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.
T
he development of mild methods for the synthesis of bulk chemicals using renewable feedstocks is a critical technological hurdle along the path to a sustainable chemical economy in the future. Carbon dioxide is one of the most attractive renewable C1 resources, and it has many practical advantages, such as abundance, economic efficiency and lack of toxicity1–9. Its favourable nature as a carbon source is, however, inextricably linked to its inherent inertness. Not surprisingly, then, reactions of carbon dioxide must be combined with a high-energy reactant to gain a thermodynamic driving force. For large-scale utilization, chemical coupling of carbon dioxide with easily available chemical feedstocks is indispensable. Thus far only a limited number of processes have succeeded in the use of carbon dioxide, examples being the industrial production of urea4, salicylic acid4, organic carbonates10–12 and polycarbonates11–17. In this context, olefins, one of the largest classes of chemicals produced today, should be suitable co-reagents because of the high potential energy of their carbon–carbon double bonds. Although extensive studies have focused on the catalytic coupling of olefins with carbon dioxide to form commodity and fine chemicals18–21, methods to produce polymers from carbon dioxide and olefins are still unexplored; all the known examples, which use dienes22, vinyl ethers23,24 or acrylonitrile25, produce only oligomers of low molecular weight. Here we provide an explanation for the intrinsic difficulty of the copolymerization reaction, on which basis we developed the copolymerization of carbon dioxide and 1,3-dienes via a lactone intermediate. As a result of the thermodynamic insight that Miller and co-workers developed, we considered a kinetic limitation for olefin/ carbon dioxide copolymerization (Fig. 1a)26. According to Miller’s protocol26, we performed thermodynamic benchmark calculations in which the CBS-4M//B3LYP/6-31G* method exhibited the best fit with reported experimental values27. The free-energy changes for the incorporation of ethylene and carbon dioxide were determined as –12.5 and 21.9 kcal mol21, respectively. Thus, alternating carbon dioxide/ethylene copolymerization is endothermic by 9.4 kcal mol21 and excess ethylene insertion (.63 mol%)
is necessary before exothermic conditions can be achieved. We propose that kinetic limitations are also important for this reaction. The free energy of TS4 is higher than that of TS1, given that the activation energy from I to IE and that from ICE to ICEE are comparable because both reactions involve alkyl chain-ends and ethylene (Fig. 1a; see caption for definitions). Therefore, even if carbon dioxide uptake (TS2) and carboxylation of ethylene (TS3) are accessible, carbon dioxide/ethylene copolymerization is kinetically unfavourable unless the difference in free energy of TS4 can be reduced. As this qualitative speculation could be valid for any mechanism of polymerization, this kinetic problem may be a general bottleneck in carbon dioxide/olefin copolymerization. In this work, the formation of a metastable intermediate has facilitated an alternative reaction pathway by which copolymerization can take place. We postulated that 3-ethylidene-6-vinyltetrahydro-2H-pyran-2one (1) (Fig. 2a), which contains 29 wt% carbon dioxide, could provide an avenue to the formal copolymerization of carbon dioxide and 1,3-butadiene if appropriate conditions for its polymerization could be developed. Lactone 1, which was first reported by Inoue and co-workers28, has attracted much attention as a platform chemical for the utilization of carbon dioxide29. As the synthesis of lactone 1 is accompanied by the formation of several side products (Fig. 2c), considerable efforts have been undertaken to achieve a high yield and selectivity for lactone 130–37. Behr et al. optimized the reaction to accomplish 95% selectivity over butadiene conversion, and the conditions were applied on a mini-plant scale38,39. The homopolymerization of 1 was investigated previously by Dinjus and coworkers40, but they reported that the reaction did not proceed in the presence of radical, cationic or anionic initiators40. They suggested that the lack of reactivity of 1 may stem from its structural similarity to either a tiglic acid ester (Fig. 2b, red) or an a-substituted allyl ester (Fig. 2b, blue), both of which are substrates known to exhibit poor reactivity when subjected to common polymerization conditions41–45. Nevertheless, we detected serendipitously that oligomerization of 1 occurred on storage under air for several 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: [email protected] NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.
1
ARTICLES
NATURE CHEMISTRY
a
ICE +
TS1
TS3
TS2
I+
ΔG4‡
IC/IC´ +
I + CO2 Energy (kcal mol–1)
TS4
P polymer chain
* Growing chain end
DOI: 10.1038/NCHEM.1882
(≈ ΔΔG1‡ + 9.4) ΔΔG1‡ 0.0 –12.5
*
*
O
O P
*O
P P
9.4
Unknown
* O
i. Ethylene homopolymerization
O
P
O
IE
ICE
IC
I
–3.1
P
O
*
P
O
* IC´
ICEE
ii. Carbon dioxide incorporation into ethylene polymerization
b
ii. Carbon dioxide incorporation into butadiene 1,4-polymerization O O
*
TS1´ TS2´
I+
ICB +
TS3´
I + CO2 Energy (kcal mol–1)
TS4´ P
IC/IC´ +
ΔG4´‡ 14.6
ΔΔG1´‡
Unknown
–9.3
P P
*
*
I
ICB
IC or IC´
0.0
P
IB
*
5.2
0.0
ICBB
I + lactone 1 (L)
+ CO2 + 2 butadiene
(≈ ΔΔG1´‡ + 14.6)
*
P O O
+ O
O 1 (L)
–26.0 IL
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 reluctance of 1 to participate in the desired transformation. The results of our efforts to promote the polymerization of 1 are summarized in Table 1. We found that heating neat 1 in the presence of a typical radical initiator, 2,2′ -azobis(isobutyronitrile) (AIBN), produced polymer poly-1 after 24 hours at 100 8C (entry 1). The presence of a radical inhibitor, such as galvinoxyl or 2,2,6,6-tetramethylpiperidine 1-oxyl (TEMPO), suppressed the 2
polymerization (Table 1, entries 2 and 3), which is consistent with propagation by a radical mechanism. Emulsion conditions using a larger quantity of initiator (Table 1, entry 4) led to an improved conversion of 1, but a similar molecular weight of the product compared to that in neat conditions. Conditions for controlled radical polymerization, such as nitroxide-mediated radical polymerization (Table 1, entry 3) or atom-transfer radical polymerization (Table 1, entry 5) did not initiate the polymerization. The yield
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.
NATURE CHEMISTRY
ARTICLES
DOI: 10.1038/NCHEM.1882
a
c
Copolymer with 33 mol% CO2 incorporation
CO2 +
Side products in the synthesis of 1 2
O
O
Pd-catalysed lactone formation
O
O
O
O
Radical polymerization
R=
1
b
ROOC R
R
Tiglate ester O O
4
3
Allylic ester
H
ROOC
5
O
6
OR
Figure 2 | The concept of this study: copolymerization of carbon dioxide and 1,3-butadiene via a lactone intermediate, 3-ethylidene-6-vinyltetrahydro-2Hpyran-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 1,1′ -azobis(cyclohexane-1-carbonitrile) (V-40), which has a higher 10 hour life-time temperature in toluene (88 8C) than that of AIBN (65 8C) (Table 1, entry 6)46. The addition of acetic acid led to higher molecular weight, 19,000 g mol21, with a low polydispersity index of 1.1 (Table 1, entry 7). NMR spectroscopic analyses indicated that poly-1 produced by the methods shown in Table 1, entries 1, 4, 6 and 7, exhibited exclusive formation of the bicyclic repeat unit a (Fig. 3b; see the Supplementary Information for additional details on the structural
characterization). The resonances of carbons a, b and e observed at 178, 57 and 81 ppm, respectively, were all downfield shifted compared to typical saturated monocyclic esters, in accordance with a strained camphor-like bicyclic structure (a). Thus, the polymerization of lactone 1 proceeded via an alternating reaction of the 2,3-disubstituted acrylate moiety and the allylic moiety to give a polymer that possessed bicyclic structures in the main chain (Fig. 4). The key to the successful utilization of the multiple-substituted acrylate moiety, which previously had been known as inert for radical polymerization, can be explained as follows. Owing to the
Table 1 | Radical polymerization of lactone 1. Me O O
Me
O
Radical initiator O
Me O
Additive 100 °C, 24 h
O
Me O
O n
1 β
α
γ Poly-1
Entry 1 2‡ 3‡ 4§ 5 6 7} 8# 9# 10# 11# 12# 13# 14# 15# 16** 17††
Initiator AIBN AIBN AIBN KPS 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
Additive – Galvinoxyl TEMPO Aqueous SDS CuBr/bpy – AcOH MgCl2 ScCl3 YCl3 ZrCl4 FeCl3 CuCl2 ZnCl2 AlCl3 ZnCl2 ZnCl2
Solvent – – – – – – – – – – – – – – – EC EC
Yield (%) 18 – – 32 – 22 17 2.3 18 Trace 15 – – 26 – 48 59
M n (103)* 2.2 – – 2.0 – 5.7 19 – Insoluble – Insoluble – – 19 – 85 62
M w /M n* 1.4 – – 1.4 – 1.3 1.1 – – – – – – 1.2 – 1.5 1.1
a:b:g† 10:0:0 – – 10:0:0 – 10:0:0 10:0:0 6:1:3 – – – – – 5:2:3 – 3:5:2 3:4:3
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 THF as 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
© 2014 Macmillan Publishers Limited. All rights reserved.
3
ARTICLES a
I
BC
H
NATURE CHEMISTRY Solvent
D E
O AO
D
G G
F
C
F H
E
I
B A
b
i O a(C)
Solvent
g
h
f c
a Ob e
n
c,d,g(CH2)
d e(CH)
b(C)
h(CH)
f(CH)
c
i(CH3)
h´ O a´
γ
b´ O
α O
e˝ f˝
e´ b˝
β a Ob e
O
O a˝
Solvent Solvent a a˝ a´ + h´ e˝ f˝ b´
200
180
160
140
120
100 ppm
n
e
80
e´ b
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).
steric hindrance between the radical chain-end and the carbon– carbon double bond of the next monomer, the continuous incorporation 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. 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 molecular weight of poly-1. The reaction mixture, however, solidified within an hour without any other additive (Table 1, entry 14). After extensive screening (see Supplementary Table 5), we found that the concurrent use of an equimolar amount of zinc chloride47 and ethylene carbonate as a solvent48 led to a significant increase in the polymer yield (48%) and Mn (85,000 g mol21; Table 1, entry 16). The yield and molecular weight did not deteriorate when the reaction was repeated on a scale of 0.20 mol (Table 1, entry 17). The product poly-1 obtained in the presence of Lewis acids possessed not only the bicyclic structure a, but also monocyclic 4
DOI: 10.1038/NCHEM.1882
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 consistent 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 elemental analyses. The glass-transition temperature was as high as 192 8C in poly-1 (zinc chloride initiator, Table 1, entry 17). 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 conditions (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 Without cyclization
On allylic ester O O
Cyclization
R O
O
β
R
α
R
1
O O
On tiglate
H Hydrogen abstraction R
H
O
γ
O
Figure 4 | Proposed reaction pathways to form repeating units of a, b and g. If the radical chain-end (R†) attacks the allylic ester moiety (upwards arrow), the upper radical intermediate can undergo intramolecular exocyclization to form unit a, or propagate without cyclization to form unit b. If the radical chain-end attacks the tiglate ester moiety (downwards arrow), the lower radical intermediate can undergo intramolecular exo-cyclization to form unit a or propagate after intramolecular hydrogen abstraction to form unit g.
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.
NATURE CHEMISTRY
ARTICLES
DOI: 10.1038/NCHEM.1882
Table 2 | One-pot/two-step co/terpolymerization of carbon dioxide and 1,3-dienes.
CO2 (86–87 mmol)
Pd(acac)2 (X µmol) PPh3 (Y µmol)
Dienes
+
Ethylene carbonate 80 °C, Z h
(70 or 75 mmol)
V-40 (0.165 mmol) ZnCl2 (16.5 mmol) O
R3
O R1
CO2/diene co/terpolymer
100 °C, 24 h
R2
(+ side products)
Entry Initial load of dienes (mmol)
First step: lactone synthesis Y Z Major lactone X products (mmol) (mmol) (h)
1
Butadiene (70)
50
150
3
1 (R1 ¼ R2 ¼ R3 ¼ H)
2
Butadiene (25) Isoprene (50)
100
300
20
3
Butadiene (25) 100 1,3-pentadiene (50)
300
20
Yield/ selectivity of lactone (%)* 44/77
Second step: polymerization CO2 Mn Mw Tg incorporation (103)§ /M n§ (88 C) (wt%)‡
Yield/ selectivity of polymer (%)† 47/81
23
19
1.6
102} 128#
1 (R1 ¼ R2 ¼ R3 ¼ H) – 7a (R1 ¼ Me, R2 ¼ R3 ¼ H) 7b (R1 ¼ R3 ¼ H, R2 ¼ Me)
46/–
20
5.5
2.5
63
8 (R1 ¼ R2 ¼ H, R3 ¼ Me)
35/–
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 ethylene carbonate as the solvent, and obtained a modest yield (44% over the initial load of 1,3-butadiene) and selectivity (.76% over a butadiene conversion of 58%). Although the crude mixture of lactone 1 contained about a 14% yield of side products, including butadiene dimer 4, and acids and esters 5 and 6, as shown in Fig. 2, the total yield of 2 and 3 was found to be ,0.3%. After venting carbon dioxide and butadiene, the resulting mixture was heated for 24 hours in the presence of V-40 and zinc chloride. The carbon dioxide/butadiene copolymer (poly-1)′ was isolated in a good yield (47% over the initial load of 1,3-butadiene) and with good selectivity (.81% based on 1,3-butadiene consumption in the first step (58%)). The yield and selectivity of the polymer were slightly higher than those of lactone 1 in the first step, because side products 4, 5 and 6 were also incorporated into the copolymer. Elemental analysis revealed that the ratio of incorporated carbon dioxide to 1,3-butadiene was 1:2.7, which corresponds to a carbon-dioxide incorporation of 27 mol% (23 wt%). The possibility of decomposition of lactone 1 under the conditions of the second step can be excluded, judging from a control experiment that the palladium catalyst did not produce lactone 1 from carbon dioxide and butadiene in the presence of a stoichiometric amount of zinc chloride. The soluble fraction of (poly-1)′ in hot dimethylformamide (DMF) was found to have a moderate molecular weight of 19,000 g mol21 and a polydispersity index of 1.6. The glass-transition temperature of (poly-1)′ was lower than that of poly-1, but still as high as 102 8C for the soluble fraction and 128 8C for the insoluble fraction in hot DMF. The one-pot/two-step protocol was further applied to the terpolymerization of carbon dioxide, butadiene and another 1,3-diene (Table 2, entries 2 and 3). Despite the establishment of the lactone synthesis from 1,3-butadiene and carbon dioxide, the use of other dienes instead of 1,3-butadiene for telomerization with carbon dioxide still remains challenging32,34. The major limitations include a slow reaction rate and low yield of lactone formation because of an augmented steric hindrance and the formation of complex mixtures that contain positional isomers. In contrast, the three-component coupling of carbon dioxide, butadiene and isoprene/1,3-pentadiene studied by Behr et al.32 demonstrates reasonable yields in the palladium acetylacetonate ((Pd(acac)2)/
triisopropylphosphine system. Our one-pot/two-step procedure was combined with this three-component coupling strategy to produce novel terpolymers of carbon dioxide, butadiene and another 1,3-diene. By using triphenylphosphine (which is sterically smaller than triisopropylphosphine) as a ligand the cross-coupling products 7a and 7b (Table 2, entry 2) and 8 (Table 2, entry 3) were predominantly formed under the conditions of the first step. After removing carbon dioxide and dienes in vacuo, the resulting mixture was heated for 24 hours in the presence of V-40 and zinc chloride to produce the carbon dioxide/butadiene/isoprene (Table 2, entry 2) or carbon dioxide/butadiene/1,3-pentadiene terpolymers (Table 2, entry 3) with high molecular weights (Mn ¼ 5,500 and 16,000, respectively) in good yields (46% and 35%, respectively, over the initial load of dienes). Elemental analysis revealed that the carbon-dioxide incorporation of these polymers were 20 wt% and 24 wt%, slightly lower than the carbon-dioxide incorporation of lactone monomers 7a, 7b and 8 (26.5 wt%). The glass-transition temperatures of these polymers were 63 and 33 8C, which shows that the introduction of methyl substituents into polylactone poly-1 causes a significant decrease in the glass-transition temperatures. A theoretical comparison of the ground-state energies for intermediates and products in both the polymerization of 1 described here (Fig. 1b, path iii) and carbon-dioxide incorporation into butadiene 1,4-polymerization (Fig. 1b, path ii) was instructive. The thermodynamic gain of each monomer incorporation was calculated using a composite CBS method (CBS-4M//B3LYP/6-31G*) according to Miller’s protocol26. The changes in free energy for butadiene incorporation via 1,4-addition and for carbon-dioxide incorporation were determined as –9.3 kcal mol21 and 23.9 kcal mol21, respectively, which indicates that the consumption of one carbon– carbon double bond of butadiene only allows ,28 mol% incorporation of carbon dioxide, which is lower than 36 mol% in the case of ethylene. Although conjugated alkenes generally have higher (kinetic) reactivity, as exemplified by some coupling reactions with carbon dioxide in the literature19–21,28, the consumption of one carbon–carbon double bond of a conjugated alkene provides less thermodynamic gain than that of a non-conjugated alkene. In the carbon-dioxide incorporation into butadiene 1,4-polymerization (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
ARTICLES
NATURE CHEMISTRY
of TS4′ over TS1′ should be 14.6 kcal mol21 (see Fig. 1b). In contrast, the formation of lactone 1 and the following cyclopolymerization 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 deactivation of the palladium catalyst suppresses the decomposition of lactone 1 back into butadiene and carbon dioxide, and so leads to poly-1 formation. Thermodynamically, the polymerization via 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. However, 33 mol% carbon dioxide was incorporated in our copolymerization 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 molecule of carbon dioxide, which provides a thermodynamic bypass and driving force sufficiently strong to overcome the endothermic penalty of carbon-dioxide incorporation. In summary, we have achieved the copolymerization of carbon dioxide and 1,3-butadiene via the lactone intermediate 3-ethylidene-6-vinyltetrahydro-2H-pyran-2-one (1). The homopolymerization 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 onepot carbon dioxide/butadiene copolymerization, and one-pot terpolymerization of carbon dioxide, butadiene and another 1,3diene. Our accomplishment reported here opens up not only a sustainable 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 conceivably lead to mass utilization of carbon dioxide for the synthesis of synthetic polymers.
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.
Received 8 July 2013; accepted 28 January 2014; published online 9 March 2014 6
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). 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). 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 the conversion of carbon dioxide into energy-rich products: stepping towards artificial photosynthesis. Phil. Trans. R. Soc. A 371, 20120111 (2012). 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). 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). 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).
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.
NATURE CHEMISTRY
ARTICLES
DOI: 10.1038/NCHEM.1882
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). 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).
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 orthoformylphenyl 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 laserlight scattering (RALLS) analyses for the molecular weights of polymers, and to B. P. Carrow (Princeton University) for careful proofreading.
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
© 2014 Macmillan Publishers Limited. All rights reserved.
7
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.
T
he development of mild methods for the synthesis of bulk chemicals using renewable feedstocks is a critical technological hurdle along the path to a sustainable chemical economy in the future. Carbon dioxide is one of the most attractive renewable C1 resources, and it has many practical advantages, such as abundance, economic efficiency and lack of toxicity1–9. Its favourable nature as a carbon source is, however, inextricably linked to its inherent inertness. Not surprisingly, then, reactions of carbon dioxide must be combined with a high-energy reactant to gain a thermodynamic driving force. For large-scale utilization, chemical coupling of carbon dioxide with easily available chemical feedstocks is indispensable. Thus far only a limited number of processes have succeeded in the use of carbon dioxide, examples being the industrial production of urea4, salicylic acid4, organic carbonates10–12 and polycarbonates11–17. In this context, olefins, one of the largest classes of chemicals produced today, should be suitable co-reagents because of the high potential energy of their carbon–carbon double bonds. Although extensive studies have focused on the catalytic coupling of olefins with carbon dioxide to form commodity and fine chemicals18–21, methods to produce polymers from carbon dioxide and olefins are still unexplored; all the known examples, which use dienes22, vinyl ethers23,24 or acrylonitrile25, produce only oligomers of low molecular weight. Here we provide an explanation for the intrinsic difficulty of the copolymerization reaction, on which basis we developed the copolymerization of carbon dioxide and 1,3-dienes via a lactone intermediate. As a result of the thermodynamic insight that Miller and co-workers developed, we considered a kinetic limitation for olefin/ carbon dioxide copolymerization (Fig. 1a)26. According to Miller’s protocol26, we performed thermodynamic benchmark calculations in which the CBS-4M//B3LYP/6-31G* method exhibited the best fit with reported experimental values27. The free-energy changes for the incorporation of ethylene and carbon dioxide were determined as –12.5 and 21.9 kcal mol21, respectively. Thus, alternating carbon dioxide/ethylene copolymerization is endothermic by 9.4 kcal mol21 and excess ethylene insertion (.63 mol%)
is necessary before exothermic conditions can be achieved. We propose that kinetic limitations are also important for this reaction. The free energy of TS4 is higher than that of TS1, given that the activation energy from I to IE and that from ICE to ICEE are comparable because both reactions involve alkyl chain-ends and ethylene (Fig. 1a; see caption for definitions). Therefore, even if carbon dioxide uptake (TS2) and carboxylation of ethylene (TS3) are accessible, carbon dioxide/ethylene copolymerization is kinetically unfavourable unless the difference in free energy of TS4 can be reduced. As this qualitative speculation could be valid for any mechanism of polymerization, this kinetic problem may be a general bottleneck in carbon dioxide/olefin copolymerization. In this work, the formation of a metastable intermediate has facilitated an alternative reaction pathway by which copolymerization can take place. We postulated that 3-ethylidene-6-vinyltetrahydro-2H-pyran-2one (1) (Fig. 2a), which contains 29 wt% carbon dioxide, could provide an avenue to the formal copolymerization of carbon dioxide and 1,3-butadiene if appropriate conditions for its polymerization could be developed. Lactone 1, which was first reported by Inoue and co-workers28, has attracted much attention as a platform chemical for the utilization of carbon dioxide29. As the synthesis of lactone 1 is accompanied by the formation of several side products (Fig. 2c), considerable efforts have been undertaken to achieve a high yield and selectivity for lactone 130–37. Behr et al. optimized the reaction to accomplish 95% selectivity over butadiene conversion, and the conditions were applied on a mini-plant scale38,39. The homopolymerization of 1 was investigated previously by Dinjus and coworkers40, but they reported that the reaction did not proceed in the presence of radical, cationic or anionic initiators40. They suggested that the lack of reactivity of 1 may stem from its structural similarity to either a tiglic acid ester (Fig. 2b, red) or an a-substituted allyl ester (Fig. 2b, blue), both of which are substrates known to exhibit poor reactivity when subjected to common polymerization conditions41–45. Nevertheless, we detected serendipitously that oligomerization of 1 occurred on storage under air for several 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: [email protected] NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.
1
ARTICLES
NATURE CHEMISTRY
a
ICE +
TS1
TS3
TS2
I+
ΔG4‡
IC/IC´ +
I + CO2 Energy (kcal mol–1)
TS4
P polymer chain
* Growing chain end
DOI: 10.1038/NCHEM.1882
(≈ ΔΔG1‡ + 9.4) ΔΔG1‡ 0.0 –12.5
*
*
O
O P
*O
P P
9.4
Unknown
* O
i. Ethylene homopolymerization
O
P
O
IE
ICE
IC
I
–3.1
P
O
*
P
O
* IC´
ICEE
ii. Carbon dioxide incorporation into ethylene polymerization
b
ii. Carbon dioxide incorporation into butadiene 1,4-polymerization O O
*
TS1´ TS2´
I+
ICB +
TS3´
I + CO2 Energy (kcal mol–1)
TS4´ P
IC/IC´ +
ΔG4´‡ 14.6
ΔΔG1´‡
Unknown
–9.3
P P
*
*
I
ICB
IC or IC´
0.0
P
IB
*
5.2
0.0
ICBB
I + lactone 1 (L)
+ CO2 + 2 butadiene
(≈ ΔΔG1´‡ + 14.6)
*
P O O
+ O
O 1 (L)
–26.0 IL
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 reluctance of 1 to participate in the desired transformation. The results of our efforts to promote the polymerization of 1 are summarized in Table 1. We found that heating neat 1 in the presence of a typical radical initiator, 2,2′ -azobis(isobutyronitrile) (AIBN), produced polymer poly-1 after 24 hours at 100 8C (entry 1). The presence of a radical inhibitor, such as galvinoxyl or 2,2,6,6-tetramethylpiperidine 1-oxyl (TEMPO), suppressed the 2
polymerization (Table 1, entries 2 and 3), which is consistent with propagation by a radical mechanism. Emulsion conditions using a larger quantity of initiator (Table 1, entry 4) led to an improved conversion of 1, but a similar molecular weight of the product compared to that in neat conditions. Conditions for controlled radical polymerization, such as nitroxide-mediated radical polymerization (Table 1, entry 3) or atom-transfer radical polymerization (Table 1, entry 5) did not initiate the polymerization. The yield
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.
NATURE CHEMISTRY
ARTICLES
DOI: 10.1038/NCHEM.1882
a
c
Copolymer with 33 mol% CO2 incorporation
CO2 +
Side products in the synthesis of 1 2
O
O
Pd-catalysed lactone formation
O
O
O
O
Radical polymerization
R=
1
b
ROOC R
R
Tiglate ester O O
4
3
Allylic ester
H
ROOC
5
O
6
OR
Figure 2 | The concept of this study: copolymerization of carbon dioxide and 1,3-butadiene via a lactone intermediate, 3-ethylidene-6-vinyltetrahydro-2Hpyran-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 1,1′ -azobis(cyclohexane-1-carbonitrile) (V-40), which has a higher 10 hour life-time temperature in toluene (88 8C) than that of AIBN (65 8C) (Table 1, entry 6)46. The addition of acetic acid led to higher molecular weight, 19,000 g mol21, with a low polydispersity index of 1.1 (Table 1, entry 7). NMR spectroscopic analyses indicated that poly-1 produced by the methods shown in Table 1, entries 1, 4, 6 and 7, exhibited exclusive formation of the bicyclic repeat unit a (Fig. 3b; see the Supplementary Information for additional details on the structural
characterization). The resonances of carbons a, b and e observed at 178, 57 and 81 ppm, respectively, were all downfield shifted compared to typical saturated monocyclic esters, in accordance with a strained camphor-like bicyclic structure (a). Thus, the polymerization of lactone 1 proceeded via an alternating reaction of the 2,3-disubstituted acrylate moiety and the allylic moiety to give a polymer that possessed bicyclic structures in the main chain (Fig. 4). The key to the successful utilization of the multiple-substituted acrylate moiety, which previously had been known as inert for radical polymerization, can be explained as follows. Owing to the
Table 1 | Radical polymerization of lactone 1. Me O O
Me
O
Radical initiator O
Me O
Additive 100 °C, 24 h
O
Me O
O n
1 β
α
γ Poly-1
Entry 1 2‡ 3‡ 4§ 5 6 7} 8# 9# 10# 11# 12# 13# 14# 15# 16** 17††
Initiator AIBN AIBN AIBN KPS 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
Additive – Galvinoxyl TEMPO Aqueous SDS CuBr/bpy – AcOH MgCl2 ScCl3 YCl3 ZrCl4 FeCl3 CuCl2 ZnCl2 AlCl3 ZnCl2 ZnCl2
Solvent – – – – – – – – – – – – – – – EC EC
Yield (%) 18 – – 32 – 22 17 2.3 18 Trace 15 – – 26 – 48 59
M n (103)* 2.2 – – 2.0 – 5.7 19 – Insoluble – Insoluble – – 19 – 85 62
M w /M n* 1.4 – – 1.4 – 1.3 1.1 – – – – – – 1.2 – 1.5 1.1
a:b:g† 10:0:0 – – 10:0:0 – 10:0:0 10:0:0 6:1:3 – – – – – 5:2:3 – 3:5:2 3:4:3
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 THF as 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
© 2014 Macmillan Publishers Limited. All rights reserved.
3
ARTICLES a
I
BC
H
NATURE CHEMISTRY Solvent
D E
O AO
D
G G
F
C
F H
E
I
B A
b
i O a(C)
Solvent
g
h
f c
a Ob e
n
c,d,g(CH2)
d e(CH)
b(C)
h(CH)
f(CH)
c
i(CH3)
h´ O a´
γ
b´ O
α O
e˝ f˝
e´ b˝
β a Ob e
O
O a˝
Solvent Solvent a a˝ a´ + h´ e˝ f˝ b´
200
180
160
140
120
100 ppm
n
e
80
e´ b
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).
steric hindrance between the radical chain-end and the carbon– carbon double bond of the next monomer, the continuous incorporation 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. 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 molecular weight of poly-1. The reaction mixture, however, solidified within an hour without any other additive (Table 1, entry 14). After extensive screening (see Supplementary Table 5), we found that the concurrent use of an equimolar amount of zinc chloride47 and ethylene carbonate as a solvent48 led to a significant increase in the polymer yield (48%) and Mn (85,000 g mol21; Table 1, entry 16). The yield and molecular weight did not deteriorate when the reaction was repeated on a scale of 0.20 mol (Table 1, entry 17). The product poly-1 obtained in the presence of Lewis acids possessed not only the bicyclic structure a, but also monocyclic 4
DOI: 10.1038/NCHEM.1882
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 consistent 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 elemental analyses. The glass-transition temperature was as high as 192 8C in poly-1 (zinc chloride initiator, Table 1, entry 17). 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 conditions (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 Without cyclization
On allylic ester O O
Cyclization
R O
O
β
R
α
R
1
O O
On tiglate
H Hydrogen abstraction R
H
O
γ
O
Figure 4 | Proposed reaction pathways to form repeating units of a, b and g. If the radical chain-end (R†) attacks the allylic ester moiety (upwards arrow), the upper radical intermediate can undergo intramolecular exocyclization to form unit a, or propagate without cyclization to form unit b. If the radical chain-end attacks the tiglate ester moiety (downwards arrow), the lower radical intermediate can undergo intramolecular exo-cyclization to form unit a or propagate after intramolecular hydrogen abstraction to form unit g.
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.
NATURE CHEMISTRY
ARTICLES
DOI: 10.1038/NCHEM.1882
Table 2 | One-pot/two-step co/terpolymerization of carbon dioxide and 1,3-dienes.
CO2 (86–87 mmol)
Pd(acac)2 (X µmol) PPh3 (Y µmol)
Dienes
+
Ethylene carbonate 80 °C, Z h
(70 or 75 mmol)
V-40 (0.165 mmol) ZnCl2 (16.5 mmol) O
R3
O R1
CO2/diene co/terpolymer
100 °C, 24 h
R2
(+ side products)
Entry Initial load of dienes (mmol)
First step: lactone synthesis Y Z Major lactone X products (mmol) (mmol) (h)
1
Butadiene (70)
50
150
3
1 (R1 ¼ R2 ¼ R3 ¼ H)
2
Butadiene (25) Isoprene (50)
100
300
20
3
Butadiene (25) 100 1,3-pentadiene (50)
300
20
Yield/ selectivity of lactone (%)* 44/77
Second step: polymerization CO2 Mn Mw Tg incorporation (103)§ /M n§ (88 C) (wt%)‡
Yield/ selectivity of polymer (%)† 47/81
23
19
1.6
102} 128#
1 (R1 ¼ R2 ¼ R3 ¼ H) – 7a (R1 ¼ Me, R2 ¼ R3 ¼ H) 7b (R1 ¼ R3 ¼ H, R2 ¼ Me)
46/–
20
5.5
2.5
63
8 (R1 ¼ R2 ¼ H, R3 ¼ Me)
35/–
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 ethylene carbonate as the solvent, and obtained a modest yield (44% over the initial load of 1,3-butadiene) and selectivity (.76% over a butadiene conversion of 58%). Although the crude mixture of lactone 1 contained about a 14% yield of side products, including butadiene dimer 4, and acids and esters 5 and 6, as shown in Fig. 2, the total yield of 2 and 3 was found to be ,0.3%. After venting carbon dioxide and butadiene, the resulting mixture was heated for 24 hours in the presence of V-40 and zinc chloride. The carbon dioxide/butadiene copolymer (poly-1)′ was isolated in a good yield (47% over the initial load of 1,3-butadiene) and with good selectivity (.81% based on 1,3-butadiene consumption in the first step (58%)). The yield and selectivity of the polymer were slightly higher than those of lactone 1 in the first step, because side products 4, 5 and 6 were also incorporated into the copolymer. Elemental analysis revealed that the ratio of incorporated carbon dioxide to 1,3-butadiene was 1:2.7, which corresponds to a carbon-dioxide incorporation of 27 mol% (23 wt%). The possibility of decomposition of lactone 1 under the conditions of the second step can be excluded, judging from a control experiment that the palladium catalyst did not produce lactone 1 from carbon dioxide and butadiene in the presence of a stoichiometric amount of zinc chloride. The soluble fraction of (poly-1)′ in hot dimethylformamide (DMF) was found to have a moderate molecular weight of 19,000 g mol21 and a polydispersity index of 1.6. The glass-transition temperature of (poly-1)′ was lower than that of poly-1, but still as high as 102 8C for the soluble fraction and 128 8C for the insoluble fraction in hot DMF. The one-pot/two-step protocol was further applied to the terpolymerization of carbon dioxide, butadiene and another 1,3-diene (Table 2, entries 2 and 3). Despite the establishment of the lactone synthesis from 1,3-butadiene and carbon dioxide, the use of other dienes instead of 1,3-butadiene for telomerization with carbon dioxide still remains challenging32,34. The major limitations include a slow reaction rate and low yield of lactone formation because of an augmented steric hindrance and the formation of complex mixtures that contain positional isomers. In contrast, the three-component coupling of carbon dioxide, butadiene and isoprene/1,3-pentadiene studied by Behr et al.32 demonstrates reasonable yields in the palladium acetylacetonate ((Pd(acac)2)/
triisopropylphosphine system. Our one-pot/two-step procedure was combined with this three-component coupling strategy to produce novel terpolymers of carbon dioxide, butadiene and another 1,3-diene. By using triphenylphosphine (which is sterically smaller than triisopropylphosphine) as a ligand the cross-coupling products 7a and 7b (Table 2, entry 2) and 8 (Table 2, entry 3) were predominantly formed under the conditions of the first step. After removing carbon dioxide and dienes in vacuo, the resulting mixture was heated for 24 hours in the presence of V-40 and zinc chloride to produce the carbon dioxide/butadiene/isoprene (Table 2, entry 2) or carbon dioxide/butadiene/1,3-pentadiene terpolymers (Table 2, entry 3) with high molecular weights (Mn ¼ 5,500 and 16,000, respectively) in good yields (46% and 35%, respectively, over the initial load of dienes). Elemental analysis revealed that the carbon-dioxide incorporation of these polymers were 20 wt% and 24 wt%, slightly lower than the carbon-dioxide incorporation of lactone monomers 7a, 7b and 8 (26.5 wt%). The glass-transition temperatures of these polymers were 63 and 33 8C, which shows that the introduction of methyl substituents into polylactone poly-1 causes a significant decrease in the glass-transition temperatures. A theoretical comparison of the ground-state energies for intermediates and products in both the polymerization of 1 described here (Fig. 1b, path iii) and carbon-dioxide incorporation into butadiene 1,4-polymerization (Fig. 1b, path ii) was instructive. The thermodynamic gain of each monomer incorporation was calculated using a composite CBS method (CBS-4M//B3LYP/6-31G*) according to Miller’s protocol26. The changes in free energy for butadiene incorporation via 1,4-addition and for carbon-dioxide incorporation were determined as –9.3 kcal mol21 and 23.9 kcal mol21, respectively, which indicates that the consumption of one carbon– carbon double bond of butadiene only allows ,28 mol% incorporation of carbon dioxide, which is lower than 36 mol% in the case of ethylene. Although conjugated alkenes generally have higher (kinetic) reactivity, as exemplified by some coupling reactions with carbon dioxide in the literature19–21,28, the consumption of one carbon–carbon double bond of a conjugated alkene provides less thermodynamic gain than that of a non-conjugated alkene. In the carbon-dioxide incorporation into butadiene 1,4-polymerization (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
ARTICLES
NATURE CHEMISTRY
of TS4′ over TS1′ should be 14.6 kcal mol21 (see Fig. 1b). In contrast, the formation of lactone 1 and the following cyclopolymerization 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 deactivation of the palladium catalyst suppresses the decomposition of lactone 1 back into butadiene and carbon dioxide, and so leads to poly-1 formation. Thermodynamically, the polymerization via 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. However, 33 mol% carbon dioxide was incorporated in our copolymerization 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 molecule of carbon dioxide, which provides a thermodynamic bypass and driving force sufficiently strong to overcome the endothermic penalty of carbon-dioxide incorporation. In summary, we have achieved the copolymerization of carbon dioxide and 1,3-butadiene via the lactone intermediate 3-ethylidene-6-vinyltetrahydro-2H-pyran-2-one (1). The homopolymerization 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 onepot carbon dioxide/butadiene copolymerization, and one-pot terpolymerization of carbon dioxide, butadiene and another 1,3diene. Our accomplishment reported here opens up not only a sustainable 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 conceivably lead to mass utilization of carbon dioxide for the synthesis of synthetic polymers.
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.
Received 8 July 2013; accepted 28 January 2014; published online 9 March 2014 6
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). 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). 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 the conversion of carbon dioxide into energy-rich products: stepping towards artificial photosynthesis. Phil. Trans. R. Soc. A 371, 20120111 (2012). 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). 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). 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).
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2014 Macmillan Publishers Limited. All rights reserved.
NATURE CHEMISTRY
ARTICLES
DOI: 10.1038/NCHEM.1882
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). 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).
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 orthoformylphenyl 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 laserlight scattering (RALLS) analyses for the molecular weights of polymers, and to B. P. Carrow (Princeton University) for careful proofreading.
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
© 2014 Macmillan Publishers Limited. All rights reserved.
7
