An innovative synthesis pathway to benzodioxanes: The peculiar reactivity of glycerol carbonate and catechol

Article abstract of DOI:10.1039/c8gc02811g

A peculiar reactivity of glycerol carbonate (GlyC) as an innovative and highly reactive alkylating agent for phenolic compounds is investigated in this article. In particular, 2-hydroxymethyl-1,4-benzodioxane (HMB), a key intermediate for the pharmaceutical industry, has been selectively synthesized by the reaction of a slight excess of GlyC with catechol in the presence of a basic catalyst (NaOCH₃, Na-mordenite, MgO), without requiring a reaction solvent. Both reagents have been quantitatively converted in just one hour at 170 °C with a HMB yield, up to 88%, in the presence of a homogeneous basic catalyst (NaOCH₃). Notably, the main side product, the HMB isomer, may be an interesting intermediate for the synthesis of calone analogues, which are important scaffolds used in fragrances. A detailed mechanistic study, supported by kinetics, GC-MS, and HMBC NMR characterization, is also reported. Accordingly, this paper describes a completely innovative and greener synthesis pathway to benzodioxanes.

Full text of DOI:10.1039/c8gc02811g

Green Chemistry
View Article Online
View Journal
PAPER
An innovative synthesis pathway to
benzodioxanes: the peculiar reactivity of
Cite this: DOI: 10.1039/c8gc02811g
glycerol carbonate and catechol†
a
T. Tabanelli, *a C. Giliberti,a R. Mazzoni, a R. Cucciniello b and F. Cavani
A peculiar reactivity of glycerol carbonate (GlyC) as an innovative and highly reactive alkylating agent for
phenolic compounds is investigated in this article. In particular, 2-hydroxymethyl-1,4-benzodioxane
(HMB), a key intermediate for the pharmaceutical industry, has been selectively synthesized by the reac-
tion of a slight excess of GlyC with catechol in the presence of a basic catalyst (NaOCH3, Na-mordenite,
MgO), without requiring a reaction solvent. Both reagents have been quantitatively converted in just one
hour at 170 °C with a HMB yield, up to 88%, in the presence of a homogeneous basic catalyst (NaOCH3).
Notably, the main side product, the HMB isomer, may be an interesting intermediate for the synthesis of
calone analogues, which are important scaffolds used in fragrances. A detailed mechanistic study, sup-
ported by kinetics, GC-MS, and HMBC NMR characterization, is also reported. Accordingly, this paper
describes a completely innovative and greener synthesis pathway to benzodioxanes.
Received 4th September 2018,
Accepted 11th December 2018
DOI: 10.1039/c8gc02811g
rsc.li/greenchem
nucleophilic substrates, such as phenols, amines, sulphones,
thiols, and methylene-active derivatives of aryl and aryloxy-
Introduction
Organic carbonates (OCs) are among the most promising acetic acids, under both batch and continuous flow
green candidates for the replacement of conventional noxious conditions.7–13 Scheme S1 in the ESI† comprehensively sum-
aprotic polar solvents because of their high boiling point and marizes the reactivity of DMC and aromatic substrates.
great solvency and also due to their non-toxicity and good
biodegradability.
On the other hand, cyclic carbonates can also be success-
fully used as alkylating agents. Of particular interest is the
For the same reasons, they are widely used in the manufac- reaction between ethylene carbonate (EC) and phenol with
ture of lithium batteries, as fuel additives, and as innovative heterogeneous catalytic systems (such as basic zeolites, e.g. Na-
reactive chemicals for the development of important intermedi- mordenites). This pathway makes possible a more selective
ates in the pharmaceutical, lubricant, and polymer industries.1
synthesis of 2-phenoxyethanol than does industrial production
In particular, OCs are successfully used as alkylating agents from ethylene oxide feedstock (Scheme 1a).14
for aromatic compounds, replacing more traditional, harmful,
Among the cyclic OCs, glycerol carbonate (4-hydroxymethyl-
and undesirable compounds. For example, alkyl iodides are 2-oxo-1,3-dioxolane, GlyC) is
a polyfunctional carbonate,
known carcinogens, while dimethyl sulphate is an extremely bearing both a carbonate moiety and a hydroxyl group repre-
hazardous liquid and vapour (which may be fatal if senting electrophilic and nucleophilic sites, respectively (Fig. 1).
inhaled).2–4
However, only the lightest members of the OC family,
namely dimethyl carbonate (DMC) and diethyl carbonate
(DEC),5,6 have been investigated for these kinds of reactions.
In more detail, DMC is capable of reacting with a number of
a“Toso Montanari” Industrial Chemistry Department, University of Bologna, Viale del
Risorgimento 4, 40136 Bologna, Italy. E-mail: tommaso.tabanelli@unibo.it
b“Adolfo Zambelli” Department of Chemistry and Biology, University of Salerno, Via
Giovanni Paolo II 132, 84084 Fisciano, SA, Italy
†Electronic supplementary information (ESI) available: Reaction schemes, sup-
plementary catalytic results, 1H- and 13C-NMR characterization of substrates,
available pure products and reaction mixtures, and GC-MS of pure compounds
and reaction mixtures. See DOI: 10.1039/c8gc02811g
Scheme 1 Alkylation of aromatic substrates with cyclic carbonates: (a)
synthesis of 2-phenoxyethanol through phenol alkylation with EC; (b)
aniline alkylation with GlyC with the formation of N-(2,3-dihydroxy)
propyl aniline as the main product.
This journal is © The Royal Society of Chemistry 2018
Green Chem.

View Article Online
Paper
Green Chemistry
Scheme 1 shows a concise overview of the cited alkylation
reactions between aromatic and cyclic OCs.
Another functionalized phenolic derivative of great impor-
tance is 2-hydroxymethyl-1,4-benzodioxane (Fig. 2), hencefor-
ward referred to as “HMB”. HMB is an important intermediate
for the pharmaceutical industry, because its moiety is present
in many active components of antidepressant, antihypertensive,
anxiolytic, and antithrombotic drugs, in addition to cardio-
vascular agents, to name a few (see Scheme S2 in the ESI†).20
The main problems related to the HMB production concern
its synthesis, which usually requires a multistep sequence of reac-
tions and the use of toxic solvents and reagents. The traditional
synthesis pathways may be divided into three main classes:
(a) glycidol derivative-based synthesis;
Fig. 1 GlyC structure and polyfunctional reactive sites.
The combination of its bio-based origin and peculiar charac-
teristics and reactivity makes GlyC a versatile and renewable
building block for a more sustainable chemistry. Because of its
high boiling point (110°–115 °C at 0.1 mmHg), high flash point
(190 °C), and low volatility (8 mbar vapour pressure at 177 °C),
GlyC and its esters are bio-based alternative solvents with a lower
vapour pressure than volatile organic compounds (VOC).15
Moreover, GlyC has been investigated for use as a component in
gas separation membranes, polyurethane foams and surfactants,
and coatings, and as a source of new hyperbranched polymers.16
Lastly, GlyC has the possibility to become a major chemical
intermediate: for instance, it can be converted into either glyci-
dol (by decarboxylation) or epichlorohydrin, a product widely
applied in industry, under mild conditions.17,18
Nevertheless, only a few studies in the literature have inves-
tigated the possibility of using GlyC as an alkylating agent for
the derivatization of aromatic compounds. For example, Selva
et al. reported on the alkylation of aromatic amines (aniline)
with GlyC, using a series of different faujasites (e.g. NaY) as
the catalyst. The reaction (performed in a 90°–160 °C tempera-
ture range) leads to a mixture of compounds in which N-(2,3-
dihydroxy)-propyl aniline is the main product (Scheme 1b).19
(b) “chiral pool” derivative-based synthesis;
(c) acrylate-based synthesis.
In the first class, glycidol must be tosylated to prevent unde-
sired side reactions. Tosylation is usually performed with
4-toluenesulfonyl chloride in the presence of a base (triethyl-
amine or/and 4-dimethylamino pyridine) at a low temperature
(0 °C) in dichloromethane (DCM) or dimethylformamide
(DMF) as a solvent.21 Subsequently, the isolated tosylated glyci-
dol is exploited as an alkylating agent for catechol (as well as
nitro-substituted or halogenated catechols), for the formation
of HMB (or substituted HMB).
This reaction is performed in DMF at 60 °C for a certain
length of time (e.g. 27 hours) in the presence of a stoichio-
metric amount of a base (K2CO3 or NaOH), leading to a vari-
able 70–75% HMB yield with the formation of a heavier by-
product, derived from the reaction of HMB with another
moiety of glycidol (Scheme 2a).22
The second synthesis strategy involves the use of chiral
building blocks from the “chiral pool”, such as D-mannitol, and
a multistep sequence of reactions, in order to obtain the haloge-
nated compound ((2,3-dibromopropoxy)methyl) benzene.
This compound subsequently undergoes a reaction with
catechol to yield the HMB intermediate (2-benzyloxymethyl-
Fig. 2 2-Hydroxymethyl-1,4-benzodioxane (HMB) structure.
Scheme 2 Traditional synthetic pathways to HMB.
Green Chem.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Green Chemistry
Paper
1,4-benzodioxane). In the final step, this intermediate is hydro-
The leaching of Mg into the reaction mixture at 200 °C was
genated in the presence of a noble metal catalyst (Pd/C), thus evaluated using microwave plasma atomic emission spec-
leading to the formation of HMB with an overall process yield troscopy (Agilent Technologies MP-EAS 4210) calibrated for the
of 42% (Scheme 2b).23
determination of Mg (in the range 0, 1–100 ppm). In particu-
Lastly, HMB analogues (HMB methyl esters) can be lar, after the reaction, the mixture was cooled at room tempera-
obtained via the bromination of methyl acrylate or other analo- ture and the organic compounds were dissolved using 10 mL
gous unsaturated esters (usually in CCl4 as the solvent), to of acetone. The mixture was filtered using a syringe equipped
obtain a halogenated compound which is able to react with with a PTFE filter (0, 20 μm) in order to remove all the hetero-
catechol in the presence of an excess of a base (e.g. K2CO3) for geneous catalyst. The mixture was diluted 10 times in water
18 hours in refluxing dry acetone (T = approx. 56 °C). This (no precipitation occurs during the dilution). To all the solu-
route leads to the formation of methyl 1,4-benzodioxan-2-car- tions (standards and samples) 100 μL of a solution of Cs2CO3
boxylate with an isolated yield of 74% (Scheme 2c).24
(100 g L−1) were added as the ionization buffer in order to
All these synthesis pathways suffer from similar drawbacks: avoid ionisation interference for Mg. The results are shown in
(i) the use of toxic and dangerous reagents (glycidol or haloge- Table S2† demonstrating that the leaching of Mg was
nated analogues), in order to enhance the condensation step negligible.
with catechol; (ii) the use of a stoichiometric amount of base
Sodium mordenite (Na-Mord). The zeolite used for hetero-
with the co-production of an equimolar amount of inorganic geneous catalytic tests was Zeocat FM-8, provided by Zeochem,
waste; (iii) the use of toxic solvents (CCl4, DMSO, DMF, DCM which was characterized by a SiO2/Al2O3 molar ratio (SAR) of 12,
and anhydrous THF); and (iv) the multistep approach which, a surface area of 350 m2 g−1 (pore size 6.5 × 7.0/2.6 × 5.7 Å), and
in the end, is detrimental to the maximum yield obtainable.
a sodium content expressed as Na2O wt% equal to 6.8.25 The
In our work, we sought an innovative synthesis procedure, zeolite was dried at 120 °C overnight before each test.
which could avoid the solvent while making use of the base-
Reaction procedure and product analysis
catalysed alkylation of catechol with GlyC, followed by intra-
molecular cyclization, thus enabling us to obtain HMB in a Catalytic tests were conducted in a round-bottom Pyrex glass
single step. In this way, the use of GlyC as an innovative, reac- flask equipped with a magnetic stirrer and fitted with a
tive, and selective alkylating agent for catechol is a completely condenser.
pioneering synthetic pathway to HMB, which is capable of
overcoming all the limitations previously reported.
In a typical experiment, the flask was charged with the aro-
matic alcohol (usually catechol from 1 to 3 mmol), catalyst
(NaOCH3 from 0.066 to 0.2 mmol or MgO/Na-Mord 5 wt% with
respect to the limiting reagent), and GlyC (from 2 to 6 mmol),
purged with a nitrogen atmosphere, and then closed. The
mixture was heated rapidly to the desired temperature (e.g.
170 °C) and stirred for 1 h, and then cooled to RT. Afterwards,
the mixture was recovered with acetone (10 mL, HPLC grade,
Experimental
Catalyst preparation and characterization
Sodium methoxide. Sodium methoxide was prepared by Sigma-Aldrich). The solution was further diluted with acetone
reacting Na(s) with an ice-cold solution of MeOH. The solvent (ten volumes) and added together with n-octane (20 µL), which
was removed under vacuum, while the resulting solid was served as an external standard. If needed (when a hetero-
stored under nitrogen before use.
geneous catalyst was used), the resulting mixture was filtered
Magnesium oxide (MgO). Magnesium oxide was synthesized and then analysed with a Thermo Focus gas-chromatograph
by a precipitation technique. A solution of Mg(NO3)2·6H2O in equipped with a HP-5 capillary column (25 m × 320 μm ×
distilled water (1 mol L−1) was added dropwise to a second aq. 1.05 μm; Tinjector: 280 °C; split ratio: 30 : 1, nitrogen flow:
solution of Na2CO3·10H2O (1 mol L−1 in distilled water) which 1.2 mL min−1). The temperature ramp was 50 °C for 2 min,
was kept, under magnetic stirring, at 60 °C. The pH was main- with subsequent heating up to 280 °C at 10 °C min−1, 280 °C
tained between 9.8 and 10.2 during the reaction. The obtained for 5 min. Each compound was calibrated with respect to
magnesium hydroxide was filtered, washed with distilled n-octane (the standard), to obtain the corresponding response
water, and dried overnight at 110 °C. The material was then factor in the appropriate range of concentration. The structure
crushed and calcined at 450 °C for 5 hours.
of the products was determined by ESI-MS and GC-MS, and
MgO powder was characterized by X-ray diffraction, using whenever possible, by comparison with pure commercial
Ni-filtered Cu Kα radiation (λ = 1.54178 Å) on a Philips X’Pert samples.
vertical diffractometer equipped with a pulse height analyser
and a secondary curved graphite-crystal monochromator.
Most of the catalytic tests were triplicated to verify their
reproducibility. In the repeated runs carried out under the
The BET surface area of MgO was determined by N2 absorp- same conditions, the values of conversion and product yields
tion–desorption at the temperature of liquid N2 using a Sorpty (determined by GC/MS) differed by less than 5% from one
1750 Fison instrument. A sample of 0.2 g was first outgassed reaction to another.
at 150 °C before N2 absorption. The surface area of MgO was
found to be equal to 170 5 m2 g−1
The spectroscopic properties of the products corresponded
to those reported in the literature.
.
This journal is © The Royal Society of Chemistry 2018
Green Chem.

View Article Online
Paper
Green Chemistry
Product separation
The reaction mixture, obtained after one hour of reaction at
170 °C with a ratio of catechol : GlyC : NaOCH3 = 1 : 2 : 0.066,
was rapidly cooled down to RT with an ice bath and then solu-
bilized and transferred into a separating funnel using 10 mL
of ethyl acetate. In this way, the heavier oligomeric hydrophilic
compounds remained in the reaction batch.
Afterwards, the products dissolved in the organic phase
were washed twice with 5 mL of slightly acidified water (e.g.
solution of diluted HCl, pH = 4): the residual unreacted cate-
chol and glycerol could thus be selectively extracted from the
aqueous solution. The organic solution was then dried with
sodium sulphate, filtered, and lastly the solvent (ethyl acetate)
was removed under vacuum, obtaining a solid in an overall iso-
lated yield of 85% (77% HMB and 8% isomer). HMB was
further purified by column chromatography on silica gel (di-
chloromethane/diethyl ether 100 : 1) with an isolated yield of
68% (see the 1H-NMR spectrum in Fig. S22†).
Scheme 3 The “one-pot” synthetic approach.
(glycerol and GlyC themselves act as solvents, and HMB has a
mp of 90 °C). The atom economy of the overall process is very
high (AEStep1 = 100%; AEStep2 = 72.8%) and the only co-pro-
ducts of the reaction (Step 2) are water and carbon dioxide.
Since the first step of this procedure was investigated in our
previous work,26,27 which enabled us to obtain a mixture
characterized by a molar ratio of catechol : GlyC equal to 1,
after 1 hour of reaction at 60 °C, in the presence of a catalytic
amount of NaOCH3 (1/15 mol with respect to CC, approx.
6.6% mol), we will focus here on the results of the second step
of the process.
NMR experiments
In a typical experiment, glycerol carbonate (40 mg) and phenol
(15 mg) or catechol (15 mg) were dissolved in 0.6 mL DMSO-d6
in a J Young valve NMR tube. Reactions were performed at the
desired temperature for 1–3.5 h without stirring. The reaction
was monitored by 1H-, 13C- and HMBC NMR using a Varian
Inova 600 (1H, 599.7; 13C, 150.8 MHz) instrument.
Screening of the best reaction temperatures and the role of
water
We first investigated the effect of temperature, in the 110°–
170 °C range, by performing Step 2 for 4 hours. The results
obtained are shown in Fig. 3.
Results and discussion
Background of this work and the “one-pot” synthesis approach
These results clearly show the importance of temperature in
speeding up the reaction rate. HMB was the main reaction
product, obtained with a yield similar to that of glycerol. This
unexpected by-product was formed by the hydrolysis of GlyC,
thanks to the water coproduced during the main reaction. The
hydrolysis also contributed to increasing GlyC conversion,
which was higher than that expected based on HMB for-
mation. On the other hand, the only other by-product of the
reaction was the HMB isomer, 3,4-dihydro-2H-benzo[b][1,4]
In a recent patent, we outlined the outstanding performance
of catechol carbonate (CC) as a novel carbonate source in the
carbonate interchange reaction (CIR) with both aliphatic alco-
hols and polyols such as glycerol.26
In particular, CC is a highly reactive compound. It is able to
react with glycerol in the presence of a basic catalyst, under
mild conditions (40° to 60 °C, atmospheric pressure, 1 hour
reaction time), leading to the selective formation of a mixture
of GlyC, with up to 96% yield, and catechol (the leaving group
of CC and the co-product of the CIR), with a molar ratio
between the two compounds equal to 1.
The separation of these two compounds, however, is not a
simple matter; the purification of the product was possible
only by means of flash chromatography. Thus GlyC was iso-
lated in a 77% yield with a purity grade >99%.27
Therefore, we decided to investigate the reactivity of GlyC
and catechol in the presence of a sodium methoxide catalyst,
using the “one-pot” synthesis strategy shown in Scheme 3:
Step 1, allow CC and glycerol to react, in order to form GlyC and
catechol in situ (Step 1 in Scheme 3), and then Step 2, change
the reaction conditions (higher temperature than in Step 1),
thus avoiding the isolation of intermediates GlyC and catechol.
It is important to note that this approach made it possible
for us to avoid the use of toxic and halogenated compounds.
Fig. 3 Catalytic results obtained with a catechol : GlyC : NaOCH3 molar
ratio of 1 : 1 : 0.066, N2 atmosphere, 4 hours at the desired temperature.
Product yields refer to catechol. C-loss is comprehensive of the CO2
released (specified in the caption). Catechol conversion (blue bar), GlyC
conversion (red bar), HMB yield (green), HMB isomer (purple), glycerol
The two reactions were performed without any reaction solvent (light blue), and atomic C-loss (grey).
Green Chem.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Green Chemistry
Paper
Scheme 4 (i) Initially proposed reaction pathway for the alkylation and condensation reactions of catechol and GlyC; (ii) GlyC hydrolysis to yield
glycerol.
dioxepin-3-ol. This compound, however, is a valuable product
because it is a reactive intermediate in the synthesis of methyl- and 5.
benzodioxepinone-based compounds (trade name “calone” or
The results of the catalytic experiments are shown in Fig. 4
Surprisingly, GlyC conversion was completed after 1 h reac-
“calone 1951”): molecules that are known in the fragrance tion at 170 °C (irrespective of its molar excess), while the reac-
industry for their distinct marine scent.28
tion required only a small excess of GlyC to enhance the HMB
The suggested reaction mechanism is shown in Scheme 4.
One of the catechol acidic hydroxyl groups is deprotonated by
the basic catalyst; the catecholate anion (a “softer” nucleophile
than aliphatic analogues due to the resonance of the negative
charge on the aromatic ring) is able to perform the nucleophi-
lic attack on the more available “soft” electrophilic site of
GlyC, leading to the formation of an unstable intermediate,
which rapidly undergoes decarboxylation followed by an intra-
molecular condensation which, in the end, leads either to the
desired HMB (Scheme 4, pathway i, a) or to its isomer
(pathway i, b), and water as the co-product. The latter is
responsible for GlyC hydrolysis back to glycerol and carbon
dioxide (Scheme 4, pathway ii).
These reactions are characterized by a non-negligible
carbon loss (Fig. 3, grey columns), which is mainly due to sub-
strate decarboxylation and, to a minor extent, to some heavier
hydrophilic by-products, such as polyol oligomers, which are
easily solubilized in water.
Fig. 4 Catalytic results obtained with a catechol : GlyC : NaOCH3 molar
ratio of 1 : 2 : X, N2 atmosphere, 1 hour at 170 °C. Product yields refer to
catechol. C-loss is comprehensive of the CO2 released (specified in the
caption). Catechol conversion (blue bar), GlyC conversion (red bar),
HMB yield (green), HMB isomer (purple), glycerol (light blue), and
atomic C-loss (grey).
The mechanisms plausibly responsible for the formation of
these by-products are summarized in Scheme S3 in the ESI.†
Surprisingly, none of the reaction intermediates suggested
in Scheme 4 has ever been observed, not even in reactions per-
formed at relatively low temperatures (110° and 140 °C) and
for short reaction times (for example, see the kinetic study at
140 °C, Fig. S1 in the ESI†). This suggests that decarboxylation
and intramolecular condensation are very fast reactions.
The reaction mechanism has been thoroughly investigated
and described in a dedicated section (see below).
Synthesis optimization
Because of the formation of an equimolar amount of water,
which is responsible for GlyC hydrolysis, a small excess of car-
bonate was necessary in order to increase the HMB yield.
Moreover, 170 °C was chosen as the best reaction temperature,
while the reaction time was decreased from 4 h to 1 h, in order
to better highlight the effect of both the reagents’ molar ratio
and the catalyst loading on HMB formation.
Fig. 5 Catalytic results obtained with a catechol : GlyC : NaOCH3 molar
ratio of 1 : X : 0.066, N2 atmosphere, 1 hour at 170 °C. Product yields
refer to catechol. C-loss is comprehensive of the CO2 released
(specified in the caption). Catechol conversion (blue bar), GlyC conver-
sion (red bar), HMB yield (green), HMB isomer (purple), glycerol (light
blue), and atomic C-loss (grey).
This journal is © The Royal Society of Chemistry 2018
Green Chem.

View Article Online
Paper
Green Chemistry
yield, as shown in Fig. 4. The effect of the catechol : GlyC
molar ratio can be summarized as follows:
At first, the effect of the reaction temperature was studied.
The results, summarized in Fig. S2,† underscore the need to
(i) A slight excess of GlyC made it possible for us to obtain work at a higher temperature, e.g. at 170 °C instead of 140 °C,
a HMB yield higher than 80% in a one-hour reaction at 170 °C; and for a longer reaction time than those used in the homo-
(ii) with a GlyC : catechol ratio of 1.5, however, the theoretical geneous system. In particular, using a Na-Mord : catechol wt
HMB maximum yield, in consideration of the concomitant ratio of 5% : 95%, a HMB yield higher than 80% was achieved
stoichiometric GlyC hydrolysis, should be equal to 75%, by performing the reaction at 200 °C for 6 hours. The conver-
suggesting an accelerating effect of GlyC also on the main reac- sion and yield profiles based on the reaction time at 200 °C
tion, compared to the parasitic ones: the hydrolysis and the are shown in Fig. 6.
formation of heavier hydrophilic compounds. This trend was
Interestingly, when performing the reaction at a lower
also confirmed in the experiment with a GlyC : catechol molar temperature (140 °C and 170 °C, Fig. S2†) or at 200 °C for
ratio of 2, which showed that a greater GlyC excess led to a shorter reaction times (1 and 2 hours), catechol conversion
further increase in the HMB yield (approx. 88% after 1 h) but was higher than GlyC conversion. This may seem meaningless
also to a lower selectivity to heavier by-products (with a lower considering the results obtained with a homogeneous catalyst;
C loss), in favour of the formation of glycerol.
however, it was due to the strong interaction of catechol with
On the other hand, some experiments showed that the use the basic sites of the catalyst (almost irreversible at low temp-
of a large excess of GlyC made it possible to obtain a complete erature), which, in the end, is responsible for the increase of
catechol conversion that, however, turned out to be detrimen- both the catechol loss and overall C-loss.
tal to the carbon balance, with a C loss higher than 50%. This
A comparison of the catalytic performances of Na-morde-
was probably due to the higher acidity of catechol hydroxyl nite and MgO at 200 °C is summarized in Table S1.†
groups compared to the GlyC and glycerol aliphatic hydroxyl
moieties; in this way, the presence of catechol in the batch pre-
Mechanism investigation
vents the basic catalyst deprotonating aliphatic –OH groups,
thus limiting the parasite reactions.
As already mentioned, no reaction intermediates were detected
under any of the conditions used. This was probably due to
the fast intramolecular reactions involved. Therefore, the reac-
tion mechanism was investigated by means of both NMR
(vide infra) and a simple, yet effective, strategy: the simplifica-
tion of functional groups.
We thus decided to investigate the influence of each func-
tional group on the reaction pathway. In particular, the follow-
ing tests were performed:
Further evidence of the importance of adjusting reaction
conditions in order to achieve a high, albeit incomplete, cate-
chol conversion, limited carbon loss, and high HMB yield was
obtained by conducting a GlyC decomposition test. In this
test, GlyC was charged with the basic catalyst (with a molar
ratio NaOCH3 : GlyC = 0.066) and heated to 170 °C.
After 10 minutes, dark, heavy, polymer-like compounds
were formed, while the complete GlyC conversion was
observed after 30 minutes, without the formation of either gly-
cerol or glycidol, and with a C-loss of 100%.
- the reaction of catechol (or other aromatics) with glycerol,
by removing the carbonate functional group from the aliphatic
counterpart (green, Fig. 7);
The best reaction conditions (catechol : GlyC : NaOCH3
=
- the reaction of phenol with GlyC, by removing the second
hydroxyl group of the catechol moiety (red, Fig. 7), thus
1 : 2 : 0.066, 170 °C, 1 h) were then chosen to further investi-
gate the effect of the basic catalyst loading; the results are
shown in Fig. 5. It is worth noting that without the catalyst, no
HMB was formed, while part of the GlyC was thermally decom-
posed into heavier by-products. An increase in the amount of
Na methoxide led to an increase in both catechol conversion
and HMB yield, while GlyC conversion became more similar to
catechol conversion, i.e. with less GlyC decomposition.
Interestingly, in all the tests conducted so far, the
HMB : isomer molar ratio was close to 10, regardless of para-
meters such as temperature.
Performance of heterogeneous catalysts
An initial screening of heterogeneous basic catalysts was per-
formed; this means the recovery of the used catalyst will be
Fig. 6 Catalytic results obtained at 200 °C with a catechol : GlyC molar
easier, thus facilitating both product purification and catalyst re- ratio of 1 : 2, Na-mordenite 5% w/w with respect to catechol, N2 atmo-
sphere. Product yields refer to catechol. Catechol conversion (blue line),
GlyC conversion (red line), HMB yield (green), HMB isomer (purple), gly-
cerol (light blue), and atomic C-loss (grey). The dotted line represents
the part of C-loss related to the CO2 released. The results showed were
cycling. In particular, we studied the catalytic activity of sodium
mordenite (Na-Mord) characterized by a silica-to-alumina ratio
(SAR) of 12, and of a synthesized “high area” magnesium oxide
(MgO); in fact, both catalysts proved to be active in other reac-
tions between carbonates and alcohols or phenolics.14,25
obtained by feeding new reagents for each reaction time, to collect the
mixture after the reaction and correctly calculate the C-loss.
Green Chem.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Green Chemistry
Paper
However, two other products were spotted. The first was 4-
(phenoxymethyl)-1,3-dioxolan-2-one, a phenolic alkylated com-
pound, which maintains the organic carbonate functionality
and behaves as
a reaction intermediate. Its formation
appeared to increase during the first hours of reaction, and
then decreased in favour of consecutive products such as 1,3-
diphenoxypropan-2-ol. A reaction mechanism is suggested in
Scheme 5.
Fig. 7 Catechol and GlyC functional groups selected for the mecha-
nism investigation.
The results obtained with phenol underscored the possi-
bility of the intramolecular rearrangement of the carbonate
group over the free –OH group of the GlyC moiety: this behav-
iour may play a key role in HMB formation. To investigate this
possibility, propylene carbonate (PC) was made to react with
catechol with the aim of forming a HMB analogue: 2-methyl-
1,4-benzodioxane (MB). The results obtained at 170° and
200 °C are shown in Fig. 9.
Interestingly, under the usual conditions required for HMB
formation with the homogeneous basic catalyst, no formation
of the target product was observed, while the main product
was the monoalkylated compound: 2-(2-hydroxypropoxy)
phenol and its isomer 2-((1-hydroxypropan-2-yl)oxy)phenol.
making the rapid intra-molecular cyclization to HMB
impossible;
- the reaction of catechol with propylene carbonate (PC), by
removing the polyfunctionality provided by the residual
hydroxyl group in GlyC (blue, Fig. 7).
- the reaction of catechol and glycidol was investigated by
using DMSO-d6 as the reaction solvent and analysing the reac-
tion mixture by means of NMR, in order to exclude the effect
of GlyC decarboxylation (green, Fig. 7).
First, the reaction of glycerol with various aromatic com-
pounds (aniline, phenol, and catechol) was performed at
170 °C in the presence of NaOCH3, to exclude the conden-
sation between the aromatic and the aliphatic –OH groups.
Interestingly, in no case were traces of condensation pro-
ducts found, demonstrating the peculiar reactivity of the
organic carbonate involved. Indeed, the reaction mechanism is
likely to start from the nucleophilic attack of the catecholate
anion on the alkylene carbon in the carbonate ring.
Furthermore, tests using the protected form of glycerol ((2,2-
dimethyl-1,3-dioxolan-4-yl)methanol, “Solketal”), in order to
somehow reproduce the ring strain in GlyC, only led to the for-
mation of the protected form of catechol (2,2-dimethylbenzo
[d][1,3]dioxole).
Afterwards, we studied the reaction between phenol and
GlyC, in order to block the intra-molecular cyclization leading
to HMB. The results are shown in Fig. 8.
Scheme 5 Proposed reaction mechanism for the phenol alkylation
with GlyC.
As expected, the main product was the O-alkylated com-
pound: 3-phenoxy-1,2-propandiol.
Fig. 9 Catalytic results obtained with a catechol : PC : NaOCH3 molar
ratio of 1 : 2 : 0.066, N2 atmosphere, 1 hour at different reaction temp-
Fig. 8 Catalytic
results
obtained
at
140
°C
with
a
phenol : GlyC : NaOCH3 molar ratio of 2 : 1 : 0.066, N2 atmosphere. eratures. Product yields refer to catechol. C-loss is comprehensive of
Product yields refer to GlyC. Phenol conversion (blue), GlyC conversion
(red), 3-phenoxy-1,2-propandiol (green), 4-(phenoxymethyl)-1,3-dioxo-
lan-2-one (purple), glycerol (light blue) and di-alkylated compounds
(pink).
the CO2 released (specified in the caption). Catechol conversion (blue
bar), PC conversion (red bar), MB yield (green), monoalkylated (purple),
1,2-propandiol (light blue), di-alkylated (orange), and atomic C-loss
(grey).
This journal is © The Royal Society of Chemistry 2018
Green Chem.

View Article Online
Paper
Green Chemistry
When the reaction temperature was increased to 200 °C, NMR experiments
however, MB formed as a minor product (approx. 5% yield),
while the increasing temperature led to a greatly enhanced
yield of dialkylated compounds.
With the aim of confirming the proposed reaction intermedi-
ates, diluted versions of the reactions between GlyC and
phenol and GlyC and catechol were performed in DMSO-d6 fol-
lowed by NMR. All identified species are summarized in
Fig. 10.
The identification of some signals ascribable to phenol, gly-
cerol carbonate, 1, 3,30 4,31 6, and 7 has been attributed either
by comparison with control experiments (Table S3,†
Fig. 11–14, Fig. S5–S21†) on the pure products (except for the
mixture of 6 + trace of 7) or with literature data.
Considering the reaction of GlyC with phenol, performed at
140 °C for 1 h, once excluded signals arising from 1, 3 and 4,
the remaining peaks and correlations (yellow lines in Fig. 11)
are in line with the structure of 2 (see also Fig. S10–S15†).
The correlations between –CH2 on the carbonate cycle with
the CvO of carbonate (approx. 155 ppm) and between the
–CH2 in the lateral chain and an ipso-C in the aromatic region
(approx. 159 ppm), which are shown in Fig. 12, further
confirm the structure of 2, demonstrating the link between
phenol and glycerol carbonate fragments. It is worth noting
that product 2 has also been detected by means of GC-MS in a
solventless reaction mixture (see the ESI†).
This behaviour proves that the intramolecular cyclization
with the formation of the benzodioxane ring is not facilitated
when PC is used instead of GlyC, clearly underscoring the
importance of the “free” –OH group of the GlyC moiety in the
promotion of a reaction pathway other than the expected con-
densation with the consequent elimination of water. The reac-
tion mechanism using PC is shown in Scheme S4 of the ESI.†
Lastly, the reaction of catechol and glycidol was investi-
gated. Indeed, glycidol is a highly reactive compound that
might be produced by the decarboxylation of GlyC at high
temperature (see Scheme S5 in the ESI†). This reaction was
performed at 170 °C in DMSO-d6 in an NMR tube by diluting
reagents and the homogeneous basic catalyst, in order to limit
the undesirable glycidol oligomerization. The mixture was ana-
1
lysed by H- and 13C-NMR (Fig. S3 and S4†) and no traces of
HMB (cf. with Table S3 and Fig. S8 and S9†) were observed,
clearly underpinning the importance of carbonate scaffolds in
the formation of benzodioxanes. On the other hand the
spectra show the formation of both mono arylglyceryl ethers
as the main products of the reaction, and of some aliphatic oli-
gomers of glycidol as expected.29
A summary of all the information obtained so far made it
possible for us to outline a new reaction mechanism for the
innovative HMB synthesis pathway proposed herein
(Scheme 6).
First, catechol is activated by the basic catalyst (otherwise
no reaction occurs) and forms the alcoholate anion, which is
now able to perform a nucleophilic attack on the more feasible
and accessible alkylene carbon of the GlyC carbonate ring. The
intermediate does not undergo decarboxylation, but an intra-
molecular closure with the consequent formation of another
cyclic carbonate, a key intermediate. Then the latter undergoes
an intramolecular nucleophilic attack by the free aromatic
–OH group, which finally leads either to the desired product
(HMB) or to its isomer.
Fig. 10 Reaction mixture obtained from the reaction of GlyC with
phenol or catechol.
Fig. 11 HMBC-NMR (600 MHz, DMSO-d6) of the mixture obtained
from the reaction between phenol and glycerol carbonate. Details of the
aliphatic region. Crosslink of the signal arising from glycerol have been
omitted. 4 is attributed by 1H-NMR and 13C-NMR by comparison with
the literature,31 although crosslink signals are not clearly attributable.
Scheme 6 Catechol and GlyC proposed reaction mechanism leading
to HMB and its isomer.
Green Chem.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Green Chemistry
Paper
The same approach was used for the identification of the
mixture obtained by the reaction of GlyC with catechol which
has been performed at 170 °C for 1 h.32
Signals arising from 5 can be identified even though the
reaction between catechol and glycerol carbonate gives much
more complicated NMR spectra (Fig. 13 and 14; Fig. S16–
S21†), due to overlapping of signals attributable to 5, 6, 7, and
other minor products. Nevertheless, HMBC and 1H- and
13C-NMR spectra, which were available for both 2 and products
6 and 7, made possible the identification of the NMR pattern
for compound 5 (yellow lines in Fig. 13 and 14). The remaining
signals could be attributed to polymeric compounds, like
those described in Scheme S3.†
It is important to stress that, as already observed for inter-
mediate 2, the correlation between –CH2 from the five-mem-
bered cyclic carbonate of 5 and CvO at approx. 155 ppm
(Fig. 14) confirms the presence of the cyclic carbonate.
Indeed, the correlation between –CH2 in the lateral chain
and an ipso C at approx. 149 ppm (chemical shift typical of a
catechol compound with an ether in the lateral chain)
made the link between the two fragments of structure 5
unambiguous.
Fig. 12 HMBC-NMR (600 MHz, DMSO-d6) of the mixture obtained
from the reaction between phenol and glycerol carbonate. Details of the
1H-NMR aliphatic region vs. 13C-NMR carbonates and the ipso-C region
(yellow lines correspond to the signals of 2).
Conclusions
An innovative and straightforward synthesis pathway to
2-hydroxymethyl-1,4-benzodioxane (HMB) and its isomer,
without the use of toxic, harmful compounds and solvents, is
proposed herein.
The peculiar reactivity of glycerol carbonate (GlyC) as an
alkylating agent for aromatic compounds, and in particular
with catechol, for the successful synthesis of heterocyclic ben-
zodioxane scaffolds has been widely discussed. Indeed, a
mechanistic investigation has underscored the unique behav-
iour of GlyC compared to other organic carbonates (e.g. propyl-
ene carbonate) and the importance of GlyC as a multi-func-
tional structure, in particular of the free aliphatic –OH group.
The latter plays a fundamental role in obtaining the reactive
carbonate intermediate responsible for intramolecular cycliza-
tion aimed at the selective formation of HMB, with only water
and carbon dioxide as benign co-products. Under the opti-
mized conditions, water is partially removed by distillation or
readily reacts with GlyC leading to the formation of glycerol.
On the other hand, CO2 represents the only light compound
produced and readily leaves the flask. These behaviours finally
prevent the formation of carbonic acid that could poison the
basic catalyst.
Fig. 13 HMBC-NMR (600 MHz, DMSO-d6) of the mixture obtained
from the reaction between catechol and glycerol carbonate. Details of
the aliphatic region.
Noteworthily, for a greater scale some tests can be per-
formed under vacuum in order to collect almost pure CO2 to
be valorised for the synthesis of organic carbonates (e.g. gly-
cerol carbonate from glycidol to reintroduce in the production
line), urea or other valuable compounds.
Future studies will be aimed at extending the reaction to
the formation of enantiomerically pure HMB starting from
chiral GlyC.
Fig. 14 HMBC-NMR (600 MHz, DMSO-d6) of the mixture obtained
from the reaction between catechol and glycerol carbonate. Details of
the 1H-NMR aliphatic region vs. 13C-NMR carbonates and ipso-C region
(yellow lines correspond to the signals of 5).
This journal is © The Royal Society of Chemistry 2018
Green Chem.

View Article Online
Paper
Green Chemistry
16 J. J. Bozell and G. R. Petersen, Green Chem., 2010, 12, 539–
554; J. R. Ochoa-Gómez, O. Gómez-Jiménez-Aberasturi,
C. Ramírez-López and M. Belsué, Org. Process Res. Dev.,
2012, 16(3), 389–399; A. Dibenedetto, F. Nocito, A. Angelini,
I. Papai, M. Aresta and R. Mancuso, ChemSusChem, 2013,
6(2), 345–352.
Conflicts of interest
There are no conflicts to declare.
Acknowledgements
17 A. Dibenedetto, A. Angelini, M. Aresta, J. Ethiraj, C. Fragale
and F. Nocito, Tetrahedron, 2011, 67, 1308–1313.
Alessandra Petroli is acknowledged for the HMBC experiments.
Maria Ricciardi is acknowledged for the glycidol NMR 18 M. O. Sonnati, S. Amigoni, E. P. Taffin de Givenchy,
experiments.
T. Darmanin, O. Choulet and F. c. Guittard, Green Chem.,
2013, 15, 283.
19 M. Selva and M. Fabris, Green Chem., 2009, 11, 1161–
1172.
Notes and references
20 S. F. Campbell, Patent US 4188390, 1980; D. R. Rao,
R. N. Kankan and D. R. Birari, Patent US 8426590B2, 2013;
C. Bolchi, L. Fumagalli, B. Moroni, M. Pallavicini and
E. Valoti, Tetrahedron: Asymmetry, 2003, 14, 3779–3785;
I. Sànchez, M. D. Pujol, G. Guillaumet, R. Massingham,
A. Monteil, G. Dureng and E. Winslow, Eur. J. Med. Chem.,
2000, 35, 663–676; C. Bolchi, E. Valoti, V. Straniero,
P. Ruggeri and M. Pallavicini, J. Org. Chem., 2014, 79, 6732–
6737; S. Wang, Y. Chen, S. Zhao, X. Xu, X. Liu, B. Liu and
G. Zhang, Bioorg. Med. Chem. Lett., 2014, 24, 1766–1770;
S. Wang, Y. Chen, X. Liu, X. Xu, X. Liu, B. F. Liu and
G. Zhang, Archiv der Pharmazie, 2014, 347, 32–41.
1 A.-A. G. Shaikh and S. Sivaram, Chem. Rev., 1996, 96, 951–
976; B. Schaffner, F. Schaffner, S. P. Verevkin and
A. Borner, Chem. Rev., 2010, 110, 4554–4581; H. Zhang,
H.-B. Liu and J.-M. Yue, Chem. Rev., 2014, 114(1), 883–898;
C. Martín, G. Fiorani and A. W. Kleij, ACS Catal., 2015, 5(2),
1353–1370; J. W. Comerford, I. D. V. Ingram, M. North and
X. Wu, Green Chem., 2015, 17, 1966–1987; J. H. Clements,
Ind. Eng. Chem. Res., 2003, 42, 663–674.
2 H. M. Bolt and B. Gansewendt, Crit. Rev. Toxicol., 1993, 23,
237–253.
3 L. A. Pokier, G. D. Stoner and M. B. Shimkin, Cancer Res.,
1975, 35, 1411.
21 M. Ilic, P. Dunkel, J. Ilaš, E. Chabielska, A. Zakrzeska,
P. Mátyus and D. Kikelj, Eur. J. Med. Chem., 2013, 62, 329–340.
22 A. Delgado, G. Leclercb, M. C. Lobato and D. Mauleon,
Tetrahedron Lett., 1988, 29(30), 3671–3674.
4 J. Mc Cann, E. Choi, E. Yamasaki and B. N. Ames, Proc.
Natl. Acad. Sci. U. S. A., 1975, 72, 5135.
5 Y. Li, B. Xue and X. He, Catal. Commun., 2009, 10, 702–707.
6 B. Xue, J. Xu, P. Liu, L. Lv and K. Zhang, J. Mol. Catal. A: 23 A. Rouf, M. A. Aga, B. Kumar and S. C. Taneja, Tetrahedron
Chem., 2012, 357, 50–58. Lett., 2013, 54, 6420–6422.
7 M. Distaso and E. Quaranta, Tetrahedron, 2004, 60(7), 1531– 24 A. Rouf, P. Gupta, M. A. Aga, B. Kumar, A. Chaubey,
1539; M. Distaso and E. Quaranta, Appl. Catal., B, 2006,
66(1–2), 72–80; M. Aresta, A. Dibenedetto and E. Quaranta,
R. Parshad and S. C. Taneja, Tetrahedron: Asymmetry, 2012,
23, 1615–1623.
Green Chem., 1999, 1, 237–242; Y. Ono, Catal. Today, 1997, 25 http://www.zettachem.com/acp/images/brochure/ZEOcat20
35(1–2), 15–25. brochure20English.pdf.
8 W. C. Shieh, S. Dell and O. Repic, Org. Lett., 2001, 3, 4279; 26 T. Tabanelli, F. Cavani and M. Selva, PCT/IB 2016/055692,
S. Ouk, S. Thiébaud, E. Borredon and P. Le Gars, 2015.
Green Chem., 2002, 4, 431; Z. L. Shen, X. Z. Jiang, 27 T. Tabanelli, E. Monti, F. Cavani and M. Selva, Green
W. M. Mo, B. X. Hu and N. Sun, Green Chem., 2005, 7, 97; Chem., 2017, 19, 1519–1528.
F. Rajabi and M. R. Saidi, Synth. Commun., 2004, 34, 28 B. Drevermann, A. Lingham, H. Hügel and P. J. Marriott,
4179. Tetrahedron Lett., 2005, 46, 39–41.
9 W. C. Shieh, S. Dell and O. Repic, J. Org. Chem., 2002, 67, 29 R. Cucciniello, M. Ricciardi, R. Vitiello, M. Di Serio,
2188.
A. Proto and C. Capacchione, ChemSusChem, 2016, 9, 3272–
10 M. Selva and P. Tundo, J. Org. Chem., 2006, 71, 1464.
3275.
11 P. Tundo, M. Selva and A. Bomben, Org. Synth., 1998, 76, 30 A. Leal-Duaso, M. Caballero, A. Urriolabeitia, J. A. Mayoral,
169–177; P. Tundo and M. Selva, Acc. Chem. Res., 2002, 35,
706–716.
12 M. Selva, P. Tundo and A. Perosa, J. Org. Chem., 2003, 68,
7374–7378.
13 P. Tundo, F. Trotta, G. Moraglio and F. Ligorati, Ind. Eng.
Chem. Res., 1988, 27, 1565–1571.
14 P. Ziosi, T. Tabanelli, G. Fornasari, S. Cocchi, F. Cavani and
P. Righi, Catal. Sci. Technol., 2014, 4, 4386.
15 Z. Zhang, D. W. Rackemann, W. O. S. Doherty and
I. M. O’Hara, Biotechnol. Biofuels, 2013, 6, 153.
J. I. García and E. Pires, Green Chem., 2017, 19, 4176–4185.
31 J. A. Kenar and G. Knothe, J. Am. Oil Chem. Soc., 2008, 85,
365–372.
32 It must be taken into account that 2 is more stable than
the intermediate 5, probably because of its rapid rearrange-
ment due to the additional aromatic –OH moiety (favoured
at high temperature (e.g. GC-MS injector at 280 °C). In
order to prevent decarboxylation and overcome the lower
stability of 5 the reaction between GlyC and catechol has
been performed in a closed J Young valve NMR tube.
Green Chem.
This journal is © The Royal Society of Chemistry 2018