A Combined Experimental–Theoretical Study on Diels-Alder Reaction with Bio-Based Furfural: Towards Renewable Aromatics

Article abstract of DOI:10.1002/cssc.202002111

The synthesis of relevant renewable aromatics from bio-based furfural derivatives and cheap alkenes is carried out by using a Diels-Alder/aromatization sequence. The prediction and the control of the ortho/meta selectivity in the Diels-Alder step is an important issue to pave the way to a wide range of renewable aromatics, but it remains a challenging task. A combined experimental-theoretical approach reveals that, as a general trend, ortho and meta cycloadducts are the kinetic and thermodynamic products, respectively. The nature of substituents, both on the dienes and dienophiles, significantly impacts the feasibility of the reaction, through a modulation on the nucleo- and electrophilicity of the reagents, as well as the ortho/meta ratio. We show that the ortho/meta selectivity at the reaction equilibrium stems from a subtle interplay between charge interactions, favoring the ortho products, and steric interactions, favoring the meta isomers. This work also points towards a path to optimize the aromatization step.

Full text of DOI:10.1002/cssc.202002111

Chemistry–Sustainability–Energy–Materials
Accepted Article
Title: A combined experimental-theoretical study on Diels-Alder
reaction with bio-based furfural: towards renewable aromatics
Authors: François Jérôme, Ivan Scodeller, Karine Vigier, Eric Muller,
Changru Ma, Frederic Guegan, and Raphael Wischert
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemSusChem 10.1002/cssc.202002111
Link to VoR: https://doi.org/10.1002/cssc.202002111
01/2020

10.1002/cssc.202002111
ChemSusChem
A combined experimental-theoretical study on Diels-Alder reaction
with bio-based furfural: towards renewable aromatics
Dr. Ivan Scodeller,[a] Prof. Karine De Oliveira Vigier,[a] Dr. Eric Muller,[b] Dr. Changru Ma,[c] Dr. Frédéric Guégan,[a]*
Dr. Raphael Wischert,[c]* and Dr. François Jérôme[a]*
Dedicated to P.H. Dixneuf for his outstanding contribution to catalysis
[a]
Institut de Chimie des Milieux et Matériaux de Poitiers, CNRS, Université de Poitiers, 1 rue Marcel Doré, 86073 Poitiers (France)..
E-mail: francois.jerome@univ-poitiers.fr, frederic.guegan@univ-poitiers.fr.
[b]
[c]
Centre de Recherche Solvay, 85 Avenue des Frères Perret, 69190 Saint-Fons (France).
Eco-Efficient Products and Processes Laboratory, UMI 3464 CNRS/Solvay, 3966 Jin Du Road, Xinzhuang, Industrial Zone, Shanghai 201108 (China). E-Mail:
raphael.wischert@solvay.com.
Supporting information for this article is given via a link at the end of the document
Abstract: We investigate the synthesis of relevant renewable from sugars and pinacol, further demonstrating the potential of
aromatics from bio-based furfural derivatives and cheap alkenes, biomass for the synthesis of aromatics.
through a Diels-Alder/aromatization sequence. The prediction and the Considering the current low price of fossil feedstocks, the
control of the ortho : meta selectivity in the Diels-Alder step is an production of bio-based aromatics with a higher value than BTX is
important issue to pave the way to a wide range of relevant renewable a more reasonable approach. Functionalized aromatics such as
aromatics, but it remains a challenging task. Through a combined benzaldehyde, benzylamine, benzonitrile, benzyl alcohol are
experimental-theoretical approach, we reveal that, as a general trend, typical examples with important applications in the field of
ortho- and meta cycloadducts are the kinetic and thermodynamic fragrances, flavors, specialty polymers, agriculture, etc.
products, respectively. The nature of substituents, both on the dienes As mentioned above, renewable aromatics can be synthesized by
and dienophiles, significantly impacts the feasibility of the reaction, reaction of bio-based furanic derivatives with olefins through a
through a modulation on the nucleo- and electrophilicity of the reagents, Diels-Alder/aromatization reaction.
In this context,
5-
as well as the ortho : meta ratio. We show that the ortho : meta hydroxymethylfufural (HMF), “the sleeping giant” among sugar-
selectivity at the reaction equilibrium stems from a subtle interplay derived chemicals, has received a growing interest, in particular for
between charge interactions, favoring the ortho products, and steric the synthesis of terephthalic acid. HMF is poorly reactive in the
interactions, favoring the meta isomers. This work also points towards Diels-Alder (DA) reaction and is often reduced to 2,5-
a path to optimize the aromatization step.
dimethylfuran[14] or oxidized to furan dicarboxylic esters[15] before
undergoing a DA reaction. Although valuable aromatics were
produced in high yields from HMF, the lack of cost-competitive
processing technologies to convert low cost sugars to HMF
currently hampers the industrial deployment of these routes. Note
that BASF, Avantium and Avabiochem have recently
communicated on the production of HMF at a pilot scale, indicating
that mature technologies may be available in a close future for such
application.
In contrast to HMF, furfural is a low cost furanic derivative (1-2€/kg)
nowadays industrially produced in large scale from biomass waste
(> 300 kT/year).[18] Like HMF however, furfural is poorly reactive in
DA reactions. The electron withdrawing -CHO group reduces the
nucleophilicity of the furanic ring (as shown by a severe lowering of
the highest occupied molecular orbital (HOMO) energy vs. the
parent furan), thus making the DA reaction unfavourable.[5] To
circumvent this problem, furfural can be decarbonylated to furan
before the DA reaction.[6,7] For instance, Hubert reported the
synthesis of toluene and benzene by co-feeding ethylene or
propylene with furfural at 450-600°C in the presence of ZSM-5, with
in situ decarbonylation of furfural.[6a]
Introduction
The production of aromatics, such as Benzene, Toluene and
Xylene (BTX) represent an enormous market, with 45 million tons
of benzene produced per year, for instance.[1] A myriad of valuable
downstream aromatic chemicals can be obtained from BTX, such
as terephthalic acid, cyclohexane, cumene, phenol, and styrene, to
mention a few only. Industrially, BTX are mainly produced by
catalytic reforming of naphtha or steam cracking of hydrocarbons.
With the exponential increase of the world population, chemistry
will face difficulty to supply aromatics in the future and alternative
routes are nowadays explored. In this context, there is a growing
interest in the synthesis of aromatics from biomass.[2]
A wide range of aromatics can be theoretically obtained from
lignocellulosic biomass. For instance, alkyl phenols can be
produced by the catalytic lignin-first depolymerization process[3] or
from tannins present in lignocellulosic biomass.[4] Aromatics can be
also produced from hemicellulose or cellulose through the furfural[5-
Alternatively, to preserve the chemical functionality of furfural while
decreasing the HOMO-LUMO gap, furfural can be derivatized to
amines,[8] acetals,[9] alcohols,[10] phenolic-type resins[11] or
amides,[12] making the DA reaction thermodynamically favorable
This strategy was employed to synthesize furfural-derived
polymers which can be depolymerized via a retro-DA reaction at
the end of life. The aromatization of these furfural-derived
13]
15]
or 5-hydroxymethylfurfural platform,[14,
using a Diels-Alder
(DA) / aromatization sequence. A similar strategy has been
reported, using pinacol[16] or limonene as other bio-based
platforms.[17] While mainly para-substituted aromatics are obtained
from lignin, ortho- and meta-substituted aromatics can be obtained
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002111
ChemSusChem
FULL PAPER
cycloadducts has been investigated only sporadically. In 2016, 73%, after 72 h of reaction (Fig. 1). The selectivity to the DA
Hailes reported the DA reaction of furfural dimethylhydrazone with cycloadducts was 98%, the missing 2% corresponding to the
highly activated dienophiles such as N-alkyl-maleimides at 80°C in deprotection of 2-(furan-2-yl)-1,3-dioxolane to furfural and ethylene
water, followed by aromatization to the corresponding phthalimide glycol, presumably due to traces of water (stabilizer) in acrylonitrile
derivatives.[13] However, starting from cheaper and more readily (0.2-0.5% w/w). Extending the reaction time to 120 h resulted in a
available olefins such as acrylonitrile or dimethyl maleate, the minor change of the yield (76%), indicating that the reaction had
reaction yields were drastically lower (< 25% yield). Guided by DFT reached the DA vs. retro-DA equilibrium (Fig. 1).
calculations, we recently reported the synthesis of meta-
xylylenediamine (MXD), an important target in polymer chemistry, Selectivity of the Diels-Alder reaction.
from furfural and acrylonitrile, two cheap and abundant chemicals.[5] At equilibrium (measured after 120 h, see Fig. S1), the ortho: meta
Furfural was first derivatized with ethylene glycol, restoring the ratio reached a value of 52 : 48. After 120 h, the endo to exo ratios
nucleophilicity of the furanic ring (upshift of the HOMO by 0.4 eV were 70 : 30 and 60 : 40 for the ortho and meta isomers,
compared to furfural), and allowing the DA cycloadducts to be respectively. These features are captured reasonably well by DFT
obtained with a yield of 75%. In addition, our strategy improved calculations (Tables S10-S11). First, activation barriers are in all
upon previous work in that (i) the subsequent aromatization step cases (ortho or meta, exo or endo) found in a narrow energy range,
was catalyzed at low temperature (30°C) by bases, thus avoiding ca 31.4-32.5 kcal/mol in free enthalpy. Thus, little to no kinetic
the competition with the retro-DA, and (ii) the reaction could be selectivity is expected. Similarly, reaction free enthalpies range
selectively driven to meta-substituted aromatics only, by playing from +3.7 to +4.6 kcal/mol, suggesting low thermodynamic
with the rate of the aromatization reaction. Although this route was selectivity. Using a Maxwell-Boltzmann statistical model on relative
an expedient protocol to synthesize renewable MXD, predicting reaction free enthalpies affords an ortho : meta ratio of 50 : 50 at
and controlling the ortho : meta selectivity during the Diels-Alder 60°C, in line with experiments. Calculated endo : exo ratios are 67 :
reaction remains highly desirable to widen the scope of this protocol 33 and 46 : 54 for the ortho and meta isomers, respectively, again
for the synthesis of renewable aromatics.
The present work addresses this fundamental scientific challenge. These findings suggests that DFT calculations can be used to
Through combined experimental-theoretical approach, we predict not only the feasibility but also the selectivity of DA reactions,
in good agreement with experiment.
a
rationalize the impact of the substituents on furfural-derived diene which will provide very helpful throughout this work. Before going
and various dienophiles on the ortho : meta selectivity of the Diels- further, it is important to mention that from a practical perspective
Alder reaction. Further, we present an improved protocol to we are interested in ortho : meta ratios. Indeed, the endo : exo
aromatize the so-obtained Diels-Alder adducts to aromatics.
selectivity is lost in the aromatization step, while the ortho : meta
selectivity is preserved. Therefore, in what follows, we will only
discuss ortho : meta ratios; additional experimental and theoretical
data on the endo : exo ratios is provided in the SI.
Results and Discussion
Effect of the reaction temperature.
Reaction between 2-(furan-2-yl)-1,3-dioxolane and acrylonitrile.
As described in our previous report,[5] furfural was first converted to
2-(furan-2-yl)-1,3-dioxolane by reaction with ethylene glycol in the
presence of a catalytic amount of Amberlyst-15. Next, the as-
obtained 2-(furan-2-yl)-1,3-dioxolane was reacted with a 5-fold
excess of acrylonitrile at 80°C, affording the corresponding DA
cycloadducts (mixture of ortho and meta regioisomers, and
corresponding endo and exo stereoisomers) in a combined yield of
Increasing the reaction temperature from 80 to 100 and 120°C
resulted in a much faster reaction, expressed as the time to reach
the highest achievable (equilibrium) DA yield, but yields were lower
with 52% after 24 h, and 42% after 8 h, at 100 and 120°C,
respectively (Table 1, entries 1-3; Fig. 1). This means that at
temperatures higher than 80°C, the retro-DA clearly became the
dominating reaction. Conversely, a decrease of the temperature to
60°C resulted in a slower reaction but the maximum yield of the DA
cycloadduct was identical to that obtained at 80°C (76%, Table 1,
entry 4; Fig. 1). A further decrease of the reaction temperature to
40°C was possible but decreased the reaction rate to an
unacceptable level (Table 1, entry 5; Fig. 1).
After ca. 24 h the ortho : meta ratio stabilized at ~50 : 50 for reaction
temperatures of 40-60°C, decreasing to ~40 : 60 at 80-120°C,
suggesting that the meta product is thermally slightly more stable
than the ortho (Fig. S1). This trend is reproduced by DFT
calculations, as temperature-dependent endergonic corrections
from electronic energies to free enthalpies are slightly smaller for
meta products than for ortho products (Table S11).
100
40°C
60°C
80°C
100°C
120°C
90
80
70
60
50
40
30
20
10
0
0
20
40
60
80
100
120
Time (h)
Figure 1. Total yield of DA cycloadducts as a function of time for the reaction of
2-(furan-2-yl)-1,3-dioxolane with a 5-fold excess of acrylonitrile at different
temperatures.
2
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002111
ChemSusChem
FULL PAPER
observed after 25 h of reaction, compared to catalyst-free
conditions (Table 2, entries 8-14). Even worse, these metal
Table 1. Effect of the reaction temperature (entries 1-5) and dienophile to diene
molar ratio (entries 6-8) on the DA reaction of 2-(furan-2-yl)-1,3-dioxolane with
acrylonitrile.
Table 2. Effect of Lewis acid catalysts on the DA reaction of 2-(furan-2-yl)-1,3-
dioxolane with acrylonitrile.[a]
Entry
Temp.
(°C)
Acrylonitrile / 2-(furan-
2-yl)-1,3-dioxolane
ratio
Time
(h)[a]
Yield
(%)[b,c]
Entry
Catalyst
DA cycloadduct
yield (%)
Deprotection
(furfural yield)
(%)
1
2
3
4
5
6
7
8
80
100
120
60
5
5
72[a]
24[a]
8[a]
76
52
1
-
GaCl3
ScCl3
InCl3
29
-
2
-
2[b]
3[b]
4[b]
5[b]
6[b]
7[b]
8
5
42
-
-
5
120[a]
120[a]
48
76
-
-
40
5
30
CuCl2
HfCl4
-
-
60
5
48[d]
44[d]
24[d]
-
-
60
20
1
48
NbCl5
MnCl2
CdCl2
SrCl2
-
-
60
48
20
23
23
29
21
18
26
51
68
57
75
21
23
39
25
4
[a] time to reach maximum (equilibrium) yield; [b] maximum (equilibrium) yield; [c]
other detected product (2%) are furfural and ethylene glycol; [d] not at reaction
equilibrium, reactions were stopped after 48 h.
9
10
11
12
13
14
15
16
17
18
19
20
BiCl3
15
34
38
15
10
11
23
12
1
Effect of the dienophile to diene molar ratio.
Reactions were performed at 60°C and stopped after 48 h to
assess the effect of the dienophile : diene molar ratio on the
reaction rate. Increasing the acrylonitrile / 2-(furan-2-yl)-1,3-
dioxolane molar ratio from 5 to 20 : 1 had no effect. However, a
decrease from 5 to 1 slowed down the reaction, with only half the
yield obtained after 48 h (Table 1, entries 6-8, Fig.S2). For practical
reasons (the boiling point of acrylonitrile is 77°C), a temperature of
60°C was selected in the following experiments, together with an
acrylonitrile : 2-(furan-2-yl)-1,3-dioxolane molar ratio of 5 : 1.
SnCl2
AlCl3
LiCl
CoCl2
ZnCl2
Zn(OTf)2
ZnI2
Effect of the catalyst.
Without catalyst, the DA reaction is quite slow at 60°C (Fig.1). For
instance, after 25 h of reaction, the DA cycloadducts were obtained
in only 26 % yield (Table 2, entry 1). To increase the formation rate
of the DA cycloadducts, different Lewis acid catalysts (10 mol%),
mostly metal chlorides, were screened (at 60 °C). The choice for
Lewis acid catalysts was motivated by the fact that Brønsted acid
catalysts led to tar-like materials as major products. In all cases,
anhydrous Lewis acid catalysts were tested to limit the deprotection
of 2-(furan-2-yl)-1,3-dioxolane to furfural and ethylene glycol.
Metal chlorides (MCln with M = Ga, In, Sc, Cu, Hf and Nb) led to the
formation of unidentified black soluble and insoluble materials as
main product (Table 2, entries 2-7). With M = Mn, Cd, Sr, Bi, Sn, Li
and Al, the reaction took place, but no improvement of the yield was
ZnI2/dppe
ZnI2/dmpe
1
[a] acrylonitrile/2-(furan-2-yl)-1,3-dioxolane molar ratio = 5,; [b] tar-like materials
were formed as major products.
chlorides catalyzed the concomitant deprotection of 2-(furan-2-yl)-
1,3-dioxolane (4-39% yield) with moisture from the air atmosphere
and/or water contained in acrylonitrile (0.2-0.5% w/w). In contrast,
in the presence of Co or Zn chloride, significantly higher yields of
51 and 68 %, respectively, were reached after 25 h (and 75% after
48 h of reaction, Fig. S3, Table 2, entries 15-16). Unlike with the
other metal chlorides, the deprotection of 2-(furan-2-yl)-1,3-
dioxolane was limited (~10% yield), although it was still higher than
3
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002111
ChemSusChem
FULL PAPER
under catalyst-free conditions (2%). To explore the effect of the
counter anion, Zn(OTf)2 and ZnI2 were also screened, affording the Effect of the protective group on furfural.
DA cycloadducts in 57 and 75% yield, respectively, after 25 h Different protective agents were used to derivatize furfural, with the
(Table 2, entries 17-18). The low yield obtained with Zn(OTf)2 is aim of assessing the effects on yield and isomer ratio. To this end,
probably due to its high hygroscopicity, promoting in turn furfural was reacted with different diols, amino-alcohols and also
deprotection to furfural (23%, Table 2, entry 17). ZnI2 was found to 1,2-ethanedithiol (see SI for details of synthesis and
be the most efficient catalyst screened, as further confirmed by a characterization). The as-obtained furfural derivatives were reacted
kinetic profile which showed that the equilibrium yield was reached with acrylonitrile under the same conditions as 2-(furan-2-yl)-1,3-
slightly earlier than with ZnCl2 (Fig. S3). Addition of strong dioxolane, i.e. 60°C, 10 mol% ZnCl2, and 5 equiv. of acrylonitrile
complexing agents such as 1,2-bis(diphenylphosphino)ethane (Table 3, entries 1-5). The reaction with 2-(furan-2-yl)-1,3-
(DPPE) or 1,2-bis(dimethylphosphino)ethane (DMPE) to ZnI2 or dioxolane is shown for comparison in entry 1. Additional
ZnCl2 inhibited their catalytic activities, further confirming the experiments were also carried out in the absence of catalysts
catalytic role of these Lewis acids (Table 2, entries 19-20). Finally, (Table S3). DFT calculations suggested the feasibility of all these
it should be noted that the ortho : meta ratio was not significantly reactions with HOMO-LUMO gaps similar to those found for the
affected by the presence of the catalyst (Fig. S4). For the remainder successful combination of 2-(furan-2-yl)-1,3-dioxolane with
of the study, ZnCl2 was preferred over ZnI2, because it is cheaper acrylonitrile (6.89-7.39 eV, Table S3).
and it does not lead to yellow coloration of the reaction products,
while its catalytic activity is very similar to that of ZnI2.
Table 3. Reaction of acetal and thioacetal derivatives of furfural with acrylonitrile.
Calculated ortho:meta ratio (%)[c]
Experimental ortho : meta
Entry
Diene
Yield (%)
ratio (%)
Thermodynamic
selectivity[d]
Kinetic selectivity[e]
1[a]
2[b]
3[b]
4[a]
75
81
85
68
50 : 50
53 : 47
52 : 48
43 : 57
50 : 50
55 : 45
56 : 44
43 : 57
63 : 37
58 : 42
59 : 41
57 : 43
67
73[f]
39 : 61
91 : 9
5[a]
37 : 63
90 : 10
[a] Reaction time = 48 h; [b] reaction time = 25 h; [c] based on a Maxwell-Boltzmann statistical model using DFT-calculated energies (see SI for details); [d] based on
relative free enthalpies of reaction at 60°C; [e] based on relative free enthalpies of activation at 60°C; [f] reaction temperature = 30 °C.
4
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002111
ChemSusChem
FULL PAPER
As expected from the DFT predictions, these furfural derivatives
underwent reaction in all cases (Table 3, entries 1-5). The yields
at equilibrium were similar to what was found for 2-(furan-2-yl)-
1,3-dioxolane (67-85%, Table 3, entries 2,5). The selectivity to
the DA cycloadducts always exceeded 89% (95% without ZnCl2),
with furfural and the protecting agent being formed as the only
side products (by deprotection of the furfural derivate). Note, that
using amino alcohols (ethanol amine, N-methylethanolamine) as
derivatizing agents of furfural led to an unexpected reaction
outcome (Table S3). Instead of forming DA adducts, the starting
oxazolidine-derived furfural was converted, but a complex
mixture of unidentified products was formed, tentatively
attributed by NMR spectroscopy to uncontrolled Michael-type
reaction with acrylonitrile.
(furan-2-yl)-4,4,5,5-tetramethyl-1,3-dioxolane
and
2-(1,3-
dithiolan-2-yl) furan, with initial amounts of ortho adduct of 65%
and 74%, respectively, at 30-40% conversion (Table S3). When
the reaction time is extended, the ortho : meta ratios converge to
the values obtained at the reaction equilibrium (Table S3, Fig.
S6), strongly suggesting that the ortho adduct is the kinetic
product. DFT modelling, based on activation free enthalpies at
60°C, nicely reproduces these observations with a kinetic
selectivity for the ortho adduct, markedly more pronounced for 2-
(1,3-dithiolan-2-yl) furan. To support this hypothesis, we carried
out the reaction of 2-(1,3-dithiolan-2-yl) furan at 30°C in the
presence of ZnCl2. As expected, under these conditions, the
ortho DA cycloadduct was the major product (91%, see Table 3,
entry 5). Note that, due to the lower temperature, leading to
inhibition of the retro-DA reaction, the yield slightly increased
from 67 to 73%, (Table 3, entry 5, Figs. S5 and S6). The
experimental and calculated endo : exo ratios were, within
experimental or computational error, identical for the meta
isomer of all DA adducts, showing a slight preference for the
endo isomer. This preference was more pronounced in the case
of the ortho isomer, Table S4). Again, a similar trend was
observed in the absence of catalyst, although the reaction
equilibrium had not been reached yet (Table S3).
At equilibrium, the ortho : meta ratio of DA adducts was around
50 : 50 for all alcohol-protected furfural derivatives (Table 3,
entries 1-4), albeit with a slightly higher amount of the meta
isomer in the case of 2-(furan-2-yl)-4,4,5,5-tetramethyl-1,3-
dioxolane (pinacol derivative) (Table 3, entry 4). For the thiol-
derivative 2-(1,3-dithiolan-2-yl) furan, an even higher amount of
meta isomer was observed (Table 3, entry 5). Note, that in the
absence of a catalyst, similar trends were observed (Table S3).
The experimental product ratios are in excellent agreement with
equilibrium product ratios obtained from DFT-calculated relative
free enthalpies of reaction (Table 3, entries 1d-5d, and Tables
S10-S19), confirming that the thermodynamic equilibrium had
been reached in all cases. In contrast, poorer agreement is found
with ratios calculated from DFT-calculated relative free
enthalpies of activation, representing kinetic conditions.
Reaction of 2-(furan-2-yl)-1,3-dioxolane with other
dienophiles.
Since the DA reaction of 2-(furan-2-yl)-1,3-dioxolane and
acrylonitrile had been successful, other relevant dienophiles
were also tested, i.e. acrolein, methyl vinyl ketone, and methyl
acrylate. The calculated HOMO-LUMO gaps for the combination
of these dienophiles with 2-(furan-2-yl)-1,3-dioxolane (6.89-7.47
eV, Table S5) were comparable to that found for the combination
with acrylonitrile (ΔE = 7.39 eV), suggesting that these reactions
We now turn our attention to isomer ratios at the beginning of the
reaction, where such kinetic selectivity should manifest.
Noteworthy, both with and without catalyst, the ortho DA
cycloadduct is initially formed in a higher amount than the meta
adduct (Table S3, Fig. S6). This effect is most evident for 2-
Table 4. Reaction of 2-(furan-2-yl)-1,3-dioxolane with different dienophiles.[a]
Experimental ortho :
Calculated ortho :
meta ratio (%)[c]
Experimental ortho : meta
ratio (%) < 15 % conv.
Calculated kinetic ortho :
meta ratio (%)[d]
Entry
-R
Yield (%)[b]
meta ratio (%)[b]
48 : 52
13 : 87
33 : 67
38 : 62
-
1
2
3
4
5
-CN
76
36
40
28
-
50 : 50
14 : 86
18 : 82
28 : 72
-
48 : 52
41 : 59
44 : 56
47 : 53
-
63:37
19:81
48:52
31:69
-
-COCH3
-COOCH3
-CHO
-CH2NH2
[a] Dienophile / 2-(furan-2-yl)-1,3-dioxolane molar ratio = 5; [b] at equilibrium; [c] based on a Maxwell-Boltzmann statistical model (see SI for details), using DFT-
calculated relative free enthalpies of reaction at 60°C; [d] based on relative free enthalpies of activation.
5
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002111
ChemSusChem
FULL PAPER
should be feasible as well. Indeed, the reactions were successful
(carried out without catalyst, see Fig. S7 for kinetic profiles).
At equilibrium, the DA reaction plateaued at 28-40% yield with
methyl acrylate, methyl vinyl ketone, and acrolein compared to
75% with acrylonitrile (Table 4 and Fig. S7). The selectivity of the
reaction remained close to 100%, suggesting that the retro-DA
reaction was less favorable for acrylonitrile than for the other
dienophiles. We also tested allyl amine as an example of an
electron-rich dienophile. The HOMO-LUMO gap calculated by
DFT was high (9.24 eV), indicating that the DA should not
proceed, as indeed observed experimentally. This result shows
that only electron poor dienophiles are eligible for DA reaction
with 2-(furan-2-yl)-1,3-dioxolane.
Reaction of other furanic dienes with acrylonitrile.
To assess the molecular diversity of DA cycloadducts that could
be obtained through this pathway, other-furfural-derived dienes
such as 2-furonitrile, furfuryl amine, furfuryl alcohol, ethylfurfuryl
ether, and 2-methylfuran, nowadays obtained on a large scale
from furfural, were screened. Like furfural, 2-furonitrile contains
an electron-withdrawing substituent (the –CN group), and
consequently the HOMO-LUMO gap calculated by DFT (ΔE =
7.77 eV) suggested an unreactive combination (for comparison,
ΔE = 7.72 eV for the unsuccessful combination of furfural and
acrylonitrile). In contrast, the other furfural derivatives contain an
electron-donating group which should permit the reaction with
acrylonitrile, as corroborated by HOMO-LUMO gaps (6.94-7.09
eV, Table S6), comparable to the reactive combination of 2-
(furan-2-yl)-1,3-dioxolane and acrylonitrile (ΔE = 7.39 eV).
Consistent with the prediction from DFT, no DA reaction was
observed for 2-furonitrile, while it was successful for furfuryl
alcohol, ethyl furfuryl ether and 2-methyl furan, with equilibrium
yields of 81%, 54% and 69%, respectively (Table 5). In the case
of furfuryl amine (Table 5, entry 2), a mixture of different products
was obtained, due to concomitant Michael addition of the -NH2
group on acrylonitrile. The DA cycloadduct was observed by
NMR (again consistent with the prediction from DFT) but, due to
significant formation of side products, it was not possible to
accurately determine the corresponding yield by 1H NMR.
Noteworthy, in the case of 2-furonitrile, attempts to perform an
inverse electron demanding DA reaction by reaction with allyl
amine also failed (HOMO-LUMO gap = 7.94 eV).
The nature of the dienophile also impacts the ortho : meta ratio.
Although in all cases the ortho cycloadduct was the kinetic
product, at equilibrium the meta product was predominant (Table
4 and Fig. S8), in particular in the case of methyl vinyl ketone,
with 87% meta isomer (Table 4, entry 2). Again, the experimental
findings are well reproduced by DFT calculations, with a clear
thermodynamic meta selectivity for methyl vinyl ketone,
methylacrylate and acrolein with an ortho : meta ratios of 14:86,
18:82 and 28:72, respectively, using free enthalpies of reaction
(Table 4, entries 2-4, and Tables S20-S32). As observed earlier,
for all ortho and meta isomers, a slight preference for the endo
isomer was observed, both experimentally and theoretically
(Table S5).
Table 5. Reaction of furan derivatives with acrylonitrile.
Experimental
ortho : meta ratio
(%) at 15 %
conv.
Calculated
kinetic ortho :
meta ratio (%)[b]
Calculated
thermodynamic ortho :
meta ratio (%)[a]
ΔE(HOMO-
LUMO) (eV)
Yield
(%)
Experimental ortho : meta
ratio (%) at equilibrium
Entry
-R
1
2
-CN
7.77
7.09
7.03
7.04
6.94
-
n.d.[c]
81
-
-
-
-
-CH2NH2
-CH2OH
-CH2OEt
-CH3
-
-
-
-
3
60 : 40
64 : 36
66 : 34
57 : 43
63 : 37[e]
61 : 39
72 : 28
66 : 34
85 : 15
82 : 18
81 : 19[e]
91 : 9
4[d]
5
54
69
[a] based on a Maxwell-Boltzmann statistical model (see SI for details), using DFT-calculated relative free enthalpies of reaction at 60°C; [b] based on relative free
enthalpies of activation; [c] n.d. means not determined, furfuryl amine successfully reacted with acrylonitrile but a complex mixture of compounds was obtained due
to parasite Michael addition of furfuryl amine to acrylonitrile, making impossible the determination of the reaction yield; [d] the reaction was stopped after 48 h
because extending the reaction produced tar-like material; [e] the calculations were performed on a simplified model (ethyl group replaced by methyl).
6
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002111
ChemSusChem
FULL PAPER
1
0,9
0,8
0,7
0,6
0,5
0,4
0,3
0,2
0,1
0
R1 = -COOCH3 (dienophile = methyl acrylate)
1
0,8
0,6
0,4
0,2
0
-CH3
R =
-CH2OH
-CH2OEt
0
10
20
30
40
50
0
20
40
60
80
Total DA cycloadducts yield (%)
Total DA cycloadducts yield (%)
1
0,9
0,8
0,7
0,6
0,5
0,4
0,3
0,2
0,1
0
R1 = -COCH3 (dienophile = methyl vinyl ketone)
Figure 2. Effect of the substitution of the furanic ring on the ortho : meta ratio.
Acrylonitrile : derivatized furfural molar ratio = 5, reaction temperature = 60°C,
without catalyst.
At equilibrium, the ortho : meta ratio varied from 60 : 40 to 66 : 34,
with 2-methyl furan showing the highest ortho selectivity (Table 5,
entries 3-5), higher than in all other cases explored so far. Kinetic
profiles (Fig. 2, Fig. S9) show that the ortho DA cycloadduct is the
kinetic product, like in all other cases. At < 15% conversion, the
ortho DA was formed in different amounts: 2-methylfuran (85%),
furfuryl alcohol (72%), ethylfurfuryl ether (66%) and 2-(furan-2-yl)-
1,3-dioxolane (52%), indicating that substituents on the furanic
ring also impact the initial formation rate of the ortho and meta DA
cycloadducts (Fig. 2, Fig. S9). DFT calculations once again nicely
retrieve the experimental findings, both in terms of
thermodynamic and kinetic selectivity (Table 5, entries 3-5,
Tables S33-S40). Again, for all DA cycloadducts, the endo
isomers were formed preferentially (Table S6).
-CH3
R =
-CH2OH
-CH2OEt
0
10
20
30
40
50
Total DA cycloadducts yield (%)
Figure 4. ortho : meta molar fraction as a function of the combined yield of DA
cycloadducts, using methyl acrylate and methyl vinyl ketone as dienophiles.
Dienophile : diene molar ratio = 5, reaction temperature = 60°C.
calculations, all tested combinations were successful, affording a
library of DA cycloadducts with different chemical functionality
(Fig. 3). As found with 2-(furan-2-yl)-1,3-dioxolane, the highest
equilibrium yields of DA cycloadducts were obtained using
acrylonitrile as a dienophile, confirming again that the -CN group
has the strongest promoting effect on the yield of the DA
cycloadduct.
In line with what was discussed earlier for the combination of
methyl vinyl ketone with 2-(furan-2-yl)-1,3-dioxolane (Table 4), at
equilibrium, the amount of meta DA cycloadducts was the highest
with methyl vinyl ketone (87%), followed by methyl acrylate (67%),
indicating that the ortho : meta ratio was mostly governed by the
substituent present on the dienophile. As observed for
acrylonitrile, the ortho cycloadduct was always the kinetic product
(Fig. 4).
Optimal combination of diene and dienophile.
Again guided by HOMO-LUMO gaps determined by DFT
calculations (Table S7), furfuryl alcohol, ethylfurfuryl ether and 2-
methylfuran were next reacted with different dienophiles (acrolein,
methyl vinyl ketone and methyl acrylate) to further improve
molecular diversity and to further assess how substituents impact
the yield and ortho or meta selectivity. As suggested by DFT
Elements of rationalization.
As we have seen, DFT calculations proved successful in
reproducing the experimental ortho : meta selectivity, both under
kinetic and thermodynamic conditions. But beyond the
reproduction and prediction of data, modelling offers insight into
the fundamentals of this selectivity. Chemical intuition suggests
that meta addition should minimize the steric repulsion between
the pending group of the diene (R1) and the pending group of the
dienophile (R2), as well as the deformation energy required to
reach the transition state and product geometries. On the other
hand, in the spirit of the Klopman-Salem model of reactivity,[19]
charge control would result in the interaction of the most charged
sites of both reagents (Scheme 1). Natural population analysis
(NPA)[20] shows that the two carbon atoms of the diene carry
positive partial charges with different relative magnitude (+ and
++), depending on the substituent R1, while the two carbon atoms
100
80
60
40
-CN
-R2
20
0
-COOCH3
-COCH3
-R1
Figure 3. Combination of various furfural-derived chemicals and dienophiles.
Dienophile : derivatized furfural molar ratio = 5 : 1, reaction time = 120 h (except
for the combination of acrylonitrile with ethylfurfuryl ether for which the reaction
time was 48 h).
7
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002111
ChemSusChem
FULL PAPER
of the dienophile carry negative partial charges (- and --), again
differing in magnitude, depending on the substituent R2.
The highest partial charges on both diene and dienophile are
always found on the carbon atoms linked to the respective
substituents (Table S8). In the DA reaction, the coulombic
interaction between the reaction partners is maximized when the
most charged atoms interact, i.e. the most positive partial charge
on the diene will attract the most negative partial charge on the
dienophile. Hence, the functional groups of the resulting DA
adduct end up in an ortho arrangement (Scheme 1).
100
90
80
70
60
50
40
30
20
10
0
0
0,02
0,04
0,06
0,08
0,1
 (charge difference on the dienophile)
ortho
Electronic interaction
meta
Steric interaction
Figure 6. Experimental equilibrium meta selectivity as a function of the charge
difference (δ) between the two C atoms of the double bond of the dienophile for
various combinations of dienophiles and dienes. For the sake of clarity, only the
trans isomer of the respective dienophile is represented. See SI for more
information on the cis isomer (Table S8).
Scheme 1. Competition between electronic and steric interactions: ++ / + and -
- / - indicate the relative magnitude of the positive and negative partial charges
of the carbon atoms involved in bond formation during the DA reaction.
For the reaction of a given dienophile with different dienes, the
higher the charge difference () between the pro-ortho and pro-
meta C atoms of the diene, the higher is the expected ortho
selectivity (and by definition the lower is the meta selectivity). This
is indeed the case, with a reasonable correlation between  and
the experimental ortho : meta ratio at equilibrium found in the
reaction of acrylonitrile with different furfural derivatives (Figure 5).
Table S8). Hence, the difference in electronic interaction between
furfural-derived chemicals and these dienophiles is expected to
be lower than with acrylonitrile, and formation of the ortho
cycloadduct should be less favored. This claim is supported by
Figure 6 showing that, irrespective of the starting diene, the lower
the  on the dienophile, the higher the amount of meta product.
Note, that while steric effects cannot be properly quantified here,
it is reasonable to assume that such effects now take over, thus
favoring the formation of the meta product.
To summarize, we may propose a “rule of thumb” to account for
the experimental observations: here, the observed selectivity
stems from a competition between steric interaction, favoring
meta addition, and electronic interactions, favoring ortho addition
(Scheme 1). The model also has the additional benefit that it can
be used by non-experts. Instead of time-consuming explicit
calculations of reaction pathways, only partial charges of the
reactants are necessary, acting as convenient descriptors for
selectivity.
55
50
45
40
35
30
Aromatization reaction
As previously reported by us, the DA cycloadducts derived from
furfural can be quantitatively aromatized, with the loss of water, in
the presence of a catalytic amount of sodium tert-butoxide and at
only 30°C, thus preventing the unwanted competitive retro-DA
reaction.[5] This protocol relies on DMSO as a solvent in order to
produce in situ a superbase, by solvation of sodium by DMSO.[21]
We found here that, except in the case of 2-(1,3-dithiolan-2-yl)
furan for which no aromatization was observed, for all other tested
protective agents, the aromatization reaction proceeded well (70-
80% yield), always with the meta aromatic being formed more
rapidly than the ortho isomers (Fig. S10, Table S9), a mean to
produce consecutively, but separately, the ortho and meta
aromatics, as previously shown by us.[5]
0,1
0,15
0,2
0,25
 (charge difference on the diene)
Figure 5. Experimental equilibrium meta selectivity as a function of the
calculated charge difference (δ) between the two C atoms of the double bond of
the diene, for the reaction of acrylonitrile with different dienes.
In the same spirit, for the reaction of a given diene with different
dienophiles, the charge difference between the two olefinic
carbons on the dienophile should impact the selectivity: for methyl
vinyl ketone (trans isomer: = 0.009 e) methyl acrylate (trans
isomer: = 0.036 e) and acrolein (trans isomer: = 0.049 e) is
significantly lower than that of acrylonitrile (= 0.091 e) (Fig. 6,
8
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002111
ChemSusChem
FULL PAPER
Table 6. Influence of solvents on the base-catalyzed aromatization of the DA
cycloadduct obtained from 2-(furan-2-yl)-1,3-dioxolane and acrylonitrile.
Total yield of
aromatics (%)
ortho : meta
ratio (%)
Entry
Solvent
1
2
3
4
DMSO
93
42
10
81
30: 70
9 : 91
5: 95
DMSO/Toluene (1/1)
DMSO/Toluene (1/3)
Toluene + 2.5% DMSO
Scheme 2. Base-catalyzed aromatization of the DA cycloadducts.
50 : 50
Although this protocol is an expedient way to produce various
functionalized renewable aromatics in nearly quantitative yields,
the use of DMSO as a solvent may raise ecological issues.
However, since the role of DMSO is to solvate tBuONa, we
surmised that it should be possible to use it in a small amount (as
a co-solvent). In what follows, the aromatization of the DA
cycloadducts obtained from 2-(furan-2-yl)-1,3-dioxolane and
acrylonitrile was selected as a case study (Scheme 2). DMSO
was first employed as a co-solvent with toluene. In a mixture of
DMSO and toluene, the aromatization was observed, but the
yields of aromatics were lower than in pure DMSO; 42% and 10%
in a mixture DMSO/toluene 1/1 and 1/3, respectively (vs 90% in
neat DMSO) (Table 6, entries 2,3). These results can be
explained by the inhibition of the DMSO/ tBuONa superbase by
released water. Indeed, in neat DMSO, the initial production rate
of aromatics (raromatic) was 1.83 mmol/h (Fig. 7). However, when 5
eq. of water (relative to the DA cycloadduct) were initially added
into DMSO, raromatic decreased to 0.42 mmol/h and the
aromatization reaction stopped at 57% conversion. A further
incremental increase of the initial amount of water in DMSO from
5 to 10 and 20 eq. stopped the reaction at yield of only 48% and
29%, respectively, and further decreased the aromatization rate
to 0.18 and 0.05 mmol/h, respectively, reflecting the deactivation
of the DMSO/ tBuONa superbase by water.
Conclusion
We show that the Diels-Alder reaction of cheap alkenes with
industrially available bio-based furfural and its derivatives is a
promising pathway for the synthesis of functionalized renewable
aromatics. Various furfural-derived dienes were successfully
coupled at 60°C to various dienophiles featuring electron-
withdrawing groups (-CHO, -CN, -COCH3, -COOCH3), affording
the corresponding DA cycloadducts with yields ranging from 28 to
85% at the thermodynamic equilibrium. In all cases, the DA
reaction proceeds with a selectivity higher than 98%. The reaction
can be accelerated by ZnCl2 or ZnI2 catalysts, while maintaining
the reaction selectivity higher than 90%. Pleasingly, it is possible
to control the ortho : meta ratio by a suitable choice of the
substituents on diene and dienophile and the reaction conditions
(kinetic or thermodynamic). The ortho product is the kinetic
product, in the sense that it forms faster than the meta adduct.
Conversely, under thermodynamic conditions, the meta product
is more stable in most cases.
DFT calculations were carried out to not only predict the feasibility
of the DA reaction for the different diene and dienophile
combinations, but also to understand how the ortho : meta ratio
depends on the substituents on the dienes and dienophiles. We
find excellent agreement between experimental and calculated
product ratios under both thermodynamic and kinetic conditions,
underlining the value of DFT calculations as predictive tool for
organic chemistry and biomass valorization.
We find that the substituents on the dienes and dienophiles
impact the reaction outcome in two principal ways. First, the
(thermodynamic) feasibility of the reaction is conditioned by the
HOMO-LUMO gap between the reactants, which in turn depends
on the electron-withdrawing or –donating nature of the
substituents. The calculated HOMO-LUMO gap can thus be used
as a convenient descriptor for the feasibility of the reaction.
Second, the ortho : meta selectivity of the reaction depends on
the balance between steric repulsion, favoring the meta isomers,
and electronic interactions (characterized by charges), orienting
towards the ortho products. As a result, the highest selectivity to
the ortho isomer is obtained for the reaction of acrylonitrile and
methyl furan, while the highest meta selectivity is observed with
methyl vinyl ketone, independent of the furan derivative. The
calculated charges of the reactants thus allow a quick estimation
of the ortho : meta selectivity, as alternative to the calculation of
full reaction pathways.
100
0 eq. H2O
90
80
70
5 eq. H2O
60
50
40
30
20
10
0
10 eq. H2O
20 eq. H2O
0
2
4
6
8
Reaction time (h)
Figure 7. Effect of water on the total yield of aromatics (30°C, DMSO, 20 mol %
tBuONa).
To maintain the catalytic activity of the DMSO / tBuONa
superbase while limiting its deactivation by released water, the
aromatization reaction was therefore conducted under reflux in
toluene, in a Dean-Stark apparatus, and in the presence of only
2.5% of DMSO. Pleasingly, with continuous removal of water,
aromatics were formed in 81% yield after only 10 min of reaction.
Here again, the meta aromatic was always produced faster than
the ortho isomer (Table 6, entry 4).
This work also points to an optimization of the aromatization step.
Notably, by continuous removal of released water, it was possible
to perform the base-catalyzed aromatization reaction is a
9
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002111
ChemSusChem
FULL PAPER
toluene/DMSO (97.5:2.5 mass ratio) mixture, instead of neat
DMSO as previously described by us.
Keywords: Aromatics • Bio-based furanics • Diels-Alder • DFT •
Renewables
By providing efficient synthetic and predictive tools to overcome
the low reactivity of bio-based furfural in Diels-Alder reaction, we
do believe this work complements the scope of biomass for the
supply of renewable aromatics. While para-substituted aromatics
are mainly produced from lignin, the present work opens the door
to ortho- and meta-substituted aromatics from sugar-derived
furfural.
[1]
[2]
https://www.lelementarium.fr/product/benzene-toluene-xylenes;
consulted on March 21st 2019.
a) http://publications.lib.chalmers.se/records/fulltext/185085/185085.pdf,
consulted on March 21st, 2019; b) A. Maneffa, P. Priecel, J. A. Lopez-
Sanchez, ChemSusChem 2016, 9(19), 2736-2748; c) H. D. Embree, T.
H. Chen, G. F. Payne, Chem. Eng. J. 2001, 84(2), 133-147.
a) T. Renders, S. Van den Bosch, S. F. Koelewijn, W. Schutyser, B. F.
Sels, Energy Environ. Sci. 2017,10, 1551-1557; b) Y. Huang, Y. Duan,
S. Qiu, M. Wang, C. Ju, H. Cao, Y. Fang, T. Tan, Sustainable Energy
Fuels 2018, 2, 637-647; c) E. M. Anderson, M. L. Stone, R. Katahira, M.
Reed, G. T. Beckham, Y. Román-Leshkov, Joule 2017, 1, 613-622; d) R.
Rinaldi, R. T. Woodward, P. Ferrini, H. J. E. Rivera, J. Braz. Chem. Soc.
2019, 30(3), 479-491; e) T. Renders, G. Van den Bossche, T. Vangeel,
K. Van Aelst, B. Sels, Curr. Opin. Biotechnol. 2018, 56, 193-201.
a) A. Arbenz, L. Avérous, Green Chem. 2015, 17, 2626-2646; b) L.
Roumeas, G. Billerach, C. Aouf, E. Dubreucq, H. Fulcrand, ACS Sust.
Chem. Eng. 2018, 6 (1), 1112-1120; c) L. Roumeas, C. Aouf, E.
Dubreucq, H. Fulcrand, Green Chem. 2013, 15(11), 3268-3275; d) A.
Zhang, J. Li, S. Zhang, Y. Mu, W. Zhang, J. Li, RSC Adv. 2017, 7(56),
35135-35146.
[3]
Experimental Section
General procedure for Diels-Alder reaction: 50 mmol of furan derivative
was mixed with 250 mmol of a dienophile and heated under N2
atmosphere at 60°C for 25-48 h in a carousel flask equipped with a
magnetic stirring bar and a condenser. At the DA equilibrium, the excess
of dienophile was removed by distillation under vacuum. The as-recovered
DA cycloadduct can be directly used as collected for the subsequent
aromatization reaction. For characterization purposes the DA cycloadduct
was purified by flash chromatography (silica gel, EtOAc/cyclohexane).
[4]
[5]
[6]
I. Scodeller, S. Mansouri, D. Morvan, E. Muller, K. De Oliveira Vigier, R.
Wischert, F. Jérôme, Angew. Chem. 2018, 130(33), 10670-10674;
Angew. Chem. Int. Ed. 2018, 57(33), 10510-10514.
Catalytic Diels-Alder reactions were performed in a similar way, with the
addition of 0.1 eq. of catalyst, preferentially ZnCl2 or ZnI2. In such case,
the catalyst needs to be removed (by purification over silica gel) before
engaging the DA cycloadduct in the aromatization step.
a) Y. Cheng, G. W. Huber, Green Chem. 2012, 14, 3114-3125; b) A. E.
Settle, L. Berstis, N. A. Rorrer, Y. Roman-Leshkóv, G. T. Beckham, R.
M. Richards, D. R. Vardon, Green Chem. 2017, 19, 3468-3492.
Y. Tachibana, S. Kimura, K. Kasuya, Sci. Rep. 2015, 5, 8249.
C. Garcia-Astrain, A. Gandini, D. Coelho, I. Mondragon, A. Retegi, A.
Eceiza, M. A. Corcuera, N. Gabilondo, Eur. Polym. J. 2013, 49, 3998-
4007.
Aromatization reaction: The reaction was performed as described in our
previous work.[5]
[7]
[8]
Computational details: All calculations were performed using Gaussian
16 rev B, at the M06-2X/6-311++G(d,p) level of theory. The M06-2X
functional indeed provides reliable reaction paths for Diels Alder additions,
both in terms of selectivity and reactivity,[22] while inclusion of polarization
and diffuse functions was shown to be necessary in the calculation of
sulphur-containing compounds.[23]
[9]
a) N. Teramoto, M. Niwa, M. Shibata, Materials 2010, 3, 369-385; b) H.
Satoh, A. Mineshima, T. Nakamura, N. Teramoto, M. Shibata, React.
Funct. Polym. 2014, 76, 49-56; c) S. Takano, F. Ito, K. Ogasawara,
Pharm. Soc. Jpn. 1982, 102(2), 153-161; d) N. Teramoto, Y. Arai, M.
Shibata, Carbohydr. Polym. 2006, 64, 78-84; e) K. Fischer, S. J. Hünig,
J. Org. Chem. 1987, 52, 564-569; f) A. Guidi, V. Theurillat-Moritz, P.
Vogel, Tetrahedron 1996, 7(11), 3153-3162.
All geometries were fully optimized and frequency calculations were
performed in order to confirm the nature of the stationary points. Unless
stated otherwise, all calculations used an implicit solvation model (PCM),
with a dielectric constant of ε = 33.0 (25°C, 1 bar value for acrylonitrile).
HOMO-LUMO energy values were extracted from the last single point
evaluation on the optimization calculations. Endo/exo selectivity were
evaluated for both ortho and meta attacks on the dienophile, and two
conformers of the diene were considered (conformation of the furan
substituting group on position 2). This led to 8 transition state and 8 product
structures in the case of acrylonitrile. For the other dienophiles (methyl
acrylate, acrolein, methyl vinyl ketone), additional geometries were
considered because of the cis/trans isomerism around the conjugated
bond, resulting in 16 transition state and 16 product structures. Atomic
charges were calculated for the optimized structures, using the Natural
Population Analysis method as implemented in NBO version 3.1 for
Gaussian. The computed natural charges for the reagents are provided in
Table S8.
[10] Y. Bai, M. De Bruyn, J. H. Clark, J. R. Dodson, T. J. Farmer, M. Honoré,
I. D. V. Ingram, M. Naguib, A. C. Whitwood, M. North, Green Chem. 2016,
18, 3945-3948.
[11] G. Çayli, S. Kusefoglu, J. Appl. Polym. Sci. 2011, 120, 1707-1712.
[12] a) J. A. Mikroyannidis, Appl. Polym. Sci. A Polym. Chem. 1992, 30, 125-
132; b) M. Oikawa, M. Ikoma, M.; Sasaki, Tet. Lett. 2005, 46, 415-418;
c) F. I. Zubkov, I. K. Airiyan, J. D. Ershova, T. R. Galeev, V. P. Zaytsev,
E. V. Nikitina, A. V. Varlamov, RSC Adv. 2012, 2, 4103-4109.
[13] S. Higson, F. Subrizi, T. D. Sheppard, H.C. Hailes, Green Chem. 2016,
18, 1855-1858.
[14] a) J. C. Kim, T. W. Kim, Y. Kim, R. Ryoo, S. Y. Jeong, C. U. Kim, Appl.
Catal. B Env. 2017, 206, 490-500; b) R. Y. Rohling, E. Uslamin, B. Zijlstra,
I. C. Tranca, I. A. W. Filot, E. J. M. Hensen, E. A. Pidko, ACS Catal. 2018,
8, 760-769; c) Y. P. Wijaya, D. J. Suh, J. Jae, Catal. Commun. 2015, 70,
12-16; d) Y. P. Wijaya, H. P. Winoto, Y. Park, D. J. Suh, H. Lee, J. Ha, J.
Jae, Catal. Today 2017, 293-294, 167-175; e) C. L. Williams, C. C.
Chang, P. Do, N. Nikbin, S. Caratzoulas, D. G. Vlachos, R. F. Lobo, W.
Fan, P. J. Dauenhauer, ACS Catal. 2012, 2, 935-939; f) C. C. Chang, S.
K. Green, C. L. Williams, P. J. Dauenhauer, W. Fan, Green Chem. 2014,
16, 585-588; g) D. Wang, C. M. Osmundsen, E. Taarning, J. A. Dumesic,
ChemCatChem 2013, 5, 2044-2050; h) I. F. Teixeira, B. T. W. Lo, P.
Kostetskyy, M. Stamatakis, L. Ye, C. C. Tang, G. Mpourmpakis, S. C. E.
Tsang, Angew. Chem. 2016, 128(42), 13255-13260; Angew. Chem. Int.
Ed. 2016, 55, 13061-13066; i) M. Shiramizu, F. D. Toste, Chem. Eur. J.
2011, 17, 12452-12457; j) J. M. Fraile, J. I. Garcia, M. A. Gomez, A. de
la Hoz, J. A. Mayoral, A. Moreno, P. Prieto, L. Salvatella, E. Vaszquez,
Eur. J. Org. Chem. 2001, 15, 2891-2899; k) X. Ding, S. T. Nguyen, J. D.
Williams, N. P. Peet, Tet. Lett. 2014, 55, 7002-7006.
Acknowledgements
The authors are grateful to the CNRS, the University of Poitiers,
the Région Nouvelle Aquitaine, and SOLVAY for financial support.
The International Consortium on Eco-conception and Renewable
Resources (FR CNRS INCREASE 3707) and the chair
“TECHNOGREEN” are also acknowledged for their funding?
10
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002111
ChemSusChem
FULL PAPER
[15] a) J. J. Pacheco, M. E. Davis, Proc. Nat. Acad. Sci. USA. 2014, 111,
83636-8367; b) J. J. Pacheco, J. A. Labinger, A. L. Sessions, M. E. Davis,
ACS Catal. 2015, 5, 5904-5913; c) M. Orazov, M. E. Davis, Chem. Sci.
2016, 7, 2264-2274.
[16] a) Y. Hu, N. Li, G. Li, A. Wang, Y. Cong, X. Wang, T. Zhang, Green Chem.
2017, 19, 1663-1667; b) Y. Hu, N. Li, G. Li, A. Wang, Y. Cong, X. Wang,
T. Zhang, ChemSusChem 2017, 10, 2880-2885.
[17] M. Colonna, C. Berti, M. Fiorini, E. Binassi, M. Mazzacurati, M. Vanini, S.
Karanam, Green Chem. 2011, 13, 2543-2548.
[18] a) R. Mariscal, P. Maireles-Torres, M. Ojeda, I. Sádaba, M. López
Granados, Energy Environ. Sci. 2016, 9, 1144-1189; b) X. Li, P. Jia, T.
Wang, ACS Catal. 2016, 6(11), 7621-7640.
[19] a) L. Salem, J. Am. Chem. Soc. 1968, 90, 543−552; b) G. Klopman, J.
Am. Chem. Soc. 1968, 90, 223−234.
[20] a) A. E. Reed, F. Weinhold, J. Chem. Phys. 1983, 78, 4066-4073; b) A.
E. Reed, R. B. Weinstock, F. Weinhold, J. Chem. Phys. 1985, 83, 735-
746.
[21] B. A. Trofimov, Sulfur Reports 1992, 11(2), 207-231
[22] D. Yepes, J. Valenzuela, J. I. Martinez-Aray, P. Pérez, P. Jaque, Phys.
Chem. Chem. Phys. 2019, 21, 7412-7428.
[23] W. M. Wah, Quantum Chemical Calculations of Sulfur-Rich Compounds
in ElementalSulfur and Sulfur-Rich Compounds II; Elemental sulfur and
sulfur-rich compounds; 2003, Springer
11
This article is protected by copyright. All rights reserved.

10.1002/cssc.202002111
ChemSusChem
FULL PAPER
Entry for the Table of Contents
- H2O
meta
ortho
Electronic interaction
Steric interaction
We report here the synthesis of renewable aromatics from bio-based furfural and various alkenes, using a Diels-Alder/aromatization
sequence. This investigation aims at rationalizing, and even predicting, not only the feasibility of the Diels-Alder reaction but also the
effect of substituents on the ortho : meta selectivity.
12
This article is protected by copyright. All rights reserved.