Facile estimation of catalytic activity and selectivities in copolymerization of propylene oxide with carbon dioxide mediated by metal complexes with planar tetradentate ligand

Article abstract of DOI:10.1021/ja5046814

Mechanistic studies were conducted to estimate (1) catalytic activity for PPC, (2) PPC/CPC selectivity, and (3) PPC/PPO selectivity for the metal-catalyzed copolymerization of propylene oxide with carbon dioxide [PPC: poly(propylene carbonate); CPC = cyclic propylene carbonate; PPO: poly(propylene oxide)]. Density functional theory (DFT) studies demonstrated that the ΔG_crb - ΔG_epx value should be an effective indicator for the catalytic activities [ΔG_epx: dissociation energy of ethylene oxide from the epoxide-coordinating metal complex; ΔG_crb: dissociation energy of methyl carbonate from the metal-carbonate complex]. In addition, metal complexes with a subthreshold ΔG_epx value were found to show low PPC/CPC selectivity. The PPC/PPO selectivity was related to the ΔG_alk - ΔG_epx value and steric environment around the metal center (ΔG_alk: dissociation energy of alkoxide ligand from the metal center). Based on the mechanistic studies, two metal complexes were designed and applied to the copolymerization to support validity of these indicators. The results presented here should be useful for brand-new catalyst candidates since these indicators can be easily calculated by DFT method without computing transition states.

Full text of DOI:10.1021/ja5046814

Article
pubs.acs.org/JACS
Facile Estimation of Catalytic Activity and Selectivities in
Copolymerization of Propylene Oxide with Carbon Dioxide Mediated
by Metal Complexes with Planar Tetradentate Ligand
Takahiro Ohkawara,† Kohei Suzuki,† Koji Nakano,‡ Seiji Mori,*,§ and Kyoko Nozaki*,†
†Department of Chemistry and Biotechnology, Graduate School of Engineering, the University of Tokyo 7-3-1 Hongo, Bunkyo-ku,
Tokyo 113-8656, Japan
‡Department of Organic and Polymer Materials Chemistry, Tokyo University of Agriculture and Technology, 2-24-16 Naka-cho,
Koganei, Tokyo 184-8588, Japan
§Faculty of Science, Ibaraki University, 2-1-1 Bunkyo, Mito, Ibaraki 310-8512, Japan
S
* Supporting Information
ABSTRACT: Mechanistic studies were conducted to estimate (1)
catalytic activity for PPC, (2) PPC/CPC selectivity, and (3) PPC/PPO
selectivity for the metal-catalyzed copolymerization of propylene oxide
with carbon dioxide [PPC: poly(propylene carbonate); CPC = cyclic
propylene carbonate; PPO: poly(propylene oxide)]. Density functional
theory (DFT) studies demonstrated that the ΔGcrb − ΔGepx value
should be an effective indicator for the catalytic activities [ΔGepx
:
dissociation energy of ethylene oxide from the epoxide-coordinating
metal complex; ΔGcrb: dissociation energy of methyl carbonate from
the metal−carbonate complex]. In addition, metal complexes with a
subthreshold ΔGepx value were found to show low PPC/CPC
selectivity. The PPC/PPO selectivity was related to the ΔGalk
−
ΔGepx value and steric environment around the metal center (ΔGalk:
dissociation energy of alkoxide ligand from the metal center). Based on the mechanistic studies, two metal complexes were
designed and applied to the copolymerization to support validity of these indicators. The results presented here should be useful
for brand-new catalyst candidates since these indicators can be easily calculated by DFT method without computing transition
states.
Cr−salen catalyst system intensively and revealed various
properties of the system such as temperature/CO2 pressure
dependency and the effect of cocatalyst and substituents on the
ligand.5 They also reported mechanistic studies focusing on the
decomposition of the product copolymer.10 Chisholm and co-
workers reported a broad range of properties of Al, Cr, and
Co−porphyrin and salen catalyst systems. They estimated the
order of Lewis acidity of the complexes to be Cr(III) ≈ Al(III)
> Co(III) by using electrospray tandem mass spectrometry.11
They also revealed that the reactivity of (TPP)AlX (TPP =
tetraphenylporphyrin and X = Cl, OR, and OCOR) for the PO
ring-opening reaction decrease in the order Cl > OR > O2CR in
the absence of DMAP (1 equiv) and in the order Cl > OCOR
> OR in the presence of DMAP, respectively.12 They also
found that the rate dependence of the PO ring-opening
reaction was first order in [Al]. Elaborate investigation of the
catalytic cycle through density functional theory (DFT) studies
were also performed for specific catalyst systems based on Zn−
INTRODUCTION
■
The epoxide/CO2 copolymerization is a promising strategy for
CO2 utilization, and thus various catalysts have been developed
for the copolymerization since the first report by Inoue using
heterogeneous zinc catalysts.1 Among the catalysts so far
investigated, recent development of homogeneous metal
catalysts is notable as shown in the following examples.2
Inoue et al. achieved living polymerization using Al−porphyrin
complex.3 Later, since 2000, various new catalyst systems with a
well-defined metal complex emerged such as Zn−diiminate4
and Cr5 and Co6−salen complexes. Thus far, the highest
catalytic performance was accomplished by Co−salen com-
plexes; for example, high catalytic activity for poly(propylene
carbonate) (PPC) generation and high selectivities for PPC
over cyclic carbonate (CPC) and for PPC over poly(propylene
oxide) (PPO) were achieved in the propylene oxide (PO)/CO2
copolymerization. We have also reported another class of metal
catalysts employing trivalent-tetradentate ligands, namely Ti,
Ge, and Sn−boxdipy7 and Fe8 and Mn9−corrole complexes.
Mechanistic studies on the copolymerization were reported
for the representative catalysts from both experimental and
theoretical viewpoints. Darensbourg and co-workers studied
Received: May 17, 2014
Published: July 15, 2014
© 2014 American Chemical Society
10728
dx.doi.org/10.1021/ja5046814 | J. Am. Chem. Soc. 2014, 136, 10728−10735

Journal of the American Chemical Society
Article
diiminate,13 Al and Cr−salen,14 and dinuclear zinc complexes
Scheme 1. Reaction Mechanism of Epoxide/CO2
Copolymerization
developed by Williams and co-workers.15
As reviewed above, each mechanistic study so far was limited
to each family of catalysts, and there has been no report to
compare different families of catalysts. In other words, it has
not been known why Co−salen catalyst systems show higher
performance compared to other systems. This prompted us to
develop a common indicator to estimate catalyst performance
(catalytic activity, selectivity for PPC/CPC, and selectivity for
PPC/PPO). In this work, catalyst systems with M−salen [M =
Co (1),6a,b Cr (2)5a,b,d−h], M−boxdipy7 [M = Ti (3), Ge (4),
Zr (5), Sn (6)], M−corrole [M = Mn (7),9 Fe (8)8], and Al−
porphyrin (9)3 were investigated. All of these complexes
possess a planar tetradentate ligand and require cocatalysts such
as [PPN]X {[PPN] = bis(triphenylphosphine)iminium} for
high catalytic activity (Figure 1). Their catalytic performances
When using metal-complex/[PPN]X catalyst systems, four
types of mechanisms may be proposed for the epoxide ring-
opening by a nucleophile (Figure 1). The first one is a
bimetallic mechanism, where two complexes work coopera-
tively. Such bimetallic system was proposed for hydrolytic
kinetic resolution of epoxides by using a Co−salen complex17
and for homopolymerization of epoxides catalyzed by a Co−
salen complex.18 The second one is a monometallic mechanism
(type i) in which a free epoxide inserts into a metal−
nucleophile (Nu) bond, as Darensbourg proposed for the
epoxide/CO2 copolymerization by using Cr−salen/cocatalyst
system.5f The third and fourth ones (monometallic mecha-
nisms, types ii and iii) are a monometallic mechanism, in which
a free nucleophile and [PPN]Nu attacks an epoxide bound to
the metal center, respectively. In order to reveal the ring-
opening mechanism with metal-complex/[PPN]X catalyst
systems, we conducted kinetic studies by using complexes 1,
3, and 4.19 Reaction orders of 1.57, 1.57, and 1.75 in catalyst
concentration were obtained for complexes 1, 3, and 4,
respectively, which agree with the result reported in Co-
salen/nBu4NX catalyst system.20 This seemed to indicate the
complexity of the propagation step, that is, some of four types
of mechanism might compete with one another. Nevertheless,
in any mechanisms, key species in the epoxide ring-opening
step should be the epoxide coordinating intermediate and the
nucleophilic carbonate (ROCOO−) whether it is free or
interacts with the metal center and [PPN] cation.
Figure 1. Structures of metal complexes investigated in this work.
for the PO/CO2 copolymerization were well-studied. Although
there are other highly active catalysts which do not require the
cocatalyst or possess different structures, such catalysts are
beyond consideration in this work.
RESULTS AND DISCUSSION
■
Indicators for Catalytic Performance. In this section, we
have proposed some indicators to rationalize catalytic activity
and selectivities. DFT studies were conducted to estimate
Gibbs energies of the key intermediates in the copolymeriza-
tion. Correlation between the relative Gibbs energy and (1)
catalytic activity for PPC generation, (2) PPC/CPC selectivity,
and (3) PPC/PPO selectivity were studied. The considerations
in this section were applied to the development of new catalysts
with higher performance in the following section.
Activity for PPC Generation. Chain propagation in the
copolymerization consists of two steps. One is the nucleophilic
attack of carbonate end on epoxide activated through
coordination to the metal center, resulting in a metal−alkoxide
intermediate. Another is CO2 insertion into the metal−alkoxide
intermediate to form a metal−carbonate intermediate (Scheme
1). Based on previous reports on several catalyst systems, it
should be reasonable to regard the ring-opening of epoxide as a
rate-determining step (RDS) in this copolymerization under a
wide range of CO2 pressures.13,15b,16
In order to derive a common indicator for catalytic activities,
the dissociation energy of ethylene oxide from the epoxide-
coordinating intermediate (A), ΔGepx, and the dissociation
energy of methyl carbonate from the metal−carbonate
intermediate (C), ΔGcrb, were computed (Figure 2). The
values of ΔGepx and ΔGcrb reflect the strength of epoxide−
metal and carbonate−metal binding, respectively. These two
parameters were compared with the experimental data of the
PO/CO2 copolymerization under the standard reaction
conditions (Scheme 2).21 The theoretical and experimental
10729
dx.doi.org/10.1021/ja5046814 | J. Am. Chem. Soc. 2014, 136, 10728−10735

Journal of the American Chemical Society
Article
Figure 3. Plot of ΔGcrb − ΔGepx value vs TOF for PPC.
ΔGcrb − ΔGepx, and the TOF was found to have a strong
negative correlation. Furthermore, the plot of logarithm of
TOF vs ΔGcrb − ΔGepx value showed linear relationship (R2 =
0.91). The parameter ΔGcrb − ΔGepx corresponds to free
energy change in carbonate-epoxide substitution on metal
center. Accordingly, it should be regarded as a quantitative
indicator for the catalytic activity of the catalyst systems
operating in monometallic mechanism. The plot of the data for
Fe−corrole catalyst system, which is expected to catalyze the
copolymerization via bimetallic mechanism,22 also follows the
linear relationship. Thus, the parameter ΔGcrb − ΔGepx can be
applied to the copolymerization via bimetallic mechanism. In
any case, the result described here well matches to our first
presumption that the key active species are epoxide-
coordinating intermediate (electrophile) and metal−carbonate
intermediate (nucleophile).
Figure 2. (a) Dissociation energies of ethylene oxide (ΔGepx) and
methyl carbonate (ΔGcrb) from metal center, (b) general structures
used in calculation for complexes 1−6, 9, and 10, and (c) general
structures used in calculation for complexes 7, 8, and 11.
PPC/CPC Selectivity. There should be four species possibly
responsible for CPC formation: metal-bound alkoxide, metal-
bound carbonate, metal-free alkoxide, and metal-free carbonate.
Based on DFT calculations, Rieger and co-workers proposed
that the backbiting reaction from a metal-free carbonate end is
the main pathway for the CPC formation in the presence of an
excess amount of CO2.14 According to our results through in
situ IR monitoring, CPC generation mainly occurred at the
early stage in the copolymerization (Figure S2). This is
probably because the backbiting occurs mainly at the initial step
of the copolymerization where the leaving group is Cl−;
namely, right after nucleophilic attack of the initiator Cl− on the
first PO and incorporation of the first CO2.10b This can be
rationalized by the fact that chloride is a much better leaving
group than alkylcarbonate which is a leaving group in
backbiting of elongated polymer chain. Accordingly, the
following discussion about the CPC generation is focused on
the initial step of the copolymerization, and the mechanism for
cyclic carbonate formation is proposed in Scheme 3.
In the proposed mechanism, polycarbonate/cyclic carbonate
selectivity should be determined in the two ways. The first one
is path (i) vs path (ii), that is, bimolecular nucleophilic attack of
metal-bound carbonate end to the activated epoxide vs
carbonate dissociation from the metal center. The rate for
path (i) depends on the value of ΔGcrb − ΔGepx (vide supra),
whereas that for path (ii) can be evaluated by ΔGcrb itself.
Accordingly, the rate difference between paths (i) and (ii)
could be evaluated by ΔGepx [ΔGcrb− (ΔGcrb − ΔGepx) =
ΔGepx]. The second one is path (iii) vs path (iv), that is, the
nucleophilic attack of metal-free carbonate end to the activated
epoxide vs the intramolecular backbiting of metal-free
carbonate end. The rate for path (iii) could be a function of
Scheme 2. Copolymerization of PO with CO2 by Using
Complexes 1−9
results are summarized in Table 1 and Figure 3. Although each
parameter ΔGepx or ΔGcrb itself did not show any correlation
with the catalytic activity [turnover frequency (TOF) for PPC
generation], the difference between these two parameters,
Table 1. Computed ΔGcrb − ΔGepx Values and Experimental
TOFs for PPC
ΔGepx
ΔGcrb
ΔGcrb − ΔGepx
TOF for PPC
[h−1
complex
[kJ/mol]
[kJ/mol]
[kJ/mol]
]
1
11.4
31.9
22.8
4.8
55.5
87.2
95.9
79.0
106.4
93.6
65.3
82.3
70.2
55.4
49.5
44.1
55.3
73.1
74.2
89.8
69.6
67.4
61.8
61.3
59.9
55.6
1144
88
2
3
28
4
14
5
16.6
24.0
−2.1
21.1
8.9
0.7
16
6
7
32
8
230
141
9
a
10
11
−4.5
−10.5
nd
a
nd
a
concentration of epoxide coordinating intermediate, i.e., ΔGepx
,
nd = not determined because the copolymer was not obtained with
complexes 10 and 11.
whereas the rate of the path (iv) should be less affected by the
10730
dx.doi.org/10.1021/ja5046814 | J. Am. Chem. Soc. 2014, 136, 10728−10735

Journal of the American Chemical Society
Article
complexes 1−9 and smaller ΔGcrb − ΔGepx values than
complexes 3−9. These results indicate that sufficiently large
ΔGepx value is concurrently necessary for PPC generation and
high PPC/CPC selectivity even if a complex shows low ΔGcrb
− ΔGepx value.26
Scheme 3. Mechanism for Cyclic Carbonate Formation
PPC/PPO Selectivity. The PO homopolymerization activity
of the complexes was investigated to estimate PPC/PPO
selectivity since PPO unit is generated via consecutive insertion
of epoxide (Scheme 4a). Catalysts with high PO homopolyme-
Scheme 4. Mechanism for Homopolymerization of Epoxide
rization activity relative to PO/CO2 copolymerization activity
should have low PPC/PPO selectivity. The ΔGalk − ΔGepx
[ΔGalk: dissociation energy of alkoxide ligand from the metal
center (Scheme 4b)] was first investigated as an indicator for
PO homopolymerization activity by analogy with the discussion
about catalytic activity in the PO/CO2 copolymerization (vide
supra). However, no trend was observed in the plot of the
ΔGalk − ΔGepx value vs the TOF of complexes 1−9 for the PO
homopolymerization (Figure 6, Table 2). Complexes 1, 2, and
catalyst system. Therefore, ΔGepx can be an indicator for the
rate difference between paths (iii) and (iv). Accordingly, ΔGepx
was regarded as a promising candidate of an indicator for PPC/
CPC selectivity.23
In contrast to the discussion about catalytic activity for PPC
generation, quantitative evaluation of PPC/CPC selectivity
seems to be difficult because the CPC generation rate decreases
at an early stage in the copolymerization, and thus the PPC/
CPC ratio changes continuously over the reaction time (vide
supra). Therefore, we compared complexes 1−9 with Al (10)24
and Zn (11)25 complexes with complete CPC selectivity
(Figure 4) to clarify the relationship between the ΔGepx values
Figure 4. Structures of completely CPC selective catalysts.
Figure 6. Plot for ΔGalk − ΔGepx vs TOF for PPO.
and the PPC/CPC selectivity. As shown in Table 1 and Figure
5, complexes 10 and 11 possess smaller ΔGepx value than
9 showed low or no activity for the homopolymerization
despite the prediction that smaller ΔGalk − ΔGepx should
provide higher PO homopolymerization activity.
A steric factor was found to correlate with the PPC/PPO
selectivity. In contrast that there are several possible
mechanisms for polycarbonate formation (Figure 1 and
Supporting Information), polyether formation has been mostly
proposed to proceed via bimetallic mechanism27 between
epoxide-coordinating species and a nucleophilic metal−
alkoxide.18,28 It should be noted that a bimetallic mechanism
in polyether formation requires shorter distance between the
two metal canters compared to the polycarbonate formation as
illustrated in Figure 7; the metal-bound alkoxide oxygen needs
to approach epoxide carbon center in PPO formation, while the
carbonate can attack the carbon center through not the metal-
bound oxygen but the carbonyl oxygen in PPC formation.
Figure 5. Plot of ΔGepx vs PPC/CPC selectivity.
10731
dx.doi.org/10.1021/ja5046814 | J. Am. Chem. Soc. 2014, 136, 10728−10735

Journal of the American Chemical Society
Article
Table 2. Calculated Energies and Experimental PPC/PPO
a b
,
Selectivity
ΔGalk
−
TOF for
PPO
ΔGalk
ΔGepx
Ester
[kJ/mol]
%VBur
[%]
PPC/
PPO
complex [kJ/mol] [kJ/mol]
[h−1
]
1
2
3
4
5
6
7
8
9
163.1
194.2
238.6
212.4
243.9
221.4
164.2
197.1
173.3
151.7
162.3
215.9
207.6
227.3
197.4
166.3
176.1
164.4
<1
3.5
33.2
31.1
25.1
25.9
18.1
15.2
9.8
0.369
0.367
0.338
0.347
0.377
0.314
0.263
0.273
0.349
>99/<1
95/5
<1
98/2
Figure 9. Parameters used for computation of the percent buried
volume for (a) NHC ligand (original one)31 and (b) metal−alkoxide
intermediate (proposed in this work).
<1
89/11
>99/<1
40/60
66/34
19/81
93/7
<1
12
1800
1150
24
12.1
30.4
He8_Steric_Parameter. The helium atoms were located in a
plane 2.14 Å apart from the alkoxide oxygen. The distance is
the average value of the calculated one between alkoxide
oxygen and methylene carbon of PO in the transition state of
the nucleophilic attack of alkoxide to activated epoxide with
bimetallic Co−salen catalyst (Figure 8b).33 The intermediates
were reoptimized in the presence of these He atoms, starting
from an optimized conformation of the original complex. The
oxygen atom of the alkoxide ligand was constrained to lie
exactly 2.14 Å above the ring centroid along the perpendicular
to its plane. On the basis of these conditions, interaction energy
Ester was computed between the intermediate in the ground-
state conformation and the ring of eight helium atoms.
a
Gibbs energy computed at the B3LYP-D(PCM)/def2-SVP//B3LYP/
BI level. Total electronic energy computed at the B3LYP(PCM)/
def2-SVP//B3LYP/BI level.
b
Figure 7. Difference in the distance of two species between polyether
formation and polycarbonate formation.
Percent Buried Volume. In the same way as for NHC
ligands, a sphere with 3.5 Å radius centered at the alkoxide
oxygen atom was built to obtain %VBur by using free program
(see Experimental Section in Supporting Information).
Accordingly, more bulky ligand would make a complex less
active for PO homopolymerization and more PPC selective by
keeping the two metal centers apart from each other.29 We
hypothesized that bulky environment of complexes 1, 2, and 9
resulted in low PO homopolymerization activity.
In order to evaluate steric environment associated with each
ligand, we used two parameters, He8_steric parameter30 and
percent buried volume.31 The former parameter was proposed
by the Harvey, Orpen, and co-workers to estimate the
interaction energy between the phosphorus ligand (PA3) and
a planar ring of eight helium atoms which is located 2.28 Å
apart from phosphorus atom (Figure 8a). This parameter has
As summarized in Table 2, complexes 1, 2, and 9 showed
relatively large value of Ester and %VBur (Figure S3). This clearly
supports our assumption that the sterically hindered environ-
ment of these complexes caused low PO homopolymerization
activity. In addition, by integrating the discussions above, the
combination of (i) small ΔGalk − ΔGepx value and (ii) low
steric hindrance around the alkoxide oxygen should be
necessary for high PO homopolymerization activity. In other
words, PPC/PPO selectivity of the complex with these
requirements should be lower. This is clearly shown in the
plot, where the vertical axis is steric parameter and horizontal
axis is the ΔGalk − ΔGepx value (Figures 10 and S4). Complexes
6−8 showed lower values for both parameters and actually
showed relatively low PPC/PPO selectivity.
The discussion so far clearly showed that the facile
estimation for PPC activity, PPC/CPC selectivity, and PPC/
Figure 8. Geometry used for computation of the He8_steric parameter
for (a) phosphorus ligand PA3 (original one)30 and (b) metal−
alkoxide intermediate (proposed in this work).
the correlation with Tolman cone angle, which is known as a
parameter of the steric environment of phosphorus ligands.32
The latter was proposed by Nolan and co-workers to estimate
the steric environment of carbene ligands defined as the percent
of the total volume of a sphere occupied by the ligand 2.10 Å
apart from the sphere center (Figure 9). These two parameters
were employed in our study to evaluate the steric environment
of the metal−alkoxide intermediate for complexes 1−9.
Figure 10. Plot of the ΔGalk − ΔGepx value vs Ester value.
10732
dx.doi.org/10.1021/ja5046814 | J. Am. Chem. Soc. 2014, 136, 10728−10735

Journal of the American Chemical Society
Article
Scheme 5. Synthesis of Complexes 12 and 13
PPO selectivity was accomplished by simple DFT calculations
without transition-state optimizations.
Development of New Catalysts. Based on the indicators
we proposed above, two types of novel complexes were
synthesized (Scheme 5): First one is Ge−boxdipy complexes
12 with substituents X (X = OMe, H) at the para position on
phenolate moiety, and second one is Fe−corrole complex 13
with bulky substituents. DFT calculations of complexes 12
demonstrated that the ΔGcrb − ΔGepx values decrease in the
order 12b > 4 > 12a, and the predicted catalytic activity
increases in the order 12b < 4 < 12a. This trend can be also
interpreted in the way that electron-donating substituents at the
para position on phenolate moiety improve the catalytic
activity. Fe(III)−corrole complex 13 was expected to show
improved PPC/PPO selectivity by increasing the steric
repulsion, although ΔGcrb − ΔGepx or ΔGalk−ΔGepx could
not be computed because of its large number of atoms. Ge−
boxdipy complexes were synthesized as shown in Scheme 5.
Arylation of pyrrole by 1-bromo-3-tert-butyl-2,5-dimethoxyben-
zene (14a)34 or 1-bromo-3-tert-butyl-2-methoxybenzene
(14b)35 afforded 15. Reaction of 15 with benzaldehyde in
the presence of catalytic amount of trifluoroacetic acid, and the
following oxidation with DDQ gave methyl-protected ligand
precursor 16. Treatment of 16 with an excess amount of
sodium thiolate resulted in the selective demethylation of
methoxy groups next to tert-butyl substituents.The structure of
demethylated product 17 was confirmed by X-ray structural
analysis (Figure 11). Deprotonation with NaH and subsequent
complexation with germanium chloride gave germanium
complexes 12. Fe−corrole complex 13 with bulky substituent
was afforded by the reaction of corrole ligand 1836 and FeCl2 in
DMF.
Figure 11. X-ray structure of compound 17a (hydrogen atoms are
omitted for clarity).
12 h]. The results are summarized in Table 3. Ge−boxdipy
complex 12a and 12b showed higher and lower TOF for PPC
generation than complex 4, respectively (Table 3, entries 1−3).
This trend was consistent with the predictions based on the
indicator ΔGcrb − ΔGepx. Accordingly, these results demon-
strated that the indicator ΔGcrb − ΔGepx may be used for
prediction of the substituent effect, while the difference is still
small. In addition, the order of PPC/PPO selectivity seemed to
reflect the steric environment of the complexes. The PPC/PPO
selectivity increased in the order 12b < 12a < 4 with increase in
steric bulkiness at the para position on phenolate moiety [H
t
(12b) < OMe (12a) < Bu (4)].
Completely alternating copolymer of PO and CO2 (Table 3,
entry 5) was obtained by using Fe−corrole complex 13 with
bulky substituents. The complete PPC/PPO selectivity
contrasts with the low PPC/PPO selectivity observed with
less bulky complex 8 in our previous study (Table 3, entry 4).
This is the first example of completely alternating copoly-
merization of PO and CO2 by using iron complexes. Further
ligand design for improvement in TOF for PPC and PPC/CPC
selectivity is currently in progress.
The copolymerization of PO with CO2 was carried out under
standard conditions [PO/complex/[PPN]Cl = 2000/1/0.5
([PPN]Cl = [Ph3PNPPh3]Cl); CO2 = 2.0 MPa; 60 °C;
10733
dx.doi.org/10.1021/ja5046814 | J. Am. Chem. Soc. 2014, 136, 10728−10735

Journal of the American Chemical Society
Article
a
Table 3. Copolymerization of Propylene Oxide with CO2 by Using Ge and Fe Complexes
b
b
c
b
d
d
entry complex yield of copolymer + CPC [%]
copolymer/CPC
TOF for PPC ΔGcrb − ΔGepx carbonate linkage [%]
Mn [g·mol−1
]
Mw/Mn
1
2
3
12a
4
12
10
7
91/9
15
14
8
72.7
74.2
75.6
61.8
−
88
4800
6000
2500
39 000
3900
1.2
1.1
1.2
1.2
1.4
95/5
89
12b
8
90/10
>99/<1
30/70
73
f
4
60
11
230
5
19
5
13
>99
a
PO (1.0 mL, 14.3 mmol), complex 4, 8, 12, or 13 (7.0−7.2 × 10−3 mmol), [PPN]Cl (3.5−3.6 × 10−3 mmol) in a 50 mL autoclave at 60 °C for 12
b
c
1
h. Determined on the basis of H NMR spectroscopy of the crude product by using phenanthrene as an internal standard. TOF = (mol of
d
f
carbonate repeating unit)·(mol of complex)−1·h−1. Determined by size-exclusion chromatography analysis using a polystyrene standard; see ref 7. 1
h.
toward Straightforward Synthesis”, No. 23105507 from MEXT
to S.M. and K. Nozaki. Partial support to K.Nozaki from
Funding Program for Next Generation World-Leading
Researchers, Green Innovation and the Global COE Program
“Chemistry Innovation through Cooperation of Science and
Engineering” from MEXT/JSPS, Japan is also acknowledged.
The generous allotment of computation time from the
Research Center for Computational Science (RCCS), the
National Institutes of Natural Sciences, Japan, is also gratefully
acknowledged. We are grateful to Mr. Masahiro Hatazawa, Mr.
Ryo Nakano, Dr. Shuhei Kusumoto (U Tokyo) and Mr.
Takayoshi Yoshimura (Ibaraki U) for data collections, helpful
suggestions, single-crystal X-ray analysis, and DFT setup,
respectively.
CONCLUSION
■
In summary, here we speculated the reaction mechanism of the
PPC formation and proposed three indicators for (1) catalytic
activity for PPC, (2) PPC/CPC selectivity, and (3) PPC/PPO
selectivity. DFT studies afforded an effective indicator, ΔGcrb
−
ΔGepx, for the evaluation of the catalytic activities for PPC
generation. Furthermore, the ΔGepx itself was found to be an
indicator for PPC/CPC selectivity. If ΔGepx is too small, the
complex tends to be CPC selective. These indicators can be
easily calculated by DFT methods without computing
transition states, so that it can be used as a standard when
the brand-new catalyst candidates were screened. An indicator,
ΔGalk − ΔGepx, and steric environment around the active center
determined the PPC/PPO selectivity. Small ΔGalk − ΔGepx and
small steric bulk around the metal center facilitate the PPO
formation, which resulted in lowering PPC/PPO selectivity.
Two types of metal complexes 12 and 13 were designed
based on the parameters above. Copolymerization results with
Ge-complexes 12 showed that the proposed indicators can be
used to predict the ligand effect. Completely alternating
copolymer was given by using Fe-complex 13 owing to its
sterically hindered ligand.
REFERENCES
■
(1) Inoue, S.; Koinuma, H.; Tsuruta, T. J. Polym. Sci., Part B: Polym.
Lett. 1969, 7, 287−292.
(2) (a) Darensbourg, D. J. Chem. Rev. 2007, 107, 2388−2410.
(b) Kember, M. R.; Buchard, A.; Williams, C. K. Chem. Commun.
2011, 47, 141.
(3) (a) Aida, T.; Inoue, S. J. Am. Chem. Soc. 1983, 105, 1304−1309.
(b) Aida, T.; Inoue, S. J. Am. Chem. Soc. 1985, 107, 1358−1364.
(c) Aida, T.; Sanuki, K.; Inoue, S. Macromolecules 1985, 18, 1049−
1055. (d) Aida, T.; Ishikawa, M.; Inoue, S. Macromolecules 1986, 19,
8−13. (e) Aida, T.; Inoue, S. Acc. Chem. Res. 1996, 29, 39−48.
(4) (a) Cheng, M.; Lobkovsky, E. B.; Coates, G. W. J. Am. Chem. Soc.
1998, 120, 11018−11019. (b) Cheng, M.; Moore, D. R.; Reczek, J. J.;
Chamberlain, B. M.; Lobkovsky, E. B.; Coates, G. W. J. Am. Chem. Soc.
2001, 123, 8738−8749. (c) Moore, D. R.; Cheng, M.; Lobkovsky, E.
B.; Coates, G. W. Angew. Chem., Int. Ed. 2002, 41, 2599−2602.
(d) Allen, S. D.; Moore, D. R.; Lobkovsky, E. B.; Coates, G. W. J. Am.
Chem. Soc. 2002, 124, 14284−14285.
(5) (a) Darensbourg, D. J.; Yarbrough, J. C. J. Am. Chem. Soc. 2002,
124, 6335−6342. (b) Darensbourg, D. J.; Yarbrough, J. C.; Ortiz, C.;
Fang, C. C. J. Am. Chem. Soc. 2003, 125, 7586−7591. (c) Darensbourg,
D. J.; Rodgers, J. L.; Fang, C. C. Inorg. Chem. 2003, 42, 4498−4500.
(d) Darensbourg, D. J.; Mackiewicz, R. M.; Rodgers, J. L.; Fang, C. C.;
Billodeaux, D. R.; Reibenspies, J. H. Inorg. Chem. 2004, 43, 6024−
6034. (e) Darensbourg, D. J.; Mackiewicz, R. M.; Rodgers, J. L.;
Phelps, A. L. Inorg. Chem. 2004, 43, 1831−1833. (f) Darensbourg, D.
J.; Mackiewicz, R. M. J. Am. Chem. Soc. 2005, 127, 14026−14038.
(g) Darensbourg, D. J.; Phelps, A. L. Inorg. Chem. 2005, 44, 4622−
4629. (h) Darensbourg, D. J.; Mackiewicz, R. M.; Billodeaux, D. R.
Organometallics 2005, 24, 144−148.
Here we presented general explanation for catalytic activity
and selectivities common for versatile metal complexes. We
believe that this result will give an impact for future catalyst
designs.
ASSOCIATED CONTENT
* Supporting Information
Experimental procedures, consideration on rate-determining
step of the copolymerization, real-time IR monitoring,
computational methods, crystallographic data, and NMR
spectra. This material is available free of charge via the Internet
at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
■
Corresponding Authors
nozaki@chembio.t.u-tokyo.ac.jp
smori@mx.ibaraki.ac.jp
Notes
The authors declare no competing financial interest.
(6) (a) Qin, Z.; Thomas, C. M.; Lee, S.; Coates, G. W. Angew. Chem.,
Int. Ed. 2003, 42, 5484−5487. (b) Cohen, C. T.; Chu, T.; Coates, G.
W. J. Am. Chem. Soc. 2005, 127, 10869−10878. (c) Cohen, C. T.;
Thomas, C. M.; Peretti, K. L.; Lobkovsky, E. B.; Coates, G. W. Dalton
ACKNOWLEDGMENTS
■
This work was supported by Grant-in-Aid for Scientific
Research on Innovative Areas “Molecular Activation Directed
10734
dx.doi.org/10.1021/ja5046814 | J. Am. Chem. Soc. 2014, 136, 10728−10735

Journal of the American Chemical Society
Article
Trans. 2006, 237−249. (d) Cohen, C. T.; Coates, G. W. J. Polym. Sci.,
Part A: Polym. Chem. 2006, 44, 5182−5191. (e) Lu, X. B.; Wang, Y.
Angew. Chem., Int. Ed. 2004, 43, 3574−3577. (f) Noh, E. K.; Na, S. J.;
Sujith, S.; Kim, S. W.; Lee, B. Y. J. Am. Chem. Soc. 2007, 129, 8082−
8083. (g) Sujith, S.; Min, K. K.; Seong, J. E.; Na, S. J.; Lee, B. Y. Angew.
Chem., Int. Ed. 2008, 47, 7306−7309. (h) Ren, W.-M.; Liu, Z.-W.;
Wen, Y.-Q.; Zhang, R.; Lu, X.-B. J. Am. Chem. Soc. 2009, 131, 11509−
11518.
contribution of other pathways for CPC generation becomes larger
which does not show large effect for other catalysts, such as a
backbiting of metal-coordinating carbonate. .
(27) For Al−porphyrin complex, monometallic mechanism was
proposed; see ref 12.
(28) Braune, W.; Okuda, J. Angew. Chem., Int. Ed. 2003, 42, 64−68.
(29) The size of the substituent on ligand was found to affect the
ring-opening transition state in (salen)Co-catalyzed epoxide hydrolysis
via bimetallic mechanism; see: Ford, D. D.; Nielsen, L. P. C.; Zuend, S.
J.; Musgrave, C. B.; Jacobsen, E. N. J. Am. Chem. Soc. 2013, 135,
15595−15608.
(7) Nakano, K.; Kobayashi, K.; Nozaki, K. J. Am. Chem. Soc. 2011,
133, 10720−10723.
(8) Nakano, K.; Kobayashi, K.; Ohkawara, T.; Imoto, H.; Nozaki, K.
J. Am. Chem. Soc. 2013, 135, 8456−8459.
(30) Fey, N.; Tsipis, A. C.; Harris, S. E.; Harvey, J. N.; Orpen, A. G.;
Mansson, R. A. Chem.Eur. J. 2006, 12, 291−302.
(9) Robert, C.; Ohkawara, T.; Nozaki, K. Chem.Eur. J. 2014, 20,
(31) Hillier, A. C.; Sommer, W. J.; Yong, B. S.; Petersen, J. L.;
Cavallo, L.; Nolan, S. P. Organometallics 2003, 22, 4322−4326.
(32) Tolman, C. A. Chem. Rev. 1977, 77, 313−348.
4789−4795.
(10) (a) Darensbourg, D. J.; Wei, S.-H. Macromolecules 2012, 45,
5916−5922. (b) Darensbourg, D. J.; Yeung, A. D. Macromolecules
2013, 46, 83−95. (c) Darensbourg, D. J.; Yeung, A. D.; Wei, S.-H.
Green Chem. 2013, 15, 1578−1583.
(33) Ahmed, S. M.; Poater, A.; Childers, M. I.; Widger, P. C. B.;
LaPointe, A. M.; Lobkovsky, E. B.; Coates, G. W.; Cavallo, L. J. Am.
Chem. Soc. 2013, 117, 18901−18911.
(34) Johansson, H.; Shanks, D.; Engman, L.; Amorati, R.; Pedulli, G.
F.; Valgimigli, L. J. Org. Chem. 2010, 75, 7535−7541.
(35) Gruza, M. M.; Chambron, J.-C.; Espinosa, E.; Aubert, E. Eur. J.
Org. Chem. 2009, 2009, 6318−6327.
(36) Liu, H.-Y.; Yam, F.; Xie, Y.-T.; Li, X.-Y.; Chang, C. K. J. Am.
Chem. Soc. 2009, 131, 12890−12891.
(11) Chen, P.; Chisholm, M. H.; Gallucci, J. C.; Zhang, X.; Zhou, Z.
Inorg. Chem. 2005, 44, 2588−2595.
(12) Chisholm, M. H.; Zhou, Z. J. Am. Chem. Soc. 2004, 126, 11030−
11039.
(13) Liu, Z.; Torrent, M.; Morokuma, K. Organometallics 2002, 21,
1056−1071.
(14) Luinstra, G. A.; Haas, G. R.; Molnar, F.; Bernhart, V.; Eberhardt,
R.; Rieger, B. Chem.Eur. J. 2005, 11, 6298−6314.
(15) (a) Lehenmeier, M. W.; Bruckmeier, C.; Klaus, S.; Dengler, J. E.;
Deglmann, P.; Ott, A.-K.; Rieger, B. Chem.Eur. J. 2011, 17, 8858−
8869. (b) Buchard, A.; Jutz, F.; Kember, M. R.; White, A. J. P.; Rzepa,
H. S.; Williams, C. K. Macromolecules 2012, 45, 6781−6795.
(16) (a) Moore, D. R.; Cheng, M.; Lobkovsky, E. B.; Coates, G. W. J.
Am. Chem. Soc. 2003, 125, 11911−11924. (b) Kember, M. R.; White,
A. J. P.; Williams, C. K. Macromolecules 2010, 43, 2291−2298.
(17) Tokunaga, M.; Larrow, J. F.; Kakiuchi, F.; Jacobsen, E. N. Science
1997, 277, 936−938.
(18) Hirahata, W.; Thomas, R. M.; Lobkovsky, E. B.; Coates, G. W. J.
Am. Chem. Soc. 2008, 130, 17658−17659.
(19) See the Supporting Information for the detail of kinetic study.
(20) Liu, J.; Ren, W.-M.; Liu, Y.; Lu, X.-B. Macromolecules 2013, 46,
1343−1349.
(21) The TOF represents a composite of the rate of initiation and
propagation. Our in situ IR monitoring of the PO/CO2 copolymeriza-
tion with complexes 1, 3, 4, and 8 revealed that the initiation periods
were negligibly short compared with the total copolymerization
periods and that the propagation proceeded at the same rate from very
early period during the copolymerization. Accordingly, the TOF are
used as the propagation rate in this work.
(22) The Fe−corrole catalyst system demonstrated the highest
activity by the addition of 0.50 equiv of [PPN]Cl to Fe, while the
addition of 1.0 equiv of [PPN]Cl diminished the catalytic activity
significantly. This result indicates the copolymerization proceeded via
bimetallic mechanism, see ref 8.
(23) Bifunctional Co−salen catalyst systems with intramolecular
ammonium moieties demonstrated high PPC/CPC selectivity.6f−h
The high selectivity should be attributed to the capture of metal-free
carbonate end with the ammonium moieties, which affect the rate of
the path (iv). Accordingly, bifunctional systems were not studied in
this work.
(24) (a) Tian, D.; Liu, B.; Gan, Q.; Li, H.; Darensbourg, D. J. ACS
Catal. 2012, 2, 2029−2035. (b) Darensbourg, D. J.; Billodeaux, D. R.
Inorg. Chem. 2005, 44, 1433−1442.
(25) (a) Decortes, A.; Martínez Belmonte, M.; Benet-Buchholz, J.;
Kleij, A. W. Chem. Commun. 2010, 46, 4580. (b) Taherimehr, M.;
Decortes, A.; Al-Amsyar, S. M.; Lueangchaichaweng, W.; Whiteoak, C.
J.; Escudero-Adan, E. C.; Kleij, A. W.; Pescarmona, P. P. Catal. Sci.
́
Technol. 2012, 2, 2231−2237. (c) Castro-Gomez, F.; Salassa, G.; Kleij,
A. W.; Bo, C. Chem.Eur. J. 2013, 19, 6289−6298.
(26) The low selectivity of complex 5 in spite of its relatively large
ΔGepx value may be attributed to its low catalytic activity. That is, the
10735
dx.doi.org/10.1021/ja5046814 | J. Am. Chem. Soc. 2014, 136, 10728−10735