Carbonates as reactants for the production of fine chemicals: The synthesis of 2-phenoxyethanol

Article abstract of DOI:10.1039/c4cy00913d

The solventless and heterogeneously catalysed synthesis of 2-phenoxyethanol (ethylene glycol monophenyl ether) via the reaction between phenol and ethylene carbonate was investigated using Na-mordenite catalysts as an alternative to the industrial process using ethylene oxide and homogeneous basic conditions. Under specific reaction conditions, it was possible to obtain total selectivity to phenoxyethanol at up to 75% phenol conversion and 82% selectivity at total phenol conversion in 5-7 hours of reaction time and using a moderate excess of ethylene carbonate. The main by-product was the linear carbonate of phenoxyethanol, bis(2-phenoxyethyl)carbonate (selectivity 15%), which could then be converted to phenoxyethanol by reacting with phenol in basic medium with 100% yield; so overall, the phenoxyethanol yield was as high as 97%. With a stoichiometric feed of phenol and ethylene carbonate, the maximum conversion of phenol was just 60%, still with 100% selectivity to phenoxyethanol. An autocatalytic phenomenon was also observed due to the higher basicity of 2-phenoxyethanol compared to phenol, which overlapped the Na-catalyzed activation of phenol. Starting from a commercial Na-mordenite, which showed significant deactivation, and by applying a post-treatment aimed at the reduction of microporosity, it was possible to minimize both the deactivation and Na leaching while keeping the selectivity enhancement effect shown by the mordenite structure.

Full text of DOI:10.1039/c4cy00913d

Catalysis
Science &
Technology
View Article Online
View Journal | View Issue
PAPER
Carbonates as reactants for the production of fine
chemicals: the synthesis of 2-phenoxyethanol†
Cite this: Catal. Sci. Technol., 2014,
4, 4386
P. Ziosi, T. Tabanelli, G. Fornasari, S. Cocchi, F. Cavani* and P. Righi*ab
ab
a
a
a
ab
The solventless and heterogeneously catalysed synthesis of 2-phenoxyethanol (ethylene glycol
monophenyl ether) via the reaction between phenol and ethylene carbonate was investigated using
Na-mordenite catalysts as an alternative to the industrial process using ethylene oxide and homogeneous
basic conditions. Under specific reaction conditions, it was possible to obtain total selectivity to
phenoxyethanol at up to 75% phenol conversion and 82% selectivity at total phenol conversion in 5–7 hours
of reaction time and using a moderate excess of ethylene carbonate. The main by-product was the linear
carbonate of phenoxyethanol, bis(2-phenoxyethyl)carbonate (selectivity 15%), which could then be
converted to phenoxyethanol by reacting with phenol in basic medium with 100% yield; so overall, the
phenoxyethanol yield was as high as 97%. With a stoichiometric feed of phenol and ethylene carbonate, the
maximum conversion of phenol was just 60%, still with 100% selectivity to phenoxyethanol. An autocatalytic
phenomenon was also observed due to the higher basicity of 2-phenoxyethanol compared to phenol,
which overlapped the Na-catalyzed activation of phenol. Starting from a commercial Na-mordenite,
which showed significant deactivation, and by applying a post-treatment aimed at the reduction of
microporosity, it was possible to minimize both the deactivation and Na leaching while keeping the
selectivity enhancement effect shown by the mordenite structure.
Received 12th July 2014,
Accepted 5th August 2014
DOI: 10.1039/c4cy00913d
www.rsc.org/catalysis
Na and Li; a heterogeneous catalyst based on hydroxide-
Introduction
2
–10
exchanged resins is also reported.
However, the most
2
-Phenoxyethanol (PE) is an organic compound that, at room
commonly used catalysts in the current processes are the
hydroxides of alkali metals, especially sodium hydroxide,
typically added in quantities of 0.1–0.3% by weight. The reac-
tion is conducted at temperatures of 110–130 °C and pres-
sures of 1–3 bar. In order to optimize the selectivity towards
the monoethoxylated product, reactants are used in equimolar
amounts; furthermore, ethylene oxide is added slowly to the
reactor. At present, the annual global capacity for PE is
temperature, is a colourless oily liquid with a mild smell of
roses. It is used as solvent for dyes, inks, and resins and is a
synthetic intermediate in the production of plasticizers, phar-
maceuticals, and fragrances. It is also used as a component
for the preparation of textile detergents and the synthesis of
paraben-free, formaldehyde-free cosmetic preservatives (in a
mixture with ethylhexylglycerin), showing a balanced spec-
trum of effectiveness against bacteria, yeasts, moulds and
fungi, as well as being a solvent, fixative, and extender for
1
1
around 170 000 tons.
One significant disadvantage of the current industrial
production is that the product obtained cannot be used as
such by the cosmetic and fragrance industry because of its
pungent metal odour, which is likely due to the residues of
alkali metal catalysts. This problem is not completely solved
even by subsequent distillation, which is carried out in order
1
perfumes, as such or after esterification with isobutyric acid.
Nowadays, PE is produced by reacting phenol with ethyl-
ene oxide (Scheme 1). The hydroxyethylation is carried out in
the presence of basic catalysts; in patents, the catalysts
reported are mainly homogeneous and include various basic
compounds, e.g. ammonia, urea, amines, and phenates of
a
Dipartimento di Chimica Industriale “Toso Montanari”, ALMA MATER
STUDIORUM Università di Bologna, Viale Risorgimento 4, 40136 Bologna, Italy.
E-mail: fabrizio.cavani@unibo.it, paolo.righi@unibo.it
b
Consorzio INSTM, Unità di Ricerca di Bologna, Firenze, Italy
†
Electronic supplementary information (ESI) available. See DOI: 10.1039/
Scheme 1 The reaction between phenol and ethylene oxide catalyzed
c4cy00913d.
by NaOH.
4386 | Catal. Sci. Technol., 2014, 4, 4386–4395
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
to separate the product from unconverted phenol and heavy
by-products. Post-treatments have been proposed whereby
Experimental
1
2
the ether is put in contact with sodium borohydride; the
alkali metal borohydride can be directly added to the reaction
The catalyst used for reactivity experiments was a “CBV 10A”
sodium mordenite molecular sieve from ZEOLYST Interna-
tional with a SiO /Al O mole ratio of 13 (Na O wt% = 6.5;
1
3
medium together with the alkali metal hydroxide. Another
drawback of the current industrial production is the forma-
tion of polyethoxylated by-products with 2 to 80 condensed
ethylene oxide molecules; polymeric glycol ethers of phenols
are formed from further reaction of the desired product
with ethylene oxide. These compounds cause the product to
darken, and post- or in situ treatments are necessary to
prevent this. Patents report maximum phenol conversions of
around 99% with variable selectivity towards 2-phenoxyethanol
2
2
3
−
2
2
surface area = 425 g m ). The zeolite was used either as
such, without any pre-treatment, or after post-treatment
using the liquid-phase deposition of tetraethyl orthosilicate
(TEOS). In the former case, experiments carried out for com-
parison with the thermally pre-treated zeolite (at 400 °C for
3 h in air flow) gave the same results as the untreated zeolite.
The liquid-phase post-treatment was carried out using a
20 mL mixture of 5 vol% TEOS in n-hexane mixed with 2.5 g
of zeolite at room temperature for 15 h. The system was
filtered, dried at 120 °C and calcined at 450 °C for 3 h. The
procedure was repeated twice.
2
–10
(from 88 to 96%), depending on the conditions used.
Alternatively, phenol is reacted with either 2-chloroethanol
1
4
or ethylene carbonate (EC), again in the presence of alkalis.
This last route was claimed in early patents to be a smooth,
controllable reaction that makes it possible to obtain
X-ray diffraction of zeolites was carried out using a Philips
PW1710 instrument (Ni-filtered CuKα radiation, λ = 0.15418 nm;
2Θ interval, 5–80°; step, 0.1°).
1
5–18
phenoxyethyl alcohols in high yields
and, more recently,
it has also been used for introducing aryl nuclei into the
chemical structure of acrylic esters (phenoxyethyl alcohols
Ar adsorption/desorption isotherms (77 K) were recorded
using a Micromeritics ASAP 2020 instrument. Samples were
previously outgassed for 120 minutes at 423 K and 30 μmHg
and then heated for 240 minutes at 623 K. Specific surface
area values were obtained using the multi-point BET equa-
1
9,20
can easily condense with acrylic acid).
The use of carbon-
ates as reactants in the synthesis of fine chemicals and inter-
mediates has now become one of the research areas of major
scientific and applied interest. The use of carbonates instead
of conventional reactants, such as alkyl halides and dialkyl
sulphates, aims not only to avoid both the use of toxic com-
pounds and the generation of waste effluents necessitating
disposal but also to develop chemistry which may offer
advantages in terms of selectivity to the desired compound;
an important example is the use of dimethyl carbonate
for the O-methylation and carboxymethylation of phenolic
tion in the 0.05–0.2 p/p range and total pore volume values
0
0
were calculated at 0.95 p/p . The micropore size distribution
was calculated using the NLDFT-statistic method.
Both in solvents used for the analytical measurement and
in solutions after reactions, atomic absorption analyses were
carefully performed to determine the Na concentration in the
reactants, with the aim of determining the amount of Na
leached during the catalytic reaction. The difference in the
Na concentration between the initial solution and the final
one was very small, just a few ppm, thus close to the analyti-
cal error. Indeed, traces of this element are always present,
thus Na is considered a ubiquitous contaminant. Because of
this, we took extreme care to carry out the analysis in such a
way as to minimize occasional errors. Due to the insolubility
of PE in water, we dissolved our samples in 2-propanol
(Sigma-Aldrich), a solvent chosen because of both its chemi-
cal–physical characteristics and its very low Na content.
The procedure adopted for the analysis was the following:
(a) 50 μL of the sample (either the reactant or the reaction
mixture after reaction) were brought to 5 mL volume with
2-propanol. (b) The sample was then analysed using a SpectraA-100
Varian instrument, equipped with a graphite furnace GTA 110.
The line at 330.3 nm was used, instead of the main one
at 589.6 nm, because the analysis of the organic solution led
to an out-of-range absorption; a further dilution of the solu-
tion would have led to a major error in the measurement,
therefore, the weaker line was used. A 10 μL sample was
injected. The furnace temperature ranged from 75 °C up
to 2000 °C, with intermediate steps at 85, 95, 120 (solvent
removal), and 700 °C (pyrolysis and incineration of organics).
2
1–26
compounds.
EC as an alkylating agent for phenol has been reported
using homogeneous catalysts such as alkali carbonates, alkaline
metal iodides, lithium hydride, and tetraethylammonium
2
7–29
iodide for the synthesis of glycol phenyl ethers.
However,
the main problems with all of these systems are the recovery
of the catalyst, the purification of the product, and – with
some catalysts – also the formation of tar compounds.
Recently, some authors have reported solid basic catalysts
made up of alkali-loaded large-pore zeolites, while an excel-
lent PE yield of 98.5% in the reaction of phenol with ethylene
3
0
carbonate has been reported for the KL zeolite. However, so
far, there has been no report on how the tuning of both acid
properties and reaction parameters for the hydroxyethylation
of phenol with EC and solid basic catalysts affects their
performance.
In the present work, we report a more sustainable process
which avoids the use of any solvent, is based on a heterogeneous
basic catalyst made up of Na-mordenite avoiding the problem
of Na contamination of the product, and with the heteroge-
3
1–34
neous catalyst which can be easily recovered and reused.
The detailed study underpinning such achievements, namely
the systematic studies on the parameters affecting the yield
and selectivity to PE, is also reported.
−
1
The analysis was carried out using an Ar flow of 3 mL min .
(c) The final Na concentration was obtained after subtracting
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 4386–4395 | 4387

View Article Online
Paper
Catalysis Science & Technology
the Na content from the solvent. For each sample, the ana-
lysis was repeated 6 times. Because of the insolubility of
Na salts in organic medium, the calibration curve was made
by means of an aqueous solution of NaNO3 (500 ppb Na),
obtained by dilution of a standard solution (1000 ppm Na);
the volume of the standard solution injected was 10 μL.
Reactivity experiments were carried out as follows: in a
round bottom Pyrex cylinder equipped with an internal
cooling circuit, phenol (3 mmol), EC (6 mmol) and Na
mordenite (0.5% weight with respect to the phenol fed, if not
otherwise specified) were added. The reaction mixture was
stirred at 210 °C for 7 h under nitrogen atmosphere. 50 μL
samples were taken without interrupting the reaction which
were then brought to a 10 mL volume with acetone (HPLC
grade, Sigma-Aldrich); then an aliquot (approx. 2 mL) was
filtered (0.45 μm PTFE filter) to separate the catalyst and then
analysed by means of GC.
Fig. 1 Reaction time needed to obtain the overall yield of PE + BPEC
equal to 95% (♦), 50% (■) and 30% (▲) as a function of the reaction
temperature used. Reaction conditions: phenol/ethylene carbonate,
1/4 (molar ratio); catalyst/phenol, 1/2 (weight ratio); catalyst:
Na-mordenite.
The analysis of the reaction mixture was performed using
an Agilent GC6850 instrument, equipped with an HP-1 capil-
lary column (30 m × 320 μm × 0.25 μm) and a FID held at
2
10 °C led to a remarkable decrease in the reaction time
−
1
−1
2
2
80 °C (H : 40 mL min , air: 450 mL min ); the carrier gas
needed. By increasing the temperature further (250 °C), the
reaction time was shorter, but at a higher conversion the
extent of consecutive reactions also increased while selectivity
declined. For example, at 210 °C the selectivity to PE was
close to 97–98% even at a very high phenol conversion, but it
dropped to less than 95% when the reaction temperature was
−
1
was H (108 mL min ). The injector was held at 250 °C, in
2
the split mode (50 : 1). The volume of the sample injected
was 1 μL. The oven temperature was as follows: 50 °C (2 min),
−
1
−1
ramp 10 °C min , 120 °C, ramp 25 °C min , final T 280 °C
3 min).
Products were isolated from the reaction mixture by
(
2
50 °C. Therefore, we decided to carry out the experiments
means of flash chromatography (230–400 mesh) by using as
the eluent a petroleum ether/ethyl acetate mixture (vol. ratios
from 8/2 to 7/3). Then, the products (PE, DPE and BPEC)
were identified by means of ESI-MS and NMR, except
for 2-[2-(2-phenoxyethoxy)ethoxy]ethanol (TPE) because the
signals due to C14 and C15 triplets overlapped the signals of
other CH2 moieties in other by-products (DPE). With this
compound, identification was achieved by means of ESI-MS
and GC-MS.
at the optimal temperature of 210 °C, which was the best
compromise between an acceptable reaction rate and lower
formation of by-products.
In regard to the latter, three side products were
identified (Scheme 2): the dihydroxyethylated compound
2
-(2-phenoxyethoxy)ethanol (DPE), the trihydroxyethylated
compound 2-(2-(2-phenoxyethoxy)ethoxy)ethanol (TPE), and
the linear carbonate of PE, bis(2-phenoxyethyl)carbonate
(
BPEC). The two by-products DPE and TPE usually formed in
ESI-MS spectra (positive or negative) were recorded using
a Waters Micromass ZQ 4000, equipped with a capillary
probe (3.54 kV), with a cone voltage of 20 volts and direct
low amounts (overall yield less than 2%), with the exception
of experiments where a large excess of ethylene carbonate
was used. Conversely, the BPEC yield was as high as 10–15%
at complete phenol conversion; this by-product may form by
means of two different reaction pathways (Scheme 3):
−
1
injection (20 μL min ).
1
13
H and C NMR spectra were recorded in deuterated
chloroform at 25 °C on a Varian Inova 300, at 300 MHz and
5 MHz, respectively. Details concerning the NMR spectra of
(
a) by successive reactions between PE and the unconverted
7
EC, and then again with another molecule of PE, with the
release of ethylene glycol (transesterification mechanism);
the products are reported in the ESI.†
(
b) by the formation of the intermediate obtained by
reacting phenol with EC (carboxyalkylation mechanism); the
former either eliminates CO2 to form PE or may be trans-
formed further by esterification with PE.
Results
Identification of the best reaction conditions
Preliminary experiments were aimed at finding the reaction
conditions necessary for obtaining high conversion of the
substrate using the Na-mordenite catalyst. Fig. 1 compares
the time needed to reach the given values of PE yield for
different reaction temperatures. At 150 °C (not shown in the
figure), there was a 10% yield of PE only after 35 h of reac-
tion time, while increasing the temperature to 180 °C and
Scheme 2 By-products in the reaction between phenol and EC.
4388 | Catal. Sci. Technol., 2014, 4, 4386–4395
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
Scheme 3 Plausible mechanisms for the formation of the BPEC by-product during the reaction between phenol and EC. Top: mechanism (a);
bottom: mechanism (b) (see text).
It is worth noting that we did not notice the formation of
diphenyl carbonate, the product of transesterification of phe-
nol with EC, which might be a parallel reaction competing
with PE formation. This has been attributed to the fact that
the softer nucleophile phenol will preferably react with the
soft alkylidene C atom of the carbonate to produce glycol
3
0
ether. In the case of PE, however, the aliphatic alcohol may
react with the hard C atom of the carbonyl to yield the trans-
esterification product; therefore, mechanism (a) is still possi-
ble. In regard to mechanism (b), the intermediate compound
2
might react preferably with PE, instead of releasing CO ,
under conditions leading to PE accumulation. Therefore, both
mechanisms (a) and (b) would be more favored under condi-
tions of high phenol conversion and are in any case promoted
at high EC concentration.
An important point is that BPEC could be completely
hydrolyzed to yield PE (Scheme 4). In fact, after separation by
flash chromatography, BPEC was completely converted into
PE (100% yield) within 5 h of reaction time, using the 2%
3
NaOCH catalyst, in refluxing methanol. The same reaction
also occurred under acidic conditions but was much slower
compared to that in the basic medium. Therefore, this prod-
uct would not be a waste compound of the process but an
intermediate compound for PE synthesis. The same is not
true for DPE and TPE, which cannot be converted to phenol
or PE.
Fig. 2 shows the effect of the reaction time on phenol and
EC conversion and on the yield of PE (the desired product) +
BPEC (the by-product which could be converted to PE in a
separate vessel), as well as on the by-products DPE and TPE,
at a temperature of 210 °C for three different phenol/EC
molar ratios, 1/1, 1/2, and 1/4. The catalyst used was
Na-mordenite SAR13 (50 wt% with respect to phenol, which
corresponds to 10 mol% Na). The results obtained highlight
the following aspects.
Fig. 2 Effect of reaction time on phenol conversion (♦), EC conversion
●), and yield of PE + BPEC (■) and DPE + TPE (▲). Reaction
(
conditions: T, 210 °C; catalyst/phenol, 1/2 (weight ratio); Na-mordenite
catalyst; phenol/EC molar ratios: 1/1 (top), 1/2 (middle) and 1/4
(bottom).
Scheme 4 BPEC transformation into PE.
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 4386–4395 | 4389

View Article Online
Paper
Catalysis Science & Technology
(
a) A decrease in the phenol/EC molar ratio led to a pro-
conditions – with an equimolar ratio between phenol and EC
and a phenol conversion no higher than 70% – the selectivity
to PE was always close to 99%, with the formation of less
than 1% BPEC and no formation of DPE and TPE.
gressive increase in the reaction rate; in fact, complete con-
version of the limiting reactant (i.e. EC with the 1/1 phenol/
EC molar ratio but phenol with 1/2 and 1/4 phenol/EC molar
ratios) was observed after 28 h in the former case, after 6–8 h
with the 1/2 ratio, and after less than 3 h with the 1/4 ratio.
The final result of this preliminary investigation is that
the best reaction conditions – i.e. those that make it possible
to obtain fast reaction rates with almost total conversion of
phenol within a few hours of reaction time while maintaining
both high selectivity (≥98%) to PE + BPEC and minimal EC
transformation into by-products – are as follows: T, 210 °C;
feed ratio phenol/EC, 1/2; and catalyst/phenol weight ratio,
1/200 (0.5 wt% catalyst). Fig. 3 summarizes the catalyst per-
formance under the optimized reaction conditions. It can be
seen that the total conversion of phenol was achieved within
9 h of reaction time; total selectivity to PE was obtained at
until ca. 85% phenol conversion, whereas at total conversion
the selectivity to the more undesired DPE (with traces of TPE)
was 4% and the selectivity to BPEC was 18%.
An important effect registered under these conditions was
the presence of an induction period, which led to negligible
PE yield in up to 2 h of reaction time. It is worth noting that
such an effect was not registered under conditions of high
catalyst loading with the same 1/2 phenol/EC feed ratio
(Fig. 2 middle), while it was apparently seen when an equimo-
lar feed ratio was used (Fig. 2 top). Therefore, the induction
period may be attributable to the use of a low catalyst amount
and may be due to problems related to the limited access of
reactants to the basic active sites located in zeolites, a problem
which can be overcome by using large catalyst amounts, probably
because of the contribution of external sites.
(
b) With the 1/1 phenol/EC feed ratio, the maximum con-
version of phenol was just 60%. Despite the use of a stoichio-
metric feed, EC was clearly the limiting reactant; this was
due to the fact that EC underwent parallel decomposition
via decarboxylation and successive oligomerisation of ethylene
glycol. This was confirmed by carrying out experiments with
EC alone, under the same reaction conditions used for the
reaction with phenol. Selectivity to PE was very high, close to
1
00%, because of the negligible formation of DPE, with no
TPE and BPEC at all.
c) When lower phenol/EC feed ratios were used, EC con-
(
version was less than 100%, even though EC still reacted
even after total phenol conversion had been reached, because
of the formation of DPE, TPE (consecutive reactions with PE,
with further consumption of EC), and ethylene glycol oligo-
mers. Total selectivity to PE was obtained only at phenol con-
version lower than 100%, because long reaction times led
to a rapid increase in by-products formation, especially with
the phenol/EC feed ratio equal to 1/4. In general, however,
when the complete phenol conversion was approached the
selectivity declined mainly due to the formation of BPEC,
since the formation of the more undesired by-products, DPE
and TPE, remained very low, especially if higher phenol/EC
ratios were used.
Fig. S1 (ESI†) shows the effect of the catalyst amount on
the catalytic behaviour under fixed reaction conditions. The
conversion of phenol and EC was proportional to the catalyst
amount in the catalyst/phenol weight ratio ranging between
On the other hand, the initial induction period was
followed at first by a slow increase in conversion and then by
an acceleration with a rapid increase in the conversion within
a short reaction time which is clearly not something one
might expect under the usual catalytic conditions. One possi-
ble explanation for the phenomenon observed is an autocata-
lytic effect, i.e. an acceleration in the rate due to the PE itself.
0
and 0.5%. For higher amounts of catalysts, the conversion
of phenol decreased while that of EC increased. As shown in
Fig. 2 (top), at an equimolar phenol/EC ratio – at which com-
plete conversion of phenol is theoretically possible – the
conversion of phenol was indeed less than 100% because of
EC transformation into non-useful products; in other words,
the decrease in the phenol conversion shown for increasing
catalyst amounts (Fig. S1†) was due to the increased contribu-
tion of EC decomposition.
Therefore, under these circumstances, the decomposition
of EC becomes kinetically more facilitated than its reaction
with phenol, which causes a decline in both PE yield and
phenol conversion. On the other hand, there was only a small
increase in phenol conversion when the catalyst amount
was increased from 0.5 to 5 wt% despite the fact that the
EC conversion was still similar to that of phenol (i.e. with
a very low EC decomposition). Therefore, it cannot be ruled
out that for high catalyst loadings, i.e. over 0.5 wt%, a poorer
contact between the reactants (because of the absence of
solvent) and the catalyst may result in slower reaction rates
compared to experiments carried out with a lower catalyst
amount. Lastly, it is also worth noting that under such
Fig. 3 Effect of reaction time on phenol conversion (♦) and on yields
of PE (■), BPEC (×), and DPE + TPE (▲). Reaction conditions: T, 210 °C;
phenol/EC molar ratio, 1/2; catalyst amount, 0.5 wt% with respect to
phenol.
4390 | Catal. Sci. Technol., 2014, 4, 4386–4395
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
The induction period and the autocatalytic effect
equal to 1/1). There was an induction period of a few hours,
after which the reaction started. It is important to note that
the possible contamination of the reactants used due to
alkali metals was ruled out based on the negligible Na
content analytically found. It is possible that the reaction
took some time (i.e. the induction period) to show a non-
negligible reaction rate in the absence of added catalysts
because of the slow formation of the phenate anion (although
it is accelerated at 210 °C compared to lower temperatures).
Then, the reaction was accelerated because of the formation
of PE (see below for the explanation of the autocatalytic
effect).
In order to confirm our hypothesis, we conducted a series of
experiments without any catalyst. The experiments shown in
Fig. 4 (top) were carried out by adding increasing amounts of
PE to the reaction medium at 210 °C (no reaction was
observed in the absence of a catalyst at 180 °C, even after
2
4 h), with a phenol/EC ratio equal to 1/2. It is worth noting
that we first treated the commercial PE using a silica plug
in order to remove any traces of alkali metal cations.
Fig. 4 (bottom) compares the PE yield shown as a function of
time in the absence of any catalyst (and without preliminary
addition of PE) and in the presence of a catalyst (0.5 wt%
Na-mordenite, with respect to phenol). The following consid-
erations may be made.
(
b) Addition of increasing amounts of PE from the begin-
ning of the reaction shortened the induction period until it
became negligible when the reaction was carried out in the
presence of a rather high amount of PE. Also, addition of PE
accelerated the reaction rate: a rapid increase in the yield as
a function of time was observed in the presence of greater
amounts of added PE.
(
a) The reaction proceeded even in the absence of any cat-
alyst but was clearly slower than in the presence of a catalyst
Fig. 4, bottom). After 7 h of reaction time, the PE yield was
(
about 4% (as also shown in Fig. S1,† for a phenol/EC ratio
(
c) In all cases, the selectivity to PE was 100% because of
the relatively low phenol conversion. Only in the case of the
experiment carried out with the highest amount of PE did we
notice the formation of BPEC, with 1.2% selectivity.
(
d) The catalytic behaviour observed with the addition of
3
± 1 ppm NaOH is also shown in Fig. 4 (bottom). It can be
seen that the homogeneous catalyst was the most efficient,
with no induction period but still with the autocatalytic effect
shown.
Assuming that the formation of phenate is the rate-
determining step (accelerated in the presence of the basic
catalyst) and because of the stronger basicity of the
deprotonated form of PE compared to the phenate, the gener-
ation of the former species within the catalytic cycle
(
Scheme 5) will lead to the rapid deprotonation of phenol,
which is activated for further transformation into PE, thus
explaining the autocatalytic effect. In other words, as long as
phenol is present in the reaction medium, the deprotonated
form of PE will readily react and generate the phenate spe-
cies, whereas at the end of the reaction, under conditions of
phenol scarcity, it will pick up the proton initially released by
phenol to the Na-mordenite catalyst. The loss of CO from
2
the adduct formed by the nucleophilic attack of phenate on
EC is also a strong driving factor for the generation of the
Fig. 4 Effect of reaction time on PE yield. Top: experiments without
any catalyst, with increasing amounts of PE added from the beginning
(
wt% with respect to phenol): 0% (♦), 4% (■), 17% (▲), and 126% (×).
Yields were calculated after subtraction of the PE amount added.
Bottom: comparison between PE yields without catalyst (♦), with
0
.5 wt% (with respect to phenol) Na-mordenite catalyst (◇) or 0.5 wt%
treated Na-mordenite (□), and with 3 ± 1 ppm NaOH (×). Reaction
conditions: T, 210 °C; phenol/EC ratio, 1/2.
Scheme 5 The autocatalytic effect of PE on the reaction rate.
Catal. Sci. Technol., 2014, 4, 4386–4395 | 4391
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
Catalysis Science & Technology
strong base, the deprotonated form of PE; the latter species,
however, is soon converted into PE because of its basic
character.
viscous, thus producing a solvent effect which may help in
facilitating diffusion in the presence of the zeolite catalyst.
On the other hand, the efficiency of the catalyst deter-
mines the length of the induction period; for example, the
use of high amounts of the catalyst (Fig. 2) resulted in a
considerable shortening of the induction time, but this is
clearly not very sustainable from the green chemistry stand-
point. The data obtained also highlight that diffusion in the
zeolite pores can play a role in the reaction. In order to inves-
tigate the effect of modification of porosity on the catalytic
behavior, we carried out a post-treatment on the Na-mordenite.
Some experimental evidence supports the hypothesis for-
mulated. The by-product BPEC forms only when the conver-
sion of phenol is very high; even though this compound is
formed by a consecutive reaction, one should in any case
expect that a small, but non-negligible, amount of it will start
to form at intermediate values of phenol conversion. This
occurs because under conditions of phenol starvation the
deprotonated form of PE may attack the carbonyl bond of EC
and start the reaction sequence leading to the formation of
+
BPEC (Scheme 3). In this regard, the basic catalyst (either Na
Post-treatment of the Na-mordenite: a comparison of
catalytic performance
(
in ppm) or Na-mordenite) acts as an initiator, rather than a
true catalyst, whereas the main role of increasing the reaction
rate is played by the deprotonated form of PE (Scheme 5).
This suggests the presence of a heterogeneously-initiated
and heterogeneously-terminated reaction (in the presence
of the zeolite), but with the fundamental contribution of
the proton exchange between phenol and PE alkoxide playing
the role of a homogeneous catalysis action. However, the
zeolite may play an important role in improving the reaction
selectivity (see below), which confirms that the reaction
occurs, at least in part, within the confined environment of
the mordenite pores.
With regard to the induction period and to the role played
by the added PE (Fig. 4, top), it must be remembered that we
operated in the absence of any solvent; therefore, the
favoured interaction between the nucleophilic O atoms of PE
and the proton of phenol might lead to the development of a
concerted weakening of the O–H bond in phenol and a facili-
tated interaction with EC (Scheme 6), thus finally leading to
a considerably shortened induction period. Moreover, we
experimentally observed that PE makes the solution less
The silanization procedure on unidimensional and
aluminum-rich zeolites, such as mordenite, showed a pore
mouth narrowing effect more significant than on other
zeolites, such as ZSM-5. In order to reduce the microporosity
through pore mouth blocking, a post-treatment on the indus-
trial mordenite was carried out using chemical liquid deposi-
tion with TEOS as the silanization agent.
3
5
The characteristics of the two samples, the original
Na-mordenite and the treated one, are summarized in
Table 1. The treatment did not lead to any change in the
XRD pattern of the sample. Indications of textural changes
are obtained from the surface area and porosity; the surface
2
−1
area of the parent mordenite was high (452 m g ) and, after
2
−1
post-treatment, decreased to 109 m g . Also, both the total
pore volume and micropore volume showed a dramatic
decrease, thus indicating clogging of the micropores, while the
mesopore volume and area were left substantially unchanged.
The catalytic behaviour of the two zeolites is compared in
Fig. 4 (bottom). It can be seen that the induction time was
considerably shorter for the treated zeolite; however, the over-
all amount of Na accessible was probably decreased in
the treated zeolite because of the lower pore volume acces-
sible to the reactants. The autocatalytic effect cannot be
responsible for the different initial behaviour shown by the
two zeolites because the amount of PE formed during the
first 1–2 h of reaction time was too low to have any effect on
the initial behaviour. This result suggests that the delay in
the start-up of the reaction is affected by the accessibility of
the Na sites to the reactants, an event which does not play an
important role only when a very high loading of catalyst is used.
Scheme 6 The hypothesis of a concerted H-bonding interaction
between phenol, PE, and EC under solventless conditions, facilitating
δ−
the attack of the Ph–O species on EC.
Table 1 Main features of the Na-mordenites used for reactivity experiments
Original Na-mordenite
Treated Na-mordenite
−
−
1
Total pore volume (cc g
Micropore volume (cc g
)
1
0.189
0.151
452
418
34
0.069
0.030
109
77
)
2
−1
Surface area (m g
)
2
−1
Micropore area (m g
)
2
−1
Mesopore area (m g
)
32
Average pore diameter (Å)
17
23
Maximum pore diameter DFT (Å)
5.9
5.9
Na content (wt% Na
2
O)
6.5
6.0
4392 | Catal. Sci. Technol., 2014, 4, 4386–4395
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
Another important effect of catalyst treatment is the differ-
ence in catalyst reusability, shown in Fig. 5, which can be
observed by comparing the conversion of phenol at 210 °C
and after 5 h of reaction time using the untreated and post-
treated zeolites. The catalysts were separated from the reac-
tion medium by filtration and then reloaded again for the
successive reactions after washing with acetone during filtra-
tion and subsequent drying in an oven at 100 °C overnight;
the tests were carried out using both low (0.5 wt% with
respect to phenol) and high (25 wt%) catalyst loadings.
The untreated catalyst already showed a clear deactivation
effect after the first use, whereas the treated catalyst showed
negligible deactivation. The deactivation of the untreated
Na-mordenite was mainly due to the accumulation of organic
residues inside the catalyst pores, as evidenced by the brown-
ish colour of the zeolite; the same effect was by far less rele-
vant in the treated catalyst, in keeping with a considerably
reduced deactivation. In fact, the IR spectrum of the used
The data shown highlight the possible role of Na ions
leached from the Na-mordenite during the reactivity experi-
ments. In order to verify this contribution, we stopped the
reaction with the untreated Na-mordenite at about 10%
phenol conversion – with the reaction carried out under the
usual conditions, i.e. T 210 °C, 0.5 wt% catalyst, phenol/EC
1/2 – filtered out the catalyst, and then allowed the reaction
to proceed in the absence of the catalyst. The PE yield in rela-
tion to time showed a trend very similar to that obtained
when using NaOH (a few ppm) as the catalyst, as shown
in Fig. 4 (bottom). Even though the interpretation of the data
is complicated by the autocatalytic effect shown by PE
(which was present in the filtered solutions, although with a
yield of just 10%), this result demonstrates that the Na leached
from the catalyst most likely also contributed to the catalytic
behaviour seen.
When the same experiment was carried out with the
treated catalyst, the reaction rate after catalyst filtration was
slower than that shown in the presence of both the Na (3 ppm)
catalyst and after filtering out the untreated Na-mordenite;
after 2 h of reaction time, the order of PE yield was as follows:
solution after filtering out the untreated Na-mordenite,
26% > fresh solution with 3 ppm Na, 20% > solution after
filtering out the treated Na-mordenite, 12%. The difference
between the three solutions decreased after 4–5 hours of
reaction time because of the autocatalytic effect due to the
formed PE. The lower leaching of Na shown with the treated
catalyst may again be related to the quicker diffusion of reac-
tants which limits the chemical interaction between the
acidic molecules and the basic sites.
−
1
untreated catalyst showed bands at 1775 and 1800 cm
,
which are attributable to CO moiety-containing organics. A
regeneration treatment of the used catalyst at 400 °C for 3 h
in flowing air did not lead to a complete recovery of the
catalytic activity. The accumulation of organic matter in the
untreated catalyst may once again be attributed to difficulties
in the diffusion of reactants, a phenomenon which might be
conducive to consecutive reactions to heavier compounds.
However, another possible reason for the deactivation is
the low, but non-negligible, leaching of Na. More specifically,
the analysis of the residual Na content in the untreated
zeolite after the first use highlighted a loss of ca. 1.6 ± 0.4%
(
relative amount) of the overall Na content, which would cor-
The better performance of the treated catalyst highlighted
that using Na-mordenite with controlled porosity makes it
possible to overcome problems related to hindered diffusion,
with reduced induction time, less deactivation, and improved
catalyst reusability. As shown in Fig. S1,† the catalytic behav-
iour was also considerably affected by the amount of the cata-
lyst used; however, the optimal amount of the catalyst to be
used depends on the catalyst features. Therefore, we repeated
some experiments by using increasing amounts of the treated
catalyst in order to see whether it was possible to optimise
both the conversion rate and the PE yield. The conditions
used were once again a phenol/EC ratio equal to 1/2 and a
reaction temperature of 210 °C because a higher feed ratio or
lower temperature would lead to a much slower reaction rate.
The results of these experiments are shown in Fig. S2,† which
shows the conversion and yields of products depending on
the reaction time with 0.5 and 5 wt% (with respect to phenol)
treated catalyst. It can be seen that with the latter catalyst,
and under the conditions where phenol is the limiting reac-
tant, 98% conversion of phenol was obtained after 5 h of reac-
tion time with 5 wt% catalyst and 97.5% conversion after 6 h
with 0.5 wt% catalyst. Remarkably, the PE yield was 86% in
the former case (with 11% BPEC and 1% DPE) and 82% in the
latter case (14% BPEC and 1.5% DPE). The use of 25 wt% cata-
lyst led to 100% phenol conversion after 5 h of reaction time
but yield to PE was only 75%, with 6% BPEC and a high yield of
respond to ca. 3 ± 1 ppm Na concentration in the reaction
medium. The leaching of Na was found to be lower in the
case of the treated zeolite after analysis of Na in the solution;
however, we would like to mention that due to the difficulties
encountered in the analysis of such a tiny amount of Na and
to the experimental error dealt with during these measure-
ments (see Experimental section), we cannot state that the
treated catalyst showed no leaching at all.
Fig. 5 Phenol conversion after 5 h of reaction time at 210 °C (phenol/EC
ratio = 1/2) using both untreated and treated Na-mordenite catalysts:
fresh catalysts and after recovery and reuse in the 2nd and 3rd reactions.
Tests were carried out using either 0.5 or 25 wt% catalyst (with respect
to phenol).
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 4386–4395 | 4393

View Article Online
Paper
Catalysis Science & Technology
DPE + TPE (19%). With regard to the conversion of EC, in the
two former cases it was close to 50% (which implies a negligi-
ble transformation of EC into waste by-products), whereas in
the latter case it was 90%, indicating an important contribu-
tion of EC decomposition. Overall, the optimized conditions
with the best catalyst were: T, 210 °C; 5 wt% catalyst; and
phenol/EC molar ratio, 1/2; under these conditions, the
induction period shown was 1 h only.
Concerning the role of microporosity in retarding the
start-up of the reaction and considering that the mesopores
are also present in the untreated Na-mordenite, we can
hypothesize that in the presence of smaller pores a higher
degree of reactant retention, both phenol and EC, leads to a
greater extent of EC transformation into heavy compounds.
The latter partially block both smaller and larger pores, thus
retarding diffusion and counterdiffusion and finally delaying
the initiation of the reaction between phenate and EC. This
hypothesis was also confirmed by the experimental evidence
that the untreated zeolite soon became brown even in the
presence of EC only due to the formation of polymeric com-
3
6
pounds. The same phenomenon was much less relevant in
the case of the treated zeolite.
Homogeneous vs. heterogeneous catalysis
Overall, under our reaction conditions and in the absence of
solvent, the use of a heterogeneous catalyst clearly suffers
from problems such as a much lower TOF (3 ppm NaOH is
enough to catalyse the reaction more efficiently than the 5 wt%
solid catalyst based on Na-mordenite) and diffusional limi-
tations which, however, can be overcome by appropriate
catalyst modification. Overcoming these problems leads to
better activity, shorter induction time (not observed with the
homogeneous catalyst), catalyst reusability, and negligible Na
leaching. On the other hand, Fig. 6 shows that the most sig-
nificant result is the selectivity to PE achieved with the opti-
mal heterogeneous catalyst under conditions which make it
possible to minimize parallel EC decomposition.
The figure shows the selectivity to both PE and PE + BPEC
of the treated catalyst (at both low and high catalyst loadings)
and the homogeneous NaOH catalyst; the better selectivity
achieved with the zeolite is evident. This difference is lower if
the comparison is made with the untreated Na-mordenite
catalyst, probably because of the relevant contribution of the
homogeneous reaction due to Na leached from the catalyst.
This also demonstrates that with the treated catalyst the reac-
tion occurred, at least in part, within the zeolite mesopores,
which made it possible to limit the formation of the bulkier
by-products and, in the end, show a better selectivity to PE at
high phenol conversion.
Fig. 6 Selectivity to PE (top) and to PE + BPEC (bottom) in relation to
phenol conversion at 210 °C (phenol/EC molar ratio, 1/2) and variation
of reaction time. Catalysts: 0.5 wt% (♦) and 5 wt% (■) treated
Na-mordenite with respect to phenol and NaOH, 3 ± 1 ppm (▲).
homogeneous (NaOH) and heterogeneous (Na-mordenite) cata-
lytic conditions without any solvent. We found that outstanding
selectivity to phenoxyethanol and bis(2-phenoxyethyl)carbonate
of over 98% could be achieved at 98% phenol conversion,
but total selectivity was shown under conditions where
phenol conversion was only 60% because ethylene carbonate
was the limiting reactant. Bis(2-phenoxyethyl)carbonate could
easily be separated and transformed with 100% yield of
phenoxyethanol. The Na-mordenite heterogeneous catalyst
proved to be reusable and to cause negligible Na leaching;
however, in order to do that it had to undergo a post-
synthesis treatment aimed at decreasing the micropore
volume.
Indeed, since a few ppm of Na is enough to catalyse the
reaction, a contribution to reactivity derived from the mini-
mal amount of Na released (below the detection limit of the
analytical method used) even with the treated catalyst cannot
be completely ruled out. Despite this, the role of the hetero-
geneous Na-zeolite is evident in finally providing better selec-
tivity than that achieved with the homogeneous NaOH
catalyst. The treated zeolite also showed a much shorter
induction period than the untreated one. An autocatalytic
Conclusions
The synthesis of phenoxyethanol by means of the reaction
between phenol and ethylene carbonate, as an alternative
to the current industrial process carried out by reaction
with ethylene oxide, was investigated in detail under both
4394 | Catal. Sci. Technol., 2014, 4, 4386–4395
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
effect was seen, which was explained by considering the
basicity of the phenoxyethanol itself.
10 G. Gelbard, Ind. Eng. Chem. Res., 2005, 44, 8468–8498.
11 Z. Wang, China Chem. Rep., 2005, 12.
Finally, it is worth noting that even though the proposed
process requires an additional step, i.e., the production of
ethylene carbonate by reaction between ethylene oxide and
carbon dioxide, a similar technology is also used in the new
Omega process for the production of ethylene glycol (MEG)
developed by Shell. MEG is produced by first reacting ethyl-
12 E. G. Harris (Emery Ind.), U.S. Pat. 4404407, 1983.
13 E. G. Harris (National Distillers and Chemical Co.), U.S. Pat.,
4533759, 1985.
14 G. Kretzschmar and P. Weissner, Zellst. Pap., 1971, 20, 70.
15 W. W. Carlson, U.S. Pat., 2448767, 1948.
16 J. C. Smith and W. L. Bressler (The Dow Chem Co.), U.S. Pat.,
3088980, 1963.
2
ene oxide with CO to produce EC, which is then transformed
to MEG. The advantage of the Omega process, compared to
conventional ethylene oxide hydrolysis, is the better final
selectivity to MEG. The same occurs in the process proposed
here: the final selectivity to PE achieved using EC is better
than that achieved using ethylene oxide as the reactant.
17 P. E. Strege (The Dow Chem Co.), U.S. Pat., 4341905, 1980.
18 K. M. Kem (The Dow Chem Co.), U.S. Pat., 4261922, 1981.
19 Y.-M. Sun and C.-S. Wang, J. Polym. Sci., Part A: Polym.
Chem., 1996, 34, 1783.
20 I. Bicu and F. Mustata, J. Appl. Polym. Sci., 2002, 83,
8
02–814.
2
2
1 M. Selva, Pure Appl. Chem., 2007, 79, 1855–1867.
2 P. Tundo and M. Selva, Acc. Chem. Res., 2002, 35, 706–716.
Acknowledgements
The INSTM (Consorzio Interuniversitario per la Scienza e
Tecnologia dei Materiali), Florence, is acknowledged for the
PhD grant to PZ. The MIUR (Ministero dell'Istruzione,
dell'Università e della Ricerca) is acknowledged for financing
this research by means of the PRIN Project “Mechanisms of
23 M. O. Sonnati, S. Amigoni, E. P. Taffin de Givenchy,
T. Darmanin, O. Choulet and F. Guittard, Green Chem.,
2013, 15, 283–306.
24 E. A. Quadrelli, G. Centi, J.-L. Duplan and S. Perathoner,
ChemSusChem, 2011, 4, 1194–1215.
CO
2
activation for the design of new materials for energy and
25 M. Peters, B. Köhler, W. Kuckshinrichs, W. Leitner,
P. Markewitz and T. E. Müller, ChemSusChem, 2011, 4,
resource efficiency” (2010A2FSS9).
1
216–1240.
6 P. P. Pescarmona and M. Taherimehr, Catal. Sci. Technol.,
012, 2, 2169–2187.
2
Notes and references
2
1
S. Budavari, The Merck Index – Encyclopedia of Chemicals,
Drugs and Biologicals, Merck and Co., Inc., Rahway, NJ, 1989.
M. De Groote and B. Keiser (Petrolite Corp.), U.S. Pat.,
27 W. W. Carlson and L. H. Cretcher, J. Am. Chem. Soc., 1947,
69, 1952–1956.
28 T. Yoshino, S. Inaba and Y. Ishido, Bull. Chem. Soc. Jpn.,
1973, 46, 553–556.
2
2
524889, 1950.
3
4
A. E. Jeltsch (Koppers Co.), U.S. Pat., 2852566, 1958.
M. Roussos and J. Lecomte (Societe de Produits Chimiques
et de Synthese), Fr. Pat., 1366067, 1964.
29 J. P. Parrish, R. N. Salvatore and K. W. Jung, Tetrahedron,
2000, 56, 8207–8237.
30 A. K. Kinage, S. P. Gupte, R. K. Chaturvedi and R. V. Chaudhari,
Catal. Commun., 2008, 9, 1649–1655.
5
M. Keisuke, K. Mikio, Y. Yoshikata, M. Kazuto and N. Hiroaki
(
1
Polyester and Ether Development Co.), Jpn. Pat., 69 21090,
969.
31 G. Busca, Ind. Eng. Chem. Res., 2009, 48, 6486–6511.
32 G. Busca, Chem. Rev., 2010, 110, 2217–2249.
33 H. Hattori, Chem. Rev., 1995, 95, 537–558.
34 A. Corma and S. Iborra, Adv. Catal., 2006, 49, 239–302.
35 R. W. Weber, K. P. Moller and C. T. O'Connor, Microporous
Mesoporous Mater., 2000, 35, 533–543.
6
7
8
K. G. Reabe and H. Dressler (Koppers Co.), U.S. Pat., 3644534, 1972.
H. Schulze (Jefferson Chem Co.), U.S. Pat. 3642911, 1972.
R. S. Boeva, K. M. Markov and S. V. Kotov, React. Kinet.
Catal. Lett., 1980, 13, 323–329.
9
R. S. Boeva, K. M. Markov and S. V. Kotov, J. Catal., 1980, 62,
36 A. G. Patil, U. R. Kapadi and D. G. Hundiwale, 364–366,
J. Sci. Ind. Res., 2005, 64.
2
31–234.
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 4386–4395 | 4395