Catalytic dehydration of sorbitol to isosorbide in the presence of metal tosylate salts and metallized sulfonic resins

Article abstract of DOI:10.1016/j.mcat.2018.11.004

Homogeneous catalytic dehydration of sorbitol to isosorbide has been performed with a series of metal tosylates as catalysts. Conversions up to 100 % and selectivities into isosorbide up to 67% were obtained with Bi(OTs)₃. The metals were exchanged with acidic sites of sulfonic resins and the resulting materials were evaluated as heterogeneous catalysts. On the contrary to their homogeneous counter parts, the heterogenized metal sites are non-active. The catalytic activity of the modified resins was systematically diminished in comparison to the native resins. The inhibition is greatly dependent on the nature of the metal and, on a larger extent, of the used resin for the cation exchange.

Full text of DOI:10.1016/j.mcat.2018.11.004

MolecularCatalysis463(2019)61–66
Contents lists available at ScienceDirect
Molecular Catalysis
journal homepage: www.elsevier.com/locate/mcat
Catalytic dehydration of sorbitol to isosorbide in the presence of metal
tosylate salts and metallized sulfonic resins
Corentin Dussennea, Hervé Wyartc, Vincent Wiatzc, Isabelle Suisseb, Mathieu Sauthierb,⁎
a Institut Français des Matériaux Agro-sourcés, Parc Scientifique de la Haute-Borne, 60 Avenue Halley, 59650, Villeneuve d’Ascq, France
b Univ. Lille, CNRS, Centrale Lille, ENSCL, Univ. Artois, UMR 8181 - UCCS - Unité de Catalyse et Chimie du Solide, F-59000, Lille, France
c Roquette Frères, 62080, Lestrem Cedex France
A R T I C L E I N F O
A B S T R A C T
Keywords:
isosorbide
sorbitol
dehydration
tosylate metal
sulfonic resins
Homogeneous catalytic dehydration of sorbitol to isosorbide has been performed with a series of metal tosylates
as catalysts. Conversions up to 100 % and selectivities into isosorbide up to 67% were obtained with Bi(OTs)3.
The metals were exchanged with acidic sites of sulfonic resins and the resulting materials were evaluated as
heterogeneous catalysts. On the contrary to their homogeneous counter parts, the heterogenized metal sites are
non-active. The catalytic activity of the modified resins was systematically diminished in comparison to the
native resins. The inhibition is greatly dependent on the nature of the metal and, on a larger extent, of the used
resin for the cation exchange.
1. Introduction
industrial scale, a drastic increase of the production of this molecule is
expected and this further strengthens the importance of developing
The use of bioresources instead of petro-sourced raw materials for
the synthesis of chemicals is a major issue for future developments.[1,
2, 3, 4] In this context, sugars are very important renewable molecules
due to their large availability at industrial scales from very diversified
vegetal sources. Sugars have since a long time found a great variety of
applications.[5, 6] In many cases, these polyhydroxylated compounds
are used without any chemical transformation, for example as energy
source (glucose) or low-calorie sweetener (xylose). Their chemical
transformation opens the way to a broader scope of applications. They
have been for example used as starting materials for the synthesis of
many industrially important chemicals such as furans (furfural and
hydroxymethylfurfural), acids (glucaric acid), alcohols (ethanol) and
alkyl polyglycosides.[7] Difunctional molecules (diols, diacids) are part
of a particularly attractive class of compounds for their use as mono-
mers in polycondensation reactions. In this context, isosorbide is a very
promising bio-derived platform molecule for the chemicals market for
which applications should dramatically increase in the future years.[8]
Derived from glucose through a multistep hydrogenation/dehydration
process, this diol is actually used for the production of high perfor-
mance polymers, for example Poly Ethylene Isosorbide Terephtalate
(PEIT), and as substitute of bisphenol A to increase properties of re-
sistance and optical clarity in polycarbonates [9,10]. This compound
has also been chemically modified to synthesize alternative plasticizers
to phthalates [11]. In view of the applications of isosorbide at the
more efficient synthetic pathways and related catalysts. The key step of
the isosorbide synthesis is the double dehydration of sorbitol. As shown
in Fig. 1, dehydration of sorbitol leads after the first dehydration step to
several different monoanhydrohexitols named sorbitans. Among them,
only the 3,6-sorbitan and the 1,4-sorbitan are precursors of isosorbide
after a second dehydration step. The others sorbitans ultimately lead,
after further intra- or inter-molecular dehydration steps, to colored
oligomeric species named “humins”. The formation of humins drasti-
cally contributes to reduce the mass balance of the reaction. More ac-
tive and selective catalysts are thus needed. The reaction has been first
performed in the presence of sulfuric acid [12]. A variety of catalysts
such as mineral [9] and organic acids [13], metal salts [14] or solid acid
catalysts such as ions exchange resins [15,16], zeolites [17,18,19] or
metal phosphates [20] have been developed. In the case of Bronsted
acid, their pKa value is a key parameter for high catalytic activities as
values below 0 are generally required [21]. Dehydration was also de-
monstrated under MW irradiation [22] and in advanced reaction media
such as sub- or supercritical solvents or in molten salts [23]. Recently,
Zhao proposed a one-step synthesis of isosorbide from cellulose in the
presence of a combined catalytic system composed of strong acid and
ruthenium supported on carbon [24]. Dehydration could also be per-
formed without any acid catalyst in high temperature liquid water [25].
Nevertheless, although this reaction gave rise to an important number
of publications [26], sorbitol dehydration to isosorbide remains a
⁎ Corresponding author.
https://doi.org/10.1016/j.mcat.2018.11.004
Received 2 July 2018; Received in revised form 1 October 2018; Accepted 5 November 2018
2468-8231/©2018ElsevierB.V.Allrightsreserved.

C. Dussenne et al.
MolecularCatalysis463(2019)61–66
Fig. 1. The different ways of dehydration of sorbitol.
challenge in terms of efficiency and selectivity.
(n = 3) in 100 mL methanol. The mixture was refluxed for 8 hours
and filtrated on celite. Methanol was evaporated under vacuum and
100 mL of acetonitrile were added. The mixture was heated 1 h at
50 °C, filtrated and washed twice with hot acetonitrile (20 mL). The
resulting solid was finally dried overnight at 150 °C under vacuum.
Recently, water-tolerant Lewis acid catalysts based on triflate metals
were found to be very efficient with, for example, an isosorbide yield up
to 71% with only 0.05 mol% of Ga(OTf)3 at 160 °C after 1 hour [27].
Metallic tosylates salts have a moderate cost in comparison to their
triflate equivalents. Their straightforward synthesis and high stability
prompted us to evaluate their efficiency as novel catalysts for the for-
mation of isosorbide. Moreover, the metals were immobilized in a
sulfonic resin through cation exchange and the synthesized materials
were further studied as potent heterogeneous catalysts for the dehy-
dration of sorbitol.
2.3. Preparation of the modified resins
The resin (5 mmol of H+) was added to a 10 mL aqueous solution of
metallic tosylate M(OTs)n (example: 0.25 mmol of metallic tosylate for
5% metal in the resin) [29,30]. After 48 h stirring at ambient tem-
perature, the resin was filtrated and washed with 10 mL of deionized
water. The filtrate was titrated by a soda solution in order to determine
the amount of PTSA released in the solution and thus to evaluate the
percent of metal loaded in the resin. The calculated maximum amount
of metal that can be loaded in the resin was in all cases in agreement
with the amount of PTSA released in the water: n(metal) = nH+/3 in
case of trivalent metals such as Cr3+ and n(metal) = nH+/2 in the case
of divalent metals such as Mn2+. In the case of iron, redox titration of
the residual iron in the filtrate according to known volumetric proce-
dures with SnCl2, HgCl2, K2Cr2O7 could be performed to further check
that all the Fe3+ have been exchanged with H+ (quantitative cation
exchange) [31,32].
2. Experimental
2.1. Materials
Sorbitol, isosorbide, isomannide were given by Roquette Frères and
used without further purification. Chemicals were purchased from
Aldrich, Alfa Aesar, Acros or Strem.
2.2. Synthesis of the metallic tosylates precursors M(OTs)n
The metallic precursors were obtained according to described pro-
cedures [28]. The purity of the compounds was assessed by elemental
analysis.
2.4. Catalytic tests
a) From the chloride precursors: the metallic precursor MCln (20 mmol)
was mixed with 44 mmol PTSA (para-toluenesulfonic acid) (when
n = 2) or 66 mmol of PTSA (n = 3) in 100 mL acetonitrile. After
dissolution of the solids under vigorous stirring, the solvent was
evaporated and the solid mixture was heated at 165 °C for 4 hours
under 1 mbar pressure. The solid was cooled and 100 mL of me-
thanol were added. The resulting mixture was refluxed for 1 h and
filtrated. The solvent was evaporated under low pressure. 100 mL of
acetonitrile were added to the solid and the mixture was stirred and
heated for 1 h at 50 °C. The solid was filtrated and washed twice
with 20 mL of hot acetonitrile then dried under vacuum at 150 °C.
b) From nitrate metallic precursor: 40 mmol of metallic precursor M
(NO3)n were dissolved in 100 mL water under stirring. Then 50 mL
of a solution of KOH (4 mol.L-1) were added. The metallic hydro-
xides precipitated immediately and were isolated by filtration. The
solid was washed with water (3 x 50 mL). The metallic tosylates
were obtained according the same procedure used for the chloride
precursors.
Dehydration reactions were performed either in a 250 mL glass re-
actor or in a 20 mL Schlenk tube from 10 g (0.055 mole) or 2.5 g
(0.0137 mole) of sorbitol respectively. Sorbitol was first completely
melted at 115 °C then the mixture was stirred under vacuum at this
temperature for 10 min in order to remove water. The reactor was filled
with nitrogen and the catalyst was added. The pressure was adjusted to
40 mbar. The temperature of the solution was maintained at 115 °C for
one hour, heated to 130 °C and stabilized for an additional hour, heated
to 145 °C and stirred for two additional hours. Aliquots of the reaction
medium were sampled every hour and analyzed by GC.
2.5. Analysis and quantification of the products
Conversions and yields were determined by gas chromatography
after derivatization of the hydroxyl groups of polyols with BSTFA (N,O-
bis(trimethylsilyl)trifluoroacetamide). α-methylglucopyranoside was
used as internal standard.
Calibration was first performed by weighing approximatively pre-
cisely 50 mg of each polyol, isomannide, isosorbide, 1,4-sorbitan, sor-
bitol and α-methylglucopyranoside which were dissolved in 50 mL
c) From the hydroxide precursors: the precursor M(OH)n (20 mmol) was
mixed with 44 mmol PTSA (when n = 2) or 66 mmol of PTSA
62

C. Dussenne et al.
MolecularCatalysis463(2019)61–66
pyridine. 1 mL of this solution was mixed with 300 μL BSTFA for one
hour at 70 °C. This mixture was injected in GC which allows de-
termining the response factor of each polyol.
67% (Table 1). The best catalysts in terms of both activity and iso-
sorbide selectivity are based on the trivalent metals rather than divalent
ones. 11 % and 24 % yields in isosorbide are obtained with respectively
Co(OTs)2 and Sn(OTs)2 (entries 3 and 4) and very few amounts of
isosorbide are obtained with Mn(OTs)2, Cu(OTs)2, Ni(OTs)2, and Zn
(OTs)2 (entries 5, 6, 7 and 8). Cr(OTs)3, Fe(OTs)3 and Bi(OTs)3 afford
respectively 64 %, 66 % and 67 % yields in isosorbide (Entries 9, 10 and
11). Ga(OTs)3 and Al(OTs)3 showed a lower catalytic activity as re-
spectively 33% and 50 % isosorbide yields were obtained along with
these catalysts (Entries 12 and 13). In these two cases, the reaction did
not reach completion within 4 hours as relatively high amounts of un-
reacted 1,4-sorbitan were obtained. The trivalent metals are likely more
electrophilic than the divalent ones and hence more acidic. This higher
acidity explains that trivalent metals are more active catalysts for the
dehydration steps. Moreover, isosorbide yields compare well with those
observed in the presence of sulfuric acid in the same catalytic condi-
tions and the amounts of other side compounds attain 19-30%.
As the homogeneous metal tosylates showed interesting catalytic
activity, we anticipated that a suitable chemical modification of acid
sulfonic resins by metals could provide efficient hybrid systems with
both Lewis and Brønsted acid sites. This choice was prompted by the
fact that in a resin, Lewis acids would bear a similar structure with their
homogeneous counterparts. In both cases, the counter-ions of the me-
tals are arylsulfonate groups. We thus synthesized metal-modified re-
sins by cation exchange between the protons of the sulfonic resins with
the metals of the metallic tosylate salts. Among the large number of
commercial sulfonic acids, we turned our attention toward highly acid
and stable at high temperatures resins. Our choice was then orientated
on different commercially available acidic resins and more particularly
: 1) the Dowex® G26, Dowex® monosphere® (Uniform Particle Size) and
Amberlyst® 70 that are polystyrene-co-divinylbenzene sulfonic acid
resins. In the Amberlyst® 70, the presence of chlorine atoms confers a
high temperature stability up to 190 °C and has been described for the
one-pot conversion of cellulose to isosorbide using metal catalysts and
ion-exchange resin [34]. 2) the Nafion® NR50 is a tetrafluoroethylene
sulfonated resin which is known as an efficient acid catalyst for a wide
scope of organic reactions [35] 3) the Aquivion® PW98, a per-
fluorinated polymer with sulfonic acid groups more recently launched
on the market. This latter resin is described as a superacid solid capable
to efficiently and selectively catalyze the oligomerization of glycerol
[36], the glycosylation of glucose with fatty alcohols [37] and the
conversion of cellulose into amphiphilic alkyl glycosides [38]. Fur-
thermore, the use of heterogeneous catalysts has several potential ad-
vantages, such as a fast and simple isolation of the reaction products as
well as recovery and recycling of the catalyst. Main characteristic of
these resins are summarized in Table 2.
The reaction mixture could then be analyzed by GC after the same
procedure: 25 mg of the crude mixture and 25 mg of α-methylgluco-
pyranoside were dissolved in 10 mL pyridine. Then 1 mL of this solution
and 300 μL of BSTFA were mixed at 70 °C for 1 h before injection in
chromatography. The use of the response factors determined from the
calibration allowed to quantify precisely each polyol. In all the tables,
“others” includes humins and other unidentified side products of re-
action. These amounts were determined from the mass balance of the
reaction [others (%) = sorbitol conversion – (yield in isosorbide +
yields in sorbitans)].
GC conditions: The GC analyses were performed on a GC Agilent
7890B equipped with a HT 5 column (30 m, 0.25 mm i.d., 0.25 μm
thickness). Conditions: injector and detector = 300 °C, gas = hydrogen
at 1.57 mL/min, temperature = 30 s at 140 °C, ramp of 5 °C/min up to
200 °C, then ramp of 10 °C/min up to 280 °C and isotherm at 280 °C
during 10 min.
3. Results and discussion
We first synthesized a large variety of metallic tosylates with dif-
ferent metals as Cr(III), Mn(II), Fe(III), Co(II), Ni(II), Cu(II), Zn(II), Al
(III), Bi(III), Ga(III) and Sn(II). The syntheses were readily implemented
thanks to published procedures [28]. The metal salts could be obtained
from the reaction of a metal hydroxide (Mn), metal nitrates (Cr and Fe)
or metal chlorides (Al, Fe, Co, Ni, Cu, Zn, Bi, Ga and Sn) with a stoi-
chiometric amount of p-toluenesulfonic acid. The yields in pure metal
tosylates ranged from 60 to 80%.
The different salts were evaluated as homogeneous catalysts in de-
hydration reaction of sorbitol to produce isosorbide. Reactions were
performed with 1 mol % of metal, at high temperatures (T = 115-
145 °C) and under vacuum (40 mbar) to eliminate water. Results are
reported in Table 1. The results are compared to those obtained with
sulfuric acid as catalyst. The reaction was stopped after 4 hours in order
to obtain a satisfactory conversion and to limit the formation of other
side compounds.
Metallic tosylates proved to be efficient in terms of activity and
selectivity for the synthesis of isosorbide although they do not offer
acidity as high as metal triflates [33]. Indeed, the aromatic sulfonic
acids are much weaker acids than triflic acid, p-toluenesulfonic acid, for
example, is about 106 times weaker acid than triflic acid [33]. The
results are nevertheless very different according to the used metal with
5% to 100% sorbitol conversions and isosorbide selectivities from 0% to
Very recently, the synthesis of sulfonic resins modified with metals
such as Nafion-Fe, Aquivion-Fe and Aquivion-Ga [39] have been re-
ported in literature. In our hands, the synthesis of the modified resins
with metals was performed via cation exchange procedures (Fig. 2, see
experimental section). An aqueous solution containing the sulfonated
resin and the adequate amount of the metallic tosylate salt were mixed
for two days at ambient temperature then filtrated to obtain the mod-
ified resins. The amount of PTSA liberated in solution was determined
by titration of the filtrate which allowed to calculate precisely the metal
percent inserted into the resin. With this general method, Amberlyst 70
resin was modified with Cr(III), Mn(II), Fe(III), Co(II), Ni(II), Cu(II), Al
(III), Bi(III) and Ga(III) which were previously evaluated as tosylate
salts under homogeneous conditions. All the metals could be easily
loaded in the resin according to cation exchange reactions by mixing
the Amberlyst with the metal tosylate salts in aqueous solutions. Ad-
vantageously in the case of iron, the amount of non-exchanged metal
remaining in the solution could be titrated in order to assess the
quantity of iron in the resin (see experimental part). Titrations showed
a quantitative immobilization of the metal in the material.
Table 1
Dehydration of sorbitol to isosorbide in the presence of tosylate salts.a
Entry Catalyst
Sorbitol
Conv. (%)
Yield (%)
Othersb
(%)
Isosorbide 1,4-Sorbitan Sorbitans
1
2
3
4
5
6
7
8
H2SO4
PTSA
Co(OTs)2
Sn(OTs)2
Mn(OTs)2 13
Cu(OTs)2
Ni(OTs)2
Zn(OTs)2
Cr(OTs)3
Fe(OTs)3
Bi(OTs)3
Ga(OTs)3
Al(OTs)3
100
100
97
68
68
11
24
0
1
0
7
64
66
67
50
33
0
0
6
9
12
9
0
26
23
23
27
13
15
5
30
29
24
21
21
19
51
39
0
30
0
51
0
0
100
53
5
7
0
100
100
100
100
100
100
12
7
10
10
10
10
9
10
11
12
13
2
19
38
a
Reaction conditions : Sorbitol : 2.5 g (0.0137 mol); tosylate salt : 1% mol;
T = 115-145 °C (115 °C for 1 h, 130 °C for 1 h, and 145 °C for 2 h), t = 4 h.
b
Other side compounds formed during the reaction, including humins.
With the catalytic procedure reported above, we first explored the
63

C. Dussenne et al.
MolecularCatalysis463(2019)61–66
Table 2
Characteristics of the sulfonated resins.a
Resin
T
(°C)
Structure
Size distribution
(mesh)
Surface
Area
Size pores (diameter)
(Å)
Total acid capacity
(mmol H+/g)
max
(m2/g)
Dowex® G26
Dowex® Monosphere® 650 C
Amberlyst® 70
Aquivion® PW98
Nafion® NR50
150
150
170
160
200
Beads
Beads
Beads
Pellets
Beads
22-25
21-25
25-40
n. d.
n. d.
n. d.
36
n. d.
n. d.
n. d.
n. d.
220
n. d.
n. d.
1.63
1.63
2.55
0.980
0.8
n. d.
a
Data from commercial source, n.d. = not determined.
sorbitol dehydration reaction in the presence of the native resins as for
most of them, nothing was found in the literature. The reactions were
performed with 1 mol % or 5 mol % of acidic site by adjusting the
weight of resin. Table 3 reports the results obtained with different
amounts of resins and comparison with sulfuric acid.
Reactions were first performed with 1 mol % of SO3H group in re-
spect to sorbitol. Nafion, monosphere, G26 and Amberlyst resins led to
limited conversions of sorbitol ranging from 50% to 61% (Entries 1 - 4,
Table 3). An increase of the resin loadings up to 5 mol % allowed ob-
taining full conversions of sorbitol (Entries 6-9, Table 3). The presence
of 1,4-sorbitan with monosphere and G26 indicate a lower catalytic
activity in comparison to Nafion and Amberlyst. With these two later
resins, no intermediate sorbitan remains after 4 hours of reaction. 1% of
the Aquivion resin proved to be sufficient to reach a full conversion
within four hours albeit with a limited selectivity due to the formation
of high amounts of undesired sorbitans and oligomers (isosorbide
yield = 58% yield, Entry 5).
Interesting features were observed thanks to GC monitoring in the
case of the resins in comparison with the homogeneous system (sulfuric
acid). Fig. 3 presents curves that give the evolutions of the amounts of
sorbitol, 1,4-sorbitan, isosorbide and side products with respect to the
reaction time in the presence of sulfuric acid and Amberlyst 70 as
catalysts. Although the maximum yields in isosorbide attained in both
cases after 4 hours of reaction are very similar, the selectivity in sor-
bitans formed in the course of the reaction is markedly different.
Whereas the two dehydration steps were consecutive in the pre-
sence of sulfuric acid, in the case of Amberlyst 70, isosorbide is pro-
duced simultaneously with the 1,4-sorbitan. This difference is clearly
evidenced by the amounts of the intermediate 1,4-sorbitan during the
reaction. The amount of 1,4-sorbitan reaches 70 % in the case of sul-
furic acid whereas the maximum yield is 30 % in the case of the resin.
Moreover, in the heterogeneous system, side products are formed from
the beginning of the reaction although the temperature does not exceed
115 °C. At this temperature in homogeneous conditions, only the first
sorbitol dehydration to sorbitans is observed. This effect can be due to
the slow diffusion of sorbitans outside the strongly acidic pores of the
resin. The local high concentration of sorbitans in this highly acid
medium explains their rapid conversion into isosorbide and side pro-
ducts. Nevertheless, at the end of the reaction after 4 hours, the iso-
sorbide yield remained very close to that obtained in the presence of
sulfuric acid. Similar profiles were observed for all tested hetero-
geneous catalytic systems.
Similar catalytic conditions have been used with the resins modified
with metals. A detailed catalytic study with Amberlyst 70 loaded with
iron revealed a strong deactivation behavior at relatively low metal
loadings as illustrated in Fig. 4. Indeed, a dramatic drop of the catalytic
activity appears from 3 % of iron loading. This shows that the cataly-
tically active acid sites have been converted in inactive sites after the
cation exchange. Clearly, these experiments show that the active mo-
lecular catalyst is completely inefficient after being grafted in the resin
as the chemical yield in isosorbide is roughly 0 % with 10 % loading of
iron in the same reaction conditions. This difference of reactivity be-
tween the homogeneous and the heterogeneous systems can arise from
the strong binding of the metal by the chelating tosylate groups within
the resin. The brutal loss of activity observed suggests a surface deac-
tivation of the material through the formation of an “egg shell like”
metal loading. The most accessible acid sites for cationic exchange with
metals are also the most accessible sites for substrates. This surface
Fig. 2. Synthesis of metallized sulfonic resins via ion exchange methods: example of iron.
64

C. Dussenne et al.
MolecularCatalysis463(2019)61–66
Table 3
Dehydration of sorbitol to isosorbide in the presence of native resins.a
Entry
Catalyst
% mol
H+
Conv.
(%)
Yield (%)
Isosorbide
Othersb
(%)
1,4-Sorbitan
Sorbitans
1
2
3
4
5
6
7
8
Nafion NR50
Monosphere
G26
Amberlyst 70
Aquivion
Nafion NR50
Monosphere
G26
Amberlyst 70
Aquivion
1
1
1
1
1
5
5
5
5
50
51
52
61
100
100
100
100
100
100
16
8
9
12
58
66
48
51
72
62
19
18
18
20
0
11
8
8
9
12
9
18
19
5
4
17
17
20
30
25
17
15
14
28
0
16
16
0
9
10
c
5
5
5
a
Reaction conditions : Sorbitol : 2.5 g (0.0137 mol); T = 115-145 °C (115 °C for 1 h, 130 °C for 1 h, and 145 °C for 2 h), 40 mbar, t = 4 h.
Other side compounds formed during the reaction, including humins.
t = 2 h; after 2 hours, the isosorbide yield decreased.
b
c
deactivation combined with the high viscosity of melted sorbitol lead to
a rapid loss of the catalytic activity.
been observed with other trivalent metals thus indicating that this ef-
fect is not metal dependent. The higher acidity of the active sulfonic
groups in Aquivion and Nafion as well as a larger accessibility of the
substrates to the acidic active sites probably explain that these two
resins are less impacted by the exchanges with metals.
A short comparative study regarding the nature of the metal (Al, Ga,
or Na) in the modified Amberlyst resins underlines similar behaviors
compared to iron (Fig. 4, right side). Metals such as Al, Fe, Ga, lead to a
very strong catalyst deactivation from 5% of metal loading and a
complete deactivation with 10% loading. The results obtained with 5%
loading allowed delineating slight differences between the metals.
Considering the amount of metal exchanged in the resin, Na+ appears
4. Conclusion
as the less deactivating metal in comparison to Fe3+, Al3+ or Ga3+
.
Metal tosylates are efficient catalysts for the conversion of sorbitol
in isosorbide according to homogeneous conditions with up to 67 %
yield of isosorbide. Heterogeneous based catalysts such as sulfonic re-
sins also efficiently promote the reaction and show different selectiv-
ities in comparison to homogeneous acidic systems as they rapidly
convert 1,4-sorbitan into isosorbide within the pores. Surprisingly,
sulfonic resins loaded with metals through cation exchange show
drastically reduced catalytic activity in comparison to the native resins.
In the case of the Amberlyst, G26 and Monosphere resins used for this
study, less than 5 % of metal is for example enough to completely de-
activate the resin. This inhibiting effect is explained by the strong
binding of the metal by the chelating tosylate groups within the resin
and the surface deactivation of the resin.
However, it should be taken into account that in the case of sodium, the
number of neutralized acidic sites is three times smaller than with tri-
valent metals.
According to the same protocol, other resins listed Table 2 were
loaded with different metals and evaluated in the catalytic synthesis of
isosorbide from sorbitol. A representative example with gallium is
shown Fig. 5.
Unexpectedly, very different behaviors are observed depending on
the nature of the resins used as support of gallium. On the one hand,
metals exchanged in Monosphere and G26 resins show similar catalytic
activities as those linked to Amberlyst 70: the catalytic activity of the
three resins is very sensitive to metal exchange with a strong deacti-
vation at low loadings. This trend is even stronger with Amberlyst resin.
On the other hand, the Aquivion resin, that has already shown better
catalytic activities toward isosorbide production in its native form
(Table 3), led also to much better yields at higher metal loadings. The
catalytic activity of the resin remains almost unchanged even with 10%
and 20 % of gallium. The Ga-Nafion resin has an intermediate behavior
in terms of isosorbide yield. No drastic drop is observed but the yield in
isosorbide gradually decreased with the amount of gallium that has
been loaded in the resin. It should be noticed that the same trend has
Acknowledgements
We are grateful to the Institut Francais des Matériaux Agrosourcés
(IFMAS) for C. Dussenne fellowship. This project is supported by the
State as part of the “Programme d’Investissement d’Avenir”, under re-
ference ANR-10-IEED-0004-01.
Fig. 3. Comparison of the kinetic monitoring
between H2SO4 (left side, 1% mol) and
Amberlyst 70 (right side, 5% mol) during de-
hydration
of
sorbitol.
Sorbitol
2.5 g
(0.0137 mol); T = 115-145 °C (115 °C for 1 h,
130 °C for 1 h, and 145 °C for 2 h), 40 mbar,
t = 4 h. Maximum yields in isosorbide are at-
tained after 4 hours of reaction in both cases.
65

C. Dussenne et al.
MolecularCatalysis463(2019)61–66
Fig. 4. Influence of the % iron in Amberlyst 70
resin on the isosorbide yield (on left side)
Influence of the % metal in Amberlyst 70 on
the isosorbide yield (on right side).
Reaction conditions
:
Sorbitol
:
10 g
(0.055 mmol) ; T = 115-145 °C (115 °C for 1 h,
130 °C for 1 h, and 145 °C for 2 h), t = 4 h. In
the case of Al3+, Fe3+ and Ga3+, the percen-
tage of H + exchanged in the resin is 3 times
the percentage of metal introduced.
[12] G. Flèche, M. Huchette, Starch-Stärke 38 (1986) 26–30.
[13] J. Defaye, A. Gadelle, C. Pedersen, Carbohydr. Res. 205 (1990) 191–202.
[14] F. Liu, K.D.O. Vigier, M. Pera-Titus, Y. Pouilloux, J.-M. Clacens, F. Decampo,
F. Jérôme, Green Chem. 15 (2013) 901–909.
[15] J.C. Goodwin, J.E. Hodge, D. Weisleder, Carbohydr. Res. 79 (1980) 133–141.
[16] (a) K. Bock, C. Pedersen, H. Thogersen, Acta Chem. Scand. B 35 (1981) 441–449;
(b) M.J. Ginés-Molina, R. Moreno-Tost, J. Santamaría-González, P. Maireles-Torres
Applied Catalysis A: General 537 (2017) 66–73.
[17] H. Kobayashi, H. Yokoyama, B. Feng, A. Fukuoka, Green Chem. 17 (2015)
2732–2735.
[18] R. Otomo, T. Yokoi, T. Tatsumi, Appl. Catal., A 505 (2015) 28–35.
[19] M. Kurszewska, E. Skorupowa, J. Madaj, A. Konitz, W. Wojnowski, A. Wisniewski,
Carbohydr. Res. 337 (2002) 1261–1268.
[20] J. Xi, Y. Zhang, D. Ding, Q. Xia, J. Wang, X. Liu, G. Lu, Y. Wang, Appl. Catal., A 469
(2014) 108–115.
[21] A.A. Dabbawala, D.K. Mishra, G.W. Huber, J.-S. Hwang, Appl. Catal., A 492 (2015)
252–261.
[22] I. Polaert, M.C. Felix, M. Fornasero, S. Marcotte, J.-C. Buvat, L. Estel, Chem. Eng. J.
222 (2013) 228–239.
[23] J. Li, A. Spina, J.A. Moulijn, M. Makkee, Cat. Sci. Tec. 3 (2013) 1540–1546.
[24] G. Liang, C. Wu, L. He, J. Ming, H. Cheng, L. Zhuo, F. Zhao, Green Chem. 13 (2011)
839–842.
[25] A. Yamaguchi, N. Hiyoshi, O. Sato, M. Shirai, Green Chem. 13 (2011) 873–881.
[26] C. Dussenne, T. Delaunay, V. Wiatz, H. Wyart, I. Suisse, M. Sauthier, Green Chem.
19 (2017) 5332–5344.
Fig. 5. Influence of the % Ga in various resins on the isosorbide yield.
Reaction conditions : Sorbitol : 2.5 g ; resin : 5 mol % ; T = 115-145 °C (115 °C
for 1 h, 130 °C for 1 h, and 145 °C for 2 h), t = 4 h. The percentage of H+ ex-
changed in the resin is 3 times the percentage of Ga3+ introduced.
[27] K. Stensrud, E. Hagberg, S. Howard, E. Rockafellow, Process for acid dehydration of
sugar alcohols (2014) WO 2014/137619 A1.
[28] S. Holmes, S. McKinley, G. Girolami, P. Szalay, K. Dunbar, Inorg. Synth 33 (2002)
91–103.
[29] S. Kanemoto, H. Saimoto, K. Oshima, H. Nozaki, Tetrahedron Lett. 25 (1984)
3317–3320.
References
[30] Y. Ma, Q. Wang, Z. Gao, X. Sun, N. Wang, R. Niu, H. Ma, Renew. Energ. 86 (2016)
643–650.
[31] D. Liu, X.-D. Du, Q. Lu, D. Ma, J.-H. Lei, J. Wuhan Univ. Technol. 6 (2012) 028.
[32] D.A. Skoog, D.M. West, F.J. Holler, S. Crouch, Fundamentals of analytical chem-
istry, Nelson Education (2013).
[33] T.N. Parac-Vogt, K. Deleersnyder, K. Binnemans, Eur. J. Org. Chem. 2005 (2005)
1810–1815.
[34] A. Yamaguchi, O. Sato, N. Mimura, M. Shirai, Catal. Commun. 67 (2015) 59–63.
[35] G. Gelbard, Ind. Eng. Chem. Res. 44 (2005) 8468–8498.
[36] A. Karam, N. Sayoud, K.D.O. Vigier, J. Lai, A. Liebens, C. Oldani, F. Jérôme, J. Mol.
Catal. A: Chem. 422 (2016) 84–88.
[1] B.M. Trost, Science (1997) 1471–1477.
[2] R.A. Sheldon, Chem. Ind. (1992) 903–906.
[3] P.T. Anastas, J.C. Warner, Green chemistry: theory and practice, Oxford University
Press, 2000.
[4] Biomass Research and Development Technical Advisory Committee, Vision for
Bioenergy and Biobased Products in the United States, (2006).
[5] H. Roeper, Starch-Stärke 54 (2002) 89–99.
[6] T. Werpy, G. Petersen, A. Aden, J. Bozell, J. Holladay, J. White, A. Manheim,
D. Eliot, L. Lasure, S. Jones, Top value added chemicals from biomass 1 (2004) Dtic
document technical report.
[7] B.R. Caes, R.E. Teixeira, K.G. Knapp, R.T. Raines, ACS Sustain, Chem. Eng. 3 (2015)
2591–2605.
[8] M. Rose, R. Palkovits, ChemSusChem 5 (2012) 167–176.
[9] (a) Y.S. Eo, H.-W. Rhee, S. Shin, J. Ind. Eng. Chem. 37 (2016) 42–46.
[10] X. Feng, A.J. East, W. Hammond, Y. Zhang, M. Jaffe, Polym. Adv. Technol. 22
(2011) 139–150.
[37] A. Karam, K. De Oliveira Vigier, S. Marinkovic, B. Estrine, C. Oldani, F. Jérôme, ACS
Catal. 7 (2017) 2990–2997.
[38] F. Jérôme, K. De Oliveira Vigier, A. Karam, B. Estrine, S. Marincovic, C. Oldani,
ChemSusChem 10 (2017) 3604–3610.
[39] R. Tassini, V.D. Rathod, S. Paganelli, E. Balliana, O. Piccolo, J. Mol. Catal. A: Chem.
411 (2016) 257–263.
[11] Y. Yang, Z. Xiong, L. Zhang, Z. Tang, R. Zhang, J. Zhu, Mater. Design 91 (2016)
262–268.
66