Insights into the organocatalyzed synthesis of urethanes in supercritical carbon dioxide: An in Situ FTIR spectroscopic kinetic study

Article abstract of DOI:10.1002/cctc.201301002

The kinetics of urethane (carbamate) formation in supercritical CO ₂ (scCO₂) has been studied by in situ FTIR spectroscopy by using the reaction between n-butanol and toluene-2,4-diisocyanate as a model for the synthesis of polyureth

Full text of DOI:10.1002/cctc.201301002

CHEMCATCHEM
FULL PAPERS
DOI: 10.1002/cctc.201301002
Insights into the Organocatalyzed Synthesis of Urethanes
in Supercritical Carbon Dioxide: An In Situ FTIR
Spectroscopic Kinetic Study
Christopher A. Smith,[a, b, c, d] Henri Cramail,[c, d] and Thierry Tassaing*[a, b]
The kinetics of urethane (carbamate) formation in supercritical
CO2 (scCO2) has been studied by in situ FTIR spectroscopy by
using the reaction between n-butanol and toluene-2,4-diiso-
cyanate as a model for the synthesis of polyurethanes. We
have evaluated the efficiency of different organocatalysts for
this transformation to eliminate the need for potentially toxic
organotin compounds. Bicyclic guanidines and amidines
(7-methyl-1,5,7-triazabicyclo[4.4.0]dec-5-ene and 1,8-diazabi-
cyclo[5.4.0]undec-7-ene) lead to a significant acceleration of
the reaction rate, but we found that their activity in scCO2 is
reduced compared to reactions in conventional media because
of the formation of guanidinium or amidinium alkylcarbonate
salts. In scCO2, these catalysts are also sensitive to the presence
of water and can lead to the formation of allophanate or iso-
cyanurate byproducts. On the other hand, we have shown that
catalysis with amine N-oxides (N-methylmorpholine-N-oxide) is
suited to the scCO2 medium.
Introduction
Polyurethanes (PUs) represent an important class of materials
and are used in a wide variety of applications that include
coatings, sealants, elastomers, adhesives, upholstery foams,
and insulation foams. To date, their synthesis has typically
been studied in bulk or organic solvents by using tertiary
amines and organotin compounds as catalysts.[1] More recently,
supercritical CO2 (scCO2) has shown great potential as an envi-
ronmentally friendly replacement for conventional organic sol-
vents. Indeed, CO2 is a readily available, inexpensive, nontoxic,
and nonflammable natural product with easily accessible criti-
cal temperature and pressure (Tc =31.18C and Pc =7.4 MPa).
The physicochemical properties of CO2, such as viscosity and
density, can be readily tuned by varying the pressure and tem-
perature, which makes it an attractive medium for polymer
synthesis.[2] As a result of the limited solubility of the large ma-
jority of polymers in CO2, most polymerizations in this medium
are heterogeneous processes such as precipitation and disper-
sion polymerization. In the latter case, it is possible to obtain
particles of a uniform size by using an appropriate stabilizer.
These are most often based on fluoropolymers or silicones,[3]
but a number of CO2-soluble poly(vinyl ester)s have also been
investigated recently.[4] In some of our previous work, we ex-
amined the synthesis of PUs by polyaddition between either
ethylene glycol (EG) or 1,4-butanediol (BD) with toluenediiso-
cyanate (TDI) by using a tin catalyst [dibutyltin dilaurate
(DBTDL)] in the presence of alkanol- or isocyanate-terminated
polydimethylsiloxane (PDMS).[5] Under the selected experimen-
tal conditions (T=608C and P=20–40 MPa), evenly sized core–
shell PU-PDMS particles with hydrophobic surface properties
could be obtained. Organotin catalysts, especially dibutyltin di-
carboxylates,[1a,6] have been used to effect this polymerization
both in conventional media and in scCO2 but their potential
toxicological effects[7] could lead to a restriction of their use in
the near future. In this context, organocatalysis could be
a promising alternative.[8] Tertiary amines,[9] amidines,[1a,10] gua-
nidines,[11] amine N-oxides,[12] and more recently carbenes[13]
have been shown to be suitable organocatalysts for the syn-
thesis of PUs, although their catalytic activity and selectivity
remain inferior to those of tin catalysts. Some of these com-
pounds show a tendency to promote secondary reactions such
as allophanate and isocyanurate formation under certain con-
ditions. In scCO2, there are relatively few studies on the use of
organocatalysts, but the existing data are encouraging as the
results are often comparable to those obtained in conventional
media.[14] In continuation of our efforts to promote the green
synthesis of PUs, we have thus investigated the synthesis of
model urethanes in scCO2 by using organocatalysts of several
different classes.
[a] Dr. C. A. Smith, Dr. T. Tassaing
Univ. Bordeaux, ISM, UMR 5255
F-33405 Talence Cedex (France)
Fax: (+33)540008402
E-mail: t.tassaing@ism.u-bordeaux1.fr
[b] Dr. C. A. Smith, Dr. T. Tassaing
CNRS, ISM, UMR 5255
F-33405 Talence Cedex (France)
[c] Dr. C. A. Smith, Prof. H. Cramail
Univ. Bordeaux, LCPO, UMR 5629
F-33607 Pessac Cedex (France)
[d] Dr. C. A. Smith, Prof. H. Cramail
CNRS, LCPO, UMR 5629
F-33607 Pessac Cedex (France)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cctc.201301002.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 1380 – 1391 1380

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
ventional media, compounds with high basicity, nucleophilicity,
or hydrogen-bonding basicity were selected. DBTDL was also
included as a reference. A catalyst loading of 5 mol% (vs. OH
functions) was chosen for the initial catalyst screening to reach
reaction completion within a short time despite the relatively
dilute conditions. In most cases, the preparation of catalyst sol-
utions in butanol was necessary because of the minute quanti-
ties of catalyst required. It was important that the catalysts
were premixed with the alcohol rather than the isocyanate to
avoid the well-known self-addition reactions of isocyanates
(e.g., dimerization, trimerization).
Results and Discussion
We decided to compare the efficacy of different catalysts on
a model urethane-forming reaction by using in situ reaction
monitoring by FTIR spectroscopy. To follow the kinetics of the
reaction in scCO2, this technique was preferred over other
methods (NMR spectroscopy,[15] HPLC[16]) as it would otherwise
be difficult to sample the reaction mixture without changing
the reaction conditions (pressure/concentration). Our model
was based on the reaction between an industrially relevant ar-
omatic diisocyanate, TDI, and a monofunctional alcohol, n-bu-
tanol (Scheme 1). A monofunctional model was preferred over
IR absorption spectra
The evolution of the IR spectra with time for the urethane-
forming reaction catalyzed by triethylamine in scCO2 at 608C
and 24 MPa is shown in Figure 2. The initial spectrum recorded
Scheme 1. Model reaction for PU synthesis.
a polymerization reaction to avoid the physical changes to the
reaction environment associated with polymer growth and the
eventual precipitation of polymer chains. Although an 80:20
mixture of 2,4- and 2,6-TDI is often used industrially (this mix-
ture has a reduced tendency to crystallize), pure 2,4-TDI was
used in this study to simplify the system. At 608C and a typical
polymerization concentration of 2.5 wt% versus CO2, this diiso-
cyanate is soluble above pressures of 17 MPa.[5c] n-Butanol was
chosen as an appropriate monofunctional model of the pri-
mary aliphatic polyols commonly used in PU synthesis and it
forms a single supercritical phase with CO2 above 10 MPa at
608C.[17] Our kinetic measurements were thus performed at T=
608C and P=24 MPa to ensure a single supercritical phase.
Organocatalysts of several different classes were investigated
in this study (Figure 1). Based on literature precedent in con-
Figure 2. In situ IR spectra of the reaction between 2,4-TDI and n-butanol in
scCO2 catalyzed by 5 mol% triethylamine. The initial spectrum in dark blue
features the absorptions of the starting alcohol and diisocyanate. Several
peaks that correspond to the urethane groups grow with the progress of
the reaction. The region between n˜ =1950 and 2650 cmÀ1 and the region
above n˜ =3500 cmÀ1 are not shown as they are saturated by the absorptions
of CO2.
at t=0 is characteristic of the TDI/BuOH/CO2 mixture (see the
Supporting Information for a detailed assignment of the IR
spectra of the solvent, reactants, and product). As a function of
time, we observe the appearance of the urethane bands of di-
butyl 4-methyl-1,3-phenylenedicarbamate (1) that are observed
at n˜ =3456 cmÀ1 (amide A, n(NH)), 1747 cmÀ1 (amide I, n(C=O)),
1530 cmÀ1 (amide II, n(CN)+d(NH)), and 1210 cmÀ1 (amide III,
n(CO)+n(CN)+d(NH)). The kinetics of urethane formation are
often monitored by following the disappearance of the asym-
metric stretch band of the isocyanate group, but this is not
possible in our case as this absorption is masked by the spec-
trum of scCO2. The same problem occurs with the OH stretch-
ing mode of the alcohol group. Therefore, to follow the kinet-
ics of the reaction, we have chosen to monitor the appearance
of the peak associated with the NÀH stretching mode of 1 ob-
served at n˜ =3456 cmÀ1.
Figure 1. Catalysts screened for their activity in the isocyanate/alcohol reac-
tion in scCO2.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 1380 – 1391 1381

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Catalyst screening with 5 mol% catalyst
with DBTDL after this time. There was nevertheless a marked
increase in the rate of reaction compared to the uncatalyzed
reaction, which was extremely sluggish.
Reaction kinetics
The evolution of the area of the urethane NÀH band with time
in the presence of different catalysts at 5 mol% loading is
shown in Figure 3. With the reference tin-based catalyst,
Reactivity of ortho- versus para-isocyanate groups
Analysis of the spectra recorded during kinetic monitoring
showed a clear difference in the shape of the NÀH peak that
depended on the catalyst employed. In the reactions with the
more active catalysts (DBTDL, MTBD, TBD, DBU, NMO), two
overlapping peaks could be observed clearly by the end of the
reaction, whereas with the less active catalysts (tertiary amines,
pyridine derivatives) only a weak shoulder was present on the
side of the main peak (Figure 4). Given that both peaks are
present in the spectrum of the pure product 1, the two peaks
are logically assigned to the two different urethane groups in
this unsymmetrical molecule. This analysis is supported by the-
oretical calculations of the vibrational frequencies of diure-
thane derivatives of TDI, which show a separation of 11 cmÀ1
between the two NÀH stretching vibrations.[19] In the solid
state or in more concentrated solutions, the NÀH bands are
broadened because of hydrogen bonding, and the two differ-
ent urethane groups are indistinguishable. The peak centered
at n˜ =3457 cmÀ1 can reasonably be assigned to the para-ure-
thane group, as this peak is formed first and the greater reac-
Figure 3. Kinetics of the reaction between 2,4-TDI and n-butanol in scCO2 in
the presence of 5 mol% of different catalysts. The absorbance of the n(NÀH)
band of the urethane group is plotted against time.
DBTDL, there was a very rapid initial rate of reaction and the
curve reached a plateau after approximately 60 min reaction
time. The cyclic amidine 1,8-diazabicyclo[5.4.0]undec-7-ene
(DBU) and guanidines 7-methyl-1,5,7-triazabicyclo[4.4.0]dec-5-
ene (MTBD) and 1,5,7-triazabicyclo[4.4.0]dec-5-ene (TBD) also
showed a rapid initial rate of reaction but the final values of
the urethane peak absorbance (i.e., yields) were slightly lower.
TBD was the only catalyst to give a slightly S-shaped kinetic
curve, which presumably results from the initial reaction of the
active hydrogen atom of TBD with isocyanate (Scheme 2). One
Scheme 2. Reaction of an isocyanate with TBD.
could also expect the activity of TBD to be hindered by carba-
mate salt formation from CO2 insertion into the NÀH bond, but
it has been shown that TBD and CO2 form a zwitterionic
adduct in which the association of CO2 is reversible[18] (see
later discussion). The amine N-oxide N-methylmorpholine-N-
oxide (NMO) showed a similar kinetic profile to the amidine/
guanidines, although there appeared to be a slight decrease in
the intensity of the urethane peak after it had reached a maxi-
mum rapidly. Tertiary amines (1,4-diazabicyclo[2.2.2]octane
(DABCO), triethylamine) and pyridine derivatives [4-dimethyl-
aminopyridine (DMAP), 4-pyrrolidinopyridine (4-PPY)] showed
a much slower rate of reaction, and the curves did not reached
a plateau after 17 h reaction time. The value of the urethane
peak area was only approximately 50–60% of that achieved
Figure 4. In situ IR spectra of the reaction between 2,4-TDI and n-butanol in
scCO2 showing the evolution of the urethane NÀH stretching peak with
time. a) Reaction catalyzed with triethylamine; only a weak shoulder is ob-
served at n˜ =3466 cmÀ1 with the less active catalysts. b) Reaction catalyzed
with DBTDL; the shoulder at n˜ =3466 cmÀ1 is much more prominent with
the more active catalysts.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 1380 – 1391 1382

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
tivity of the para-isocyanate versus the ortho-isocyanate group
in 2,4-TDI has been well documented.[20] The shoulder at n˜ =
3466 cmÀ1 then corresponds to the ortho-urethane group,
which is formed more slowly, and in the case of the less active
catalysts, it is formed to a much smaller extent.
The presence of two nonequivalent isocyanate groups ac-
tually leads to two possible reaction pathways with four associ-
ated reaction rates (Scheme 3). Two factors influence the rate
Figure 5. In situ IR spectra of the reaction between 2,4-TDI and n-butanol in
scCO2, catalyzed with DBTDL: the evolution of bands in the region between
n˜ =1240 and 1440 cmÀ1 appear to highlight a two-step reaction pathway.
Peaks at n˜ =1276, 1301, and 1382 cmÀ1 reach a maximum after approxi-
mately 10 min reaction time (dark green spectrum) and then diminish some-
what as peaks at n˜ =1323 and 1417 cmÀ1 become more prominent.
thus presumably correspond to the intermediate 2, whereas
the other two peaks correspond to the final product 1. In the
reactions catalyzed by the less active catalysts (triethylamine,
DABCO, DMAP, and 4-PPY), there appeared to be much less
formation of the ortho-urethane group, and significant
amounts of intermediate 2 thus remain unreacted after 17 h
reaction time.
Scheme 3. The two possible reaction pathways for the reaction of 2,4-TDI
with n-butanol.
The nature of any catalyst used clearly has an influence on
the relative rates of the individual reaction steps. Boitsov and
Trub reported that organotin catalysts such as DBTDL en-
hanced the difference in reactivity between the ortho- and
para-NCO groups of 2,4-TDI (k1/k2) compared to the uncata-
lyzed reaction, whereas tertiary amines had an equalizing ef-
fect.[15b] In the spectra recorded in the present study, the rela-
tive selectivity of the different catalysts for the ortho- and
para-isocyanate groups can be examined by comparing the
shape of the NÀH peak at low isocyanate conversion
(Figure 6). In the reaction catalyzed by MTBD, the two isocya-
constants of these reaction steps. First of all, as mentioned
above, the ortho-isocyanate group is less reactive than the
para-isocyanate group (i.e., k1 >k2, k4 >k3) because of the steric
and/or electronic effects of the methyl group.[20b] Secondly, for
a given isocyanate group, the reactivity is greater if the ring is
substituted with another isocyanate group than with a ure-
thane group (i.e., k1 >k4, k2 >k3) because of the more electron-
withdrawing nature of the isocyanate group. Several studies
have attempted to calculate the relative rates of these four re-
action steps by fitting kinetic data obtained through NMR
spectroscopy, IR spectroscopy, HPLC, and/or back titra-
tion,[15,16b,21] or by theoretical computations.[22] As k1 is greater
than k2 and all the reactions can be considered to be irreversi-
ble under the experimental conditions, the overall reaction
passes predominantly through the intermediate 2. Additionally,
as k4 is greater than or equal to k2 (according to most reports),
there should not be an accumulation of intermediate 3 during
the reaction.
Other than the changes in the form of the NÀH stretching
peak during the reaction, changes to the absorption bands in
the region between n˜ =1240–1440 cmÀ1 also appear to high-
light the two-step reaction pathway. The evolution of the spec-
tra with time in a reaction catalyzed by DBTDL is shown in
Figure 5. Peaks at n˜ =1276, 1301, and 1382 cmÀ1 grow rapidly
at the beginning of the reaction to reach a maximum after ap-
proximately 10 min (dark green spectrum) and then slightly di-
minish afterwards, whereas peaks at n˜ =1323 and 1417 cmÀ1
grow gradually throughout the reaction. The first three peaks
Figure 6. Urethane NÀH peak in the reaction of 2,4-TDI with n-butanol in
scCO2 at approximately 40% isocyanate conversion, which shows the de-
pendence of the peak shape on the catalyst used. In the MTBD-catalyzed re-
action, the intensities of the main peak and the high-frequency shoulder are
more even, which suggests reduced selectivity between the ortho- and
para-isocyanate groups.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 1380 – 1391 1383

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
nate groups appear to react somewhat more simultaneously
(greater intensity of the shoulder at n˜ =3466 cmÀ1) compared
to the reaction catalyzed by DBTDL. NMO and triethylamine
show intermediate selectivity. DBU gave a similar result to that
presented for MTBD. A more quantitative analysis of the reac-
tion rates would appear to be difficult because of the presence
of three different urethane species with overlapping absorp-
tions.
In the spectra of the reaction residues, several additional
peaks could be identified other than those that belonged to
the principal product. The region around d=2 ppm (which
contains the aromatic methyl singlets such as f) was particular-
ly useful for the identification of the compounds present, by
comparison with the spectra of model compounds (Figure 9).
Amino-substituted urethanes 4 and 5 were clearly identified in
all the spectra (CH3 singlets at d=2.02 and 1.96 ppm, respec-
tively). The amino groups are formed by reaction of the residu-
al isocyanate groups with moisture present in the NMR sol-
vent. These compounds thus correspond to unreacted isocya-
nate groups from the reactions in scCO2 (compounds 2 and 3).
Several generally weak peaks could be observed between
d=2.12 and 2.19 ppm and appear to correspond to diarylureas
6–8. Ureas are formed if an amine, generated from the reac-
tion of water with an isocyanate, goes on to react with another
isocyanate group. The three structural isomers each show dif-
ferent chemical shifts for the methyl singlets. It is difficult to
assign all of these peaks with certainty in the spectra of the re-
action residues because of their weak concentration. These
ureas could be formed either during the reaction in scCO2 or
during the subsequent analysis, but the former seems more
likely given the low concentration of isocyanate in the NMR so-
lution (notably, in one of the analyses there is a relatively high
amount of the amino-substituted urethane but very little urea).
Two other types of expected byproducts were isocyanurates
(isocyanate trimers) and allophanates. The isocyanurate 11 and
allophanates 9 and 10 were prepared as model compounds.
The isocyanurate 11 showed a broad, highly shifted methyl sin-
glet at d=2.24 ppm. The allophanate 9 shows two different
methyl singlets at d=2.23 and 2.05 ppm and has a characteris-
tic high-field NH shift of d=10.55 ppm. However, it was found
that the allophanate linkage dissociated back to urethane and
isocyanate in [D6]DMSO (and [D7]DMF). The isocyanate was
then converted to the amino group by reaction with water.
This was a surprising result as none of NMR spectroscopic
studies in these solvents in the literature mentioned this as an
issue. Additionally, Okuto[24] and Furukawa and Yokoyama[25]
Reaction kinetics with 1 mol% catalyst
The compounds with the best performance from the initial cat-
alyst screen were further tested at lower catalyst loading. With
1 mol% catalyst, the superior rate of reaction with DBTDL com-
pared to MTBD, DBU, and NMO was more marked (Figure 7).
Figure 7. Kinetics of the reaction between 2,4-TDI and n-butanol in scCO2 in
the presence of 1 mol% of different catalysts. The absorbance of the n(NÀH)
band of the urethane group is plotted against time.
Nevertheless, the initial reaction rate observed with the orga-
nocatalysts remained relatively rapid, and NMO showed
a faster rate than MTBD and DBU. The most striking difference
between DBTDL and the organocatalysts was in the final ure-
thane peak areas, that is, yields of the reactions. With
the organocatalysts, the kinetic curves reached a pla-
teau quickly, but at much lower yield than with
DBTDL, which suggests either a decrease in catalytic
activity or the consumption of reactants in a secon-
dary reaction. To further investigate these possibili-
ties, analysis of the reaction residues was performed
by NMR and Raman spectroscopy.
NMR and Raman spectroscopy of product residues
Several of the kinetic runs were repeated with NMR
and/or Raman spectroscopic analysis of the product
residue to identify possible byproducts. [D6]DMSO
was selected as the solvent as there is already some
existing data in the literature on the NMR spectra of
TDI derivatives in this solvent.[23] The NMR spectrum
of the main reaction product 1 is shown in Figure 8. Figure 8. 1H NMR spectrum of 1 in [D6]DMSO.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 1380 – 1391 1384

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Figure 9. Possible byproducts in the reaction of 2,4-TDI with n-butanol.
have reported that reactive amines are necessary to degrade
allophanate (and biuret) linkages in DMSO at low tempera-
tures. Traces of peaks at d=10.53 and 10.56 ppm were ob-
served in the residues of reactions catalyzed by 1 mol% MTBD,
DBU, and NMO (although not in reactions with 5 mol% cata-
lyst). It is difficult to determine if only traces of allophanate
were formed in the reactions in scCO2 or if higher levels of al-
lophanate were formed and only the remaining traces can be
observed because of dissociation during the analysis. Biurets
could also show peaks in this region and are reportedly more
stable to dissociation.[26]
plained by the elimination of a larger proportion of the more
CO2-soluble compounds if the CO2 pressure is released. Howev-
er, even if the analysis is not strictly quantitative, it provides
good qualitative information on each reaction.
In the reaction catalyzed with 5 mol% MTBD, the kinetic
curve was somewhat different to that obtained in the original
catalyst screening (compare Figures 3 and 10). There was an in-
itial period of rapid reaction before the rate slowed suddenly
and finally reached a yield of around 60% of that observed in
the DBTDL-catalyzed reaction. This “kinked” curve shape was
reproduced in a second repeat experiment. In the NMR spec-
trum of the reaction residue (Figure 11), small amounts of both
the ortho- and para-amino-substituted urethanes 4 and 5 were
present, which represent unreacted isocyanate groups. There
appeared to be small amounts of the diarylureas, and more
notably, an isocyanurate peak at d=2.24 ppm. The presence of
isocyanurate groups was confirmed by the presence of a peak
at n˜ ꢀ1780 cmÀ1 in the Raman spectrum of the reaction resi-
due (Figure 12). This band corresponds to the symmetric car-
Further analysis of the NMR spectrum from each reaction is
provided below, and the associated kinetic curves are shown
in Figures 7 and 10. It is not always easy to account for the
mass balance in these reactions, that is, in which the kinetic
curve shows that the yield was less than quantitative, the cor-
responding amount of byproducts or unreacted groups should
be observed in the NMR spectrum. This could perhaps be ex-
Figure 10. Kinetics of the reaction between 2,4-TDI and n-butanol in scCO2
with 5 mol% catalyst. The traces for NMO and MTBD are the kinetic runs in
which the reaction residue was analyzed by NMR/Raman spectroscopy.
Figure 11. NMR spectrum of the residue from the reaction between 2,4-TDI
and n-butanol catalyzed by 5 mol% MTBD.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 1380 – 1391 1385

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
and 10). In the NMR spectrum of the reaction residue (Fig-
ure 13a), a significant amount of the ortho-amino species 5
could be observed as well as a little of the diarylurea 6. How-
ever, there appeared to be few other byproducts present in
the mixture, unlike the reaction with 5 mol% catalyst loading.
The fact that the kinetic curve reached a plateau while a signifi-
cant amount of unreacted isocyanate groups remained sug-
gests that deactivation of the catalyst may have occurred. This
could be because of the presence of traces of water, which re-
sults in MTBD tied up in the form of a guanidinium bicarbon-
ate salt (vide infra).
The reaction catalyzed with 1 mol% DBU showed a similar
rate of reaction to that catalyzed with 1 mol% MTBD with
a marginally lower yield obtained (Figure 7). The NMR spec-
trum of the reaction residue (Figure 13b) showed a little of
both the ortho- and para-amino species 4 and 5, which sug-
gests that catalysis with DBU is somewhat less selective than
MTBD between the ortho- and para-isocyanate groups. There
were few other byproduct peaks observable in the spectrum,
and it is difficult to account for the mass balance in this reac-
tion.
Figure 12. Raman spectra of 1 (c) and the residue from the reaction of
2,4-TDI with n-butanol catalyzed by 5 mol% MTBD (c).
bonyl vibration (A1’, n(C=O)), which is IR inactive.[27] Thus, the
reduced yield observed during in situ IR monitoring of this re-
action seems primarily because of isocyanurate formation,
which as well as consuming isocyanate groups could, perhaps,
also render any urethane groups attached to the isocyanurate
ring unobservable because of the poor solubility of the isocya-
nurate in scCO2.
In the reaction catalyzed by 5 mol% NMO, the kinetic curve
reached a plateau rapidly before it descended slightly with
prolonged reaction time (Figure 10). The final yield was greater
than that achieved with 5 mol% MTBD but still significantly
lower than that achieved with DBTDL. In the NMR spectrum of
the reaction residue (Figure 13c) there was a reasonable
In the reaction catalyzed by 1 mol% MTBD, although the ini-
tial rate of reaction was slower, there was no kink in the shape
of the kinetic curve and the final yield was actually higher than
that attained with 5 mol% catalyst loading (compare Figures 7
Figure 13. NMR spectrum of the residue from the reaction between 2,4-TDI and n-butanol catalyzed by a) 1 mol% MTBD, b) 1 mol% DBU, c) 5 mol% NMO,
and d) 1 mol% NMO.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 1380 – 1391 1386

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
amount of the ortho-amino species 5 present as well as the di-
arylurea 6. The relatively high levels of 6 observed could be re-
lated to the highly hygroscopic nature of the amine oxide. As
with MTBD at 1 mol% loading, the fact that the kinetic curve
reached a plateau while a significant amount of unreacted iso-
cyanate groups remained suggests that deactivation of the
catalyst may have occurred, although we have not been able
to identify a mechanism for this process. Small amounts of
other byproducts can be observed in the spectrum, which are
difficult to identify with certainty.
10 mol% water was added to a reaction catalyzed with
5 mol% MTBD (Figure 15). This led to the complete suppres-
sion of catalysis (dotted dark blue line) with the rate of reac-
tion decreased to that of the uncatalyzed reaction. The same
protocol was followed for a reaction catalyzed by 5 mol%
NMO, which resulted only in a reduction in yield with no clear
effect on the reaction rate (dotted brown line). Indeed, the re-
duced yield could be because of the formation of (poorly solu-
ble) ureas. The inhibiting effect of water on MTBD catalysis is
probably attributable to the formation of the guanidinium bi-
carbonate salt from the guanidine/CO2/water ternary mixture.
The formation of carbonate salts and the consequences for
catalyst activity are described in further detail in the following
sections. For completeness, the reaction of the water present
in these reactions with isocyanate would lead to amines that
could themselves go on to form carbamate salts with CO2, but
we have not been able to observe this.
In the reaction catalyzed by 1 mol% NMO, the initial rate of
reaction was greater than that for MTBD and DBU catalysts at
the same loading, but the curve reached a plateau at a lower
reaction yield (Figure 7). In the NMR spectrum of the reaction
residue (Figure 13d), there was a ratio of approximately 1:1 of
the diurethane 1 and the ortho-amino-substituted species 5,
which suggests once again that the catalyst may have been
deactivated. This also shows the strong selectivity for the para-
isocyanate group.
Robustness of organocatalysts versus DBTDL
Some variability in the results of kinetic runs with the amidine/
guanidine catalysts indicated that there might be some sensi-
tivity to the reaction conditions. A series of experiments was
thus undertaken to investigate the robustness of these organo-
catalysts compared to DBTDL. The use of ordinary “wet” buta-
nol rather than anhydrous butanol as well as standard industri-
al-quality rather than N45-quality CO2 (<7 ppm H2O) led to
a dramatic decrease in the rate of the DBU-catalyzed reaction
(Figure 14, compare solid and dotted pink lines). In fact, after
a reduced period of initial rapid reaction, the rate was similar
to that of the uncatalyzed reaction (green line). Under the
same conditions, MTBD showed no initial period of rapid reac-
tion at all (dotted dark blue line), although the rate remained
greater than that of the uncatalyzed reaction. In comparison,
the kinetic runs with DBTDL were highly reproducible regard-
less of the dryness of the butanol or the quality of CO2 em-
ployed (solid and dotted light blue lines). To quantify some-
what the effect of water on the MTBD-catalyzed reaction,
Figure 15. Kinetics of the reaction between 2,4-TDI and n-butanol in scCO2
catalyzed by 5 mol% catalyst under different conditions. c: anhydrous
conditions. g: 10 mol% water added to the reaction.
The behavior of amidines/guanidines in the presence of
CO2: An explanation of catalytic activity
We envisaged that the interaction of guanidines and amidines
with CO2 could have a major impact on their catalytic activity.
These molecules have been studied for use in CO2-storage ap-
plications and CO2-fixing synthetic transformations, and
a number of studies have aimed to elucidate the mechanism
of CO2 fixation.[28] There has been some debate as to whether
binary CO2–guanidine or CO2–amidine adducts exist or if labile
hydrogen-containing compounds such as alcohols, amines, or
water are necessary to form a carbonate, carbamate, or bicar-
bonate salt, respectively.[29] A TBD–CO2 adduct has been isolat-
ed in crystal form,[18] but equivalent adducts with peralkylated
guanidines such as MTBD could be less favorable because of
a lack of hydrogen-bond stabilization. In addition, it was found
that the TBD–CO2 adduct was transformed readily into a bicar-
bonate salt in the presence of water. The formation of alkylcar-
bonate salts from guanidine/amidine and CO2 in the presence
of an alcohol has shown particular interest for “switchable sol-
vents” and CO2-capture agents as the association of CO2 is re-
versible with increased temperature.[30]
Figure 14. Kinetics of the reaction between 2,4-TDI and n-butanol in scCO2
catalyzed by 1 mol% catalyst under different conditions. c: anhydrous
butanol and N45-quality CO2. g: standard “wet” butanol and industrial-
quality CO2.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 1380 – 1391 1387

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
We have found that the formation of alkylcarbonate salts
can be followed conveniently by in situ FTIR spectroscopy
(Figure 16). If a stirred, equimolar solution of MTBD and metha-
Figure 17. Plot of the methyl carbonate n(C=O) peak height versus tempera-
ture in the equilibrium between the guanidinium methyl carbonate salt and
MTBD/methanol/CO2. The n(C=O) absorbance is plotted as a percentage of
the initial value at 238C.
Figure 17 (data recorded while the solution was heated from
ambient temperature are shown in red, and data recorded
while the solution was allowed to cool are shown in blue). The
curves that correspond to heating and cooling are not exactly
superimposed, and there seems to be some loss of the methyl
carbonate upon cooling. This is perhaps a result of the replace-
ment of methanol with adventitious water to form the much
less soluble guanidinium bicarbonate salt. At 608C (the tem-
perature used for the model reaction in scCO2), the equilibrium
ratio of guanidinium methyl carbonate to guanidine was ap-
proximately 70:30. However, the CO2 concentration in this ex-
periment is of course much lower than if scCO2 is used as a sol-
vent. The MTBD free base is soluble in scCO2, and the IR spec-
trum of MTBD at 608C and 24 MPa is shown in Figure 18 (red).
The guanidine n(CN) stretch can be observed at around n˜ =
1620 cmÀ1, and the peak has a similar appearance to that ob-
served in organic solvents. However, if a mixture of MTBD and
Figure 16. In situ IR spectra following the formation of a guanidinium alkyl-
carbonate salt in acetonitrile solution. At 0 min, a solution of methanol and
MTBD is subjected to 0.2 MPa CO2 pressure. The methanol n(OÀH) band dis-
appears, and a new band that corresponds to the methyl carbonate ion ap-
pears. In addition, the band at around n˜ =1606 cmÀ1 grows in intensity as
the guanidine becomes protonated. A peak corresponding to dissolved CO2
at n˜ =2342 cmÀ1 appears later (after the formation of the methyl carbonate
salt), which suggests that the process is rate limited by diffusion of CO2 into
the solution.
nol in acetonitrile was subjected to approximately 0.2 MPa CO2
pressure, the methanol OÀH stretching band centered at n˜ =
3539 cmÀ1 disappeared, and new peaks emerged that corre-
spond to the methyl carbonate ion at n˜ =1675 cmÀ1 (C=O
stretch), 1286, and 1075 cmÀ1. There was a reduction in the in-
tensity of the peak at n˜ =1511 cmÀ1 (CÀN/C=N/CÀC stretch
combination of MTBD[31]), and the MTBD peak at n˜ =1606 cmÀ1
(CÀN stretch combination[31]) was shifted to n˜ =1604 cmÀ1 and
saw a significant increase in intensity (protonated MTBD shows
two bands at around n˜ =1600 cmÀ1 in dissociated salts, but
a single band may be observed if it is involved in hydrogen
bonding[32]). The appearance of a peak that corresponds to dis-
solved CO2 at n˜ =2342 cmÀ1 only after the complete formation
of the methyl carbonate salt suggested that the process was
rate limited by the diffusion of CO2 into the solution, in agree-
ment with the findings of Heldebrant and co-workers.[30b]
Equivalent results were obtained if MTBD was replaced by
DBU. However, if “wet” acetonitrile was used, the bicarbonate
salt was formed, which precipitated from the solution.
Figure 18. IR spectra recorded in scCO2 (P=24 MPa, T=608C). c: MTBD
(c=8.4ꢁ10À3 moldmÀ3). c: mixture of MTBD and n-butanol (c=8.4ꢁ10À3
and 0.17 moldmÀ3, respectively). If there were no interactions in the MTBD/
butanol mixture, the black spectrum would have the greatest intensity
across all wavenumbers (i.e., the spectrum would be the sum of the spectra
of MTBD and butanol). However, this is not the case and the n(CN) band of
guanidine is notably missing from the spectrum of the MTBD/butanol mix-
ture. The two broad, weak peaks in its place suggest that the guanidinium
butylcarbonate salt has been formed but is only poorly soluble in scCO2.
If the guanidinium methyl carbonate solution was heated
under fixed volume conditions, the reversibility of the reaction
could be established by following the decrease in the intensity
of the methyl carbonate peaks and the reappearance of the
methanol OÀH stretching peak. The absorbance of the methyl
carbonate C=O stretch is plotted against temperature in
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 1380 – 1391 1388

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
n-butanol in a molar ratio of 1:20 (i.e., 5 mol% MTBD) is dis-
solved in scCO2 under the same conditions of temperature and
pressure, the spectrum of the mixture is not simply the spec-
trum of MTBD with additional absorptions from butanol
(Figure 18, black). Two broad, weak peaks are all that can be
observed in the region of the guanidine n(CN) stretch. This
suggests that the equilibrium is highly shifted towards the al-
kylcarbonate salt, which is apparently poorly soluble in scCO2.
onium cation, which renders it effective for the catalysis of Mi-
chael reactions, for example.[33] Alkylcarbonate salts (with
a metal or quaternary ammonium cation) have also been pa-
tented as moderately active catalysts for the preparation of
polyisocyanurates or poly(urethane-isocyanurates).[34] Equally,
the salts of DBU with weak acids have been used as catalysts
in PU formulations.[10] The IR spectra presented in Figure 18
suggest that the guanidinium alkylcarbonate salt is the de fac-
to catalytic species for reactions conducted in supercritical
CO2. In the presence of water, the more stable bicarbonate salt
should be favored over the alkylcarbonate salt,[30b,33] but it
does not appear to show any catalytic activity. This difference
in activity is presumably because of a difference in solubility in
scCO2.
Effect of CO2 on the reaction kinetics in organic solvent
To investigate the consequences of guanidinium alkylcarbon-
ate salt formation on catalysis, the kinetics of the urethaniza-
tion reaction were investigated in dichloromethane solution at
248C, conditions under which the formation of the alkylcar-
bonate salt is more or less quantitative under a low pressure
of CO2. The reaction between phenyl isocyanate and n-butanol
was chosen because of its simple kinetic scheme.
The effect of CO2 on the urethanization reaction catalyzed
by NMO was also tested by using the same procedure, but in
this case there was no clear effect on the rate of reaction
(Figure 19, solid and dotted brown lines). The key result is that
although the MTBD-catalyzed reaction is faster in the absence
of CO2, the NMO-catalyzed reaction is superior in the presence
of CO2.
A dichloromethane solution of butanol with 5 mol% MTBD
was first subjected to 0.2 MPa CO2 pressure, and the formation
of the guanidinium alkylcarbonate salt was verified by the ex-
pected changes in the IR spectrum. A phenyl isocyanate solu-
tion was then added to the mixture, and the formation of the
urethane product was followed by the appearance of the C=O
stretching band. For comparison, the reaction was performed
under the same conditions in the absence of CO2 (Figure 19).
Conclusions
In situ FTIR spectroscopy is a valuable method for reaction
monitoring in supercritical CO2 (scCO2), which allows a compari-
son of various organocatalysts with regards to the kinetics of
urethane formation in a model reaction. Analysis of the spectra
has also enabled the identification of the reaction intermedi-
ates and a comparison of the selectivity of the catalysts to-
wards the ortho- and para-isocyanate groups of toluene-2,4-di-
isocyanate. Guanidine and amidine catalysts as well as amine
oxides show clearly higher catalytic activity than traditional ter-
tiary amine catalysts for the urethanization reaction in scCO2.
We have been able to highlight a change in the order of activi-
ty of these catalysts if the reaction is conducted in the pres-
ence or absence of CO2, which can be attributed to the forma-
tion of guanidinium or amidinium alkylcarbonate salts. Most
importantly, amine oxides have emerged as the most suitable
class of organocatalysts for the formation of urethanes in
scCO2, as the guanidines and amidines are sensitive to the
presence of water and are prone to the formation of allopha-
nate and particularly isocyanurate byproducts. Although the
organocatalysts we have studied show lower activity than the
classic tin-based catalyst, they have enabled the development
of a metal-free alternative for the synthesis of polyurethane
particles in scCO2.[35]
Figure 19. Kinetics of the reaction of phenyl isocyanate with n-butanol cata-
lyzed by 5 mol% catalyst in dichloromethane solution. c: reaction con-
ducted in the absence of CO2. g: reaction conducted under 0.2 MPa CO2
pressure.
There was a clear difference between the MTBD-catalyzed
reactions performed in the presence or absence of CO2 (solid
and dotted dark blue lines, respectively), in which CO2 brought
about a reduction in the rate of reaction. Given that the rate of
reaction in the presence of CO2 was still relatively rapid com-
pared to that of the uncatalyzed reaction (green curve), it
seems unlikely that the reaction is catalyzed only by the very
small equilibrium concentration of MTBD free base, and it is
conceivable that a different catalytic mechanism is in opera-
tion, which involves the butyl carbonate ion as either a nucleo-
phile or base. It has been reported that the basicity of the al-
kylcarbonate anion is enhanced if it is associated with an
Experimental Section
Materials
2,4-TDI>99% (analytical standard grade) and phenyl isocyanate
98% were purchased from Sigma–Aldrich and used as received. n-
Butanol (Sigma–Aldrich) was distilled over CaH2 and stored over
molecular sieves. CO2 (N45 quality, <7 ppm H2O) was supplied by
Air Liquide. DBTDL 95% (Alfa Aesar), MTBD 98% (Sigma–Aldrich),
TBD 98% (Sigma–Aldrich), DBU 99% (Alfa Aesar), triethylamine
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 1380 – 1391 1389

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
99% (Alfa Aesar), DABCO 98% (TCI), DMAP 99% (Sigma–Aldrich),
4-PPY 98% (TCI), and NMO 97% (Sigma–Aldrich) were used as re-
ceived. Dichloromethane was purified through a Grubbs apparatus
before use. Acetonitrile was distilled over CaH2 before use.
with a resolution of 4 cmÀ1). For experiments conducted in the
presence of CO2, the butanol/catalyst solution was subjected to
0.2 MPa CO2 pressure before the addition of the phenyl isocyanate
solution. In the case of the MTBD-catalyzed reaction, the formation
of the guanidinium butyl carbonate salt was followed by the corre-
sponding changes to the IR spectrum.
IR spectroscopy set-up
The IR absorption measurements were performed by using a Ther-
moOptek Nicolet 6700 FTIR spectrometer equipped with a globar
as the IR source, a KBr/Ge beamsplitter, and a deuterated triglycine
sulfate (DTGS) detector to investigate the spectral range 400–
4000 cmÀ1. Single-beam spectra recorded with 2 cmÀ1 resolution
were obtained after the Fourier transformation of 32 accumulated
interferograms (the resolution and number of scans were reduced
during kinetic reaction monitoring to reduce the time required to
collect each spectrum). The IR absorption experiments were per-
formed by using a stainless-steel cell built in-house[36] equipped
with two cylindrical silicon windows and a variable path length.
The seal was obtained by using the unsupported area principle.
The windows were positioned on the surface of a stainless-steel
plug with a 100 mm Kapton foil placed between the window and
the plug to compensate for any imperfections between the two
surfaces. Teflon O-rings were used to ensure the seal between the
plug and the cell body. The cell was heated by using cartridge
heaters in the periphery of the body of the cell. Two thermocou-
ples were used, the first one located close to a cartridge heater for
the temperature regulation and the second one close to the
sample area to measure the temperature of the sample with an ac-
curacy of approximately 28C. The cell was connected by a stain-
less-steel capillary tube to a hydraulic pressurizing system that al-
lowed the pressure to be increased to 50 MPa with an absolute un-
certainty of Æ0.1 MPa and a relative error of Æ0.3%.
Raman spectroscopy
The Raman measurements were performed by using a Horiba
Jobin–Yvon HR800-UV confocal spectrometer with a resolution of
0.7 cmÀ1 in the spectral range 200–3300 cmÀ1. The laser excitation
wavelength was 752 nm and the laser beam power was 3.7 mW. To
improve the signal-to-noise ratio, each spectrum was the result of
two accumulated spectra for typical times of 60 s.
NMR/Raman spectroscopic analysis of reaction residues
To obtain samples of the reaction residues for NMR spectroscopy,
the NMR solvent was added directly to the IR cell after cooling and
depressurization. To obtain samples for Raman spectroscopy, the
cell was washed through with dichloromethane, and the solvent
was evaporated before analysis of the residue.
CO2 absorption experiments
A 1:1 solution of methanol and MTBD in freshly distilled acetoni-
trile was prepared under Ar (0.15 moldmÀ3). The high-pressure cell
(configured with a path length of 0.12 mm) was placed under
vacuum and filled with 4.0 mL of the solution. The IR spectrum of
the solution was recorded then the cell was opened to 0.2 MPa
CO2 pressure and the solution agitated with a magnetic stirrer (T=
238C). Further spectra were recorded every 30 s and CO2 absorp-
tion was found to be complete after 90 s. The closed cell (fixed
volume) was heated to 308C then to 1008C in steps of 108C, and
the IR spectrum of the solution was recorded at each temperature.
Successive spectra were recorded until the spectrum stabilized.
The process was then repeated while the cell was allowed to cool
in steps of 108C.
Kinetic monitoring in scCO2
The high-pressure cell (configured with a path length of 5.05 mm)
was dried by heating under vacuum and then rinsed with scCO2.
Butanol (9.2 mL, 0.10 mmol) or a solution of catalyst in butanol
(1:20 mol/mol for 5 mol% catalyst) was added into the capillary
tube above the cell then injected with ꢀ6 MPa CO2. The tempera-
ture was increased to 608C, and the pressure increased to 20 MPa.
A spectrum was recorded of the butanol/catalyst solution in scCO2.
2,4-TDI (7.1 mL, 5.0ꢁ10À2 mmol) or PhNCO (11.0 mL, 0.10 mmol) was
then injected from a small length of capillary tube in between two
valves with a pressure gradient of ꢀ10 MPa. The pressure was
then adjusted to 24 MPa. A first spectrum was recorded immedi-
ately after the addition of the isocyanate. Further spectra were re-
corded every 2 min at the beginning of the reaction then every
10 min up to approximately 17 h reaction time. Each spectrum was
Acknowledgements
The authors acknowledge the Ministꢀre de la Recherche for the
PhD fellowship of C.A.S. and the Agence Nationale de la Recher-
che (ANR-09-CP2D-15-04) for support for working costs. The au-
thors are pleased to thank Dr. Jean-Luc Bruneel for his help in
the Raman experiments. The authors would also like to thank
Dr. Eric Cloutet, Prof. Yannick Landais, Dr. Frꢁdꢁric Robert, Prof.
Daniel Taton, and Dr. Jꢁrꢂme Alsarraf for fruitful discussions.
the result of four accumulated scans with a resolution of 4 cmÀ1
.
Kinetic monitoring in dichloromethane solution
Keywords: green chemistry
polymers · supercritical fluids
· kinetics · organocatalysis ·
A mixed solution of butanol and catalyst (0.50 moldmÀ3 in butanol
and 2.5ꢁ10À2 moldmÀ3 in catalyst) and a solution of phenyl isocya-
nate (0.50 moldmÀ3) in dry dichloromethane were prepared under
Ar. The high-pressure cell (configured with a path length of
0.29 mm) was dried by heating under vacuum, then the butanol/
catalyst solution (2.0 mL) was added directly. The IR spectrum of
the solution was recorded, the PhNCO solution (2.0 mL) was
added, and the mixture was agitated with a magnetic stirrer at
248C. Spectra were recorded every 30 s (five accumulated scans
[1] a) A. L. Silva, J. C. Bordado, Catal. Rev. Sci. Eng. 2004, 46, 31–51; b) G.
Wegener, M. Brandt, L. Duda, J. Hofmann, B. Klesczewski, D. Koch, R. J.
Kumpf, H. Orzesek, H. G. Pirkl, C. Six, C. Steinlein, M. Weisbeck, Appl.
Catal. A 2001, 221, 303–335.
[2] a) P. G. Jessop, T. Ikaria, R. Noyori, Science 1995, 269, 1065–1069; b) J. L.
Kendall, D. A. Canelas, J. L. Young, J. M. DeSimone, Chem. Rev. 1999, 99,
543–563; c) A. Baiker, Chem. Rev. 1999, 99, 453–473; d) E. J. Beckman,
J. Supercrit. Fluids 2004, 28, 121–191.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 1380 – 1391 1390

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
[3] a) A. I. Cooper, J. Mater. Chem. 2000, 10, 207–234; b) P. B. Zetterlund, F.
Aldabbagh, M. Okubo, J. Polym. Sci. Part A 2009, 47, 3711–3728; c) K. J.
Thurecht, S. M. Howdle, Aust. J. Chem. 2009, 62, 786–789.
[18] a) C. Villiers, J.-P. Dognon, R. Pollet, P. Thuꢃry, M. Ephritikhine, Angew.
Chem. Int. Ed. 2010, 49, 3465–3468; Angew. Chem. 2010, 122, 3543–
3546.
[4] a) H. Lee, E. Terry, M. Zong, N. Arrowsmith, S. Perrier, K. J. Thurecht, S. M.
Howdle, J. Am. Chem. Soc. 2008, 130, 12242–12243; b) E. J. Park, A. P.
Richez, N. A. Birkin, H. Lee, N. Arrowsmith, K. J. Thurecht, S. M. Howdle,
Polymer 2011, 52, 5403–5409; c) N. A. Birkin, N. J. Arrowsmith, E. J. Park,
A. P. Richez, S. M. Howdle, Polym. Chem. 2011, 2, 1293–1299; d) N. A.
Birkin, O. J. Wildig, S. M. Howdle, Polym. Chem. 2013, 4, 3791–3799;
e) E. Girard, T. Tassaing, J. D. Marty, M. Destarac, Polym. Chem. 2011, 2,
2222–2230; f) E. Girard, T. Tassaing, S. Camy, J. S. Condoret, J. D. Marty,
M. Destarac, J. Am. Chem. Soc. 2012, 134, 11920–11923; g) E. Girard, T.
Tassaing, C. Ladaviꢂre, J.-D. Marty, M. Destarac, Macromolecules 2012,
45, 9674–9681.
[5] a) P. Chambon, E. Cloutet, H. Cramail, Macromolecules 2004, 37, 5856;
b) P. Chambon, E. Cloutet, H. Cramail, T. Tassaing, M. Besnard, Polymer
2005, 46, 1057–1066; c) B. Renault, T. Tassaing, E. Cloutet, H. Cramail, J.
Polym. Sci. Part A 2007, 45, 5649–5661; d) B. Renault, E. Cloutet, P. La-
croix-Desmazes, H. Cramail, Macromol. Chem. Phys. 2008, 209, 535–543;
e) B. Renault, E. Cloutet, H. Cramail, T. Tassaing, M. Besnard, J. Phys.
Chem. A 2007, 111, 4181–4187; f) B. Renault, E. Cloutet, H. Cramail, Y.
Hannachi, T. Tassaing, J. Phys. Chem. A 2008, 112, 8379–8386.
[6] W. J. Blank, Z. A. He, E. T. Hessell, Prog. Org. Coat. 1999, 35, 19–29.
[7] a) H. Autrup, W. Dekant, B. Jansson, J. Linders, E. Testai, M. Vighi, M. Vi-
luksela, P. Verger, P. W. Wester, B. Bilitewski, Revised assessment of the
risks to health and the environment associated with the use of the four or-
ganotin compounds TBT, DBT, DOT and TPT, The Scientific Committee on
Health and Environmental Risks (SCHER), 2006; b) C. Harper, F. Llados,
G. Diamond, L. L. Chappell, Toxicological Profile for Tin and Tin com-
pounds, ATSDR (Agency for Toxic Substances and Disease Registry),
2005.
[8] M. K. Kiesewetter, E. J. Shin, J. L. Hedrick, R. M. Waymouth, Macromole-
cules 2010, 43, 2093–2107.
[9] R. L. Zimmerman, T. L. Renken, J. Cell. Plast. 1989, 25, 259–269.
[10] S. Hashimoto in International Progress in Urethanes, Vol. 3 (Eds.: K.
Ashida, K. C. Frisch), Technomic Publishing Co., Westport, 1981, pp. 43–
82.
[11] a) K. G. Flynn, D. R. Nenortas, J. Org. Chem. 1963, 28, 3527–3530; b) J. T.
Geoghegan, R. W. Roth, J. Appl. Polym. Sci. 1965, 9, 1089–1093; c) H.
Ulrich, H. E. Reymore Jr., A. A. R. Sayigh, J. Polym. Sci. Polym. Chem. Ed.
1975, 13, 2031–2033; d) J. Alsarraf, Y. A. Ammar, F. Robert, E. Cloutet, H.
Cramail, Y. Landais, Macromolecules 2012, 45, 2249–2256.
[12] a) J. Burkus, J. Org. Chem. 1962, 27, 474–476; b) G. D. Titskii, I. P. Garku-
sha-Bozhko, Ukr. Khim. Zh. (Russ. Ed.) 1993, 59, 94; c) R. K. Adams, J. R.
Cooper, US Patent No.3173897, 1965; d) E. E. Frauendorfer, W. B. Klock-
emann, US, Vol. 10052228, 2003.
[19] G. G. Suchkova, L. I. Maklakov, Vib. Spectrosc. 2009, 51, 333–339.
[20] a) J. A. Parker, J. J. Thomas, C. L. Zeise, J. Org. Chem. 1957, 22, 594–596;
b) M. E. Bailey, V. Kirss, R. G. Spaunburgh, Ind. Eng. Chem. 1956, 48, 794–
797; c) D. M. Simons, R. G. Arnold, J. Am. Chem. Soc. 1956, 78, 1658–
1659.
[21] a) A. Di Giacomo, J. Phys. Chem. 1961, 65, 696–697; b) M. Bertoldo, C.
Cappelli, S. Catanorchi, V. Liuzzo, S. Bronco, Macromolecules 2005, 38,
1385–1394; c) A. A. Caraculacu, S. Coseri, Prog. Polym. Sci. 2001, 26,
799–851.
[22] M. C¸oban, F. A. S. Konuklar, Comput. Theor. Chem. 2011, 963, 168–175.
[23] a) M. Lesar, M. Zigon, T. Malavasic, J. Appl. Polym. Sci. 1993, 47, 805–
814; b) H. R. Kricheldorf, W. E. Hull, Makromol. Chem. 1981, 182, 1177–
1196; c) B. Fu, C. Feger, W. J. MacKnight, N. S. Schneider, Polymer 1985,
26, 889–894; d) M. Sumi, Y. Chokki, Y. Nakai, M. Nakabayashi, T. Kanza-
wa, Makromol. Chem. 1964, 78, 146–156.
[24] H. Okuto, Makromol. Chem. 1966, 98, 148–163.
[25] M. Furukawa, T. Yokoyama, J. Polym. Sci. B 1979, 17, 175–180.
[26] a) M. Spirkova, M. Kubin, K. Dusek, J. Macromol. Sci. Chem. Part A 1987,
24, 1151–1166; b) M. Spirkova, M. Kubin, K. Dusek, J. Macromol. Sci.
Chem. Part A 1990, 27, 509–522.
[27] V. D. Vorotyntseva, T. N. Roginskaya, V. P. Moleva, T. N. Tikhonova, A. I.
Finkel’shtein, V. I. Zagranichnyi, J. Appl. Spectrosc. 1972, 17, 1317–1320.
[28] a) A. Barbarini, R. Maggi, A. Mazzacani, G. Mori, G. Sartori, R. Sartorio,
Tetrahedron Lett. 2003, 44, 2931–2934; b) F. S. Pereira, E. R. deAzevedo,
E. F. da Silva, T. J. Bonagamba, D. L. da Silva Agostꢄni, A. Magalh¼es, A. E.
Job, E. R. Pꢃrez Gonzꢅlez, Tetrahedron 2008, 64, 10097–10106; c) E. R.
Pꢃrez, M. O. Da Silva, V. C. Costa, U. P. Rodrigues-Filho, D. W. Franco, Tet-
rahedron Lett. 2002, 43, 4091–4093; d) E. R. Pꢃrez, R. H. A. Santos,
M. T. P. Gambardella, L. G. M. De Macedo, U. P. Rodrigues-Filho, J. C.
Launay, D. W. Franco, J. Org. Chem. 2004, 69, 8005–8011; e) S. Huang, J.
Ma, J. Li, N. Zhao, W. Wei, Y. Sun, Catal. Commun. 2008, 9, 276–280;
f) N. Della Ca’, B. Gabriele, G. Ruffolo, L. Veltri, T. Zanetta, M. Costa, Adv.
Synth. Catal. 2011, 353, 133–146.
[29] D. J. Heldebrant, P. G. Jessop, C. A. Thomas, C. A. Eckert, C. L. Liotta, J.
Org. Chem. 2005, 70, 5335–5338.
[30] a) P. G. Jessop, D. J. Heldebrant, X. Li, C. A. Eckert, C. L. Liotta, Nature
2005, 436, 1102; b) D. J. Heldebrant, C. R. Yonker, P. G. Jessop, L. Phan,
Energy Environ. Sci. 2008, 487–493; c) L. Phan, D. Chiu, D. J. Heldebrant,
H. Huttenhower, E. John, X. Li, P. Pollet, R. Wang, C. A. Eckert, C. L.
Liotta, P. G. Jessop, Ind. Eng. Chem. Res. 2008, 47, 539–545.
[31] B. Brzezinski, G. Schroeder, V. I. Rybachenko, L. I. Kozhevin, Spectrochim.
Acta Part A 2001, 57, 405–410.
[32] a) B. Brzezinski, G. Zundel, J. Mol. Struct. 1996, 380, 195–204; b) T. Glo-
wiak, I. Majerz, L. Sobczyk, B. Brzezinski, G. Wojciechowski, E. Grech, J.
Mol. Struct. 2000, 526, 177–184; c) P. Przybylski, G. Wojciechowski, B.
Brzezinski, G. Zundel, F. Bartl, J. Mol. Struct. 2003, 661–662, 171–182.
[33] M. Fabris, V. Lucchini, M. Noꢂ, A. Perosa, M. Selva, Chem. Eur. J. 2009,
15, 12273–12282.
[13] O. Coutelier, M. El Ezzi, M. Destarac, F. Bonnette, T. Kato, A. Baceiredo, G.
Sivasankarapillai, Y. Gnanou, D. Taton, Polym. Chem. 2012, 3, 605–608.
[14] a) P. G. Jessop, J. Supercrit. Fluids 2006, 38, 211–231; b) A. J. Parratt, D. J.
Adams, A. A. Clifford, C. M. Rayner, Chem. Commun. 2004, 2720–2721;
c) A. G. Nigmatov, I. V. Kuchurov, D. E. Siyutkin, S. G. Zlotin, Tetrahedron
Lett. 2012, 53, 3502–3505; d) I. Blakey, A. Yu, S. M. Howdle, A. K. Whit-
taker, K. J. Thurecht, Green Chem. 2011, 13, 2032–2037.
[15] a) P. F. Yang, Y. D. Han, T. D. Li, J. Y. Li, Chin. Chem. Lett. 2010, 21, 853–
855; b) E. N. Boitsov, E. P. Trub, Kinet. Catal. 1984, 25, 221–224.
[16] a) K. Schwetlick, R. Noack, J. Chem. Soc. Perkin Trans. 2 1995, 395–402;
b) B. Grepinet, F. Pla, P. Hobbes, P. Swaels, T. Monge, J. Appl. Polym. Sci.
2000, 75, 705–712.
[34] M. G. Allen, G. Van Dyke Tiers, US Patent No. 3817939, 1974.
[35] C. A. Smith, E. Cloutet, T. Tassaing, H. Cramail, Green Chem. 2013, 15,
2769–2776.
[36] V. Martinez, S. Mecking, T. Tassaing, M. Besnard, S. Moisan, F. Cansell, C.
Aymonier, Macromolecules 2006, 39, 3978–3979.
Received: November 25, 2013
[17] T. Suzuki, N. Tsuge, K. Nagahama, Fluid Phase Equilib. 1991, 67, 213–
226.
Published online on March 18, 2014
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 1380 – 1391 1391