Angewandte

A Journal of the Gesellschaft Deutscher Chemiker

International Edition Chemie www.angewandte.org

Accepted Article

Title: Chemo- and regio-selective additions of nucleophiles to cyclic

carbonates for the preparation of self-blowing non-isocyanate
polyurethane foams

Authors: Florent Monie, Bruno Grignard, Jean-Michel Thomassin,

Raphaël Mereau, Thierry Tassaing, Christine Jerome, and
Christophe Detrembleur

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: Angew. Chem. Int. Ed. 10.1002/anie.202006267

Link to VoR: https://doi.org/10.1002/anie.202006267

Angewandte Chemie International Edition 10.1002/anie.202006267

RESEARCH ARTICLE
Chemo- and regio-selective additions of nucleophiles to cyclic
carbonates for the preparation of self-blowing non-isocyanate
polyurethane foams
Florent Monie, [a] Bruno Grignard, [a] Jean-Michel Thomassin, [a] Raphael Mereau, [b] Thierry Tassaing, [b]
Christine Jerome, [a] Christophe Detrembleur *[a]
[a] Center for Education and Research on Macromolecules (CERM), CESAM Research Unit, University of Liège, Department of Chemistry, Sart-Tilman, B6A,
4000 Liège, Belgium

Accepted Manuscript
E-mail: [email protected]
[b] Institut des Sciences Moléculaires (ISM), UMR5255 CNRS, Université de Bordeaux, 351 Cours de la libération, F-33405 Talence Cedex, France

Supporting information for this article is given via a link at the end of the document.((Please delete this text if not appropriate))

Abstract: Polyurethane (PU) foams are indisputably daily copolymerizing CO2-based building blocks.3 Amongst them,
essential materials found in many applications, notably for comfort the step-growth copolymerization of di- or polyamines with 5-
(e.g. matrasses) or energy saving (e.g. thermal insulation). Today, membered polycyclic carbonates, easily synthesized by the
greener routes for their production are intensively searched for [3+2] coupling of CO2 with epoxides,4 is growing rapidly and
avoiding the use of toxic isocyanates. We now describe an easily gives access to poly(hydroxyurethane)s (PHUs).5 Unlike
scalable process for the simple construction of self-blown isocyanates, cyclic carbonates however do not decompose
isocyanate-free PU foams by exploiting the organocatalyzed spontaneously by simple addition of water to release CO2 as
chemo- and regio-selective additions of amines and thiols to easily the blowing agent, rendering impossible the formation of self-
accessible cyclic carbonates. These reactions are first validated on blowing PHU foams. To date, foaming procedures for PHUs
model compounds and rationalized by DFT calculations. Various are rare and very limited in scope. The first strategy was
foams are then prepared and characterized in terms of morphology reported by Caillol by thermal decomposition of Momentive
and mechanical properties, and the scope of the process is MH15, i.e. poly(methylhydrogenosiloxane), into dihydrogen (a
illustrated by modulating the composition of the reactive highly flammable gas) that expanded the PHU matrix during its
formulation. Due to the impressive diversity and accessibility of the formation.6 Another approach reported by Mulhaupt consisted
main components of the formulations, this new robust and solvent- in adding Solkane, i.e. a fluorocarbon physical blowing agent,
free process is opening avenues for constructing more sustainable that foamed the PHU matrix by evaporation during the
PU foams, and offers the first realistic alternative to the traditional polymerization.7 Rigid PHU foams with open cells porosity
isocyanate route. were designed by using sodium bicarbonate as chemical
blowing agent at high temperature (>200°C)8, or by the addition
of maleic acid and glutaraldehyde to a PHU formulation (the
foaming mechanism is however not clear in this case).9
Introduction Only two relevant works have described the use of CO2
as blowing agent for the production of PHU foams without
Polyurethanes (PUs) foams, with a world-wide requiring the addition of external chemical foaming agents. The
production expected to reach 12.7 million tons by 2024,1 are first one reported the fabrication of microcellular PHU foams
key materials of our modern life with a broad range of with promising heat insulating properties by the supercritical
applications for comfort and wellness (furniture, mattresses, CO2-assisted batch foaming technology.10 In that case, pre-
seats…), shock absorption, thermal/acoustic insulation or as synthesized PHUs thermoplastics were impregnated by CO2 at
sealants. Industrially, PUs are made by polyaddition of di- or the supercritical state and, upon release of the pressure, CO2
polyisocyanates and di- or polyols.2 Their foaming is generally was expanding the polymer matrix. In 2018, the first CO2 self-
induced by hydrolysis of the isocyanates with concomitant blowing PHU foam was reported by North by exploiting a
formation of CO2. The in-situ release of this gas expands the specific bis(cyclic carbonate) monomer, i.e. sorbitan bis-
polymer during its formation, leading to foams with properties carbonate, that underwent an intramolecular side-reaction
that depend on the composition of the reactive formulation.2 provoking the partial decarboxylation of the monomer during
Due to the large diversity of polyols that can be used, rigid to thermally driven aminolysis.11 The CO2 release was however
soft foams can be easily produced by this simple process. observed at a low level and this strategy was limited to sorbitan
However, the health concerns and regulation changes bis-carbonate, therefore strongly limiting the scope of the
associated to the use of toxic isocyanates are motivating the process. The quest for a universal strategy to producing self-
industries and the scientists to explore novel routes to design blowing PHUs that is versatile, simple, compatible to industrial
isocyanate-free PU foams. The recent advances in the processes (reactive extrusion foaming, reactive molding,
chemical transformations of CO2 opened new opportunities to foaming in place, etc.) and that mimics the conventional PUs
solve this challenge and multiple pathways are now accessible foaming methodology (by the in-situ generation of CO2 from
to fabricate PUs by directly using CO2 as a comonomer or by

1
This article is protected by copyright. All rights reserved.
Angewandte Chemie International Edition 10.1002/anie.202006267

RESEARCH ARTICLE
one of the comonomers) is highly desirable for the next The self-blowing PHU foaming method lies on two
generation of greener PU foams. competitive reactions, i.e. the aminolysis of 5CC to form the
To tackle this challenge, we took inspiration from the urethane linkages and to construct PHU, and the S-alkylation
dialkylcarbonates chemistry responding to the Pearson’s Hard- of thiols to generate the blowing agent (CO2) while creating
Soft Acid-Base (HSBA) theory. Dialkylcarbonates display two thioether linkages. The rates and selectivity of the reactions
electrophilic sites, the carbonyl and methylene carbons, with a have to be adjusted for controlling both the PHUs synthesis
reactivity that can be discriminated by the Hard-Soft character and foaming. Ideally, both reactions have to take place in a
of the nucleophile.12 Hard nucleophiles (e.g. amines) are range of temperature between 80 °C and 120 °C, which are
undergoing carbonyl attack whereas methylene attack occurs common operative conditions to synthesize PHUs.5 The main
with soft nucleophiles (e.g. thiols) (Scheme 1). The same challenge in self-blowing processes is to form polymer chains
situation is also noted for 5-membered cyclic carbonates with a sufficient viscosity and a crosslinking degree prior to the
(5CCs) but under harsher reaction conditions. We therefore release of the blowing agent (CO2) in order to avoid the
postulated that combining two types of nucleophiles with collapse of the foam. In our case, this means that the 5CC ring-

Accepted Manuscript
distinct hard and soft characters within a single PHU opening by the amine should be faster that the decarboxylation
formulation should lead to two competitive and selective 5CC promoted by the thiol. The S-alkylation of thiols with
ring-opening pathways, i.e. the acylation of the amine to form dialkylcarbonates has been reported but was slow and
β-hydroxyurethane linkages (via carbonyl attack, Scheme 1) generally realized under demanding conditions (T = 100-
and the alkylation of the soft nucleophile with concomitant in- 180 °C, K2CO3 as catalyst).12c These operative conditions are
situ release of CO2 (via methylene attack, Scheme 1). Although not compatible with the PHU chemistry as they will favor the
the S-alkylation of thiols was sparsely reported with occurrence of side-reactions such as the urea formation and
dialkylcarbonates,12c we identified thiols as ideal soft provoke a deviation of the stoichiometry which is detrimental
nucleophiles to combine with amines (hard nucleophiles) for for the step-growth polymerization.5i, 13
fabricating PHU foams from reactive formulations. Identifying a single catalyst that is able to catalyze both
the S-alkylation and the aminolysis is then a prerequisite for
Additions to dialkylcarbonates
H
the formation of self-blowing PHU. Organobases such as DBU,
carbonyl attack Nu = NH2
O R
N O R’ + R’ OH DBN, TBD, DMAP, DABCO are potential candidates and some
R’ O O R’ Nu-R
O urethane alcohol of them are known to catalyze the aminolysis of 5CC.14 We
therefore investigated and compared the organobase
Nu = SH S R’
methylene attack
R + R’ OH + CO2 catalyzed additions of a diamine (2,2'-(Ethylenedioxy)bis-
thioether alcohol
ethylamine) and a dithiol (2,2'(ethylenedioxy)diethanethiol) of
Additions to cyclic carbonates
R’
similar structure on propylene carbonate (PrC) as a model 5CC
H H
carbonyl attack Nu = NH2
R
N O
OH + O
R
N
OH
(Figure 1). These reactions were carried out under equimolar
O
O O R’ conditions between the reactive groups without any solvent
O O Nu-R hydroxyurethane regioisomers
R’ and under ambient atmosphere at 80 °C in order to approach
R’ Nu = SH
R
S
OH + CO2 reactions conditions that will be implemented for the foaming.
methylene attack
hydroxythioether The reactions were monitored by 1H-NMR spectroscopy as
Scheme 1. (Regio)selective addition of amines or thiols to dialkylcarbonates described in Supporting Information (Figure S1 and S2).
or 5-membered cyclic carbonates

In this work, we explore the formation of self-blown PHU

foams by directing the concomitant selective aminolysis of the
cyclic carbonate and the decarboxylative S-alkylation of a thiol
100 100
from reactive formulations containing amines, thiols and cyclic
carbonates. First, we investigate the organocatalyzed regio- 80 80
PrC conversion (%)

PrC conversion (%)

no catalyst
selective addition of model thiols to five-membered cyclic
60 60 DMAP
carbonates that leads to the decarboxylative formation of
DABCO
hydroxythioethers or poly(hydroxythioether)s, and we 40 40 DBN
approach the reaction mechanism by combining kinetics DBU
20 20
studies and DFT modeling. We then report how to exploit this TBD

reaction for the preparation of microcellular PHU foams and we 0 0

0 50 100 150 0 50 100 150
illustrate the scope of this process. We also show that, beside Time (min) Time (min)
the generation of the blowing agent during the PHU formation, Figure 1. Conversion of propylene carbonate (PrC) with the reaction time
the thiol/cyclic carbonate reaction provides hydroxythioether for (a) the S-alkylation of thiol and (b) the aminolysis catalyzed by
linkages that contribute to the polymer matrix construction and organobases (DMAP, DABCO, TBD, DBN or DBU). Conditions: [SH] or
improve the thermal properties of the final material. [NH2]/[PrC] = 1, organobase = 5 mol% compared to PrC, 80 °C, bulk.

Under catalyst-free conditions, the thiol did not react with PrC,
Results and Discussion which is in sharp contrast to the amine that provided the
expected two hydroxy-urethane regioisomers (80 % PrC
Model reactions conversion in 1h).14 By adding weak bases such as DMAP or
DABCO (5 mol% compared to PrC), the S-alkylation of the thiol

This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition 10.1002/anie.202006267

RESEARCH ARTICLE
by decarboxylative addition onto PrC was observed, however ring-opening by the amine is easier and faster than the
was slow with respectively 10 or 30 % conversion in 2 h. decarboxylation promoted by the thiol.
Importantly, this reaction was accelerated by using superbases
with conversions as high as 90 % in 2 h with DBN or 96 % with
DBU in 1 h, with the selective formation of the S-alkylated
regioisomer that originates from the thiol attack at the
electrophilic methylene site of PrC. Although DMAP did not
affect the PrC aminolysis, the addition of DABCO, DBU and
DBN significantly accelerated the reaction with the production
in all cases of the two hydroxyurethane regioisomers. The
comparison of the kinetics profiles for the two considered
reactions, i.e. the S-alkylation of thiol and the aminolysis of PrC,
gave clear insights into the feasibility of the PHU foaming

Accepted Manuscript
methodology. First, whatever the catalyst, the aminolysis of
PrC was always faster than the S-alkylation of the thiol which
is a prerequisite to tailor PHU chains before the in-situ
formation of the blowing agent (CO2). Second, by the choice of
the catalyst, the rates of these two competitive reactions can
be adjusted, which constitutes an additional lever for tuning the
viscosity of the reaction medium prior PHU expansion. From Figure 2. Mechanistic insight of the DBU-catalyzed aminolysis (black) and
S-alkylation (red) of propylene carbonate at the M062X/6-311G(d,p) level.
the comparative kinetic studies, DBU was selected in the next The Gibbs free energies are expressed in kcal.mol-1 (black: aminolysis; red:
studies as it catalyzed both reactions under reasonable period S-alkylation).
of time (Figure S3). It is also operative over a broad range of
temperature (25-100 °C) for the S-alkylation of the thiol (Figure Copolymerization of bis(cyclic carbonate) with dithiols
S4). and/or diamines
We then investigated the influence of adding a dithiol to
DFT calculations a formulation producing PHU, thus containing a mixture of a
bis(cyclic carbonate) and a diamine. Indeed, although the step-
Model reactions carried out in the presence of DBU were growth copolymerization of a diamine with a bis(cyclic
confronted to mechanistic DFT calculations at the M062X/6- carbonate) is largely described in the literature for the
311G(d,p) level of theory (see ESI5 for details). Figure 2 synthesis of PHU, the terpolymerization with dithiols or even
depicts the two competitive pathways (i.e. aminolysis and S- the direct copolymerization of dithiols with bis(cyclic
alkylation) catalyzed by DBU using simplified model amine and carbonate)s was never reported. Series of DBU-catalyzed
thiol of similar structure (CH3-O-CH2-NH2 and CH3-O-CH2-SH). polymerizations were performed under solvent-free conditions
In the case of the aminolysis, the overall reaction proceeds in at 80 °C under ambient atmosphere for 24 h (5 mol% DBU vs
2 steps. From the van der Walls complex (-2.3 kcal.mol-1), the the bis(cyclic carbonate)). Note that ambient atmosphere was
initial step corresponds to the nucleophilic attack of the amine used here to fit the conditions that are conventionaly used for
onto the carbonate group of 5CC accompanied by a proton foaming PUs.
transfer from the amine towards the sp2 N atom of DBU. This Scheme 2 illustrates the structure of the monomers used
step is the rate determining step with a barrier height of 22.3 for this study, i.e. the bis(cyclic carbonate) C1 (4,4’-[1,4-
kcal.mol-1. The resulting cyclic amino alkoxide anion Butanediylbis(oxymethylene)]bis(1,3-dioxolan-2-one)), the
(interacting with DBUH+ by H-bond) is then encountering a ring diamine A1 (1,2-bis(2-aminoethoxy)ethane) and the dithiol T1
opening by C-O bond cleavage and a simultaneous proton (2,2’-(ethylenedioxy)diethanethiol). Various comonomer molar
transfer between DBUH+ and the adjacent O atom of the ratios [C1]/[A1]/[T1] were tested from 1/1/0 (to produce PHU)
carbonate to product the target hydroxyurethane and the to 1/0/1 (to yield poly(hydroxythioether)), with intermediate
regeneration of DBU with a barrier height of 6.9 kcal.mol-1. compositions in order to tentatively synthesize unprecedented
The S-alkylation reaction is also proceeding in two elementary poly(hydroxyurethane-co-hydroxythioether)s. The results are
steps. The first one consists in the initial van der Walls complex collected in Table 1. The macromolecular characteristics of the
(-1.6 kcal.mol-1) that reacts by a methylene attack of the thiol polymers were determined by size exclusion chromatography
onto the non-substituted carbon atom of the 5CC while the thiol (SEC) on the crude product in order to avoid any polymer
proton is transferred to the sp2 N atom of DBU. This step fractionation during purification (Figure S5).
exhibits a barrier height of 30.1 kcal.mol-1, making it the rate
determining step of the S-alkylation reaction. In a second step,
a decarboxylation of the resulting linear carbonate anion
occurs together with the formation of the hydroxythioether by
proton transfer from DBUH+ and then the regeneration of DBU
with a barrier height of 5.5 kcal.mol-1.
Therefore, the barrier heights of the rate determining step
of the aminolysis (22.3 kcal.mol-1) and the S-alkylation (30.1
kcal.mol-1) reactions obtained from DFT calculations are in
good agreement with the kinetic trends observed, i.e the 5CC

This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition 10.1002/anie.202006267

RESEARCH ARTICLE
A1 O
OH
OH
H H
Catalyst Comonomers the 13C-NMR spectrum (Figure S6). The 1H-NMR spectrum of
O N N
Cat. (5 mol%) O R1 R2 N PC1T1 highlighted the formation of thioether moieties with the
O n R2
80 °C H 2N NH2 A1
Poly(hydroxyurethane) PC1A1 N two signals at 2.67 and 2.78 ppm corresponding to the
O O
HS
R2
SH T1 methylenes linked to the sulfur atom as well a methyne
O
R1
T1
R1 R2 resonance (-CH-OH) at 3.9 ppm. The S-alkylation of T1 via
S S
Cat. (5 mol%)
O
80 °C OH OH m
R1 = -CH2-O-(CH2)4-O-CH2- decarboxylative addition onto C1 was further confirmed by the
O
C1
O - CO2 Poly(hydroxythioether) PC1T1 R2 = -(CH2)2-O-(CH2)2-O-(CH2)2- absence in the 13C-NMR spectrum of a resonance typical of a
OH
carbonyl group that could have been produced via the
O OH
A1 + T1
O
H
N
H
N
formation of a thiocarbonate linkage. This characterization
O R1 R2 R1 R2
O R1 R2 S S S S
Cat. (5 mol%)
O OH OH OH
confirmed the selective addition of the thiol onto the methylene
80 °C O m
n
- CO2
OH site of the cyclic carbonate ring. The organocatalyzed S-
Poly(hydroxyurethane-co-hydroxythioether) PC1A1T1
alkylation of dithiols with bis(cyclic carbonate) is therefore
Scheme 2. Synthesis of poly(hydroxyurethane), poly(hydroxythioether) and
poly(hydroxyurethane-co-hydroxythioether) by DBU-catalyzed step-growth highly attractive to construct regio-regular

Accepted Manuscript
copolymerization. poly(hydroxythioether)s that are currently only accessible by
the exothermic LiOH driven thiol-epoxy click reaction.15
As expected the poly(hydroxyurethane) PC1A1 was
PC1A1T1 displayed the typical signals of the methylene or
formed by copolymerizing C1 with A1, however the relative
methyne groups of both the hydroxyurethane and
number average molar mass was rather low (Mn = 4,200
hydroxythioethers units at similar chemical shifts as the one
g/mol) as the result of the uncomplete C1 conversion (Table 1,
observed in PC1A1 and PC1T1. This observation confirms that
entry 1). By adding dithiol T1 to the C1/A1 formulation, almost
both A1 and T1 were inserted in the growing chains when the
full C1 conversion was noted under identical conditions, with
terpolymerization of C1, A1 and T1 was considered.
the formation of polymers of higher molar masses for all
compositions (Mn = 8,400-8,800 g/mol; Table 1, entries 2-4).
Thermal analysis of the copolymers
Importantly, the copolymerization of the dithiol T1 with the
The thermal stability and glass transition temperature of
bis(cyclic carbonate) C1 was also efficient with the formation
the polymers were investigated by thermogravimetric analysis
of a polymer characterized by a Mn of 9,200 g/mol (Table 1,
(TGA) and differential scanning calorimetry (DSC),
entry 5). No C1/T1 copolymerization was observed in the
respectively (Figure 4). Interestingly, the introduction of
absence of DBU, highlighting the importance of the
thioether units within the chains strongly improved the thermal
organobase for the thiol/cyclic carbonate reaction in agreement
stability of the polymers. The thermal degradation temperature
with the previous model reactions.
(Td5%) increased progressively with the thioether linkages
content and evolves from 222 °C for the pure PHU PC1A1 to
289 °C for the pure poly(hydroxythioether) PC1T1 (Table 1,
Figure 4A). This observation is consistent with a recent study
reporting the higher thermal stability of sulfur-containing
polyurethanes prepared by the isocyanate chemistry.16 DSC
curves (Figure 4B) showed clear endothermic transitions
corresponding to the Tg of the different polymers. The Tg values
were dependent on the A1/T1 molar content of the reactive
formulation and evolved from -15 °C for the pure PHU (PC1A1)
to -46 °C for the poly(hydroxythioether) (PC1T1) (Table 1). This
observation is linked to the interchain hydrogen-bonding
interactions favored at high content of urethane linkage which
reduced the chain mobility. The presence of only one Tg for
each polymer, whatever its A1/T1 composition, suggests that
the urethane and thioether linkages were relatively randomly
distributed in the polymer chains.
1
Figure 3. H-NMR spectrum of poly(hydroxyurethane) PC1A1,
poly(hydroxythioether) PC1T1 and a representative poly(hydroxyurethane-
co-hydroxythioether) PC1A1T1 with [A1]/[T1] = 1.

The microstructure of the poly(hydroxyurethane) PC1A1,

poly(hydroxythioether) PC1T1 and a representative
poly(hydroxyurethane-co-hydroxythioether) PC1A1T1 were
elucidated by 1H- (Figure 3) and 13C-NMR (Figure S6)
spectroscopy. PC1A1 displayed the typical resonances of the
two methylene groups -CH2-N(C(O)O- and -CH2-O(C(O)N-
adjacent to the urethane moiety at δ = 4.10-4.17 ppm and 3.37 Figure 4. (A) Thermogravimetric (TGA) and (B) dynamic scanning
ppm, as well as the methylene and the methyne groups close calorimetry (DSC) analyses of poly(hydroxyurethane) PC1A1,
to the alcohol function at 3.65 and 4.88 ppm, respectively. The poly(hydroxythioether) PC1T1 and poly(hydroxyurethane-co-
microstructure of PC1A1 was further confirmed by the hydroxythioether) PC1A1T1 with various [A1]/[T1] compositions.
presence of the carbonyl signal of the urethane at 157 ppm in

This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition 10.1002/anie.202006267

RESEARCH ARTICLE
Table 1. Poly(hydroxyurethane) PC1A1, poly(hydroxyurethane-co-hydroxythioether) PC1A1T1 and poly(hydroxythioether) PC1T1 prepared at 80 °C in the
presence of DBU: reaction conditions, macromolecular characteristics and thermal properties.

Entry [C1]/[A1]/[T1] Conv.a Mnb Mwb Đ Td5% Tg

(%) (g/mol) (g/mol) (°C) (°C)
1 1/1/0 95 4,200 6,000 1.4 222 -15
2 1/0.75/0.25 99 8,400 14,000 1.7 220 -18
3 1/0.5/0.5 99 8,600 14,700 1.7 230 -28
4 1/0.25/0.75 99 8,800 15,400 1.7 242 -38
5 1/0/1 99 9,200 17,400 1.9 289 - 46
Conditions: 80 °C, 24 h, catalyst loading = 5 mol% compared to the carbonate moiety, bulk.
a
Cyclic carbonate conversion determined by 1H-NMR spectroscopy on the crude samples
b
Determined by SEC using a polystyrene standard calibration and DMF with LiBr as eluant.

Preparation and characterizations of PHU foams O O

R

Accepted Manuscript
O S S
O O HO HO
O OH OH
O
In order to prepare CO2 self-blowing PHU foams with O
O O hydroxythioether linkage O
dimensional stability, crosslinking of the polymer chains has to O O
C2 NH NH
occur during the expansion of the foam. We therefore added a O DBU
R R
O NH NH
tris(cyclic carbonate) C2 (trimethylolpropane triscarbonate) to O
4 h, 100 °C
O O
foaming
the PHU formulation containing the diamine A1 and the dithiol H 2N
O
O
NH2
O O

hydroxyurethane linkage
T1. We worked under stoichiometric conditions in the A1
OH OH
OH
complementary reactive groups ([C2]/[A1]/[T1] = O O
O SH R
O
1/1.125/0.375; thus [5CC]/[NH2]/[SH] = 1). After HS O
OH
N
H
N
H
O

T1
homogenization of the mixture under solvent-free conditions
with DBU (5 mol% compared to C2) at room temperature for 2
min, the formulation was poured in a silicone mold. By heating
this formulation at 100 °C, a tacky-free foam was collected after
6 h. The foam was however heterogeneous with a large
distribution of pores (from 800 µm to 2.7 mm) (Figure 5), which
was assigned to a too low initial viscosity of the reactive
formulation. Analysis of the PHU foam by Raman spectroscopy
evidenced the disappearance of the S-H stretching band
around 2550 cm-1 of the thiol T1 and carbonate band at 1800
cm-1 of the cyclic carbonate C2 present in the initial formulation
(Figure S7). This analysis also demonstrated the presence of
two specific linkages in the foam, thioether and urethane, with
the new bands at 500 cm-1 for the C-S-C thioether bending and
at 1715 cm-1 for the C=O urethane one.
To improve the foaming, the formulation was pre-reacted
at 25 °C for a longer time in order to pre-polymerize PHU and
increase the viscosity prior to the decarboxylation promoted by
the thiol. Figure S8 shows that the viscosity indeed increased
progressively up to 715 Pa.s in 16 h due to the slow aminolysis
of C2 with the formation of oligohydroxyurethanes. The Figure 5. PHU formulation and SEM characterizations of self-blown foams
absence of a gel point (corresponding to the crossover point in made from C2, A1 and T1. Conditions: [C2]/[A1]/[T1] = 1/1.125/0.375, DBU
the time evolution of G’ and G" values) after 16 h suggested (5 mol% compared to 5CCs), mixing for 2 min at rt, curing at 100 °C for 4 h.
that the polymer was not crosslinked under these conditions.
The foaming was then thermally induced by a curing for 2 h at To demonstrate the robustness of the process but also to
100 °C followed by a temperature increase to 120 °C in 30 min tune the characteristics and properties of the foams, some
and an additional 1 h post-curing treatment at 120 °C (this additives were added to the formulation, such as a synthetic
optimal procedure has been identified and selected after clay (Laponite S482) or an epoxy-functionalized PDMS
screening different conditions). This foaming methodology oligomer (TEGOMER E-Si 2330). Laponite is a hydrophilic
enabled us to design self-blown PHU foams in a reproducible nanosheet clay that was chosen to reinforce the polymer matrix
manner (Figure 6). Although some shrinkage was observed at and to act as a nucleating agent that is expected to favor the
the end of the foaming process as soon as the foam was formation of homogeneous foams with small cells size. The
removed from the oven, the material recovered its expended PDMS oligomers was selected as a surface acting agent for
shape either spontaneously after 1 to 2 h at 25 °C or by heating the stabilization of the cells during their growth. All formulations
the sample at 80 °C for 1 min. At this stage, we do not added by clay (1.5 wt%), PDMS (0.01 eq compared to C2) or
understand this phenomena but the process is perfectly by the two additives (clay + PDMS) provided foams after curing
reproducible and systematically furnished homogeneous PHU
foams.

This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition 10.1002/anie.202006267

RESEARCH ARTICLE

Accepted Manuscript
Figure 6. SEM characterization of self-blown PHU foams a) additives-free PHU foam, b) PHU foam loaded with 1.5 wt% clay, c) PHU foam loaded with 0.1 eq
compared to C2 of Tegomer Si-C 2230 and d) PHU foam loaded with both PDMS and clay.

Table 2. Composition and characteristics of the self-blown PHU foams.

Entry PHU Foam formulation PHU Foam characteristics Thermo-mechanical properties of PHU foams

[C2]/[A1]/[T1]a Additives Density Pore size Ah/Ac Cell density (Ncell) Td5% Tg Stress c Recovery
b
(g/cm³) (mm) (cells/cm³) (°C) (°C) (kPa) Time d (s)
1 1/1.125/0.375 None 0.166 +/- 0.007 0.98 +/- 0.35 17.2 1,620 +/- 110 248 3.6 8.2 +/- 0.2 30 +/- 2
2 1/1.125/0.375 Clay 0.207 +/- 0.011 0.62 +/- 0.23 23.7 4,570 +/- 50 246 7.9 11.4 +/- 0.7 17 +/- 0.5
3 1/1.1325/0.3775 PDMS 0.196 +/- 0.005 0.87 +/- 0.23 9.7 2,280 +/- 310 247 3 11.2 +/- 0.9 10 +/- 0.2
4 1/1.1325/0.3775 Clay + PDMS 0.180 +/- 0.005 0.60 +/- 0.21 11.5 5,070 +/- 240 244 2.4 8.7 +/- 0.5 8.4 +/- 0.1
a
molar ratio adjusted to take into account the epoxy groups of Tegomer in entries 3 and 4
b
Ncell = (nM/A)3/2ρs/ρp with n = the number of cells in the SEM image (average value on 6 SEM images), M the magnification, A the surface area of the image
(mm²) and ρs and ρp the solid and the foamed sample densities. The solid density was estimated to 0.98 +/- 0.02 g/cm³ on non-foamed samples
c
at 50 % of strain
d
for a 99 % shape recovery

This article is protected by copyright. All rights reserved.

Angewandte Chemie International Edition 10.1002/anie.202006267

RESEARCH ARTICLE
following the procedure described above. The foams
morphology, the cells size and distribution, as well as the open-
to-close cell ratio were determined by scanning electron
microscopy (SEM) characterizations. Figure 6 illustrates the
morphologies of the self-blown PHU foams and their
characteristics are summarized in Table 2. In all cases,
homogeneous microcellular foams with an open cell
morphology were formed with a density ranging from 0.166 to
0.207 g/cm3. In the absence of additives, foams with a
homogeneous cells size distribution and a large average cell
diameter of 0.98 ± 0.35 mm were obtained. The addition of
PDMS reduced the pore size to 0.87 ± 0.23 mm and their Figure 7. Shape recovery rate of the PHU foams after a 50% deformation
distribution. As expected, loading the formulation with clay

Accepted Manuscript
as determined by dynamic mechanical analysis (DMA).
further decreased the cell size to 0.62 +/- 0.23. PHU foams
loaded with both clay and PDMS were also highly Finally, other commercially available diamines, i.e. 1,6-
homogeneous with a pore size of 0.60 ± 0.21 mm (Figure S9). hexamethylene diamine (HMDA, A2) and m-xylylenediamine
The PHU foam with the lowest density of 0.166 g/cm³ was (MXDA, A3) were tested in combination with C2 and T1 in
obtained with the additive-free formulation. The other foams order to illustrate the robustness of the new methodology and
were characterized with slightly higher densities ranging from the possible variations of the PHU structures. As the intrinsic
0.207 g/cm3 (with clay) to 0.180 g/cm3 (with clay + PDMS). reactivity of the amine with the cyclic carbonate is dependent
The open-to-close cells ratio was then evaluated by SEM on their respective structural features, the foaming protocols
images treatment following the methodology reported by (curing steps) were adjusted to identify the most appropriate
Fitzgeral.17 Cell-faces area Ac and cell-face holes area Ah were foaming window of the reactive PHU formulations (these
measured, and the Ah/Ac ratio was calculated[31] and is procedures are detailed in Supporting information, ESI 13).
reported in Table 2. The Ah/Ac ratio values show that PDMS Under optimized foaming conditions, the reactive mixture
reduced the open porosity of the foams from 17.2 for an [C2]/[A2]/[T1] = 1/1.125/0.375 added by 5 mol% of DBU
additive-free PHU foam to 9.7 in the presence of PDMS. At the (compared to the carbonate moieties) provided a PHU foam
opposite, clay increased the content of open pores to 23.7. For that did not collapse after preparation. The PHU material
PHU foams containing both additives, the Ah/Ac ratio was displayed an open cell microcellular morphology with a cell size
estimated to 11.2 showing that the increase of open porosity average of 0.26 ± 0.085 mm (Figure 8A , Table S1, entry 1)
induced by clay was counter-balanced by PDMS. and a foam density of 0.320 ± 0.012 g/cm3. For a similar
The thermal degradation of the foams was evaluated by formulation with A3 in replacement of A2 ([C2]/[A3]/[T1] =
TGA. All samples depicted a similar degradation profile (Figure 1/1.125/0.375), a foam with an open cells size gradient from
S10) with a degradation temperature (Td5%) at around 245 °C 0.3 mm to 3.5 mm was collected with a density of 0.210 +/-
(Table 2). There was therefore no significant influence of the 0.041 g/cm³ (Figure 8B, Table S1, entry 2). Blending A3 with
additive on the degradation temperature of the foam, at least A2 (with a [A3]/[A2] molar ratio of 80/20) enabled to preparing
within the additive loading range that was considered in this homogeneous foams that displayed open cells of 0.62 +/- 0.11
work. DSC analysis showed an endothermic transition for all mm and a density of 0.250 +/- 0.02 g/cm³ (Figure 8C, Table S1,
samples, with Tg values between 2.4 and 7.9 °C. The rather entry 3). All PHU foams made from A2, A3 or a blend of A2
similar values are not surprising as the Tg is mainly guided by and A3 showed similar thermal behavior with Tg and Td5%
the choice and content of the comonomers that are almost values of 20.5-24.6 °C and 232-238 °C, respectively (Table S1).
identical in all formulations. Preliminary mechanical properties Note that the Tg values close to the room temperature
were realized by dynamic mechanical analysis (DMA). Figure prevented us to determine the mechanical properties of the
S11 illustrates the compression stress-strain curves of the foams with sufficient accuracy.
different PHU foams. Whatever the foam formulation, the
stress to apply on the PHU foams to reach a 50 % strain was Conclusion
found in the same range with values between 8.2 +/- 0.2 to
11.4 +/- 0.7 kPa which reflected the similar foam densities This work describes a general strategy to prepare
values of the foams (0.166 – 0.208 g/cm³) (Table 2). Creep microcellular self-blown isocyanate-free polyurethane foams
experiments were then realized to evaluate the ability of the by exploiting chemo- and regio-selective additions of amines
foams to recover their initial shape after compression. For each and thiols to 5-membered cyclic carbonates (5CCs), resulting
sample, three compression-release cycles at a fixed strain of in an attractive alternative to conventional PU foams. The main
50 % were applied and the time for the foams to recover their challenge of this work was to demonstrate that poly(cyclic
initial shape was evaluated (Figure 7). After the third carbonate)s used in the preparation of poly(hydroxyurethane)s
compression, all samples recovered over 99 % of their initial (PHU) by copolymerization with diamines could also be
shape in less than 30 s, qualifying these materials as flexible involved in the matrix foaming by generation of the blowing
(Table 2). The PHU foams containing the PDMS additive agent (CO2) by decarboxylation. We first investigated on model
displayed the fastest shape recovery rate in only 8.4 to 10 compounds the reaction conditions that promoted the
seconds, in line with the high flexibility of silicones, even at a formation of urethane linkages by aminolysis of 5CC and the
very low PDMS loading. release of CO2 by the attack of the thiol to the methylene site
of 5CC with the formation of thioether

7
This article is protected by copyright. All rights reserved.
Angewandte Chemie International Edition 10.1002/anie.202006267

RESEARCH ARTICLE

Accepted Manuscript
Figure 8. SEM characterizations of self-blown PHU foams. A) PHU foam made from [C2]/[A2]/[T1] = 1/1.125/0.375, B) PHU foam made from [C2]/[A3]/[T1] =
1/1.125/0.375, C) PHU foam made from [C2]/[Ax]/[T1] = 1/1.125/0.375 with [Ax] being a blend of A3 and A2 with a molar ratio [A3]/[A2] = 80/20.

linkages. While the aminolysis was possible at 80°C without cost effective fashion. Due to the impressive diversity and
any catalyst, the decarboxylation promoted by the thiol accessibility of these components and all possible additives,
required the use of an organobase (e.g. DBU) to occur. When this new solvent-free process is opening new avenues for
this organocatalyst was added at 5 mol%, the two constructing more sustainable PU foams with various
chemoselective additions were possible, with a faster properties, and offers the first realistic alternative to the
aminolysis. DFT calculations of the mechanism of the two traditional isocyanate route.
competitive reactions enabled to rationalize the difference of
reactivity between the amine and the thiol for the addition to
5CC. Applied to di(cyclic carbonate)s, these reactions give Acknowledgements
access to new polymers containing both urethane and
thioether linkages by DBU-catalyzed copolymerization with
The authors of Liège thank the Fonds National de la
diamines and dithiols.
Recherche Scientifique (FNRS) and the Fonds
The PHU foaming was then considered from reactive
formulations containing a tri(cyclic carbonate), a diamine and Wetenschappelijk Onderzoek – Vlaanderen (FWO) for funding
a dithiol for the proof of concept. In the presence of DBU and the EOS project n°O019618F (ID EOS: 30902231). C.
an appropriate mild thermal curing, the aminolysis of 5CC Detrembleur is FNRS Research Director. The authors thank
provided the PHU matrix that expanded by the decarboxylation Cedric Malherbe for providing the Raman spectra.
of 5CC promoted by the thiol. The formation of thioether bonds
during the formation of the blowing agent participated in the Keywords: non-isocyanate polyurethane • foams • CO2 •
polymer network construction and fixed the thiol to the polymer cyclic carbonate. thiols
matrix. Homogeneous microcellular PHU foams with open cell
morphology were easily prepared under ambient atmosphere. [1] www.grandviewresearch.com/press-release/global-polyurethane-foam-
Highly flexible foams were prepared as a proof of concept, and market
[2] a) E. Delebecq, J.-P. Pascault, B. Boutevin, F. Ganachaud, Chem. Rev.
we demonstrate that the process is compatible to the addition 2013, 113, 80-118; b) K. Ashida, in Polyurethane and Related Foams:
of additives such as clay and/or PDMS. The main components Chemistry and Technology, (Eds. B. Raton), CRC press, Taylor & Francis,
of the formulation (the amines and the thiols) are accessible at 2007
[3] a) B. Grignard, S. Gennen, C. Jérôme, A.W. Kleij, C. Detrembleur, Chem.
large scale and low cost, and the cyclic carbonates are easily Soc. Rev., 2019, 48, 4466-4514; b) N. Yadav, F. Seidi, D. Crespy, V. D'Elia,
prepared by coupling CO2 to the corresponding epoxides in a ChemSusChem, 2019, 12, 724-754, c) Y. Li, Y.-Y. Zhang, L.-F. Hu, X.-H.
Zhang, B.-Y. Du, J.-T. Xu, Prog. Polym. Sci., 2018, 82, 120–157 ; d) A.

8
This article is protected by copyright. All rights reserved.
Angewandte Chemie International Edition 10.1002/anie.202006267

RESEARCH ARTICLE
Kamphuis, F. Picchioni, P. Pescarmona, Green Chem., 2019, 21, 406–448 ;
e) Y. Xu, L. Lin, M. Xiao, S. Wang, A. T. Smith, L. Sun, Y. Meng, Prog.
Polym. Sci., 2018, 80, 163–182 ; f) S. Gennen, B. Grignard, T. Tassaing, C.
Jérôme, C. Detrembleur, Angew. Chem. Int. Ed., 2017, 35, 10394-10398
[4] a) A.J. Kamphuis, F. Picchioni, P.P. Pescarmona, Green Chem., 2019,
21, 406-448; b) M. Alves, B. Grignard, R. Mereau, C. Jerome, T. Tassaing,
C. Detrembleur, Catal. Sci. Technol. 2017, 7, 2651-2684; c) G. Fiorani, W.
Guo, A. W. Kleij, Green Chem. 2015, 17, 1375-1389; d) C. Martín, G.
Fiorani, A. W. Kleij, ACS Catal. 2015, 5, 1353-1370; e) H. Büttner, L.
Longwitz, J. Steinbauer, T. Werner, Top. Curr. Chem. 2017, 375:50, 1-56
[5] a) L. Maisonneuve, O. Lamarzelle, E. Rix, E. Grau, H. Cramail, Chem.
Rev., 2015, 115, 12407-12439 ; b) F. Bobbink, A. van Muyden, P. Dyson,
Chem. Commun., 2019, 55, 1360-1373 ; c) S. Panchireddy, B. Grignard, J.-
M. Thomassin, C. Jerome, C. Detrembleur, Polym. Chem., 2018, 9, 2650-
2659 ; d) G. Beniah, D. Fortman, W. Heath, W. Dichtel, J. Torkelson,
Macromolecules , 2017, 50, 4425-4434 ; e) S. Gennen, B. Grignard, J.-M.
Thomassin, B. Gilbert, B. Vertruyen, C. Jerome, C. Detrembleur, Eur. Polym.
J., 2016, 84, 849-862 ; f) C. Carre, Y. Ecochard, S. Caillol, L. Averous,

Accepted Manuscript
ChemSusChem, 2019, 12, 3410-3430 ; g) F. Magliozzi, G. Chollet, E. Grau,
H. Cramail, ACS Sustainable Chem. Eng., 2019, 7, 17282-17292 ; h) O.
Lamarzelle, P.-L. Durand, A.-L. Wirotius, G. Chollet, E. Grau, H. Cramail,
Polym. Chem., 2016, 7, 1439-1451 ; i) A. Bossion, R. Aguirresarobe, L.
Irusta, D. Taton, H. Cramail, E. Grau, D. Mecerreyes, C. Su, G. Liu, A.
Muller, H. Sardon, Macromolecules, 2018, 51, 5556-5566 ; j) L. Poussard,
J. Mariage, B. Grignard, C. Detrembleur, C. Jérôme, C. Calberg, B.
Heinrichs, J. Winter, P. Gerbaux, J.-M. Raquez, P. Dubois, Macromolecules
2016, 49, 2162–2171
[6] a) A. Cornille, C. Guillet, S. Benyahya, C. Negrell, B. Boutevin, S. Caillol,
Sylvain, Eur. Polym. J., 2016, 84, 873-888 ; b) A. Cornille, S. Dworakowska,
D. Bogdal, B. Boutevin, S. Caillol, Eur. Polym. J., 2015, 66, 129-138
[7] H. Blattmann, M. Lauth, R. Muelhaupt, Macromol. Mater.
Eng., 2016, 301, 944-952
[8] X. Xi, A. Pizzi, C. Gerardin, J. Du, J. Renew. Mater., 2019, 7, 301–312
[9] X. Xi, A. Pizzi, X. Chen, S. Amirou, H. Lei, X. Chen, C. Gerardin,
Polymers, 2019, 11, 1802-1816
[10] B. Grignard, J.-M. Thomassin, S. Gennen, L. Poussard, L. Bonnaud, J.-
M. Raquez, P. Dubois, M.-P. Tran, C. Park, C. Jerome, C. Detrembleur,
Green Chem, 2016, 18, 2206–2215
[11] J. Clark, T. Farmer, I. Ingram, Ian Y. Lie, M. North, Eur. J. Org. Chem.,
2018, 31, 4265-4271
[12] a) P. Tundo, M. Musolino, F. Aricò, Green Chem., 2018, 20, 28-85;
b) M. Selva, A. Perosa, D. Rodríguez-Padrón, R. Luque, ACS Sustainable
Chem. Eng., 2019, 7, 6471-6479; c) M. Selva, P. Tundo, J. Org. Chem.
2006, 71, 1464-1470; d) P. Tundo, Ind. Eng. Chem. Res., 1988, 27, 1565-
1571
[13] V. Besse, F. Camara, F. Mechin, E. Fleury, S. Caillol, J.-P. Pascault, B.
Boutevin, Eur. Polym. J., 2015, 71, 1-11
[14] a) M. Blain, H. Yau, L. Jean-Gerard, R. Auvergne, D. Benazet, P.
Schreiner, S. Caillol, B. Andrioletti, ChemSusChem, 2016, 9, 2269-2272; b)
M. Blain, L. Jean-Gerard, R. Auvergne, D. Benazet, S. Caillol, B. Andrioletti,
Green Chem., 2014, 16, 4286-4291
[15] A. Brändle, A. Khan, Polym. Chem., 2012, 3, 3224-3227
[16] S. Oprea, D. Timpu,V. Oprea, J. Polym. Res., 2019, 26:117
[17] C. Fitzgerald, I. Lyn, N. Mills, J. Cell. Plastics, 2004, 40, 89–110

Entry for the Table of Contents

Microcellular self-blown non-isocyanate polyurethanes foams

were constructed from reactive formulations by guiding the
chemo- and regioselective ring-opening and decarboxylation
of cyclic carbonates by amines and thiols. This novel concept
is versatile, easy accessible and scalable.

9
This article is protected by copyright. All rights reserved.