the adsorption states of CO2 on a metal oxide surface can be summarized as
five configurations: (1) one oxygen atom of CO2 is perpendicularly connected to the
surface metal atoms, CO2 remains linear structure; (2) the carbon atom of CO2 is
adsorbed on the top site of surface metal atoms to form monodentate carbonate; (3)
both the carbon atom and one oxygen atom of CO2 are interacted with the surface to
form the bidentate carbonate; (4) both of the oxygen atoms of CO2 are linked to
metal atoms respectively, while the carbon atom is directly down to generate a CÁÁÁO
bond with the surface oxygen atoms; (5) ditto, except that the carbon atom of CO2 is
directly above to form a bridge carbonate structure. In this study, each possible
configuration of adsorbed CO2 is adopted as initial structure for optimization.
123

M. Gu et al.
Fig. 8 Optimized structures of CO2 (a–c) adsorbed to the K2CO3 (001)
Table 4 Optimized structure parameters and adsorption energies for adsorbed CO2
˚
Configurations Bond length/A
Bond angle DEads
of CO2/°
/eV
C
CO2
À Osurf aOa
À Ksurf aOb
À Ksurf
C
CO2
À O
CO2
CO2
CO2
A
B
C
a
2.31
2.45
3.91
3.04
2.88
1.19
3.12
1.18
1.19
1.17
157.4
169.8
155.1
-0.35
-0.26
-0.17
3.37
3.36
OaCO and ObCO refers to the two oxygen atoms of CO2, respectively
2
2
Fig. 9 A possible catalytic mechanism for 3,6-DCSA synthesis from 2,5-Cl2PhOK and CO2 on the
surface of K2CO3
Finally, three convergent structures are obtained, as shown in Fig. 8a–c. Each
structural parameter is listed in Table 4, as well as the corresponding adsorption
energy.
123

Hybrid catalytic effects of K2CO3 on the synthesis of…
0.20
3,6-DCSA
2,5-DCP
0.15
0.10
0.05
0.00
0
2
4
6
8
Reaction time/hr
(A) Without K2CO3 added.
0.25
3,6-DCSA
2,5-DCP
0.20
0.15
0.10
0.05
0.00
0
1
2
3
4
5
6
7
8
9
Reaction time/hr
(B) With K2CO3 added.
Fig. 10 A comparison between the amount of 3,6-DCSA in aqueous phase and 2,5-DCP without and
with K2CO3 added(K2CO3:2,5-DCP = 1:1). a Without K2CO3 added. b With K2CO3 added
It can be found that the C–O bond lengths of the adsorbed CO2 are 1.18, 1.19,
˚
˚
1.17 A, somewhat elongated compared to gas CO2 (1.17 A), which is caused by the
transfer process of the partial electrons of the surface O atom to C–O anti-bonding
orbital of the adsorbate CO2. The negative adsorption energy indicates that the
adsorption of CO2 on K2CO3 (001) surface is a chemical adsorption, and the
monodentate carbonate structure (A) is thought to be the most possible adsorption
configuration owing to its highest adsorption energy. The C–O bond is elongated
and the C atom is highly activated with a high electroaffinity. Usually, the bent CO2
is an activated precursor of different chemical reactions. A possible catalytic
mechanism of K2CO3 on the synthesis of 3,6-dichloro salicylate is depicted in
123

M. Gu et al.
Fig. 9. CO2 is activated on the basic site with a bent angle, forming an intermediate
2,5-Cl2PhOK-CO2 with 2,5-DCP on the surface of the K2CO3, then the C atom
electrophilically attacks the ortho carbon of phenoxide; meanwhile, the hydrogen in
ortho carbon shifts to the phenoxyl oxygen via 1,3-proton transfer. The reaction
route is almost the same as the ‘‘3-intermediate and 3-transition states’’ mechanism
proposed by Markovic et al. [17].
The deprotonate effect of K2CO3 as a co-reactant
As is known, the side reaction will generate equal amounts of phenol and salicylate
if no catalyst is added (Fig. 10a), yet the addition of K2CO3 makes a different
product distribution (Fig. 10b). In that case, the amount of 2,5-DCP is significantly
less than that of 3,6-DCSA, while the latter has an enhanced yield. It can be
explained by the fact that 2,5-DCP is deprotonated by K2CO3 to form 2,5-Cl2PhOK,
which can be carboxylated to 3,6-dichloro salicylate. This is another important role
of K2CO3: as the co-reactant likes alkali to deprotonate 2,5-DCP to 2,5-dichloro
phenoxide, which is the prerequisite to generate 3,6-DCSA. The chemical
stoichiometry equation of K2CO3 reacting with 2,5-DCP is written as follows [39]:
ð6Þ
A quantitative measurement on the deprotonated effect on 2,5-DCP by K2CO3 is
implemented. The conversion of 2,5-DCP is investigated by reacting with K2CO3 at
60
40
20
0
0
2
4
The mole ratio of K2CO3/2,5-DCP
Fig. 11 The conversion of 2,5-DCP to 2,5-Cl2PhOK under the different usage amount of K2CO3
123

Hybrid catalytic effects of K2CO3 on the synthesis of…
different mole ratios, under the Kolbe–Schmitt reaction conditions (160 °C, 6 MPa,
4 h, xylene as solvent). The result is shown in Fig. 11. It can be seen that K2CO3 is
quite a weak alkali due to its un-dissociation in organic solvents. Only 11 %
conversion of 2,5-DCP can be obtained when the mole ratio of K2CO3/2,5-DCP is
1:1, while the yield of 3,6-DCSA can be improved by 13 % in a practical reaction.
Conclusions
Hybrid effects of K2CO3 have been observed both in raising the reaction rate and in
enhancing the equilibrium yield in the Kolbe–Schmitt reaction. On one hand, it acts
as a catalyst since the CO2-TPD of K2CO3 exhibits two weak basic centers on the
surface of K2CO3 for adsorption of the CO2 molecules, consistent with DFT
calculations showing that CO2 could be adsorbed and activated over K2CO3(001)
surface with an adsorption energy of 0.35 eV. The adsorbed and highly activated
CO2 is then converted into monodentate carbonate structure. On the other hand,
K2CO3 acts as a deprotonate agent, dehydrating the auxiliary product 2,5-DCP to
2,5-Cl2PhOK, which can be carboxylated to 3,6-dichloro salicylate.
Acknowledgments Financial support from the Fundamental Research Funds for the Central Univer-
sities is gratefully acknowledged.
References
1. C.R. Worthing, S.B. Walker, Pesticides Manual: A World Compendium (British Crop Protection
Council, London, 1987), p. 693
2. A.S. Lindsey, H. Jeskey, Chem. Rev. 57, 583 (1957)
3. F. Wessely, K. Benedikt, H. Benger, G. Friedrich, F. Prillinger, Monatsh. Chem. 81, 1071 (1950)
4. X. Yan, Z. Cheng, Z. Yue, P. Yuan, Res. Chem. Intermed. 40, 3059 (2014)
5. H. Hoch, Q. Seeger, US 4,002,675 (1977)
6. O.F. Schulz, H. Martens, K. Burgdorf, US 2,808,434 (1957)
7. J.C. Wygant, US 3,089,905 (1960)
8. J.L. Hales, J.I. Jones, A.S. Lindsey, J. Chem. Soc. 3145 (1954). doi:10.1039/JR9540003145
9. S.E. Hunt, J.I. Jones, A.S. Lindsey, D.C. Killoh, H.S. Turner, J. Chem. Soc. 3152 (1958). doi:10.
1039/JR9580003152
10. M. Kunert, E. Dinjus, M. Nauck, J. Sieler, Chem. Ber./ Recl. 130, 1461 (1997)
11. O. Baine, G.F. Adamson, J.W. Barton, J.L. Fitch, D.R. Swayampati, H. Jeskey, J. Org. Chem. 19, 510
(1954)
12. Z. Markovic, S. Markovic, N. Begovic, J. Chem. Inf. Model. 46, 1957 (2006)
13. I. Stanescu, R.R. Gupta, L.E.K. Achenie, Mol. Simul. 32, 279 (2006)
14. I. Stanescu, L.E.K. Achenie, Chem. Eng. Sci. 61, 6199 (2006)
15. Z. Markovic, S. Markovic, N. Manojlovic, J. Predojevic-Simovic, J. Chem. Inf. Model. 47, 1520
(2007)
16. Y. Kosugi, Y. Imaoka, F. Gotoh, M.A. Rahim, Y. Matsui, K. Sakanishi, Org. Biomol. Chem. 1, 817
(2003)
17. Z. Markovic, J.P. Engelbrecht, S. Markovic, Z. Naturfors, Sect. A-J. Phys. Sci. 57, 812 (2002)
18. X. Yan, Z. Cheng, Z. Yue, P. Yuan, Res. Chem. Intermed. 40, 3045 (2014)
19. N. Aljaar, J. Conrad, U. Beifuss, J. Org. Chem. 78, 1045 (2013)
20. V. Eta, P. Ma¨ki-Arvela, A.R. Leino, T. Salmi, D.Y. Murzin, J.P. Mikkola, Ind. Eng. Chem. Res. 49,
9609 (2010)
21. Q. Cai, B. Lu, L. Guo, Y. Shan, Catal. Commun. 10, 605 (2009)
123

M. Gu et al.
22. T. Iijima, T. Yamaguchi, Appl. Catal. A 345, 12 (2008)
23. T. Yamaguchi, Y. Wang, M. Komatsu, M. Ookawa, Catal. Surv. Jpn. 5, 81 (2002)
24. S. Kurth, J.P. Perdew, P. Blaha, Int. J. Quantum Chem. 75, 889 (1999)
25. J.P. Perdew, Y. Wang, Phys. Rev. B. 45, 13244 (1992)
26. V. Balachandran, S. Rajeswari, S. Lalitha, J. Mol. Struct. 1007, 63 (2012)
27. M.A. Rahim, Y. Matsui, Y. Kosugi, Bull. Chem. Soc. Jpn. 75, 619 (2002)
28. U. Krtschil, V. Hessel, D. Reinhard, A. Stark, Chem. Eng. Technol. 32, 1774 (2009)
29. M. Kashid, A. Renken, L. Kiwi-Minsker, Chem. Eng. Sci. 66, 1480 (2011)
30. A. Yeadon, T.A. Turney, G. Ramsay, J. Chem. Educ. 62, 518 (1985)
31. A. Sclafani, L. Palmisano, G. Farneti, Chem. Commun. 529 (1997). doi:10.1039/A607705F
32. I. Lukic, J. Krstic, D. Jovanovic, D. Skala, Bioresour. Technol. 100, 4690 (2009)
33. F. Massoth, L. Doraiswamy, Ind. Eng. Chem. Process Des. Dev. 5, 351 (1966)
34. S. Pereda, E.A. Brignole, S.B. Bottini, J. Supercrit. Fluids 47, 336 (2009)
35. Y. Idemoto, J.W. Richardson Jr, N. Koura, S. Kohara, C.-K. Loong, J. Phys. Chem. Solids 59, 363
(1998)
36. H.Y. Becht, R. Struikmans, Acta Crystallogr. Sect. B: Struct. Sci. 32, 3344 (1976)
37. H. Gao, S. Pishney, M.J. Janik, Surf. Sci. 609, 140 (2013)
38. O. Seiferth, K. Wolter, B. Dillmann, G. Klivenyi, H.J. Freund, D. Scarano, A. Zecchina, Surf. Sci.
421, 176 (1999)
39. V.L. Mil’to, V.Y. Orlov, G.S. Mironov, V.V. Kopeikin, Kinet. Catal. 42, 471 (2001)
123