Copolymerization of carbon dioxide and butadiene via a lactone intermediate

Article abstract of DOI:10.1038/nchem.1882

Although carbon dioxide has attracted broad interest as a renewable carbon feedstock, its use as a monomer in copolymerization with olefins has long been an elusive endeavour. A major obstacle for this process is that the propagation step involving carbon dioxide is endothermic; typically, attempted reactions between carbon dioxide and an olefin preferentially yield olefin homopolymerization. Here we report a strategy to circumvent the thermodynamic and kinetic barriers for copolymerizations of carbon dioxide and olefins by using a metastable lactone intermediate, 3-ethylidene-6-vinyltetrahydro-2H- pyran-2-one, which is formed by the palladium-catalysed condensation of carbon dioxide and 1,3-butadiene. Subsequent free-radical polymerization of the lactone intermediate afforded polymers of high molecular weight with a carbon dioxide content of 33 mol% (29 wt%). Furthermore, the protocol was applied successfully to a one-pot copolymerization of carbon dioxide and 1,3-butadiene, and one-pot terpolymerizations of carbon dioxide, butadiene and another 1,3-diene. This copolymerization technique provides access to a new class of polymeric materials made from carbon dioxide.

Full text of DOI:10.1038/nchem.1882

ARTICLES
PUBLISHED ONLINE: 9 MARCH 2014 | DOI: 10.1038/NCHEM.1882
Copolymerization of carbon dioxide and butadiene
via a lactone intermediate
Ryo Nakano, Shingo Ito and Kyoko Nozaki
*
Although carbon dioxide has attracted broad interest as a renewable carbon feedstock, its use as a monomer in
copolymerization with olefins has long been an elusive endeavour. A major obstacle for this process is that the
propagation step involving carbon dioxide is endothermic; typically, attempted reactions between carbon dioxide and an
olefin preferentially yield olefin homopolymerization. Here we report a strategy to circumvent the thermodynamic and
kinetic barriers for copolymerizations of carbon dioxide and olefins by using a metastable lactone intermediate,
3-ethylidene-6-vinyltetrahydro-2H-pyran-2-one, which is formed by the palladium-catalysed condensation of carbon
dioxide and 1,3-butadiene. Subsequent free-radical polymerization of the lactone intermediate afforded polymers of high
molecular weight with a carbon dioxide content of 33 mol% (29 wt%). Furthermore, the protocol was applied successfully
to a one-pot copolymerization of carbon dioxide and 1,3-butadiene, and one-pot terpolymerizations of carbon dioxide,
butadiene and another 1,3-diene. This copolymerization technique provides access to a new class of polymeric materials
made from carbon dioxide.
he development of mild methods for the synthesis of bulk is necessary before exothermic conditions can be achieved. We
chemicals using renewable feedstocks is a critical technological propose that kinetic limitations are also important for this reaction.
T
hurdle along the path to a sustainable chemical economy in The free energy of TS4 is higher than that of TS1, given that the acti-
the future. Carbon dioxide is one of the most attractive renewable vation energy from I to IE and that from ICE to ICEE are compar-
C1 resources, and it has many practical advantages, such as abun- able because both reactions involve alkyl chain-ends and ethylene
dance, economic efficiency and lack of toxicity1–9. Its favourable (Fig. 1a; see caption for definitions). Therefore, even if carbon
nature as a carbon source is, however, inextricably linked to its dioxide uptake (TS2) and carboxylation of ethylene (TS3) are acces-
inherent inertness. Not surprisingly, then, reactions of carbon sible, carbon dioxide/ethylene copolymerization is kinetically
dioxide must be combined with a high-energy reactant to gain a unfavourable unless the difference in free energy of TS4 can be
thermodynamic driving force. For large-scale utilization, chemical reduced. As this qualitative speculation could be valid for any mech-
coupling of carbon dioxide with easily available chemical feedstocks anism of polymerization, this kinetic problem may be a general bot-
is indispensable. Thus far only a limited number of processes have tleneck in carbon dioxide/olefin copolymerization. In this work, the
succeeded in the use of carbon dioxide, examples being the indus- formation of a metastable intermediate has facilitated an alternative
trial production of urea4, salicylic acid4, organic carbonates10–12 reaction pathway by which copolymerization can take place.
and polycarbonates11–17. In this context, olefins, one of the largest
We postulated that 3-ethylidene-6-vinyltetrahydro-2H-pyran-2-
classes of chemicals produced today, should be suitable co-reagents one (1) (Fig. 2a), which contains 29 wt% carbon dioxide, could
because of the high potential energy of their carbon–carbon double provide an avenue to the formal copolymerization of carbon
bonds. Although extensive studies have focused on the catalytic dioxide and 1,3-butadiene if appropriate conditions for its polymer-
coupling of olefins with carbon dioxide to form commodity and ization could be developed. Lactone 1, which was first reported by
fine chemicals18–21, methods to produce polymers from carbon Inoue and co-workers28, has attracted much attention as a platform
dioxide and olefins are still unexplored; all the known examples, chemical for the utilization of carbon dioxide29. As the synthesis of
which use dienes22, vinyl ethers23,24 or acrylonitrile25, produce lactone 1 is accompanied by the formation of several side products
only oligomers of low molecular weight. Here we provide an expla- (Fig. 2c), considerable efforts have been undertaken to achieve a high
nation for the intrinsic difficulty of the copolymerization reaction, yield and selectivity for lactone 130–37. Behr et al. optimized the reac-
on which basis we developed the copolymerization of carbon tion to accomplish 95% selectivity over butadiene conversion, and
dioxide and 1,3-dienes via a lactone intermediate.
the conditions were applied on a mini-plant scale38,39. The homo-
As a result of the thermodynamic insight that Miller and polymerization of 1 was investigated previously by Dinjus and co-
co-workers developed, we considered a kinetic limitation for olefin/ workers40, but they reported that the reaction did not proceed in
carbon dioxide copolymerization (Fig. 1a)26. According to Miller’s the presence of radical, cationic or anionic initiators40. They
protocol26, we performed thermodynamic benchmark calculations suggested that the lack of reactivity of 1 may stem from its structural
in which the CBS-4M//B3LYP/6-31G* method exhibited the best similarity to either a tiglic acid ester (Fig. 2b, red) or an a-substituted
fit with reported experimental values27. The free-energy changes allyl ester (Fig. 2b, blue), both of which are substrates known to
for the incorporation of ethylene and carbon dioxide were exhibit poor reactivity when subjected to common polymerization
determined as –12.5 and 21.9 kcal mol21, respectively. Thus, conditions41–45. Nevertheless, we detected serendipitously that
alternating carbon dioxide/ethylene copolymerization is endother- oligomerization of 1 occurred on storage under air for several
mic by 9.4 kcal mol21 and excess ethylene insertion (.63 mol%) weeks. This encouraging observation led us to pursue the
Department of Chemistry and Biotechnology, Graduate School of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan.
e-mail: nozaki@chembio.t.u-tokyo.ac.jp
*
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.
1

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1882
ARTICLES
TS4
a
Growing chain end
P
polymer chain
*
ICE +
TS1
TS3
TS2
I +
ΔG4‡
(≈ ΔΔG1‡ + 9.4)
IC/IC´ +
I + CO2
ΔΔG1‡
9.4
Unknown
O
O
0.0
P
–3.1
P
P
O
O
*
*
–12.5
ICE
IC
O
*
I
P
P
O
P
O
*
*
*
IE
IC´
ICEE
O
i. Ethylene homopolymerization
ii. Carbon dioxide incorporation into ethylene polymerization
ii. Carbon dioxide incorporation into butadiene 1,4-polymerization
b
O
TS4´
ICB +
P
O
*
TS1´
I +
TS2´
TS3´
I + CO2
IC/IC´ +
ΔG4´‡
(≈ ΔΔG1´‡ + 14.6)
14.6
ΔΔG1´‡
Unknown
ICB
5.2
IC or IC´
0.0
0.0
I + lactone 1 (L)
ICBB
–9.3
+ CO2
P
P
*
+ 2 butadiene
*
I
O
P
*
O
P
+
IB
–26.0
*
O
O
IL
1 (L)
iii. Polymerization via lactone 1
i. Butadiene homopolymerization
Figure 1 | Tentative energy diagram of carbon dioxide/ethylene copolymerization and carbon dioxide/butadiene copolymerization. I, C, E, B and L are the
initiation from primary alkyl species, carbon dioxide, ethylene, butadiene and lactone 1, respectively. For example, IC or IC′ represents the isomeric reaction
products from I and C. The growing chain end is shown simply with a purple asterisk because the discussion is applicable to any radical, cationic and anionic
polymerization. Values DG (kcal mol21) are from the starting primary alkyl species I, approximated by the values calculated (CBS-4M//B3LYP/6-31G*)
according to Price and colleagues26. a, DDG1‡ is the assumed reaction barrier for the alkyl chain-end and ethylene. As successive incorporation of ethylene
is indispensable to make the reaction exothermic, direct carbon dioxide/ethylene copolymerization (path ii, blue) has to compete with ethylene
homopolymerization (path i, black). DDG4‡ would be comparable to DDG1‡ because both are the reaction of an alkyl chain-end with ethylene. Therefore,
DG4‡ should be much higher than DDG1‡, which reflects the thermodynamic stability of the intermediates I and ICE. b, DDG1′‡ is the assumed reaction
barrier for the alkyl chain-end and butadiene. The direct carbon-dioxide incorporation into butadiene 1,4-polymerization (path ii, blue) is kinetically
overwhelmed by butadiene homopolymerization (path i, black), as described in the case of ethylene. The formation of lactone 1 is isothermic, and the
polymerization of 1 is highly exothermic (path iii, red). A kinetic explanation of the copolymerization here is as follows. In the first step, lactone 1 synthesis,
the absence of any polymerization initiator leads to the formation of lactone 1. In the second step, lactone polymerization, the absence or deactivation of
the palladium catalyst suppresses the decomposition of lactone 1 back into butadiene and carbon dioxide, and the lactone 1 can be converted into poly-1.
Further depiction of the kinetic explanation is given in the Supplementary Information.
polymerization of 1 further, despite the prior report detailing the polymerization (Table 1, entries 2 and 3), which is consistent with
reluctance of 1 to participate in the desired transformation.
propagation by a radical mechanism. Emulsion conditions using a
The results of our efforts to promote the polymerization of 1 are larger quantity of initiator (Table 1, entry 4) led to an improved con-
summarized in Table 1. We found that heating neat 1 in the pres- version of 1, but a similar molecular weight of the product com-
ence of a typical radical initiator, 2,2′-azobis(isobutyronitrile) pared to that in neat conditions. Conditions for controlled radical
(AIBN), produced polymer poly-1 after 24 hours at 100 8C polymerization, such as nitroxide-mediated radical polymerization
(entry 1). The presence of a radical inhibitor, such as galvinoxyl (Table 1, entry 3) or atom-transfer radical polymerization
or 2,2,6,6-tetramethylpiperidine 1-oxyl (TEMPO), suppressed the (Table 1, entry 5) did not initiate the polymerization. The yield
2
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1882
ARTICLES
a
c
Side products in the synthesis of 1
Copolymer
with 33 mol%
CO2 incorporation
CO2
+
2
3
4
O
O
O
O
O
O
Pd-catalysed
lactone formation
Radical
polymerization
R =
1
H
ROOC
ROOC
b
R
R
Tiglate
ester
Allylic
ester
5
6
O
O
O
OR
Figure 2 | The concept of this study: copolymerization of carbon dioxide and 1,3-butadiene via a lactone intermediate, 3-ethylidene-6-vinyltetrahydro-2H-
pyran-2-one (1). a, The polymerization of lactone 1 affords a carbon dioxide/butadiene copolymer with a carbon dioxide content of 33 mol%, which could
not be accessed via direct copolymerization. b, The two carbon–carbon double bond moieties of lactone 1 have structural similarities with a tiglate ester (red)
and allylic ester (blue). c, Side products in the synthesis of lactone 1 under palladium catalysis.
and molecular weight of poly-1 increased slightly with the use of characterization). The resonances of carbons a, b and e observed at
1,1′-azobis(cyclohexane-1-carbonitrile) (V-40), which has a higher 178, 57 and 81 ppm, respectively, were all downfield shifted com-
10 hour life-time temperature in toluene (88 8C) than that of pared to typical saturated monocyclic esters, in accordance with a
AIBN (65 8C) (Table 1, entry 6)46. The addition of acetic acid led strained camphor-like bicyclic structure (a). Thus, the polymeriz-
to higher molecular weight, 19,000 g mol21, with a low polydisper- ation of lactone 1 proceeded via an alternating reaction of the
sity index of 1.1 (Table 1, entry 7).
2,3-disubstituted acrylate moiety and the allylic moiety to give a
NMR spectroscopic analyses indicated that poly-1 produced by polymer that possessed bicyclic structures in the main chain
the methods shown in Table 1, entries 1, 4, 6 and 7, exhibited exclu- (Fig. 4). The key to the successful utilization of the multiple-substi-
sive formation of the bicyclic repeat unit a (Fig. 3b; see the tuted acrylate moiety, which previously had been known as inert for
Supplementary Information for additional details on the structural radical polymerization, can be explained as follows. Owing to the
Table 1 | Radical polymerization of lactone 1.
Me
O
O
Me
Radical
Me
initiator
O
O
Me
Additive
100 °C, 24 h
O
O
O
O
n
1
β
γ
α
Poly-1
Entry
Initiator
Additive
Solvent
Yield (%)
Mn (103)*
Mw /Mn*
a:b:g†
1
AIBN
AIBN
AIBN
KPS
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
EC
EC
18
–
–
32
–
22
17
2.3
18
Trace
15
–
–
26
–
2.2
–
–
2.0
–
5.7
19
–
Insoluble
–
Insoluble
–
–
19
–
85
62
1.4
–
–
1.4
–
1.3
1.1
–
–
–
–
–
–
1.2
–
1.5
1.1
10:0:0
–
–
10:0:0
–
10:0:0
10:0:0
6:1:3
–
–
–
–
–
2‡
Galvinoxyl
TEMPO
Aqueous SDS
CuBr/bpy
–
AcOH
MgCl2
ScCl3
YCl3
ZrCl4
FeCl3
CuCl2
ZnCl2
AlCl3
ZnCl2
ZnCl2
3‡
4§
5ꢀ
MBrP
V-40
V-40
V-40
V-40
V-40
V-40
V-40
V-40
V-40
V-40
V-40
V-40
6
7}
8#
9#
10#
11#
12#
13#
14#
15#
16**
17††
5:2:3
–
3:5:2
3:4:3
48
59
A mixture of 1 (1.0 mmol) and initiator (0.010 mmol, 1.0 mol%) was stirred at 100 8C for 24 hours in a 5 ml vial, unless otherwise stated. KPS, potassium persulfate; SDS, sodium dodecylsulfate; MBrP, methyl
2-bromopropionate; bpy, 2,2′-bipyridine; EC, ethylene carbonate. *Determined by RALLS analysis with size-exclusion chromatography (SEC) using THFas eluent. †Determined by 1H NMR analysis. ‡0.020 mmol
(2.0 mol%) additive. §5.0 mmol scale. Water (12.5 ml), SDS (0.188 mmol, 15 mM) and KPS (0.250 mmol, 5.0 mol%). ꢀ0.010 mmol (1.0 mol%) CuBr and 0.020 mmol (2.0 mol%) bpy. }2.0 mmol (200 mol%)
additive. #0.50 mmol (50 mol%) additive. **1.0 mmol (100 mol%) zinc salt and 4.0 mmol EC. ††0.20 mol of 1, 0.20 mol of ZnCl2, and 0.80 mol of EC, 48 hours.
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
3
© 2014 Macmillan Publishers Limited. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1882
ARTICLES
structures b and g, as deduced from NMR spectroscopic data
(Fig. 3c) and infrared spectra (see the Supplementary Information
for additional details concerning structural characterization). The
resonances of the sp2 carbon observed at 125–130 ppm (b′) and
130–140 ppm (h′), and the resonance of the carbonyl carbon
observed at 165–169 ppm (a′) in the 13C NMR spectra (see also
Supplementary Figs 10 and 12) suggest that poly-1 contains an
internal conjugated ester unit (b). The proposed structure for b
gives a relatively sharp signal, observed at 63 ppm, as the secondary
carbon at the a-position of ester (e′) and other resonances.
Presumably, the unit b is formed by propagation through an allyl
ester radical intermediate prior to cyclization (Fig. 4), which is con-
sistent with the reported acceleration effect of zinc chloride on the
polymerization of allyl acetate47. Additionally, the broad signal
observed at 175 ppm (a′′) in the 13C NMR spectra is assignable to
a non-conjugated ester, which is different from the repeating unit
a. Given that there is no unassigned 13C NMR resonance at
60–100 ppm, and that the infrared absorption at 1,713 cm21 should
be from the six-membered lactone, we conceived the structure g,
which may be formed by intramolecular hydrogen abstraction
from an a-carbonyl radical intermediate (Fig. 4). The coordination
of zinc chloride to the carbonyl oxygen would be expected to
enhance the electrophilicity of an a-carbonyl radical formed by
radical attack on the tiglate moiety. Subsequent intramolecular
hydrogen abstraction to form an allyl radical species is thus expected
to be accelerated49. Most of the polymer is free of decarboxylation, as
confirmed by quantitative NMR spectroscopy, matrix-assisted laser
desorption/ionization–time of flight mass spectroscopy and elemen-
tal analyses. The glass-transition temperature was as high as 192 8C
in poly-1 (zinc chloride initiator, Table 1, entry 17).
a
I
Solvent
C
B
A
D
G
H
D
E
C
F
G
O
O
H
E
F
I
B
A
b
Solvent
i
g
h
f
n
c,d,g(CH2)
h(CH)
O
c
b
a
O
d
e
b(C)
f(CH)
e(CH)
a(C)
i(CH3)
h´
a´
b´
c
O
γ
O
f˝
α
a
e˝
e´
b˝
O
β
O
a˝
b
O
O
n
e
Solvent
a´
Solvent
e´ b
a˝
a
h´+ e˝
f˝
b´
e
The formation and polymerization of 1 was adapted successfully
into a one-pot/two-step process (Table 2, entry 1). We found that
the presence of the constitutional isomers 2 and 3, formed during
the reaction of 1,3-butadiene and carbon dioxide to give 1, retarded
the subsequent polymerization. Indeed, isolated lactone 3 could not
be polymerized even under zinc chloride/ethylene carbonate con-
ditions (see the Supplementary Information). As such, optimization
of the reaction conditions to minimize the formation of 2 or 3 was
essential for the development of a one-pot reaction. According to
200 180 160 140 120
100
ppm
80
60
40
20
0
Figure 3 | 13C NMR spectra and the assignments for poly-1. a, 13C NMR
spectrum of lactone 1 (C6D6). b, 13C NMR spectrum of poly-1 (CDCl3)
obtained with KPS (Table 1, entry 4). The consumption of the alkenyl
carbons and downfield signals of a, e and b indicate the exclusive formation
of a bicyclic unit a. c, 13C NMR spectrum of poly-1 (trifluoroacetic acid)
obtained with ZnCl2 (Table 1, entry 17). In addition to the unit a, new signals
are observed in the carbonyl and alkenyl region (see the Supplementary
Information for the full details).
Without
cyclization
On allylic
ester
β
α
steric hindrance between the radical chain-end and the carbon–
carbon double bond of the next monomer, the continuous incorpor-
ation of substituted acrylate, such as tiglate, is impossible. In
contrast, the alternating incorporation with the allylic ester moiety
liberalizes the steric hindrance between the chain end and the
next monomer. The glass-transition temperature of poly-1 (KPS
initiator, Table 1, entry 4) was found to be 178 8C, which suggests
its potential application as a new engineering plastic.
O
R
O
Cyclization
R
O
O
R
1
O
O
On tiglate
H
The effects of Lewis-acid additives47 and solvents on the
polymerization were also probed (Table 1, entries 8–15; see also
Supplementary Tables 3 and 4). Whereas most of the Lewis acids
we examined retarded the polymerization of 1 (Table 1, entries 8,
10, 12, 13 and 15) or gave insoluble products (Table 1, entries 9
and 11), the addition of zinc chloride was found to improve the mol-
ecular weight of poly-1. The reaction mixture, however, solidified
within an hour without any other additive (Table 1, entry 14).
Hydrogen
abstraction
R
O
H
γ
O
After extensive screening (see Supplementary Table 5), we found Figure 4 | Proposed reaction pathways to form repeating units of a, b and
that the concurrent use of an equimolar amount of zinc chloride47 g. If the radical chain-end (R†) attacks the allylic ester moiety (upwards
and ethylene carbonate as a solvent48 led to a significant increase in arrow), the upper radical intermediate can undergo intramolecular exo-
the polymer yield (48%) and Mn (85,000 g mol21; Table 1, entry cyclization to form unit a, or propagate without cyclization to form unit b.
16). The yield and molecular weight did not deteriorate when the If the radical chain-end attacks the tiglate ester moiety (downwards arrow),
reaction was repeated on a scale of 0.20 mol (Table 1, entry 17).
the lower radical intermediate can undergo intramolecular exo-cyclization to
The product poly-1 obtained in the presence of Lewis acids pos- form unit a or propagate after intramolecular hydrogen abstraction to
sessed not only the bicyclic structure a, but also monocyclic form unit g.
4
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1882
ARTICLES
Table 2 | One-pot/two-step co/terpolymerization of carbon dioxide and 1,3-dienes.
Pd(acac)2 (X µmol)
PPh3 (Y µmol)
V-40 (0.165 mmol)
ZnCl2 (16.5 mmol)
Dienes
CO2
+
CO2/diene
co/terpolymer
R3
O
O
Ethylene carbonate
100 °C, 24 h
(86–87 mmol)
(70 or 75 mmol)
R1
R2
(+ side products)
80 °C, Z h
First step: lactone synthesis
Second step: polymerization
Entry Initial load of
dienes (mmol)
X
(m
Y
(m
Z
Major lactone
products
Yield/
Yield/
CO2
Mn
Mw
Tg
§
mol)
mol) (h)
selectivity selectivity incorporation
of lactone of polymer (wt%)‡
(%)*
(103)§
/Mn (8C)ꢀ
(%)†
1
Butadiene (70)
50
150
3
1 (R1 ¼ R2 ¼ R3 ¼ H)
1 (R1 ¼ R2 ¼ R3 ¼ H)
44/77
47/81
23
20
19ꢀ
5.5
1.6ꢀ
2.5
102}
128#
2
Butadiene (25)
Isoprene (50)
100
300
20
–
–
46/–
35/–
63
7a (R1 ¼ Me, R2 ¼ R3 ¼ H)
7b (R1 ¼ R3 ¼ H, R2 ¼ Me)
3
Butadiene (25)
1,3-pentadiene (50)
100
300
20
8 (R1 ¼ R2 ¼ H, R3 ¼ Me)
24
16
2.0
33
A mixture of carbon dioxide, dienes, Pd(acac)2, PPh3 and EC was stirred at 80 8C in a 50 ml stainless-steel autoclave. After removing redundant gaseous materials, the reaction mixture was stirred at 100 8C for
24 hours with V-40 and ZnCl2. *Yield determined by GC analysis versus EC as the internal standard. Selectivity for lactone 1 over butadiene conversion. †Yield based on the amount of dienes in the initial feed.
Selectivity based on the amount of dienes consumed in the first step. ‡Determined by elemental analysis. §Determined by RALLS analysis with SEC using THF as eluent. ꢀDetermined by differential scanning
calorimetry. }Only the fraction of (poly-1)′ soluble in hot DMF was analysed. #Only the fraction of (poly-1)′ insoluble in hot DMF was analysed.
Behr’s report39, we performed the preparation of lactone 1 using triisopropylphosphine system. Our one-pot/two-step procedure
ethylene carbonate as the solvent, and obtained a modest yield was combined with this three-component coupling strategy to
(44% over the initial load of 1,3-butadiene) and selectivity produce novel terpolymers of carbon dioxide, butadiene and
(.76% over a butadiene conversion of 58%). Although the crude another 1,3-diene. By using triphenylphosphine (which is sterically
mixture of lactone 1 contained about a 14% yield of side products, smaller than triisopropylphosphine) as a ligand the cross-coupling
including butadiene dimer 4, and acids and esters 5 and 6, as shown products 7a and 7b (Table 2, entry 2) and 8 (Table 2, entry 3)
in Fig. 2, the total yield of 2 and 3 was found to be ,0.3%. After were predominantly formed under the conditions of the first step.
venting carbon dioxide and butadiene, the resulting mixture was After removing carbon dioxide and dienes in vacuo, the resulting
heated for 24 hours in the presence of V-40 and zinc chloride. mixture was heated for 24 hours in the presence of V-40 and zinc
The carbon dioxide/butadiene copolymer (poly-1)′ was isolated in chloride to produce the carbon dioxide/butadiene/isoprene
a good yield (47% over the initial load of 1,3-butadiene) and with (Table 2, entry 2) or carbon dioxide/butadiene/1,3-pentadiene
good selectivity (.81% based on 1,3-butadiene consumption in terpolymers (Table 2, entry 3) with high molecular weights
the first step (58%)). The yield and selectivity of the polymer were (Mn ¼ 5,500 and 16,000, respectively) in good yields (46% and
slightly higher than those of lactone 1 in the first step, because 35%, respectively, over the initial load of dienes). Elemental analysis
side products 4, 5 and 6 were also incorporated into the copolymer. revealed that the carbon-dioxide incorporation of these polymers
Elemental analysis revealed that the ratio of incorporated carbon were 20 wt% and 24 wt%, slightly lower than the carbon-dioxide
dioxide to 1,3-butadiene was 1:2.7, which corresponds to a incorporation of lactone monomers 7a, 7b and 8 (26.5 wt%). The
carbon-dioxide incorporation of 27 mol% (23 wt%). The possibility glass-transition temperatures of these polymers were 63 and
of decomposition of lactone 1 under the conditions of the second 33 8C, which shows that the introduction of methyl substituents
step can be excluded, judging from a control experiment that the into polylactone poly-1 causes a significant decrease in the
palladium catalyst did not produce lactone 1 from carbon dioxide glass-transition temperatures.
and butadiene in the presence of a stoichiometric amount of zinc
A theoretical comparison of the ground-state energies for inter-
chloride. The soluble fraction of (poly-1)′ in hot dimethylforma- mediates and products in both the polymerization of 1 described
mide (DMF) was found to have a moderate molecular weight of here (Fig. 1b, path iii) and carbon-dioxide incorporation into buta-
19,000 g mol21 and a polydispersity index of 1.6. The glass-tran- diene 1,4-polymerization (Fig. 1b, path ii) was instructive. The ther-
sition temperature of (poly-1)′ was lower than that of poly-1, but modynamic gain of each monomer incorporation was calculated
still as high as 102 8C for the soluble fraction and 128 8C for the using a composite CBS method (CBS-4M//B3LYP/6-31G*) accord-
insoluble fraction in hot DMF.
The one-pot/two-step protocol was further applied to the terpo- incorporation via 1,4-addition and for carbon-dioxide incorpor-
lymerization of carbon dioxide, butadiene and another 1,3-diene ation were determined as –9.3 kcal mol21 and 23.9 kcal mol21
ing to Miller’s protocol26. The changes in free energy for butadiene
,
(Table 2, entries 2 and 3). Despite the establishment of the respectively, which indicates that the consumption of one carbon–
lactone synthesis from 1,3-butadiene and carbon dioxide, the use carbon double bond of butadiene only allows ,28 mol% incorpor-
of other dienes instead of 1,3-butadiene for telomerization with ation of carbon dioxide, which is lower than 36 mol% in the case of
carbon dioxide still remains challenging32,34. The major limitations ethylene. Although conjugated alkenes generally have higher
include a slow reaction rate and low yield of lactone formation (kinetic) reactivity, as exemplified by some coupling reactions
because of an augmented steric hindrance and the formation of with carbon dioxide in the literature19–21,28, the consumption of
complex mixtures that contain positional isomers. In contrast, the one carbon–carbon double bond of a conjugated alkene provides
three-component coupling of carbon dioxide, butadiene and less thermodynamic gain than that of a non-conjugated alkene. In
isoprene/1,3-pentadiene studied by Behr et al.32 demonstrates the carbon-dioxide incorporation into butadiene 1,4-polymeriz-
reasonable yields in the palladium acetylacetonate ((Pd(acac)2)/ ation (Fig. 1b, path ii), again the surplus activation energy DDG′‡
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.
5

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1882
References
1. Aresta, M. Carbon Dioxide as Chemical Feedstock (Wiley-VCH, 2010).
2. Sakakura, T., Choi, J. C. & Yasuda, H. Transformation of carbon dioxide. Chem.
Rev. 107, 2365–2387 (2007).
3. Arakawa, H. et al. Catalysis research of relevance to carbon management:
progress, challenges, and opportunities. Chem. Rev. 101, 953–996 (2001).
4. Behr, A. Use of carbon dioxide in industrial organic syntheses. Chem. Eng.
Technol. 10, 16–27 (1987).
5. Omae, I. Recent developments in carbon dioxide utilization for the production
of organic chemicals. Coord. Chem. Rev. 256, 1384–1405 (2012).
6. Sumida, K. et al. Carbon dioxide capture in metal–organic frameworks. Chem.
Rev. 112, 724–781 (2012).
ARTICLES
of TS4′ over TS1′ should be 14.6 kcal mol21 (see Fig. 1b). In con-
trast, the formation of lactone 1 and the following cyclopolymeriza-
tion of 1 to form the structure a were calculated to be isothermic by
0.0 kcal mol21 and exothermic by –26.0 kcal mol21, respectively
(Fig. 1b, path iii). Thus, the polymerization via lactone 1 proceeded
without endothermic intermediates to afford a short cut to avoid
TS4′ and kinetic competition with butadiene homopolymerization
(Fig. 1b, path i). In the first step, lactone 1 synthesis, the absence
of any polymerization initiator results in the formation of lactone
1. In the second step, lactone polymerization, the absence or deac-
tivation of the palladium catalyst suppresses the decomposition of
lactone 1 back into butadiene and carbon dioxide, and so leads to
7. Huang, K., Sun, C-L. & Shi, Z-J. Transition-metal-catalyzed C–C bond formation
through the fixation of carbon dioxide. Chem. Soc. Rev. 40, 2435–2452 (2011).
poly-1 formation. Thermodynamically, the polymerization via 8. Tsuji, Y. & Fujihara, T. Carbon dioxide as a carbon source in organic
transformation: carbon–carbon bond forming reactions by transition-metal
catalysts. Chem. Commun. 48, 9956–9964 (2012).
9. Aresta, M., Dibenedetto, A. & Angelini, A. The use of solar energy can enhance
lactone was also advantageous, as follows. Based on our calculations,
direct carbon dioxide/butadiene copolymerization (Fig. 1b, path ii)
can only occur with a carbon-dioxide incorporation ratio up to
28 mol%, upon which the reaction becomes endothermic.
the conversion of carbon dioxide into energy-rich products: stepping towards
artificial photosynthesis. Phil. Trans. R. Soc. A 371, 20120111 (2012).
However, 33 mol% carbon dioxide was incorporated in our copoly-
merization via the lactone intermediate (Fig. 1b, path iii). The high
content of carbon dioxide arises from the consumption of four or
three double bonds from two equivalents of 1,3-butadiene per mol-
ecule of carbon dioxide, which provides a thermodynamic bypass
and driving force sufficiently strong to overcome the endothermic
penalty of carbon-dioxide incorporation.
10. Shaikh, A-A. G. & Sivaram, S. Organic carbonates. Chem. Rev. 96,
951–976 (1996).
11. Darensbourg, D. J. & Holtcamp, M. W. Catalysts for the reactions of epoxides
and carbon dioxide. Coord. Chem. Rev. 153, 155–174 (1996).
12. Lu, X-B. & Darensbourg, D. J. Cobalt catalysts for the coupling of CO2 and
epoxides to provide polycarbonates and cyclic carbonates. Chem. Soc. Rev. 41,
1462–1484 (2012).
13. Fukuoka, S. et al. A novel non-phosgene polycarbonate production process using
by-product CO2 as starting material. Green Chem. 5, 497–507 (2003).
14. Coates, G. W. & Moore, D. R. Discrete metal-based catalysts for the
copolymerization of CO2 and epoxides: discovery, reactivity, optimization, and
mechanism. Angew. Chem. Int. Ed. 43, 6618–6639 (2004).
15. Darensbroug, D. J. Making plastics from carbon dioxide: salen metal complexes
as catalysts for the production of polycarbonates from epoxides and CO2. Chem.
Rev. 107, 2388–2410 (2007).
16. Sugimoto, H. & Inoue, S. Copolymerization of carbon dioxide and epoxide.
J. Polym. Sci. A 42, 5561–5573 (2004).
In summary, we have achieved the copolymerization of carbon
dioxide and 1,3-butadiene via the lactone intermediate 3-ethyli-
dene-6-vinyltetrahydro-2H-pyran-2-one (1). The homopolymeriza-
tion of lactone 1 provided a breakthrough for the inherent kinetic
and thermodynamic obstacles in the long-standing challenge of
carbon dioxide/olefin copolymerization. We thus obtained carbon
dioxide/butadiene copolymers with a carbon-dioxide incorporation
ratio of 33 mol%, which is comparable to a theoretical maximum
value by typical direct copolymerization of carbon dioxide and
1,3-butadiene. Also, the protocol was adapted successfully to one-
pot carbon dioxide/butadiene copolymerization, and one-pot
terpolymerization of carbon dioxide, butadiene and another 1,3-
diene. Our accomplishment reported here opens up not only a sus-
tainable route for polymer synthesis, but also potential access to a
new class of polymeric materials. Butadiene is already produced
on very large scales (.7 × 106 t yr21)50, which suggests that the
present carbon dioxide/butadiene copolymerization could concei-
vably lead to mass utilization of carbon dioxide for the synthesis
of synthetic polymers.
17. Kember, M. R., Buchard, A. & Williams, C. K. Catalysts for CO2/epoxide
copolymerisation. Chem. Commun. 47, 141–163 (2011).
18. Lejkowski, M. L. et al. The first catalytic synthesis of an acrylate from CO2 and an
alkene—a rational approach. Chem. Eur. J. 18, 14017–14025 (2012).
19. Takaya, J., Sasano, K. & Iwasawa, N. Efficient one-to-one coupling of easily
available 1,3-dienes with carbon dioxide. Org. Lett. 13, 1698–1701 (2011).
20. Williams, C. M., Johnson, J. B. & Rovis, T. Nickel-catalyzed reductive
carboxylation of styrenes using CO2. J. Am. Chem. Soc. 130,
14936–14937 (2008).
21. Takimoto, M., Nakamura, Y., Kimura, K. & Mori, M. Highly enantioselective
catalytic carbon dioxide incorporation reaction: nickel-catalyzed asymmetric
carboxylative cyclization of bis-1,3-dienes. J. Am. Chem. Soc. 126,
5956–5957 (2004).
22. Soga, K., Hosoda, S. & Ikeda, S. Copolymerization of carbon dioxide and some
diene compounds. Die Makromol. Chemie 176, 1907–1911 (1975).
23. Soga, K., Hosoda, S., Tasuka, Y. & Ikeda, S. Copolymerization of carbon dioxide
and ethyl vinyl ether. J. Polym. Sci. Polym. Lett. 13, 265–268 (1975).
24. Soga, K., Sato, M., Hosoda, S. & Ikeda, S. Copolymerization of methyl vinyl ether
and carbon dioxide. J. Polym. Sci. Polym. Lett. 13, 543–548 (1975).
25. Chiang, W-Y. Copolymerization of carbon dioxide and acrylonitrile. Proc. Natl
Sci. Council Republic of China 2, 170–176 (1978).
26. Price, C. J., Reich, B. J. E. & Miller, S. A. Thermodynamic and kinetic
considerations in the copolymerization of ethylene and carbon dioxide.
Macromolecules 39, 2751–2756 (2006).
27. Darensbourg, D. J. & Yeung, A. D. Thermodynamics of the carbon dioxide–
epoxide copolymerization and kinetics of the metal-free degradation: a
computational study. Macromolecules 46, 83–95 (2013).
28. Sasaki, Y., Inoue, Y. & Hashimoto, H. Reaction of carbon dioxide with butadiene
catalysed by palladium complexes. Synthesis of 2-ethylidenehept-5-en-4-olide.
J. Chem. Soc. Chem. Commun. 605–606 (1976).
29. Behr, A. & Henze, G. Use of carbon dioxide in chemical syntheses via a lactone
intermediate. Green Chem. 13, 25–39 (2011).
30. Musco, A. Co-oligomerization of butadiene and carbon dioxide catalysed by
tertiary phosphine–palladium complexes. J. Chem. Soc. Perkin Trans. 1
693–698 (1980).
31. Behr, A. & Juszak, K-D. Palladium-catalyzed reaction of butadiene and carbon
dioxide. J. Organomet. Chem. 255, 263–268 (1983).
Methods
General procedure for the polymerization of lactone 1. A mixture of lactone 1
(1.0 mmol) and initiator (0.010 mmol) was stirred at 100 8C for one day in a 5 ml
vial. The resulting mixture was diluted with excess methanol, and the formed
precipitate was collected and washed with methanol and water. After drying under
vacuum, the crude materials were dissolved in hot DMF or in trifluoroacetic acid and
reprecipitated with aqueous hydrochloric acid (1.0 M) to afford poly-1.
General procedure for one-pot co/terpolymerization of carbon dioxide and
dienes. To a 50 ml stainless-steel autoclave with a magnetic stirring bar, Pd(acac)2
(0.050–0.100 mmol) and triphenylphosphine (PPh3) (0.150–0.300 mmol) were
charged with ethylene carbonate (7.50 g). To the mixture was added dienes
(70–75 mmol in total) with stirring. Then the autoclave was pressurized with carbon
dioxide (85–86 mmol) and stirred at 80 8C for 4–20 hours. The gas pressure was
released and the reaction mixture sampled for gas chromatography (GC) analysis.
After removing the redundant dienes, the reaction mixture was stirred at 100 8C for
24 hours with V-40 (40.1 mg, 0.165 mmol) and ZnCl2 (2.25 g, 16.5 mmol). The
resulting mixture was diluted with excess methanol (about 150 ml), and the
remaining precipitates were collected and washed with methanol and water. The
obtained solid product was poured into hot DMF (approximately 30 ml) and
reprecipitated with aqueous hydrochloric acid (1.0 M) to afford the copolymer
or terpolymer.
32. Behr, A., He, R., Juszak, K-D., Kru¨ger, C. & Tsay, Y-H. Steuerungsmo¨glichkeiten
bei der u¨bergangsmetall-katalysierten umsetzung von 1,3-dienen mit
kohlendioxid. Chem. Ber. 119, 991–1015 (1986).
Received 8 July 2013; accepted 28 January 2014;
published online 9 March 2014
6
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1882
ARTICLES
33. Braunstein, P., Matt, D. & Nobel, D. Carbon dioxide activation and catalytic
lactone synthesis by telomerization of butadiene and carbon dioxide. J. Am.
Chem. Soc. 110, 3207–3212 (1988).
34. Dinjus, E. & Leitner, W. New insights into the palladium-catalysed synthesis of
d-lactones from 1,3-dienes and carbon dioxide. Appl. Organomet. Chem. 9,
43–50 (1995).
35. Pitter, S. & Dinjus, E. Phosphinoalkyl nitriles as hemilabile ligands: new aspects
in the homogeneous catalytic coupling of CO2 and 1,3-butadiene. J. Mol.
Catal. A 125, 39–45 (1997).
36. Behr, A. & Heite, M. Telomerisation von kohlendioxid und 1,3-butadien:
verfahrensentwicklung via miniplant-technik. Chem. Ing. Tech. 72,
58–61 (2000).
37. Behr, A. & Becker, M. The telomerisation of 1,3-butadiene and carbon dioxide:
process development and optimisation in a continuous miniplant. Dalton Trans.
4607–4613 (2006).
38. Behr, A., Bahke, P. & Becker, M. Palladium-katalysierte telomerisation von
kohlendioxid mit butadien im labor- und miniplantmaßstab. Chem. Ing. Tech.
76, 1828–1832 (2004).
39. Behr, A., Bahke, P., Klinger, B. & Becker, M. Application of carbonate solvents in
the telomerisation of butadiene with carbon dioxide. J. Mol. Catal. A 267,
149–156 (2007).
40. Haack, V., Dinjus, E. & Pitter, S. Synthesis of polymers with an intact lactone
ring structure in the main chain. Die Angew. Makromol. Chem. 257,
19–22 (1998).
47. Iio, K., Kobayashi, K. & Matsuo, M. Radical polymerization of allyl alcohol and
allyl acetate. Polym. Adv. Technol. 18, 953–958 (2007).
48. Matyjaszewski, K., Nakagawa, Y. & Jasieczek, C. B. Polymerization of n-butyl
acrylate by atom transfer radical polymerization: remarkable effect of ethylene
carbonate and other solvents. Macromolecules 31, 1535–1541 (1998).
49. Sato, T., Shimooka, S., Seno, M. & Tanaka, H. Radical polymerization of ortho-
formylphenyl crotonate involving intramolecular hydrogen-abstraction. Eur.
Polym. J. 28, 1357–1364 (1992).
50. Grub, J. & Lo¨ser, E. Butadiene. Ullmann’s Encyclopedia of Industrial Chemistry
(Wiley-VCH, 2011).
Acknowledgements
This work was supported by the Funding Program for Next Generation World-Leading
Researchers, Green Innovation and the Global COE Program ‘Chemistry Innovation
through Cooperation of Science and Engineering’ from the Ministry of Education, Culture,
Sports, Science and Technology (MEXT)/Japan Society for the Promotion of Science,
Japan, and a Grant-in-Aid for Innovative Areas ‘Molecular Activation Directed toward
Straightforward Synthesis’ from MEXT, Japan. The theoretical calculations were performed
using computational resources provided by the Research Center for Computational
Science, National Institutes of Natural Sciences, Okazaki, Japan. We are grateful to
H. Sugimoto, H. Goto and S. Honda (Tokyo University of Science) for right-angle laser-
light scattering (RALLS) analyses for the molecular weights of polymers, and to B. P.
Carrow (Princeton University) for careful proofreading.
41. Laible, R. C. Allyl polymerizations. Chem. Rev. 58, 807–843 (1958).
42. Miller, M. L. & Skogman, J. Polymerization of tert-butyl crotonate. J. Polym. Sci.
A 2, 4551–4558 (1964).
43. Kassi, E. & Patrickios, C. S. Well-defined polymers from biosourced monomers:
the case of 2-(methacryloyloxy)ethyl tiglate. Macromolecules 43,
1411–1415 (2010).
44. Kitayama, T. et al. Synthesis and polymerization of methyl
3-methylcyclobutene-1-carboxylate. Macromolecules 35, 1591–1598 (2002).
45. Chen, X-P., Sufi, B. A., Padias, A. B. & Hall, H. K. Jr. Controlled/‘living’ reverse
atom transfer radical polymerization of a monocyclic olefin, methyl
1-cyclobutenecarboxylate. Macromolecules 35, 4277–4281 (2002).
46. Brandrup, J., Immergut, E. H. & Grulke, E. A. Polymer Handbook 4th edn
(Wiley, 1999).
Author contributions
All authors designed the studies, discussed the results and wrote the paper. R.N. performed
all the experimental and computational work.
Additional information
Supplementary information and chemical compound information are available in the
online version of the paper. Reprints and permissions information is available online at
www.nature.com/reprints. Correspondence and requests for materials should be addressed
to K.N.
Competing financial interests
The authors declare no competing financial interests.
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
7
© 2014 Macmillan Publishers Limited. All rights reserved.