Synthesis of cyclic carbonates from epoxides and carbon dioxide catalyzed by an easy-to-handle ionic iron(iii) complex

Article abstract of DOI:10.1039/c3dt32961e

We report the successful utilization of monometallic, ionic iron(ii)- and iron(iii)-N₂O₂-ligand-systems as highly active homogeneous catalysts for the conversion of CO₂ with different epoxides to cyclic carbonates. The cat

Full text of DOI:10.1039/c3dt32961e

Dalton
Transactions
View Article Online
View Journal | View Issue
PAPER
Synthesis of cyclic carbonates from epoxides and
carbon dioxide catalyzed by an easy-to-handle ionic
Cite this: Dalton Trans., 2013, 42, 5322
iron(III) complex†
a
a
b
a
a
a
a
M. A. Fuchs, T. A. Zevaco,* E. Ember, O. Walter, I. Held, E. Dinjus and M. Döring*
Received 10th December 2012,
Accepted 27th January 2013
We report the successful utilization of monometallic, ionic iron(II)– and iron(III)–N
2 2
O –ligand-systems as
highly active homogeneous catalysts for the conversion of CO with different epoxides to cyclic carbo-
2
DOI: 10.1039/c3dt32961e
nates. The catalytic tests were performed using propylene oxide (PO) and a range of nine substituted
epoxides. Terminal monosubstituted oxides react quantitatively.
www.rsc.org/dalton
Introduction
Being the final product of all the energy-producing processes
involving carbon-based substances (e.g. burning fossil- and
biofuels), carbon dioxide is omnipresent in our atmosphere.
Its concentration is continuously increasing since the begin-
ning of the industrial revolution, representing in 2004 77% of
the total anthropogenic greenhouse gases with an emission of
1
3
8 gigatonnes. This nearly endless availability and de facto
“
renewability” together with a low toxicity make it a perfect
2
,3
C1-synthon for organic synthesis.
Scheme 1 Possible products of the reaction of CO2 and epoxides.
In recent years many attempts have been made to use the
thermodynamically stable CO in the synthesis of valuable
2
hand cyclic carbonates and on the other hand aliphatic poly-
carbonates (Scheme 1).
chemicals and several processes could be actually industrial-
ized. Some of them with a promising future like e.g. the syn-
thesis of urethane derivatives, cyclic carbonates and
polycarbonates. Older industrial processes like the synthesis of
hydroxy-benzoic acids, the production of urea and methanol
5
Cyclic carbonates increasingly find an industrial appli-
cation like e.g. their use as non-toxic, polar, high boiling-point
solvents, as electrolytes in lithium ion batteries or as reactive
intermediates (e.g. in ring-opening polymerizations, reactions
still play an important role in the “positive image” of the CO
chemistry.
2
-
6
with amines, alcohols, thiols, and carboxylic acids).
4
For a successful industrial utilization, the conversion of
Epoxides can be seen as high-energy starting materials
which are ideal for engaging a reaction (catalytic or not) with
the inert CO . One class of fully “atom-efficient”, biodegrad-
2
CO and epoxides to cyclic carbonates requires, as a catalytic
2
system, mostly a synergetic combination of a Lewis acid and a
Lewis base. Most of the catalytic systems reported earlier in
the literature needed high temperatures and pressures to
achieve reasonable conversions, limiting in some way the prac-
ticability of the synthesis. However, new promising approaches
published within the last decade showed that high cyclic car-
bonate yields could be reached using mild operating
able products gained from this reaction includes on the one
a
Institut für Katalyseforschung und -Technologie (IKFT), Karlsruher Institut für
Technologie (KIT), Hermann-von-Helmholtz-Platz 1, 76344 Eggenstein-
Leopoldshafen, Germany. E-mail: thomas.zevaco@kit.edu, manfred.doering@kit.edu;
Fax: +49 721 608-22244
7
–10
conditions.
Generally speaking a wide range of compounds
can be used as catalysts, encompassing onium salts (with or
without supplementary metal salts), some ionic liquids,
b
Department Chemie, Lehrstuhl II für Technische Chemie, Technische Universität
München, Lichtenbergstraße 4, 85748 Garching, Germany.
organocatalysts (bases), metal halides, metal oxides and tran-
E-mail: erika.ember@mytum.de; Fax: +49-89-289-13544; Tel: +49-89-289-13545
11,12
sition metal complexes.
For the latter a wide range of
†
Electronic supplementary information (ESI) available. CCDC 906827. For ESI
and crystallographic data in CIF or other electronic format see DOI:
0.1039/c3dt32961e
elements have been successfully tested as catalysts in the
1
“CO
2
-involving” formation of cyclic carbonates. They are
5
322 | Dalton Trans., 2013, 42, 5322–5329
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
ranging from complexes of the “less toxic” zinc and alu-
minium to the more problematic tin, cobalt and
1
1–13
chromium.
The two latter metals are currently the most
promising candidates despite a high toxicity causing an
effective separation of the catalyst to be paramount.
Iron-systems have the advantage of displaying a low intrin-
sic toxicity but also a high catalytic activity in the conversion
1
4–16
of CO2 and epoxides to carbonates.
we investigated the catalytic activity of the iron–N
, 3 and 4 towards different epoxides and CO
In this contribution
2 2
O
-systems
2
2
.
Results and discussion
Synthesis of the catalysts
3
Fig. 1 ESR spectra of complex 3 in the solid state and in CH CN solution
recorded at 123 K.
The synthesis of the catalytic Fe(II)- and Fe(III)-systems 2 and 3
is shown in Scheme 2. The ligand 1 is formed by condensation
of ortho-phenylene diamine and ethyl 2-cyano-3-ethoxyacrylate
in yields of up to 92%. According to a procedure introduced by
two intense bands at λ (ε) = 300 nm (32 739 M− cm−1) and
3
1
max
53 nm (22 272 M− cm−1) which can be tentatively assigned to
π–π* intra-ligand transitions localized predominantly on
1
1
7,18
Jäger
2
ligand 1 reacts with Fe(OAc) in DMF, pyridine being
2
0,21
phenyl ring and conjugated –CN moieties.
In addition the
added afterwards as a weak coordinating ligand.
analysis of powder ESR spectra of complex 3 reveals features
regularly found in the literature for iron(III) complexes even
though a definitive attribution of the Landé g effective factors
The obtained iron(II) complex 2 is air-stable for weeks.
Although it can be taken from elemental analysis that the iron
should be pentacoordinated, it is known from the literature to
be six-fold coordinated via an intermolecular interaction invol-
ving the nitrile substituent of a neighbouring molecule, as
suggested by the splitting of the ν(CN) absorption band in the
2
2
remains complicated.
The X-band spectrum (Fig. 1) displays a strong isotropic
signal at geff = 4.1793 and two broad signals of lower intensity
at geff = 2.1383 and geff = 2.0056. According to the literature
these data suggest the presence in the solid of two iron(III)
complexes with different coordination geometries around the
metal. For instance, the presence of a high-spin Fe -complex
3d , S = 5/2) in a distorted environment would account for the
isotropic signal around g ≈ 4. This coordination geometry
around the iron can be either octahedral (with one weak axial
ligand) or square pyramidal/trigonal bipyramidal. In contrast,
a second low spin Fe (S = 1/2) complex in a more symmetric
environment could be responsible for the broad signal around
g ≈ 2.1. This species would display an octahedrally coordinated
iron(III) with two pyridines in axial positions. The shoulder at
2
3–25
1
7–19
IR spectra.
Compound 2 can be further oxidized under
-system 3.
mild conditions by iodine to afford the iron(III)–N
2 2
O
As many iron(III)-based systems are paramagnetic, reliable
spectroscopic NMR data are not easily available and the struc-
ture of the complexes has to be ascertained via elemental
analysis and MS-ESI. UV-visible-, infrared- and electron spin
resonance spectroscopy provide also complementary infor-
mation on ligand and coordination geometry. For instance
complex 3 forms a deep brown solution in acetonitrile with
3+
5
(
2
5
3
+
g
eff = 2.0056 can be attributed to a low-spin, octahedrally co-
ordinated iron(III) with one pyridine and one nitrile in axial
positions. This hypothesis is further backed by the spectra of
complex 3 in acetonitrile measured at 123 K. The isotropic
signal at geff = 4.1793 together with a second signal of lower
intensity at geff = 2.0056 speaks in favour of two different
coordination geometries around the iron(III) centers. The weak-
ening/broadening of the shoulder at g = 2.0056 supports its
attribution to an iron(III) complex with pyridine and ligand’s
nitrile in axial positions, the competition between nitrile of a
neighbouring ligand and acetonitrile being displaced in favour
of the latter.
Changing the base in favour of N-methylimidazole
(
N-MeIm) yields the related iron(II)–N O /N-MeIm-complex 4.
2 2
The synthesis is shown in Scheme 3.
Single crystals suitable for a structure determination were
obtained from the reaction mixture, by leaving it standing at
Scheme 2 Synthesis of the iron catalysts 2 and 3.
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 5322–5329 | 5323

View Article Online
Paper
Dalton Transactions
Table 2 Selected bond lengths (Å) and angles (°) for 4
Bonds
Distances (Å) and angles (°)
Fe(1)–N(1)
Fe(1)–N(2)
Fe(1)–N(5)
Fe(1)–N(7)
Fe(1)–O(1)
Fe(1)–O(3)
C(4)–N(3)
C(1)–O(1)
C(1)–O(2)
C(2)–C(3)
C(1)–C(2)
C(2)–C(4)
2.1285(17)
2.1070(17)
2.1827(17)
2.2217(17)
2.0647(14)
2.0834(14)
1.149(3)
1.244 (2)
1.346(2)
1.413(3)
1.427(3)
1.423(3)
1.311(3)
1.416(3)
107.23(6)
165.39(6)
87.21(6)
87.17(6)
165.39(6)
78.32(6)
89.84(6)
90.05(6)
92.46(7)
92.51(7)
87.13(6)
87.21(6)
91.42(7)
91.14(7)
175.13(7)
Scheme 3 Synthesis of the iron catalyst 4.
room temperature. After one night small deep red crystals were
formed. Crystals of compound 4 are triclinic, crystallizing in
the space group P ˉ1 (no. 2). (Crystallographic data of the struc- C(3)–N(1)
ture have been deposited at the Cambridge Crystallographic
Database Centre, supplementary publication Nos. CCDC
9
whereas crystallographic data, selected bond lengths and
angles of 4 are listed in Tables 1 and 2, respectively.
C(7)–N(1)
O(1)–Fe(1)–O(3)
O(1)–Fe(1)–N(2)
06827.) The thermal ellipsoid plot of 4 is shown in Fig. 2 O(3)–Fe(1)–N(2)
O(1)–Fe(1)–N(1)
O(3)–Fe(1)–N(1)
N(2)–Fe(1)–N(1)
O(1)–Fe(1)–N(5)
O(3)–Fe(1)–N(5)
N(2)–Fe(1)–N(5)
N(1)–Fe(1)–N(5)
O(1)–Fe(1)–N(7)
O(3)–Fe(1)–N(7)
N(2)–Fe(1)–N(7)
N(1)–Fe(1)–N(7)
N(5)–Fe(1)–N(7)
Complex 4 displays octahedral coordination geometry
around the metal center with two N-methylimidazole mole-
cules in the axial position and the N O –ligand occupying the
2 2
equatorial plane. One molecule of dimethylformamide per
iron complex is present in the molecular structure, showing
weak interactions with the ligands’ nitrile groups as suggested
in Fig. 2 (more details, with intermolecular distances, can be
found in the ESI†). The diverse interactions with the CN
groups have a logical effect on the IR-spectrum, leading to an
overlap in the related CvN stretching band region. Owing to
the fact that compound 4 is coordinatively saturated and that
the DMF–ligand interactions remain weak, the splitting of the
resulting ν(CN) measured in the spectrum is smaller than in
Fig. 2 ORTEP drawings (50% probability) of 4. For clarity, hydrogen atoms
have been removed.
Table 1 Crystallographic data for 4
−
1
the case of the pyridine complex 2, Fe(N O )Py (8 cm vs.
2
2
Chemical formula
Formula mass
C
29
H
35FeN
O
9 5
−1
1
7 cm ).
645.51
Crystal system
a/Å
b/Å
Triclinic
9.6875(7)
10.7398(8)
The overall structure of 4 is comparable to the iron(II)–
2
6
N O /Py complex, Fe(N O )Py *Py, investigated by Jäger et al.
2
2
2
2
2
The iron atom occupies the center of an octahedron, slightly
distorted from the ideal geometry (distortion from the mean
N O -plane: 3.1 pm), the N O ligand being almost planar as
c/Å
α/°
β/°
γ/°
15.2491(11)
100.4530(10)
97.6860(10)
91.2450(10)
1544.5(2)
173(2)
P ˉ1 (No. 2)
2
2
2 2
27
3
in Jäger’s complex and in Fe–salen derivatives. The bond
Unit cell volume/Å
Temperature/K
Space group
lengths found within the N O coordination sphere are within
2 2
a small range, the Fe–O bonds being a tad shorter than the
Fe–N ones: Fe(1)–N(1) and Fe(1)–N(2) with 2.1285(17) Å and
2.1070(17) Å while Fe(1)–O(1) and Fe(1)–O(3) amount to
No. of formula units per unit cell, Z
No. of reflections measured
No. of independent reflections
2
23 122
7165
R
int
0.0294
0.0406
0.1094
0.0501
0.1159
2
.0647(14) Å and 2.0834(14) Å respectively. As expected from
Final R
Final wR(F ) values (I > 2σ(I))
Final R values (all data)
Final wR(F ) values (all data)
1
values (I > 2σ(I))
2
the literature, the Fe–N apical bonds are, with Fe(1)–N(7),
2.2217 Å, and Fe(1)–N(2), 2.18127 Å, longer than the equatorial
ones and similar to those found in the connected homoleptic
1
2
5
324 | Dalton Trans., 2013, 42, 5322–5329
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
2
8
36,37
complex [Fe(NMi) ][PF ]. The four (NO)–Fe angles within the onium salts are on their own active towards this reaction.
6
6
plane suggest an influence of the ethyl ester groups on the A synergetic effect can be seen considering the reaction with a
coordination environment at the metal center. The angle O(3)– combined catalyst and a co-catalyst (entry 3 + 4). 0.2 mol% of
Fe–O(1) which involves both ester groups is the widest with the catalyst–co-catalyst mixture showed a yield of 37% cyclic
1
07.2° whereas the opposite angle, N(1)–Fe–N(2), can be seen carbonate (entry 3) and 1 mol% a conversion of even 91%
as compressed with only 78.3°. (entry 4). After oxidation with iodine, the iron(III) complex 3
The two remaining N–Fe–O angles are similar, around displays a much higher activity than the iron catalyst in the
7.2°. The angle involving iron and the two axial nitrogens of lower oxidation state 2 (entries 5 and 6). The reasons for
8
the imidazoles amounts to 175.13(7)°, towards values found in the observed increased reactivity leading to a quantitative con-
N4–iron/N-methylimidazole systems like e.g. iron porphyrins version of propylene oxide are manifold: (i) two pyridine
2
9
or phthalocyanins, whereas the Jäger’s Fe–pyridine complex molecules in the coordination sphere of the complex are able
displays a N–Fe–N angle of 170.31°. The interatomic distances to act as co-catalysts, (ii) a higher Lewis acidity of the
3
0
measured within this ligand class illustrate the strong inter- metal centre and (iii) the ionic character of complex 3 which
action with the metal center: as an example, all the C–C dis- comprises an iodide anion also able to interact with an
tances involving C(2) are in a narrow range (around 1.42 Å) epoxide. It can be easily seen that within complex 3, all the
while the C–N involving the coordinating nitrogens (e.g. catalytic components needed for a successful insertion of CO
C3–N1) are significantly shorter than the standard C–N bonds into an epoxidic ring are present and that no supplementary
2
3
1
found in amines (C(3)–N(1): 1.311(3) Å instead of 1.45 Å).
Unfortunately attempts to oxidize complex 4 with iodine in anchored to the ligand is needed.
order to isolate an iron(III)(N-MeIm) complex analogue to 3, optimal reaction conditions, different reaction temperatures
also with solvents other than DMF (e.g. THF and CHCl ), did and catalyst concentrations were tested. At 50 °C with
co-catalyst or the challenging synthesis of a co-catalyst
3
8–40
In order to find the
3
not succeed, delivering only complex mixtures. Likewise, 0.2 mol% of 3 the reaction only produces 76% of propylene
trying to oxidize compound 3 by bubbling air through the carbonate (entry 7). The reaction at 100 °C is not that effective
solution and vigorous stirring delivered no clear results.
as at 80 °C with a yield of 96%, caused by a slow degradation
of the catalyst, which makes catalyst recycling problematic
(
entry 8). At 80 °C the reaction can be easily performed to
Catalytic tests with propylene oxide
deliver high yields of cyclic carbonates permitting potential
The catalytic screening of iron(II) complex 2 and iron(III) com- catalyst recycling.
plexes 3 and 4 involving propylene oxide and carbon dioxide
Lowering the catalyst concentration to only 0.1 mol% leads
was performed in 70 ml autoclaves under solvent-free con- to a lower yield of propylene carbonate of 69% but logically
ditions. Standard reaction time was 20 h. The results are sum- rises the TON up to 690 (entry 9). Best conversion was achieved
3
2–34
marized in Table 3. In agreement with similar studies,
the with a catalyst concentration of 0.2 mol% and a temperature of
1
conversions were directly determined using H-NMR spec- only 80 °C (entry 6).
troscopy with an internal standard, a swift and reliable analyti-
Low-pressure experiments were carried out in a 70 ml auto-
cal method for such samples if one considers the sometimes clave continuously connected to a gas cylinder via a back-
tedious separation of the final cyclic carbonate (e.g. a complex pressure regulating valve, pressing 2 bar of CO into the reac-
2
3
4,35
distillation step).
tion mixture. After heating the reaction mixture for 20 h, 36%
Catalyst 2 alone showed no reactivity towards the CO con- of the propylene oxide was converted to propylene carbonate
2
version (entry 2). The co-catalyst TBAB alone showed a conver- (entry 10).
sion of 35% (entry 1), which is not surprising knowing that
Complex 4 as iron(II) complex shows no catalytic activity
without the addition of a cocatalyst, as already shown with
complex 2 (entries 2 and 11). Only the addition of a cocatalyst
a
2
Table 3 CO /PO coupling catalysed by iron-systems
(
in this case again TBAB) leads to a conversion to propylene
b
c
Entry
Catalyst
Mol%
p/bar
T/°C
% Yield
TON
carbonate with a satisfactory yield (75%) (entry 12). The higher
PC yield observed with compound 4 and TBAB at low catalyst
loadings (entries 12 and 3) can be tentatively explained by
the higher nucleophilicity of N-methylimidazole (compared to
pyridine) and/or an active role in the ring opening of the
epoxide.
The reactions with propylene oxide and 3 proceed fast. In
Fig. 3 a typical diagram displaying temperature and pressure
against reaction time is represented. The dashed line shows
the reaction with catalyst 3, the continuous one shows the reac-
tion with 2 and TBAB as a co-catalyst. The reaction in the pres-
ence of 2 needs 20 h to reach a comparable conversion to the
reaction with 3 after 3 hours.
1
2
3
4
5
6
7
8
9
TBAB
2
2 TBAB
2 TBAB
3
3
3
3
1
1
0.2
1
0.5
0.2
0.2
0.2
0.1
0.2
0.2
0.2
1
50
50
35
50
50
50
50
50
35
2
80
80
80
80
80
80
50
100
80
80
80
80
80
35
—
37
91
99
99
76
96
69
36
—
75
94
35
—
185
91
198
495
380
480
690
180
—
3
3
4
10
11
12
13
50
50
50
4 TBAB
4 TBAB
375
94
a
b
c
Reaction time 20 h. Yields by NMR. TON = molPC/molKAT
.
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 5322–5329 | 5325

View Article Online
Paper
Dalton Transactions
Fig. 4 Yields in recycling experiments of catalyst 3 (0.2 mol%) with propylene
Fig. 3 Comparison of the recorded pressures during the reaction of propylene
oxide at 80 °C and 50 bar.
2
oxide and CO with the catalysts 2 (Fe(II) + TBAB) and 3 (Fe(III)).
Tests with other epoxides
Compared to other iron catalysts in the literature, the
amount of iron used in our conversions is lower. Moreover, the
examples from the literature used co-catalysts like tetrabutyl-
ammonium bromide (TBAB), tetrabutylammonium iodide
To point out the broad applicability of the ionic catalyst 3, we
performed a broader screening using a range of commercially
available epoxides. The results of the cyclization with CO , ran
2
under an optimized reaction temperature of 80 °C and a cata-
lyst concentration of 0.2 mol%, are listed in Table 4. Yields
were estimated by H NMR using an internal standard (tetra-
(
(
TBAI)
and
bis(triphenylphosphine)-iminium
chloride
PPNCl), which are not necessary with our most effective cata-
1
lyst system 3. Related studies with iron catalysts were pub-
lished by Rieger and co-workers (quantitative conversion of
PO: 1 mol% Fe(II)-catalyst, TBAB as a co-catalyst, 100 °C, 15 bar
1
5
chloroethane and, if not possible, chloroform) to quantify the
catalytic activity. Under the selected conditions no conversion
was found using epoxides with more than one bulky substitu-
ent like 1,2-epoxy-2-methylpropane, 2,3-epoxy-2-methylbutane
and cyclohexene oxide (entries 8–10). But also a sterically more
hindered mono-substituted epoxide like 3,3-dimethyl-1,2-
epoxybutane showed no conversion to the corresponding
cyclic carbonate (entry 7). This suggests either an unfavourable
competition between coordination of the epoxide and de-
coordination of a pyridine molecule at the iron centre or, if a
coordination indeed takes place, an ineffective nucleophile
attack of the iodide anion at the epoxide. The mechanism of
the reaction is under further study and details on its exact
nature together with kinetic data will be reported in due
course.
1
4
CO
2
, 2 h), the research group of Williams (91% conversion of
PO: 1 mol% iron in Fe(II)-catalyst, 1 mol% PPNCl, 25 °C, 1 atm
CO , 48 h) and Kleij et al. (96% conversion of PO: 1 mol%
iron in Fe(III)-catalyst, 5 mol% TBAI, 25 °C, 2 bar CO , 18 h).
2
Both catalysts show a very high selectivity in the formation
of cyclic carbonates. IR analysis of the products displayed
1
6
2
neither
ν
as(CvO) vibration typical for polycarbonate
)
−
1 41
(1752 cm
nor a massive stretching ν(C–O) band (around
050 cm− ) typical for polyethers, potential products of a ring
1
1
opening polymerisation of the epoxide. Absorption bands
related to potential traces of 1,2-diols as hydrolysis’ products
(
−
1
−1
ν(OH) and δ(OH) at 3600–3650 cm and around 1600 cm
4
2
respectively) were not observed.
The other monosubstituted epoxides showed very high
yields of their cyclic carbonates, 1,2-epoxyhexane and 2,3-
epoxypropyl benzene could be for instance quantitatively con-
verted (entry 2 + 6). The other mono-substituted epoxides like
styrene oxide, epichlorhydrine, 2,3-epoxypropyl phenyl ether
and allyl glycidyl ether delivered also the desired cyclic carbon-
ates in nearly quantitative yields (entries 1, 3–5).
Time resolved in situ IR spectroscopy, ran with a Mettler
Toledo React-IR 45m, supports that the formation of cyclic car-
bonate from propylene oxide using our catalysts does not
4
3,44
involve the degradation of a prior formed polycarbonate.
To check the possibility of recycling catalyst 3, five succes-
sive runs with propylene oxide were started. After every exper-
iment the resulting propylene carbonate was distilled and the
residue catalyst was collected and reused in the next batch fol-
2
lowing the standard procedure (80 °C, 50 bar CO , 20 h, cata-
lyst concentration 0.2 mol%). The catalyst loss was evened up
by proportionally lowering the added amount of propylene
Conclusion
oxide in order to maintain the catalyst-to-epoxide ratio We reported a highly selective iron-based, easy-to-handle cata-
1 : 500).
lytic system for the conversion of epoxides with CO to cyclic
Fig. 4 shows the yields per experiment. The first use of cata- carbonates. Based on a ligand system reported by Jäger et al.
(
2
lyst 3 is the most successful one, with a yield of 99%. The fol- in the early eighties, we were able to synthesize iron(II) and
lowing batches display yields between 78% and 91%, showing iron(III) complexes on a multi-gram scale from commercially
that 3 can be reasonably recycled after removing the reaction available starting materials (Fe(III) only with pyridine as a
product.
ligand). The three complexes were firstly tested for the
5
326 | Dalton Trans., 2013, 42, 5322–5329
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
a
Table 4 CO
2
/epoxide coupling catalysed by iron-system 3
with a Varian Inova instrument equipped with an Oxford
magnet (9.4 T, 400 MHz for H). IR spectra were obtained with
1
b
c
Entry Epoxide
1
Product
p/bar % Conversion
TON
a Varian 660-IR FT-IR Spectrometer. Elementary analysis was
performed with a CHNS-Analyser from the company Elemen-
tar. The ESR measurements were recorded in perpendicular
mode on an X-band Joel Jes Fa 200 spectrometer equipped
with a cylindrical mode cavity and a liquid helium cryostat.
The ESR measurements were performed in quartz tubes at
50
96
480
2
3
50
35
100
98
500
490
123 K (9.45 GHz, 1 mW microwave power). The analysis and
simulation of the ESR data was done with the Jes-Fa Series
software package (Version 2.2.0). Mass spectra were measured
with a Bruker ApexQe hybrid 9.4 T FT-ICR. UV-Vis spectra were
measured with an MCS 501 UV-NIR (Zeiss). The X-ray analyses
were performed on a Bruker Apex II Quazar diffractometer.
2124 frames were collected with an irradiation time of 1 s per
frame. Integration of the data proceeded with SAINT, the data
were corrected for Lorentz- and polarisation effects, and an
experimental absorption correction with sadabs was per-
4
50
99
495
5
6
35
50
95
475
500
100
4
5
formed. For searches relating to single-crystal X-ray diffrac-
4
6
7
8
50
50
—
—
—
—
tion data, the Cambridge Structural Database was used.
Crystallographic data have been deposited at the Cambridge
Crystallographic Database Centre (CCDC), supplementary pub-
lications No. CCDC 906827. Figures were prepared with the
appropriate software of the CCDC, MERCURY 3.0 (Build RC5)
4
7
for Windows . Ligand 1 and the iron(II) catalyst 2 were syn-
9
50
50
—
—
—
—
1
7
thesized according to a literature procedure.
Fe(III) complex 3. Complex 2 (3.0 g, 6.16 mmol) was dis-
solved in 30 ml of pyridine. Afterwards iodine (780 mg,
1
0
3.07 mmol) was added and the solution was stirred for 48 h at
room temperature. The solvent was removed under reduced
pressure leaving 3 as a dark brown solid (4.24 g, 99%). IR
(
KBr): 1611 cm− (ν(CvC), ν(CvN)), 2205 cm (ν(CuN)). MS:
1
−1
a
Reaction conditions: catalyst loading 0.2 mol%, reaction time 20 h,
+
+
b
1
(ESI+) m/z (%): 408.1 (58) [M-2Py-I] , 487.1 (100) [M-Py-I] ,
816.1 (31) [2(M-2Py-I)] , 847.1 (63) [2(M-2Py-I) OMe] . HR-MS:
ESI+) m/z (%) calcd for C H FeN O : 408.0516, found:
08.0516 (58); calcd for
reaction temperature 80 °C. Conversion by H-NMR with an internal
standard. TON = molcycl.carb/molKAT
+
+
+
c
.
+
(
4
1
8
16
4 4
+
C
19
H
19FeN
4
O
5
Na : 462.0597,
+
conversion of propylene oxide. The iron(II) complexes 2 and 4
need a rather high catalyst concentration and the addition of a
co-catalyst to fulfil the reaction. The more active iron(III)
system 3 demonstrates a high activity without co-catalysts in a
short reaction time. The screening tests were extended to a
wider range of epoxides and showed the very high catalytic
activity of this system towards terminal, mono-substituted
epoxides and CO . Kinetics of the reaction and spectroscopic
2
investigation of the mechanism are the topic of ongoing
studies.
found: 462.0599 (29); calcd for C23
H
21FeN
5
O
4
: 487.0938,
found: 487.0939 (100); calcd for C H Fe N O : 816.1038,
3
6
32
2 8 8
found: 816.1041 (31); calcd for C37
H
35Fe
2
N
8
O
9
: 847.1222,
found: 847.1221 (63). Anal. calcd for C28
H
26FeIN
6
O
4
: C, 48.51;
H, 3.78; N, 12.12, found: C, 49.11; H, 3.95; N, 12.61, UV/Vis
−
1
(
acetonitrile): λmax (ε) = 300 (32 739), 353 nm (22 272 mol
3
−1
dm cm ).
Fe(II) complex 4. The ligand 1 (3.12 g, 8.81 mmol) was dis-
solved in 35 ml DMF and Fe(OAc) (1.53 mg, 8.81 mmol) was
2
added. The reaction mixture was heated to 75 °C and stirred
for half an hour. Afterwards N-methylimidazole (1.45 g,
1
7
7.62 mmol) was added and the reaction mixture was stirred at
5 °C for 1 h. Then it was cooled to room temperature and left
Experimental section
General
standing without stirring overnight. Deep red crystals were col-
lected by filtration and dried (4.09 g, 6.33 mmol, 72%).
All reactions were performed under argon using standard IR (KBr): 1612 cm− (ν(CvC), ν(CvN)), 2196 cm , 2187 cm
1
−1
−1
Schlenk techniques. The chemicals were purchased from (ν(CuN)). MS: (ESI+) m/z (%): 408.0 (10) [M-2N-MeIm], 409.0
+
Sigma Aldrich except Fe(OAc)
2
(abcr), pyridine (AppliChem) (9) [M-2N-MeIm + H] , 490.0 (8) [M-N-MeIm]. HR-MS: (ESI+)
16FeN : 408.05156, found: 408.0518
and iodine (Roth). Propylene oxide and pyridine were dried m/z (%) calcd for C18
H
4 4
O
1
over CaH2 before their use. H-NMR spectra were obtained (3); calcd for C H FeN O Na: 431.0413, found: 431.0417 (13);
18
16
4
4
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 5322–5329 | 5327

View Article Online
Paper
Dalton Transactions
calcd for C19
H
19FeN
4
O
5
Na: 462.0597, found: 462.0601 (12); This vessel is equipped with a “sidewise-attached” diamond
calcd for C H FeN O : 490.1046, found: 490.1049 (100); calcd probe sensor (DiComp™ window) and coupled via a Sentinel™
2
2
22
6 4
for C36
for C2
H, 5.606; N, 19.37.
H
32Fe
35FeN
2 8 8
N O : 816.1036, found: 816.1044 (22). Anal. calcd K4 conduit to a ReactIR 45m from Mettler-Toledo. The reaction
9
H
9
O
5
: C, 53.96; H, 5.47; N, 19.53, found: C, 53.81; was monitored with the dedicated software iCIR (version 4.2).
The autoclave was flushed with argon. Afterwards 10 ml of
dichloromethane and 40 ml of propylene oxide were added.
The autoclave was closed and heated to 80 °C. A cartridge was
Autoclave tests
The tests were performed in a screening test-bench consisting installed at one of the autoclave’s accesses and also flushed
of 70 ml autoclaves (mat. 1.4571 (SS 316Ti), 70 ml, pmax 200 with argon. Then the catalyst was dissolved in 15 ml of
bar, Tmax 250 °C) equipped with magnetic stirrers and P, dichloromethane, and injected into the cartridge. When the
T-Acquisition via multi-meters (Agilent 34970A Data Acqui- reaction mixture reached the desired temperature, the back-
sition/Data Logger Switch Unit + 34901A 20 Channel Multi- ground for the measurement was collected. Afterwards the
plexer) and PCs. The heating was performed with an measurement was started and the catalyst solution was
aluminium heating block controlled via the electronic contact pressed into the reaction mixture with 50 bar of CO2.
thermometer of the magnetic stirrers. The dried autoclaves
were filled with the catalyst and set under vacuum for 1 h.
Afterwards the epoxide was added under argon via one of the
autoclave’s ports and the reaction mixture was pressurized
Acknowledgements
2
with CO . This procedure was chosen owing to the low reacti- The authors would like to thank the BMBF-Project “Dream
vity of complexes 2, 3 and 4 with epoxides alone (no ring Reactions – Stoffliche CO -Verwertung” (Förderkennzeichen
2
opening polymerisation of the epoxides was observed under 01RC0901D) and the Helmholtz Research School “Energy-
the chosen experimental conditions). With all epoxides except Related Catalysis” for financial support. Likewise the technical
propylene oxide the autoclave was cooled to 0 °C during the support of Gilbert Zwick (MS) and Tina Klinger (synthesis and
CO
2
loading ensuring that enough CO
2
was available for the tests) is greatly acknowledged.
reaction (CO
2
/epoxide molar ratio around 3). Once the auto-
clave was heated, pressure and temperature were recorded
every 30 seconds during the whole reaction time. After the
reaction the autoclaves were cooled down, slowly vented in a
fume hood, opened and the products collected. The analysis
Notes and references
1 IPCC 2007: Climate Change 2007: Synthesis Report. Contri-
bution of Working Groups I, II and III to the Fourth Assess-
ment Report of the Intergovernmental Panel on Climate
Change, ed. R. K. Pachauri and A. Reisinger, IPCC, Geneva,
Switzerland, 1961.
2 M. Peters, T. Müller and W. Leitner, TCE Today (Chem.
Eng.), 2009, 813, 46.
3 N. A. M. Razali, K. T. Lee, S. Bhatia and A. R. Mohamed,
Renew. Sust. Energ. Rev., 2012, 16, 4951.
1
was done via H-NMR and FT-IR spectroscopy. The reproduci-
bility of a typical reaction was assessed via performing numer-
ous tests with complex 3 and PO (2–3 runs for each entry/
entries 5–10) and found an overall very good reproducibility.
The rest of the runs, performed with the other epoxidic sub-
strates and catalysts, are single experiments performed within
the same screening campaign.
Low-pressure experiment
4
T. Sakakura, J.-C. Choi and H. Yasuda, Chem. Rev., 2007,
107, 2365.
A first low pressure screening test was performed in a standard
glass equipment with two high performance reflux condensers
at −20 °C, partly filled with metallic Raschig rings to enhance
the cooling efficiency, with moderate success. Hence the low-
pressure experiments were carried out in a 70 ml autoclave per-
manently connected during the reaction to a gas cylinder
5 X.-B. Lu and D. J. Darensbourg, Chem. Soc. Rev., 2012, 41,
1462.
6 J. H. Clements, Ind. Eng. Chem. Res., 2003, 42, 663.
7 X.-B. Lu, X.-J. Feng and R. He, Appl. Catal., A, 2002, 234, 25.
8 T. Chang, H. Jing, L. Jin and W. Qiu, J. Mol. Catal. A: Chem.,
2007, 264, 241.
2
pressing 2 bar CO into the mixture. The gaseous propylene
oxide is kept in the autoclave by means of a back pressure
regulator. The dried autoclave is operated the same way as for
the high pressure experiments: filled with catalyst 3 and propy-
9 X.-B. Lu, B. Liang, Y.-J. Zhang, Y.-Z. Tian, Y.-M. Wang,
C.-X. Bai, H. Wang and R. Zhang, J. Am. Chem. Soc., 2004,
126, 3732.
2
lene oxide under an argon atmosphere, at 2 bar CO pressur- 10 X. Zhang, Y.-B. Jia, X.-B. Lu, B. Li, H. Wanga and L.-C. Sun,
ized and heated for the same reaction time. Afterwards the
products were analyzed by H-NMR spectroscopy.
Tetrahedron Lett., 2008, 49, 6589.
11 M. North, R. Pasquale and C. Young, Green Chem., 2010,
1
1
2, 1514.
In situ IR experiments
1
2 A. Decortes, A. M. Castilla and A. W. Kleij, Angew. Chem.,
The experiments were carried out in a “home-made” stainless
Int. Ed., 2010, 49, 9822.
steel 316Ti autoclave (mat. 1.4571, 150 ml, pmax 300 bar, Tmax 13 T. A. Zevaco, A. Janssen and E. Dinjus, ARKIVOC, 2007, 3,
2
00 °C, magnetic stirring, aluminium heating block with TIC).
151.
5
328 | Dalton Trans., 2013, 42, 5322–5329
This journal is © The Royal Society of Chemistry 2013

View Article Online
Dalton Transactions
Paper
1
1
1
1
1
1
2
2
2
2
2
2
2
2
4 A. Buchard, M. R. Kember, K. G. Sandeman and
C. K. Williams, Chem. Commun., 2011, 47, 212.
5 J. E. Dengler, M. W. Lehenmeier, S. Klaus, C. E. Anderson,
E. Herdtweck and B. Rieger, Eur. J. Inorg. Chem., 2011, 336.
A. Chaudhary, S. K. Ghosh and S. P. Rath, Inorg. Chem.,
2010, 49, 2057; CCDC-Name: ASESOH; L. A. Yatsunyk,
M. D. Carducci and F. A. Walker, J. Am. Chem. Soc., 2003,
125, 15986; CCDC-Name: AQAKEI.
6 C. J. Whiteoak, E. Martin, M. M. Belmonte, J. Benet- 30 B. Weber, E. Kaps, J. Obel and W. Bauer, Z. Anorg. Allg.
Buchholz and A. W. Kleij, Adv. Synth. Catal., 2012, 354, 469. Chem., 2008, 634, 1421; CCDC-Name: SIZVIH.
7 E.-G. Jäger, E. Häussler, M. Rudolph and A. Schneider, 31 F. H. Allen, O. Kennard, D. G. Watson, L. Brammer and
Z. Anorg. Allg. Chem., 1985, 525, 67. A. G. Orpen, J. Chem. Soc., Perkin Trans. 2, 1987, 1.
8 E.-G. Jäger, B. Kirchhof, E. Schmidt, B. Remde, A. Kipke 32 W. Clegg, R. W. Harrington, M. North and R. Pasquale,
and R. Müller, Z. Anorg. Allg. Chem., 1982, 485, 141. Chem.–Eur. J., 2010, 16, 6828.
9 M. Kröger, C. Folli, O. Walter and M. Döring, Adv. Synth. 33 B. R. Buckley, A. P. Patel and K. G. U. Wijayantha, Chem.
Catal., 2006, 348, 1908. Commun., 2011, 47, 11888.
0 J. Sima, L. Horvathova and M. Izakovic, Monatsh. Chem., 34 G. F. Pauli, B. U. Jaki and D. C. Lankin, J. Nat. Prod., 2005,
004, 135, 5. 68, 133.
1 A. A. A. Abou-Hussein and W. Linert, Spectrochim. Acta, Part 35 S. K. Bharti and R. Roy, Trends Anal. Chem., 2012, 35, 5.
2
A, 2012, 95, 596.
2 A. C. Vajpei, A. Rousset, N. Favillier and G. Ablart, J. Mater.
Sci. Lett., 1985, 4, 1403.
3 T. Beissel, K.-S. Bürger, G. Voigt, K. Wieghardt, C. Butzlaff
and A. X. Trautwein, Inorg. Chem., 1993, 32, 124.
4 M. Braun, I. García Rubio and L. Thöny-Meyer, Appl. Micro-
biol. Biotechnol., 2005, 67, 234.
5 P. Das, P. P. Sarmah, M. Borah and A. K. Phukan, Inorg.
Chim. Acta, 2009, 362, 5001.
36 V. Caló, A. Nacci, A. Monopoli and A. Fanizzi, Org. Lett.,
2002, 4, 2561.
37 H.-Y. Ju, M.-D. Manju, K.-H. Kim, S.-W. Park and
D.-W. Park, J. Ind. Eng. Chem., 2008, 14, 157.
38 K. Nakano, T. Kamada and K. Nozaki, Angew. Chem., Int.
Ed., 2006, 45, 7274.
39 E. K. Noh, S. J. Na, S.-W. Kim and B. Y. Lee, J. Am. Chem.
Soc., 2007, 129, 8082.
40 X.-B. Lu, L. Shi, Y.-M. Wang, R. Zhang, Y.-J. Zhang,
X.-J. Peng, Z.-C. Zhang and B. Li, J. Am. Chem. Soc., 2006,
128, 1664.
6 H. Görls and E.-G. Jäger, Cryst. Res. Technol., 1991, 26, 349;
CCDC-Refcode: JIRLUR.
7 M. M. Bhadbhade and D. Srinivas, Polyhedron, 1998, 17, 41 D. J. Darensbourg, J. C. Yarbrough, C. Ortiz and C. C. Fang,
699; CCDC-Name: HOGQEZ; B. J. Kennedy, A. C. McGrath, J. Am. Chem. Soc., 2003, 125, 7586.
K. S. Murray, B. W. Skelton and A. H. White, Inorg. Chem., 42 L. Wang, W. Xu, A. F. Khalizov, J. Zheng, C. Qiu and
987, 26, 483; CCDC-Name: FESYAD01, CCDC-Name: FESYIL. R. Zhang, J. Phys. Chem. A, 2011, 115, 8940.
2
1
2
8 D. S. Jacob, V. Kahlenberg, K. Wurst, L. A. Solovyov, 43 Spectrumprovided in the ESI.†
I. Felner, L. Shimon, H. E. Gottlieb and A. Gedanken, 44 D. J. Darensbourg, J. L. Rodgers, R. M. Mackiewicz and
Eur. J. Inorg. Chem., 2005, 522; CCDC-Name: FIVQUI; A. L. Phelps, Catal. Today, 2004, 98, 485.
T. B. Higgins, M. K. Safo and W. R. Scheidt, Inorg. Chim. 45 Bruker, Bruker AXS Inc., Madison, Wisconsin, USA, 2007;
Acta, 1990, 178, 261; CCDC-Name: SIYRUN; R. Patra,
A. Chaudhary, S. K. Ghosh and S. P. Rath, Inorg. Chem.,
G. M. Sheldrick, Acta Crystallogr., Sect. A: Fundam. Crystal-
logr. A, 2008, 64, 112; L. Zsolnai, XPMA, University of
Heidelberg, Germany, 1996.
2
010, 49, 2057; CCDC-Name: ASESOH; A. Yatsunyk,
M. D. Carducci and F. A. Walker, J. Am. Chem. Soc., 2003, 46 C. R. Groom and F. H. Allen, Wiley Interdiscip. Rev.:
25, 15986; CCDC-Name: AQAKEI. Comput. Mol. Sci., 2011, 1, 368.
9 T. B. Higgins, M. K. Safo and W. R. Scheidt, Inorg. Chim. 47 MERCURY 3.0 (Build RC5); http://www.ccdc.cam.ac.uk/
Acta, 1990, 178, 261; CCDC-Name: SIYRUN; R. Patra, mercury
1
2
This journal is © The Royal Society of Chemistry 2013
Dalton Trans., 2013, 42, 5322–5329 | 5329