Combined experimental and computational study on catalytic cyclocoupling of epoxides and CO2 using porphyrin-based Cu(II) metal-organic frameworks with 2D coordination networks

Full text of DOI:10.1246/bcsj.20170371

Selected Paper
Combined Experimental and Computational Study on Catalytic
Cyclocoupling of Epoxides and CO Using Porphyrin-Based Cu(II)
2
Metal-Organic Frameworks with 2D Coordination Networks
1
1
2
2
Tsukasa Murayama, Masayuki Asano, Tetsushi Ohmura, Arimitsu Usuki,
1
1
Takeshi Yasui, and Yoshihiko Yamamoto*
1Department of Basic Medicinal Sciences, Graduate School of Pharmaceutical Sciences, Nagoya University,
Chikusa, Nagoya, Aichi 464-8601, Japan
2Toyota Central R&D Laboratories, Inc., Nagakute, Aichi 480-1192, Japan
E-mail: yamamoto-yoshi@ps.nagoya-u.ac.jp
Received: November 16, 2017; Accepted: December 6, 2017; Web Released: December 19, 2017
Yoshihiko Yamamoto
Yoshihiko Yamamoto received his Ph.D. from Nagoya University in 1996, where he was appointed as an
Assistant Professor in 1996 and promoted to an Associate Professor in 2003 and Full Professor in 2012. He
also worked for Tokyo Institute of Technology as an Associate Professor from 2006 to 2008. His research
interests are focused on the development of organometallic catalysts and their application to the synthesis of
biologically important molecules.
Abstract
using MOFs has been a topic of intensive studies. The molec-
ular recognition owing to the distinct pore size and intrinsic
coordination ability of metal components enable adsorption
The cyclocoupling of epoxides and CO was investigated
using porphyrin-based Cu(II) metal-organic frameworks with
2
4
2
D coordination networks. A variety of mono- and disubsti-
and catalytic conversion of small molecules. Moreover, robust
tuted epoxides were transformed into cyclic carbonates under
mild and neat conditions. Several control experiments were
carried out to elucidate that the catalytically active site is the
dicopper paddle wheel unit rather than the copper porphyrin
complex moiety. The proposed mechanism was corroborated
by density functional theory calculations of a model paddle
wheel unit.
coordination networks allow the efficient separation and reuse
of MOFs. Although these features are similar to those of
conventional porous solid materials such as zeolites, MOFs are
advantageous because their topologies, pore sizes, and func-
tions can be readily tunable through rational design of the
2
starting components. For example, the development of CO
2
fixing technologies that utilize the CO adsorption ability of
2
5
MOFs is fundamental to harnessing the major greenhouse gas
6
as an abundant renewable C1 resource. Various MOFs have
been investigated as catalysts for cyclocoupling of epoxides
1
. Introduction
7
and CO . In a pioneering study, Cu (btc) (HKUST-1, btc:
2
3
2
8
Metal-organic frameworks (MOFs) are organic-inorganic
hybrid materials, in which characteristic uniform porous struc-
tures are assembled by coordination networks of metal ions
benzene-1,3,5-tricarboxylate, Figure 1a) catalyzed cyclocou-
pling of epichlorohydrin under CO2 pressure (7 bar) at elevated
temperatures (70­100 °C) to furnish the corresponding cyclic
1
9
and organic linkers. MOFs have garnered enormous attention
carbonate in a modest yield. Later, the catalytic performance
from the scientific community over the last decade, because the
judicious combination of a wide variety of metals with ration-
ally designed organic linkers enabled access to diverse new
was greatly improved using different types of Cu MOFs,
MMPF-9 and MMCF-2, in which tetrakis(3,5-dicarboxybi-
phenyl)porphine (H tdcbpp) or 1,4,7,10-tetrazacyclodode-
1
0
2
MOFs. Moreover, because of their robust porous structures,
cane-N,N¤,N¤¤,N¤¤¤-tetra-p-methylbenzoic acid (H tactmb) were
4
1
0
MOFs find extensive applications in gas storage and separation,
removal of toxic chemicals, (opto-)electronic devices, energy
storage, heterogeneous catalysts, and drug-delivery systems
used as polycarboxylate linkers, respectively. The cyclo-
coupling of propylene oxide proceeded under atmospheric
pressure CO at ambient temperature in the presence of tetra-
2
1
,3
among others.
butylammonium bromide (TBAB) as a cocatalyst, producing
the corresponding cyclic carbonate in higher yields than those
obtained with the benchmark catalyst, HKUST-1. In these
Among these applications, the selective transformation of
small organic raw materials into valuable functional molecules
Bull. Chem. Soc. Jpn. 2018, 91, 383–390 | doi:10.1246/bcsj.20170371
© 2018 The Chemical Society of Japan | 383

O
O
CO2H
O
O
Table 1. Cyclocoupling of styrene oxide and CO using Cu
MOFs
(
a)
2
Cu
Cu
R
R
R
R
O
O
O
O
O
Cu MOF (10 mg), 10 mol % TBAB
HO2C
CO2H
O
O
H3btc (HKUST-1)
Ph
a 1 mmol
CO
2
, 28 °C, 48 h
O
paddle wheel units
CO2H
Ph
2
CO2H
CO2H
3a
HO2C
Entry
Cu MOF
CO2/MPa
Yield/%a
HO2C
HO2C
CO2H
1
HKUST-1
0.1
0.1
0.1
0.1
2.5
5.0
47
52
66
46
77
74c
N
H
N
N
N
N
2
1a
1b
1b
1b
1b
3
4b
N
N
H
N
5
CO2H
6
a
Yields of isolated products. bTBAI was used instead of
CO2H
CO2H
CO2H
HO2C
c
1
H4tactmb (MMCF-2)
TBAB. Crude yield determined by H NMR analysis.
H10tdcbpp (MMPF-9)
(
b)
N
N
MOFs 1a and 1b have a 2-dimensional 2.2 nm square-grid
coordination network consisting of a copper paddle wheel unit,
Cu (OCOR) , and a copper tetra(4-pyridyl)porphyrin ligand
R
R
2
4
O O
Cu Cu
O O O O
N
N
N
N
N
N
N
N
O
O
N
Cu
N
N
Cu
N
(CuTPyP). In 1a, the 2D sheets are slip-stacked through
hydrogen bonds between the H atoms of the methyl groups
of one 2D sheet and the carbonyl oxygens of an adjacent 2D
sheet to form an overall 3D structure, in which the square
cavities are not aligned. In 1b, acetate ligands are replaced by
1-naphthoates. In contrast to the 3D structure of 1a, that of 1b
has 2D sheets stacked through CH­π interactions between 1-
naphthoate groups on the faces and edges of the adjacent 2D
sheets. Because of the bulky 1-naphthyl groups, 1b has lower
gas adsorption ability, albeit with higher CO2/N2 selectivity
than 1a. We hypothesized that the porous structures of 1a and
R
R
N
N
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
Cu
Cu
R
R
R
R
R
R
R
R
Cavity
Cu
N
Cu
N
R
O
R
N
N
N
O
O O
Cu Cu
O O O O
N
N
N
N
N
Cu
N
N
N
Cu
N
N
N
1
b, together with two different copper environments, could
R
R
lead to unprecedented molecular recognition and catalytic
ability. Unlike conventional Cu MOFs, the networks of 1a and
1b are assembled by coordination between the pyridyl moiety
on the porphyrin ligand and the copper paddle wheel units.
Therefore, no apparent vacant site of the paddle wheel unit
is available for activation of epoxides. Herein, we report the
results of our study on catalytic cyclocoupling of epoxides and
2.2 nm
R = Me in 1a
R = 1-naphthyl in 1b
CO using Cu-MOFs 1a and 1b with particular emphasis on the
2
substrate scope and elucidation of catalytic activity origin.
2
. Results and Discussion
1a slip-stack alignment
1b stack alignment
The cyclocoupling of styrene oxide 2a with CO was con-
2
ducted using Cu MOFs 1a and 1b along with commercial
HKUST-1 (Basolite 300 ) as a benchmark (Table 1). The reac-
Figure 1. (a) Copper paddle wheel units and polycarbox-
ylic acid ligands used for previous Cu MOFs, HKUST-1,
MMPF-9, and MMCF-2; (b) Porphyrin-based Cu MOFs
developed by Toyota Central R&D Labs.
μ
tion was performed by adding 10 mg of MOFs and 10 mol %
n
Bu NBr (TBAB) to 1 mmol of 2a without a solvent under
4
atmospheric pressure (0.1 MPa) of CO at 28 °C for 48 h. When
2
examples, copper complexes of the porphyrin or cyclic poly-
amine ligands boosted the catalytic efficiency, although copper
paddle wheel units, Cu (OCOR) , have been recognized as
catalytically active sites. Moreover, a copper porphyrin-based
MOF has been reported to be effective for CO2 capture and
subsequent activation.11
HKUS1 was used as the reference catalyst, the expected car-
bonate 3a was obtained in 47% yield (Entry 1). A slightly
higher yield of 52% was obtained using 2D-MOF 1a (Entry 2).
The 1-naphthoate analog 1b was used as the catalyst under the
2
4
same conditions, affording 3a in an improved yield of 66%
n
(Entry 3). When Bu NI (TBAI) instead of TBAB was used as
4
We investigated the catalytic activity of porphyrin-based Cu
MOFs 1a and 1b, developed by Toyota Central R&D Labs,
the cocatalyst together with 1b, the yield of 3a decreased to
46% (Entry 4). The effect of CO pressure was then examined
2
1
2
toward the cyclocoupling of epoxides and CO (Figure 1b).
using 1b. The yield of 3a increased to 77% when the reaction
2
384 | Bull. Chem. Soc. Jpn. 2018, 91, 383–390 | doi:10.1246/bcsj.20170371
© 2018 The Chemical Society of Japan

Table 2. Cyclocoupling of styrene oxide and CO2 using Cu
catalystsa
Table 3. Catalyst recycling experiments using 1b for cyclo-
coupling of 2aa
O
1
b (10 mg), 10 mol % TBAB
O
Cycle
1b amt/mg
10.0
8.7
Yield/%
TON
26.0
27.4
29.1
30.5
30.1
O
Ph
a 1 mmol
CO (0.1 MPa), 28 °C, 48 h
2
O
1
2
3
4
5
74
68
54
47
42
Ph
2
3
a
6.5
Entry
Variation from the standard conditions
None
Yield/%
5.4
1
2
3
4
5
6
7
66
4.9
Without TBAB
0
aAll reactions were performed on a 1 mmol scale using 10
mol % TBAB under 2.5 MPa CO2 at 28 °C for 48 h. Yields of
isolated products are shown.
22b
Without 1b
2.85 mol % Cu(OAc)2 instead of 1b
2.85 mol % Cu(Naph)2 instead of 1b
2.85 mol % CuTPP instead of 1b
AcOEt extract was used instead of 1b
45
67
15b
25b
reaction of styrene oxide 2a was conducted using 1b under
.5 MPa CO . After the completion of the first cycle, the
2
2
a
Naph: 1-naphthoate, TPP: 5,10,15,20-tetraphenylporphinate.
reaction mixture was diluted with AcOEt and 1b was separated
by centrifugation. The recovered 1b was dried under vacuum
and reused in the next cycles. As summarized in Table 3,
Cu MOF 1b could be reused over five cycles; however, the
amounts of recovered 1b gradually decreased, and accordingly,
the yields of 3a lowered from 74% for the 1st cycle to 42% for
the fifth cycle. Nevertheless, turnover number (TON) gradu-
ally increased over reaction cycles. These results indicate that
recovered 1b maintained its catalytic activity, although the
amount of recovered 1b decreased by partial degradation with
unsticking of the 2D MOF sheets. The slight increase in TON
can be attributed to the accumulated TBAB cocatalyst. In strik-
ing contrast, copper (II) 1-naphthoate (Naph), Cu(Naph)2,
could not be recycled in a similar manner, which indicates an
advantage of the MOF catalyst regarding recyclability.
b
Yields of isolated products are shown. Crude yield determined
by H NMR analysis.
1
was carried out under 2.5 MPa CO (Entry 5). An increase in
CO pressure to 5.0 MPa did not lead to further improvement
Entry 6).
2
2
(
Several control experiments were conducted to investigate
the origin of the observed catalytic activity of 1a and 1b, as
summarized in Table 2. No reaction occurred in the absence
of TBAB, which implies that this cocatalyst is imperative
Entry 2). In contrast, TBAB alone catalyzed the cyclocoupling
to afford 3a with a low yield of 22% (Entry 3). Several
copper compounds were tested as catalysts. Because 10 mg of
(
13
1
b corresponds to 2.85 mol % Cu, 2.85 mol % of Cu(O2CR)2
was used as the catalyst. Cu(II) acetate and 1-naphthoate exhib-
ited significant catalytic activity, affording 3a in 45% and 67%
yields, respectively (Entries 4 and 5). The use of CuTPP as a
catalyst led to the formation of 3a in a lower yield than that
obtained with TBAB alone (Entry 6).14 Therefore, the catalytic
activity of the copper porphyrin was negligible, and the effect
of the carboxylate ligands on the efficiency was the same as
that observed for 1a/1b. These results suggest that the catal-
ytically active sites of 1a and 1b are the copper paddle wheel
units rather than the copper porphyrin moiety. This consid-
eration is in striking contrast to the previous study that implied
a significant contribution of the copper porphyrin moiety to the
Previous studies on epoxide-CO cyclocoupling with Cu
2
MOF catalysts have reported limitations in epoxide scope.
In general, epoxides larger than propylene oxide lowered
the product yields. Functional-group compatibility was also
neglected in those studies. We performed the reactions of vari-
ous epoxides to establish the substrate scope as summarized in
Table 4. The reaction of 2-(tert-butoxymethyl)oxirane (2b) was
conducted under the standard conditions and at atmospheric
pressure of CO furnished cyclic carbonate 3b in high yield
2
(82%), indicating that the bulky tert-butoxymethyl substituent
was tolerated by 1b (Entry 1). Acetoxymethyl derivative 3c,
propargyloxymethyl derivative 3d, and hydroxymethyl deriv-
ative 3e were also obtained in 56­85% yields (Entries 2­4).
The oxygen functionality in the substituent was unnecessary
because both chloromethyl derivative 3f and but-3-enyl deriv-
ative 3g were obtained in 68% yields (Entries 5 and 6). Cyclic
carbonates with perfluoroalkyl substituents have been claimed
to be superior solvents for the electrolyte salt of lithium second-
ary batteries as compared to conventional ethylene and pro-
catalytic performance of MMPF-9 in epoxide-CO cyclocou-
2
pling. Moreover, the efficiency of 1a and 1b was greater than
that of HKUST-1 despite the paddle wheel units in 1a and 1b
with saturated coordination sites. Thus, the mechanism for
the activation of epoxides deserves further discussion (vide
infra). Finally, potential catalyst leaching was examined as
follows: 10 mg of 1b was stirred in AcOEt at 28 °C for 24 h,
and after centrifugation, the AcOEt supernatant was decanted
and concentrated. Subsequently, 10 mol % TBAB and 1 mmol
of 2a were added to the obtained extract, and the reaction
mixture was stirred at 28 °C for 48 h to obtain 3a in 25% NMR
yield (Entry 7). The contribution of catalyst leaching can be
neglected because the observed yield was comparable to that
of TBAB alone (Entry 2).
1
5
pylene carbonates. The perfluoroalkyl derivative 3h was
obtained in 71% yield from commercial epoxide 2h using our
method (Entry 7). Accordingly, various functional groups
including ester, alcohol, ether, terminal alkene, terminal alkyne,
chloroalkyl, and perfluoroalkyl, were well tolerated, although
unidentifiable byproducts were detected as trace impurities in
1
6
some cases. In addition, 2,2-disubstituted epoxide 2i and
cyclopentane-fused epoxide 2j could be used as substrates
(Entries 8 and 9), but the yield decreased for bicyclic carbonate
We then investigated the recyclability of the catalyst by
taking advantage of the heterogeneous nature of the MOFs. The
Bull. Chem. Soc. Jpn. 2018, 91, 383–390 | doi:10.1246/bcsj.20170371
© 2018 The Chemical Society of Japan | 385

Table 4. Scope of epoxide substratesa
(a)
LA
LA
LA
O
LA
O
O
– Br–
CO2
Yield/
%
O
Entry
1
Epoxide
Carbonate
O
R
O
O
R
O
R
O
O
Br–
Br
R
Br
O
O
tBuO
O
4
5
tBuO
82
85
56
70
2
b
c
(b)
O
3
b
c
d
1b (10 mg), 10 mol % TBAB
O
O
Cl
O
O
CO2 (0.1 MPa), 28 °C, 48 h
O
Cl
O
AcO
(S)-2f (97% ee)
mmol
2
3
4
5
6
AcO
(S)-3f 70% (82% ee)
α]D23 = +30.7° in MeOH
lit. [α]D20 = +37.6° (ref. 18)
1
2
[
3
O
O
O
O
Scheme 1. (a) General mechanism for cyclocoupling of
epoxides and CO2 using LA/TBAB binary catalysts;
O
O
2d
(b) cyclocoupling of non-racemic epoxide (S)-2j using
3
1
b as the catalyst.
O
O
O
HO
Cl
O
CO into the LA­O bond of resultant intermediate 4 generates
HO
Cl
2
2e
carboxylate intermediate 5. The final ring closure from 5 occurs
with concomitant restoration of a bromide ion, affording cyclic
carbonates. Because the 1-naphthoate ligand is more electron-
withdrawing than the acetate ligand, as indicated by the pKa
values of acetic acid and 1-naphthoic acid (4.76 and 3.70,
respectively), the 1-naphthoate ligand renders the copper
center more Lewis acidic than the acetate ligand. This predic-
tion is consistent with the trend in catalytic efficiency found for
3
e
O
O
O
O
68b
2f
3f
1
8
O
O
O
68b
O
2g
the cyclocoupling of epoxides and CO using Cu MOFs 1a and
b or Cu(II) carboxylate salts.
2
3
g
1
O
O
According to this scenario, the attack of a bromide ion
F
F
F
F
O
O
F
F
F
F
should take place at the less-substituted oxirane carbon in the
SN2 mode with retention of configuration at the adjacent chiral
center. To verify this assumption, we employed commercial
chiral non-racemic 2-(chloromethyl)oxirane (S)-2f (97% ee,
GC) as a probe (Scheme 1b). The reaction of (S)-2f under the
standard conditions afforded the corresponding product (S)-3f
in 70% yield. As expected, the optical rotation of the isolated
F
F
7
8
9
71
F
F
F
F
F
F
F
F
2h
3
h
O
O
O
O
Cl
52b
Cl
2i
2
3
3
i
product ([α]D = +30.7° in MeOH) was in good accordance
2
0
with that previously reported for optically pure (S)-3f ([α]D
=
O
1
9
O
O
+
37.6° in MeOH). Although the slight erosion of the enan-
O
27
41)
tiomeric purity was observed for the obtained sample (82% ee),
the stereochemistry of the chiral center was mostly preserved
during the reaction, indicating that the bromide ion attacked the
less-substituted carbon.
According to the results from the control experiments sum-
marized in Table 2, the catalytic activity of Cu MOFs 1a and
1b is ascribed to the copper paddle wheel units rather than
the copper porphyrin moiety. However, the vacant sites of the
paddle wheel units are occupied by the pyridyl ligands, and
no Lewis acid sites are available for electrophilic activation
of epoxides. This situation is also consistent with a recently
c
(
2j
3
j
aAll reactions were performed under standard conditions: 1b
10 mg), TBAB (10 mol %), epoxide (1 mmol), CO2 (0.1 MPa),
(
2
b
8 °C, 48 h. Yields of isolated products are shown. Trace
c
impurities derived from naphthoate were detected. The yield
of the reaction performed at 50 °C.
3
j. Because the steric hindrance is assumed to hamper the
nucleophilic attack of a bromide anion to 2j, this reaction was
repeated at an elevated temperature of 50 °C under otherwise
identical conditions to improve the yield of 3j to 41%.
Scheme 1a shows an accepted mechanism for the cyclo-
reported 1D Cu MOF, Cu(Hip) Bpy (Hip: 5-hydroxyisoph-
2
thalic acid, Bpy: 4,4¤-bipyridyl), in which the coordination sites
of each Cu(II) center are saturated with two carboxylate and
2
0
two pyridyl ligands. To avoid discrepancy between saturation
of the coordination sites and high catalytic performance, the
authors proposed that the coordination mode of the carboxylate
coupling of epoxides and CO using Lewis acid (LA)/TBAB
2
1
7
binary catalysts. The electrophilic activation of epoxides by
LA, followed by the nucleophilic attack of a bromide ion, leads
to opening of the oxirane ring. Subsequent incorporation of
2
1
ligand changed from ¬ to ¬ to open a vacant site (Scheme 2).
A similar change in coordination mode is possible in the
386 | Bull. Chem. Soc. Jpn. 2018, 91, 383–390 | doi:10.1246/bcsj.20170371
© 2018 The Chemical Society of Japan

O
O
O
O
O
O
O
Cu
Ar
Cu
Cu
Ar
Ar
κ2 mode
κ1 mode
=
vacant sites
L
L
O
O
O
Cu
O
O
O
R
R
Cu
R
R
O
R
R
O
O
O
O
Cu
L
O
O
O
R
Cu
L
O
R
O
Scheme 2. Proposed coordination-mode changes in active
copper sites.
dinuclear situation of the paddle wheel units: one of the four
2
1
bridging ¬ carboxylate ligands is converted into a ¬ ligand to
provide an open Lewis acid site for the activation of epoxides.
A possibility that Cu species derived from decomposition of
Cu MOF 1b promote cyclocoupling is less likely because the
leaching species have virtually no catalytic activity (Table 2,
Entry 7).
To evaluate the proposed possibility, density functional
theory (DFT) calculations were performed for a model system
corresponding to a single paddle wheel unit. Triplet states were
calculated at the unrestricted B3LYP21 level because singlet
states were unstable (see SI for details). Single-point energy
calculations were carried out using the Truhlar’s M06L func-
tional, because this functional is more accurate in estimating
2
2
medium-range correlation energy than the B3LYP functional.
A smaller ligand system comprising formates and pyridines
was used for calculation efficiency. Among the carboxylate
ligands, one closely relevant to the reaction site was replaced
by a benzoate ligand. A tetramethylammonium ion replaced a
tetrabutylammonium ion, and ethylene oxide was chosen as an
epoxide substrate.
As discussed above, the opening of the benzoate bridge
would make a vacant coordination site. However, no transition
state (TS) could be located for this change in coordination
mode. Instead, we could find an alternative mechanism for
opening a vacant site (Scheme 3). The bridging benzoate
Scheme 3. Benzoate ligand slippage step starting from
model complex A. Relative Gibbs free energies (298 K,
1
atm) are indicated in parentheses. Atom distances (¡) are
given in italic numbers.
oxide mediated by Me NBr only has a higher activation energy
4
‡
of ¦G = +32.0 kcal/mol and is endergonic by 19.0 kcal/mol
(Figure S1 in SI). Therefore, the activation with the copper
center is imperative for the efficient oxirane ring opening. In
addition, the control experiment showed that CuTPP was
totally inefficient as a catalyst (Table 2, Entry 6). The geom-
etry optimization of Cu porphyrin complex with oxiran and
tetramethylammonium bromide afforded a stationary point
(K, Figure S2 in SI). However, no transition state for the
nucleophilic ring opening of oxiran from K could be located.
This is probably because the interaction between the Cu center
and oxiran is very week; the Cu­O distance is significantly
longer in K (2.522 ¡) than in D (2.197 ¡). Moreover, Cu por-
phyrin complex has no appropriate low-lying vacant orbital to
interact with the HOMO of oxiran, whereas a model dinuclear
ligand underwent slippage via TS with an activation energy
AB
‡
of ¦G = +12.2 kcal/mol to generate unsymmetrical complex
B. In B, the benzoate ligand coordinates to Cu in a κ coordi-
2
b
nation mode, and one of its two oxygen atoms still coordinates
to Cu . Because the C(carbonyl)­O bond lengths (1.307 and
a
1
.238 ¡) and Cu ­O distances (2.048 and 2.623 ¡) are signifi-
b
cantly different, the carbonyl oxygen readily dissociated to
1
generate complex C with the κ benzoate ligand. The ligand
slippage process is kinetically feasible, even though the forma-
tion of C from A is endergonic by 8.5 kcal/mol.
1
Cu complex with a κ -benzoate ligand has a suitable LUMO on
The ring opening of ethylene oxide occurred from complex
D, via nucleophilic attack of the epoxide by a bromide ion
the unsaturated Cu center (Figure S3 in SI).
Subsequently, a CO molecule was inserted into the Cu­O
2
(
Scheme 4). This process was estimated to proceed with an
bond of the resultant Cu bromoethoxide (F ¼ TSFG ¼ G,
Scheme 5). The insertion proceeded via four-center transition
state TSFG, in which the Cu­O(alkoxide) bond was elongat-
ed and, the C(CO )­O(alkoxide) and Cu­O(CO ) distances
‡
activation energy of ¦G = +11.2 kcal/mol, and the formation
of copper bromoethoxide intermediate E is slightly endergonic.
Therefore, this step is thermodynamically feasible and rever-
2
2
1
sible, and it is concluded that the paddle wheel unit with a ¬
were simultaneously shortened. In the CO fragment, one of the
2
carboxylate ligand has sufficient Lewis acidity to promote
oxirane-ring opening. Notably, the ring opening of ethylene
two C=O bonds is longer than the other (1.247 and 1.216 ¡,
respectively), and the bent angle is 137.8°. The activation
Bull. Chem. Soc. Jpn. 2018, 91, 383–390 | doi:10.1246/bcsj.20170371
© 2018 The Chemical Society of Japan | 387

Scheme 4. Epoxide ring-opening step starting from model
complex D. Relative Gibbs free energies (298 K, 1 atm) are
indicated in parentheses. Atom distances (¡) are given in
italic numbers.
‡
barrier was calculated as ¦G = +18.7 kcal/mol and the for-
mation of Cu carboxylate intermediate G from F is slightly
endergonic, which indicates that this process is also thermo-
dynamically feasible and reversible. Subsequent ring closure
of the 2-bromoethoxy carbonate moiety occurred via transition
‡
state TSGH with an activation energy of ¦G = +8.6 kcal/mol,
which is lower than that of the CO insertion. The formation of
2
carbonate complex H from G is exergonic by 9.6 kcal/mol.
Therefore, the final ring closure step is important as a thermo-
dynamic sink to irreversibly generate cyclic carbonates.
The overall energy surface is shown in Figure 2. As
already discussed, each step is kinetically feasible as the activa-
tion barriers are lower than 20 kcal/mol. The most energetically
Scheme 5. Final CO2 insertion and ring-closure steps start-
ing from model complex F. Relative Gibbs free energies at
2
3
2
98 K, 1 atm are indicated in parentheses. Atom distances
(¡) are given in italic numbers.
TSFG
32.7
severe step is that of CO insertion. This is in good accordance
2
+
TSGH
with the fact that the yield of the product from styrene oxide
+28.6
was improved by increasing the CO pressure. In contrast,
TSDE
2
ΔG, kcal/mol
+
8.6
previous DFT calculations on closely relevant cyclocouplings
using binary catalyst systems suggested that epoxide ring open-
+22.3
+18.7
+
20.0
G
TSAB
+12.2
+11.2
2
4
ing rather than CO insertion is the most difficult step. The
2
+
14.0
formation of cyclic carbonate complex H from the starting
complex A is endergonic by 10.4 kcal/mol. This thermody-
namic penalty can be mostly attributed to the benzoate ligand
slippage step with endergonicity of 11.1 kcal/mol. Thus, the
nature of the carboxylate ligand plays a critical role as 1b with
the 1-naphthoate ligand superior to 1a with the acetate ligand.
The efficiency of the cyclocoupling can further be improved by
choosing other ligands that facilitate the ligand slippage.
+
11.1
B
+11.1
D
E/F
+10.4
H
+8.5
C
0
A
.0
Ligand slippage
Epoxide
ring opening
CO2 insertion
ring closure
Figure 2. Calculated energy surface for the overall process,
with relative Gibbs free energies (298 K, 1 atm).
3
. Conclusion
We demonstrated that the combination of porphyrin-based
Cu MOFs 1b with tetrabutylammonium bromide efficiently
catalyzed cyclocoupling of epoxides and CO2 under mild
conditions without solvents. This catalyst system has a good
substrate scope as epoxides with a long and bulky substituent
could be used. A wide range of functional groups, including
388 | Bull. Chem. Soc. Jpn. 2018, 91, 383–390 | doi:10.1246/bcsj.20170371
© 2018 The Chemical Society of Japan

alcohol, ether, ester, alkene, alkyne, and haloalkyl groups, were
compatible with this catalytic system. However, epoxides with
two substituents gave lower yields. The catalyst recycling
studies revealed that 1b could be reused several times for the
cyclocoupling of styrene oxide, although the product yields
gradually decreased.
Soc. Rev. 2017, 46, 3242. p) Z. Kang, L. Fan, D. Sun, J. Mater.
Chem. A 2017, 5, 10073.
4
a) J. Y. Lee, O. K. Farha, J. Roberts, K. A. Scheidt, S. T.
Nguyen, J. T. Hupp, Chem. Soc. Rev. 2009, 38, 1450. b) J. Liu,
L. Chen, H. Cui, J. Zhang, L. Zhang, C.-Y. Su, Chem. Soc. Rev.
2
014, 43, 6011. c) S. Bhattacharjee, Y.-R. Lee, P. Puthiaraj, S.-M.
Cho, W.-S. Ahn, Catal. Surv. Asia 2015, 19, 203. d) A.
Dhakshinamoorthy, A. M. Asiri, H. Garcia, Chem. Soc. Rev.
The control experiments suggest that the catalytically active
sites comprised the Cu paddle wheel unit, although no vacant
site is seemingly available. It was proposed that the bridging
carboxylate ligand dissociates from one of the two copper
2
015, 44, 1922. e) A. H. Chughtai, N. Ahmad, H. A. Younus, A.
Laypkov, F. Verpoort, Chem. Soc. Rev. 2015, 44, 6804. f) A.
Dhakshinamoorthy, A. M. Asiri, H. Garcia, Catal. Sci. Technol.
1
centers to assume a ¬ coordination mode. As such, a vacant
2
016, 6, 5238. g) Y.-B. Huang, J. Liang, X.-S. Wang, R. Cao,
Chem. Soc. Rev. 2017, 46, 126.
a) K. Sumida, D. L. Rogow, J. A. Mason, T. M. McDonald,
E. D. Bloch, Z. R. Herm, T.-H. Bae, J. R. Long, Chem. Rev. 2012,
12, 724. b) D. Andirova, C. F. Cogswell, Y. Lei, S. Choi,
coordination site is available for the electrophilic activation of
epoxides. This assumption was corroborated by the DFT calcu-
lations of a single paddle wheel unit as a model system. The
computed energy surfaces suggest that the epoxide ring open-
5
1
ing, CO insertion, and the final carbonate ring closure occur at
2
Microporous Mesoporous Mater. 2016, 219, 276. c) Y.
Belmabkhout, V. Guillerm, M. Eddaoudi, Chem. Eng. J. 2016,
296, 386. d) J. Yu, L.-H. Xie, J.-R. Li, Y. Ma, J. M. Seminario,
P. B. Balbuena, Chem. Rev. 2017, 117, 9674.
the single copper center and are both kinetically and thermo-
dynamically feasible.
This research is partially supported by JSPS KAKENHI
6
a) H. Arakawa, M. Aresta, J. N. Armor, M. A. Barteau, E. J.
(
Grand Number JP 16KT0051).
Beckman, A. T. Bell, J. E. Bercaw, C. Creutz, E. Dinjus, D. A.
Dixon, K. Domen, D. L. DuBois, J. Eckert, E. Fujita, D. H.
Gibson, W. A. Goddard, D. W. Goodman, J. Keller, G. J. Kubas,
H. H. Kung, J. E. Lyons, L. E. Manzer, T. J. Marks, K. Morokuma,
K. M. Nicholas, R. Periana, L. Que, J. Rostrup-Nielson, W. M. H.
Sachtler, L. D. Schmidt, A. Sen, G. A. Somorjai, P. C. Stair, B. R.
Stults, W. Tumas, Chem. Rev. 2001, 101, 953. b) R. S. Haszeldine,
Science 2009, 325, 1647. c) M. Cokoja, C. Bruckmeier, B. Rieger,
W. A. Herrmann, F. E. Kühn, Angew. Chem., Int. Ed. 2011, 50,
8510. d) I. Omae, Coord. Chem. Rev. 2012, 256, 1384.
Supporting Information
Experimental procedures and compound characterization
data. This material is available on http://dx.doi.org/10.1246/
bcsj.20170371.
References
1
a) S. Natarajan, P. Mahata, Chem. Soc. Rev. 2009, 38, 2304.
b) B. Chen, S. Xiang, G. Qian, Acc. Chem. Res. 2010, 43, 1115.
c) H. Furukawa, K. E. Cordova, M. O’Keeffe, O. M. Yaghi,
Science 2013, 341, 1230444. d) H. Furukawa, U. Müller, O. M.
Yaghi, Angew. Chem., Int. Ed. 2015, 54, 3417. e) P. Silva, S. M. F.
Vilela, J. P. C. Tomé, F. A. A. Paz, Chem. Soc. Rev. 2015, 44,
7
a) A. C. Kathalikkattil, R. Babu, J. Tharun, R. Roshan,
D.-W. Park, Catal. Surv. Asia 2015, 19, 223. b) J. W. Maina, C.
Pozo-Gonzalo, L. Kong, J. Schütz, M. Hill, L. F. Dumée, Mater.
Horiz. 2017, 4, 345.
8
S. S.-Y. Chui, S. M.-F. Lo, J. P. H. Charmant, A. G. Orpen,
I. D. Williams, Science 1999, 283, 1148.
E. E. Macias, P. Ratnasamy, M. A. Carreon, Catal. Today
6
774. f) A. Schoedel, M. Li, D. Li, M. O’Keeffe, O. M. Yaghi,
Chem. Rev. 2016, 116, 12466.
a) O. K. Farha, J. T. Hupp, Acc. Chem. Res. 2010, 43, 1166.
9
2
2012, 198, 215.
b) N. Stock, S. Biswas, Chem. Rev. 2012, 112, 933. c) W. Lu, Z.
Wei, Z.-Y. Gu, T.-F. Liu, J. Park, J. Park, J. Tian, M. Zhang, Q.
Zhang, T. Gentle, III, M. Bosch, H.-C. Zhou, Chem. Soc. Rev.
10 a) W.-Y. Gao, L. Wojtas, S. Ma, Chem. Commun. 2014, 50,
5316. b) W.-Y. Gao, Y. Chen, Y. Niu, K. Williams, L. Cash, P. J.
Perez, L. Wojtas, J. Cai, Y.-S. Chen, S. Ma, Angew. Chem., Int. Ed.
2014, 53, 2615. Also, see: c) W. Jiang, J. Yang, Y.-Y. Liu, S.-Y.
Song, J.-F. Ma, Chem.®Eur. J. 2016, 22, 16991.
2
014, 43, 5561.
3
a) J.-R. Li, R. J. Kuppler, H.-C. Zhou, Chem. Soc. Rev.
2
2
009, 38, 1477. b) J.-R. Li, J. Sculley, H.-C. Zhou, Chem. Rev.
012, 112, 869. c) J. B. DeCoste, G. W. Peterson, Chem. Rev. 2014,
11 Y. Liu, Y. Yang, Q. Sun, Z. Wang, B. Huang, Y. Dai, X.
Qin, X. Zhang, ACS Appl. Mater. Interfaces 2013, 5, 7654.
12 a) T. Ohmura, A. Usuki, K. Fukumori, T. Ohta, M. Ito, K.
Tatsumi, Inorg. Chem. 2006, 45, 7988. b) T. Ohmura, A. Usuki, Y.
Mukae, H. Motegi, S. Kajiya, M. Yamamoto, S. Senda, T.
Matsumoto, K. Tatsumi, Chem.®Asian J. 2016, 11, 700. c) T.
Ohmura, N. Setoyama, Y. Mukae, A. Usuki, S. Senda, T.
Matsumoto, K. Tatsumi, CrystEngComm 2017, 19, 5173.
13 V. Caló, A. Nacci, A. Monopoli, A. Fanizzi, Org. Lett.
2002, 4, 2561.
114, 5695. d) E. Barea, C. Montoto, J. A. R. Navarro, Chem. Soc.
Rev. 2014, 43, 5419. e) V. Stavila, A. A. Talin, M. D. Allendorf,
Chem. Soc. Rev. 2014, 43, 5994. f) B. Li, M. Chrzanowski, Y.
Zhang, S. Ma, Coord. Chem. Rev. 2016, 307, 106. g) M. Giménez-
Marqués, T. Hidalgo, C. Serre, P. Horcajada, Coord. Chem. Rev.
2
016, 307, 342. h) L. Wang, Y. Han, X. Feng, J. Zhou, P. Qi, B.
Wang, Coord. Chem. Rev. 2016, 307, 361. i) B. Li, H.-M. Wen, W.
Zhou, J. Q. Xu, B. Chen, Chem 2016, 1, 557. j) I. Nath, J.
Chakraborty, F. Verpoort, Chem. Soc. Rev. 2016, 45, 4127. k) F.-Y.
Yi, D. Chen, M.-K. Wu, L. Han, H.-L. Jiang, ChemPlusChem
14 The reason for the inferior result using the CuTPP/TBAB
system is unclear, but the week coordination of bromide anions to
CuTPP might lower the turnover of the TBAB catalyst.
15 M. Koh, A. Yamauchi, WO2006106655 A1.
2
016, 81, 675. l) C. L. Jones, A. J. Tansell, T. L. Easun, J. Mater.
Chem. A 2016, 4, 6714. m) X. Cao, C. Tan, M. Sindoro, H. Zhang,
Chem. Soc. Rev. 2017, 46, 2660. n) K. Adil, Y. Belmabkhout, R. S.
Pillai, A. Cadiau, P. M. Bhatt, A. H. Assen, G. Maurin, M.
Eddaoudi, Chem. Soc. Rev. 2017, 46, 3402. o) W. P. Lustig, S.
Mukherjee, N. D. Rudd, A. V. Desai, J. Li, S. K. Ghosh, Chem.
1
16 H NMR analysis inferred that the byproducts were derived
from 1-naphthyl ligand as absorption due to a naphthyl moiety was
observed.
17 a) D. J. Darensbourg, Chem. Rev. 2007, 107, 2388. b) T.
Bull. Chem. Soc. Jpn. 2018, 91, 383–390 | doi:10.1246/bcsj.20170371
© 2018 The Chemical Society of Japan | 389

Sakakura, K. Kohno, Chem. Commun. 2009, 1312.
23 Model complex F was located higher in energy than
E + CO2 by +3.9 kcal/mol. This is probably because the ¦G
of CO2 was underestimated. Therefore, the ¦Gs of F­H were
adjusted in such a way that F is located at the same energy level
with E + CO2.
24 a) T. Ema, Y. Miyazaki, J. Shimonishi, C. Maeda, J.
Hasegawa, J. Am. Chem. Soc. 2014, 136, 15270. b) D. Adhikari,
S. T. Nguyen, M.-H. Baik, Chem. Commun. 2014, 50, 2676.
c) K. R. Roshan, A. C. Kathalikkattil, J. Tharun, D. W. Kim, Y. S.
Won, D. W. Park, Dalton Trans. 2014, 43, 2023. d) C. Maeda, J.
Shimonishi, R. Miyazaki, J. Hasegawa, T. Ema, Chem.®Eur. J.
2016, 22, 6556.
1
095.
1
090.
8
T. Ivšić, Z. Hameršak, Tetrahedron: Asymmetry 2009, 20,
H. Kisch, R. Millini, I.-J. Wang, Chem. Ber. 1986, 119,
A. C. Kathalikkattil, D.-W. Kim, J. Tharun, H.-G. Soek, R.
1
1
9
2
0
Roshan, D.-W. Park, Green Chem. 2014, 16, 1607.
a) W. Kohn, A. D. Becke, R. G. Parr, J. Phys. Chem. 1996,
00, 12974. b) P. J. Stephens, F. J. Devlin, C. F. Chabalowski,
2
1
1
M. J. Frisch, J. Phys. Chem. 1994, 98, 11623. c) A. D. Becke,
J. Chem. Phys. 1993, 98, 5648. d) C. Lee, W. Yang, R. G. Parr,
Phys. Rev. B 1988, 37, 785.
2
2
Y. Zhao, D. G. Truhlar, Acc. Chem. Res. 2008, 41, 157.
390 | Bull. Chem. Soc. Jpn. 2018, 91, 383–390 | doi:10.1246/bcsj.20170371
© 2018 The Chemical Society of Japan