New cobalt, iron and chromium catalysts based on easy-to-handle N 4-chelating ligands for the coupling reaction of epoxides with CO2

Article abstract of DOI:10.1039/c3dt53084a

In this contribution, we report the successful utilization of several transition metal complexes based on substituted N₄-N,N-bis(2- pyridinecarboxamide)-1,2-benzene chelating ligands as catalysts in the coupling of epoxides with carbon dioxide.

Full text of DOI:10.1039/c3dt53084a

Dalton
Transactions
View Article Online
View Journal | View Issue
PAPER
New cobalt, iron and chromium catalysts based on
easy-to-handle N4-chelating ligands for the
coupling reaction of epoxides with CO2†
Cite this: Dalton Trans., 2014, 43,
3285
M. Adolph, T. A. Zevaco,* C. Altesleben, O. Walter and E. Dinjus
In this contribution, we report the successful utilization of several transition metal complexes based on
substituted N4-N,N-bis(2-pyridinecarboxamide)-1,2-benzene chelating ligands as catalysts in the coup-
ling of epoxides with carbon dioxide. The complexes were tested towards cyclohexene oxide and pro-
pylene oxide. Additionally the recyclability of the catalytic system was evaluated and a broader catalytic
screening involving several commercially available epoxides was carried out with selected catalysts.
Received 31st October 2013,
Accepted 9th December 2013
DOI: 10.1039/c3dt53084a
www.rsc.org/dalton
generates cyclic carbonates – the thermodynamic products – as
well as aliphatic polycarbonates – the kinetic products – is
Introduction
The quest for renewable resources and sustainable systems has attracting growing interest from the “CO2-community” mainly
a foreground position in the contemporary economy and plays due to a promising range of industrial applications.2 Aliphatic
an ever-increasing role in different sectors of modern chem- polycarbonates formed along this synthesis route are an
istry, as well in academia and in industry, particularly if one important addition to the common aromatic polycarbonates
considers the fact that mineral oil-based resources will most owing to, amongst others, a higher intrinsic biodegradability3
probably be exhausted within the coming century. One de facto (e.g. polypropylene carbonate). On the other hand, cyclic car-
renewable resource of major interest is carbon dioxide. Carbon bonates are industrially used as electrolytes in Li-ion cells, as
dioxide1 is indisputably abundant, easy-to-handle, and non- polar, non-toxic solvents in cleaning processes or as substrates
toxic and is, once the problem of its thermodynamic stability for the synthesis of high value-added molecules.4
is solved, an attractive raw material for organic syntheses. The
Since the first catalytic system for the copolymerization of
coupling of epoxides (Scheme 1) with carbon dioxide which CO2 with epoxides described by Inoue et al.5 in 1969, varieties
of heterogeneous or homogeneous catalytic systems have been
reported.6–8 These well-documented studies have in common
that, in the case of homogeneous catalysts, three complemen-
tary features are required for a successful reaction: an acidic
metal centre able to coordinate the epoxide, a chelating ligand
with suitable geometry and a leaving group able to easily
form a reactive metal–alkoxide bond. Only this bond is able to
interact easily with CO2 and keep the activation–insertion
steps running. A wide range of metal/leaving group/ligand con-
figurations has been tested and many transition metal com-
plexes displaying these valuable characteristics were reviewed
in recent years.9–11 Catalysts involving zinc(II), aluminum(III)
and cobalt(III) seem to be currently the most promising candi-
dates with one particular trend as far as the ligand design is
concerned: the grafting of ionic moieties (ammonium or phos-
phonium) on the ligand backbone able to perform as co-
catalysts.
Perusing the literature dealing with “renowned” chelating
ligands like, e.g., salen (N2O2) investigated a.o. by Coates,12
tetra-phenylporphyrin (N4) pioneered by Inoue,13 tetramethyl-
tetraazaannulene (N4) used by Darensbourg14 and salan (N2O2,
Scheme 1 Possible products of the reaction of CO2 with epoxides.
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, eckhard.dinjus@kit.edu;
Fax: +49 721 608-22244; Tel: +49 721 608-24385
†Electronic supplementary information (ESI) available. CCDC 958114, 958115
and 958116. For ESI and crystallographic data in CIF or other electronic format
see DOI: 10.1039/c3dt53084a
This journal is © The Royal Society of Chemistry 2014
Dalton Trans., 2014, 43, 3285–3296 | 3285

View Article Online
Paper
Dalton Transactions
salen’s reduced form) by Rieger,15 some trends can be identi- were measured using a Merck gel permeation chromatograph
fied: on the one hand, a relatively planar geometry at the metal (Lichograph Gradient pump L-6200 with thermostat, LaChrom
center guaranteed by an adequate ligand design. On the other RI detector L-7490) equipped with a pre-column and two
hand, a good interplay between electron-donating/withdrawing different columns (PSS SDV 5 m, 1000 and 100 Å). Toluene was
capabilities of the ligand and Lewis acidity of the metal ion. used as an eluent, and calibration was performed using poly-
These properties together with a good solubility of the metal styrene standards. Elementary analysis was performed using a
complex in the epoxide–CO2 mixture go along with the pre- CHN-Analyser Vario EL III of the company Elementar. The
requisites of a copolymerization reaction. As is usual with X-ray analyses were performed using a Bruker Apex II Quazar
many efficient ligand systems, the tuning of the backbone diffractometer. Integration of the data proceeded with SAINT,
ligand should remain easy, in the same way as the up-scaling the data were corrected for Lorentz- and polarisation effects,
of the ligand synthesis. Additionally the tuning of the ligand and an experimental absorption correction with SADABS was
framework plays a significant role in the heterogenization of performed.24 For searches relating to single-crystal X-ray diffr-
the catalysts, the presence of reactive groups enabling a further action data, the Cambridge Structural Database was used.25
grafting of an efficient homogeneous catalyst to an organic Figures were prepared with the appropriate software of the
(e.g. merrifield polymer16) or inorganic matrix17 (silica, CCDC, MERCURY 3.0 (Build RC5) for Windows.26 Crystallo-
alumina); this was revealed to be particularly interesting in the graphic data of the structures have been deposited at the
synthesis of the cyclic carbonates.
Cambridge Crystallographic Database Centre, supplementary
In this contribution, we report a further development of the publication no.: CCDC 958116 for complex 23 (LMe2), 958114
bpb (bpb: N,N-bis(2-pyridinecarboxamide)-1,2-benzene) ligand for complex 6 (LCl2) and 958115 for 22 (LNO2).
system18 dealing with the variation of the diamino-linker in
Synthesis of 6–8. 2.0 mmol of CoOAc2·4H2O and 2.0 mmol
the ligand framework. Surprisingly, although based on a of the appropriate ligand (2–4) were dissolved in 15 ml of DMF
straightforward synthetic method, many derivatives are still and stirred at RT under air for 10 min. Then 2.88 mmol of
unknown and missing in the “bpb-collection”. Hence we NEt4OAc·4H2O were added and the solution was stirred at
focused on new chromium, cobalt and iron complexes display- 70 °C for 12 h. After cooling, the solution was stirred for an
ing ligand frameworks containing either electron donating additional 6 hours. DMF was removed in vacuo and the residue
groups (Me) or electron withdrawing groups (Cl and NO2). The was dissolved in 20 ml of acetonitrile. A precipitate was
resulting air-stable metal complexes were thoroughly charac- formed on adding 20 ml of diethyl ether to the solution and
terized and, in three cases, further investigated via X-ray struc- cooling down to 2 °C. The product was isolated after filtering,
ture determination. These complexes have been tested for the washing two times with 20 ml of a 1 : 1 acetonitrile–diethyl
first time in catalytic screening tests involving carbon dioxide ether mixture and drying under vacuum. For complex 6 green
and several epoxides, mainly cyclohexene oxide and propylene crystals were obtained by slowly adding a small excess of
oxide.
diethyl ether into an acetonitrile solution of 5 and sub-
sequently cooling down to 2 °C. After 2 days small green crys-
tals were formed.
6: green powder/crystals, Yield: 78%. IR (cm−1): 2985 (w),
2924 (w), 2853 (w), 1625 (s), 1598 (s), 1557 (m), 1470 (s), 1400 (m),
1314 (m), 1229 (w), 1102 (w), 1002 (w), 977 (w), 926 (w),
Experimental section
General procedure
All manipulations were carried out under normal atmosphere. 764 (w), 683 (m), 670 (m), 548 (w). 1H-NMR δ (ppm, dmso-d6) =
The chemicals were of reagent grade, obtained from commer- 1.01 (s, 6 H), 1.10 (t, J = 7.3 Hz, 12 H), 3.15 (q, J = 7.3 Hz, 8 H),
cial sources such as Aldrich, abcr or Alfa Aesar and used 7.65–7.69 (m, 2 H), 7.84 (dd, J = 7.6 Hz, J = 1.5 Hz, 2 H), 8.07
without further purification. Cyclohexene oxide and propylene (t, J = 7.3 Hz, 2 H), 8.88 (d, J = 0.6 Hz, 2 H), 10.09 (d, J = 5.8 Hz,
oxide were dried over CaH2 before their use. The ligands 1–4 2 H). 13C{1H}NMR/DEPT135 (prim. = primary carbon; sec. =
and the complexes 5, 9, 10, 13, 14 and 18 were synthesized secondary carbon; tert. = tertiary carbon; quat. = quaternary
according to the literature procedures reported by Yama- carbon): δ (ppm, dmso-d6) = 7.1 (prim., 4 C), 24.1 (prim., 2 C),
zaki,19a Vagg et al.,19b Mukherjee et al.,20,21 Zhou et al.22 and 51.4 (sec., 4 C), 121.0 (tert., 2 C), 122.5 (tert., 2 C; quat., 2 C),
Kim et al.23 Compound 20 was synthesized according to our 124.3 (tert., 2 C), 139.5 (tert., 2 C), 143.7 (quat., 2 C), 156.8
own contribution.18
(tert., 2 C), 158.8 (quat., 2 C), 166.5 (quat., 2 C), 175.9 (quat.,
The NMR spectra of the catalysts and copolymer solutions 2 C). MS: (ESI−) m/z = 561 [M − NEt4]−. Anal. Calcd for
were recorded using a Varian Inova 400 spectrometer (1H: C30H36Cl2CoN5O6: C, 52.03; H, 5.24; N, 10.11. Found: C, 51.27;
399.81 MHz, 13C: 100.54 MHz). TMS was used as an internal H, 5.49; N, 10.05.
standard (13C, 1H) with different deuterated solvents. The
7: brown powder, Yield: 63%. IR (cm−1): 3081 (w), 2985 (w),
chemical shifts δ (in ppm) are given relative to the residual 2926 (w), 1629 (vs), 1601 (s), 1557 (m), 1482 (s), 1427 (m),
signal of the solvent. Mass spectra were measured with an 1386 (s), 1315 (vs), 1297 (s), 1146 (m), 1062 (w), 1002 (w),
Agilent 1100 MSD. Infrared spectra (KBr pellets and thin films 957 (w), 898 (w), 756 (w), 683 (m), 616 (m), 505 (w), 464 (w).
between KBr plates) were recorded on a Varian 660-IR FT-IR 1H-NMR δ (ppm, CD3CN) = 1.01 (s, 6 H), 1.09 (tt, J = 7.3 Hz, J =
Spectrometer. Molecular weights and MWD of the polymers 1.7 Hz, 12 H), 3.13 (q, J = 7.2 Hz, 8 H), 7.65–7.73 (m, 2 H), 7.88
3286 | Dalton Trans., 2014, 43, 3285–3296
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
(dt, J = 8.6 Hz, J = 1.6 Hz, 3 H), 8.06–8.12 (m, 2 H), 8.82 (d, J = 1003 (w), 943 (w), 889 (w), 765 (m), 681 (m), 651 (w), 509 (w),
1
9.0 Hz, 1 H), 9.54 (d, J = 2.6 Hz, 2 H), 10.10 (dd, J = 12.2 Hz, J = 406 (w). H-NMR δ (ppm, CD3CN) = 1.16 (t, J = 7.3 Hz, 12 H),
5.3 Hz, 2 H). 13C{1H}-NMR/DEPT135: δ (ppm, CD3CN) = 7.5 2.29 (s, 6 H), 3.10 (q, J = 7.3 Hz, 8 H), 7.76 (t, J = 6.4 Hz, 2 H),
(prim., 4 C), 24.5 (prim., 2 C), 51.8 (sec., 4 C), 116.0 (tert., 1 C), 8.05–8.16 (m, 4 H), 8.68 (s, 2 H), 9.64 (d, J = 5.5 Hz, 2 H). 13C
119.5 (tert., 1 C), 120.1 (tert., 1 C), 123.0 (tert., 1 C), 123.3 {1H}-NMR/DEPT135: δ (ppm, CD3CN) = 7.6 (prim., 4 C), 20.0
(tert., 1 C), 124.8 (tert., 1 C), 125.1 (tert., 1 C), 140.0 (tert., 2 C), (prim., 2 C), 52.9 (sec., 4 C), 123.9 (tert., 2 C), 124.4 (tert., 2 C),
141.8 (quat., 1 C), 144.2 (quat., 1 C), 151.3 (quat., 1 C), 157.1 127.5 (tert., 2 C), 131.9 (quat., 2 C), 140.0 (tert., 2 C),
(tert., 1 C), 157.2 (tert., 1 C), 158.7 (quat., 1 C), 159.2 (quat., 142.1 (quat., 2 C), 151.7 (tert., 2 C), 162.4 (quat., 2 C). MS:
1 C), 167.2 (quat., 1 C), 167.7 (quat., 1 C), 176.4 (quat., 2 C). (ESI−) m/z = 473 [M − NEt4]−. Anal. Calcd for C28H36Cl2-
MS: (ESI−) m/z
C30H37CoN6O8·H2O: C, 53.89; H, 5.73; N, 12.24. Found: C, H, 6.39; N, 11.38.
= 538 [M −
NEt4]−. Anal. Calcd for CoN5O2·H2O: C, 54.03; H, 6.15; N, 11.25. Found: C, 54.26;
52.48; H, 5.72; N, 12.49.
Synthesis of 15 and 16. 2.0 mmol of FeCl3, 2.0 mmol of the
8: green powder, Yield: 51%. IR (cm−1): 3087 (w), 2978 (w), corresponding ligand (3–4), 4.0 mmol of triethylamine and
2921 (w), 2857 (w), 1621 (s), 1596 (vs), 1485 (s), 1456 (m), 4.42 mmol of NEt4Cl·H2O were dissolved in 15 ml of DMF and
1405 (s), 1371 (s), 1316 (m), 1172 (w), 1095 (w), 1001 (m), 886 (w), stirred at RT under air for 12 hours. DMF was removed
759 (m), 681 (m), 619 (w), 509 (w). 1H-NMR δ (ppm, dmso- in vacuo and 50 ml of acetonitrile for 15 and for 16 20 ml of a
d6) = 1.02 (s, 6 H), 1.15 (t, J = 7.3 Hz, 12 H), 2.24 (s, 6 H), 3.20 1 : 1 acetonitrile–diethyl ether mixture were added to the
(q, J = 7.2 Hz, 8 H), 7.64 (ddd, J = 7.4 Hz, J = 5.6 Hz, J = 1.6 Hz, residue. The resulting precipitate was filtered and washed
2 H), 7.83 (dd, J = 7.7 Hz, J = 1.2 Hz, 2 H), 8.07 (dt, J = 7.6 Hz, twice (for 16 thrice) with 20 ml of a 1 : 1 acetonitrile–diethyl
J = 1.3 Hz, 2 H), 8.62 (s, 2 H), 10.16 (d, J = 5.5 Hz, 2 H). 13C ether mixture and eventually dried under vacuum.
{1H}-NMR/DEPT135: δ (ppm, dmso-d6) = 7.1 (prim., 4 C), 19.8
15: green powder, Yield: 80%. IR (cm−1): 3122 (w), 3068 (w),
(prim., 2 C), 24.2 (prim., 2 C), 51.4 (sec., 4 C), 122.1 (tert., 2 C), 2983 (w), 2925 (w), 1625 (vs), 1594 (s), 1559 (s), 1506 (m),
122.4 (tert., 2 C), 123.5 (tert., 2 C), 129.5 (quat., 2 C), 139.1 1471 (m), 1421 (w), 1346 (s), 1323 (vs), 1296 (s), 1136 (m),
(quat., 2 C), 141.7 (tert., 2 C), 156.8 (quat., 2 C), 160.2 (tert., 1113 (m), 1045 (w), 1000 (w), 969 (w), 924 (w), 827 (w), 760 (m),
2 C), 165.3 (quat., 2 C), 175.8 (quat., 2 C). MS: (ESI−) m/z = 521 696 (m), 647 (w), 488 (m). MS: (ESI−) m/z = 487 (100%), 489
[M − NEt4]−. Anal. Calcd for C32H42CoN5O6·H2O: C, 57.39; H, (65%) [M − NEt4]−. Anal. Calcd for C26H31Cl2FeN6O4·H2O: C,
6.62; N, 10.46. Found: C, 56.22; H, 6.89; N, 10.34.
Synthesis of 11 and 12. 2.0 mmol of CoCl2, 2.0 mmol of
49.08; H, 5.23; N, 13.21. Found: C, 49.18; H, 4.95; N, 13.40.
16: dark-green powder, Yield: 68%. IR (cm−1): 3045 (w),
the corresponding ligand (3–4), 4.0 mmol of triethylamine 2985 (w), 2915 (w), 2865 (w), 1615 (s), 1582 (vs), 1563 (s),
and 4.42 mmol of NEt4Cl·H2O were dissolved in 15 ml of DMF 1478 (s), 1455 (m), 1401 (m), 1354 (s), 1287 (m), 1254 (w),
and stirred at RT under normal atmosphere for 12 hours. 1182 (w), 1093 (w), 1044 (m), 1002 (w), 954 (w), 893 (w),
DMF was removed in vacuo and 50 ml of acetonitrile was 764 (m), 695 (m), 648 (w), 504 (w), 470 (w). MS: (ESI−) m/z =
added to the residue. A precipitate was formed, which was 470 [M − NEt4]−. Anal. Calcd for C28H36Cl2FeN5O2·2H2O:
filtered and washed with 20 ml of a 1 : 1 acetonitrile–diethyl C, 52.76; H, 6.33; N, 10.99. Found: C, 53.03; H, 6.62; N, 11.19.
ether mixture. After drying under vacuum the product was
isolated.
Synthesis of 17 and 19. 2.0 mmol of CrCl3·6H2O, 2.0 mmol
of the ligand (1 or 4), 4.0 mmol of triethylamine and
11: red-brown powder/crystals, Yield: 91%. IR (cm−1): 4.42 mmol of NEt4Cl·H2O were dissolved in 15 ml of DMF and
3070 (w), 2986 (w), 2949 (w), 1632 (vs), 1600 (s), 1558 (s), stirred at 120 °C for 12 hours. After cooling the solution was
1481 (vs), 1426 (m), 1390 (s), 1319 (vs), 1296 (s), 1142 (m), stirred at RT and under air for an additional 6 hours. DMF was
1062 (m), 999 (w), 958 (w), 895 (w), 758 (m), 684 (m), 615 (w), removed in vacuo and 20 ml of a 1 : 1 acetonitrile–diethyl ether
1
508 (w), 465 (w), 352 (w). H-NMR δ (ppm, CD3CN) = 1.12–1.19 mixture was added to the residue yielding a precipitate, which
(m, 12 H), 3.11 (q, J = 7.3 Hz, 8 H), 7.80–7.86 (m, 2 H), was filtered, washed twice with 20 ml of a 1 : 1 acetonitrile–
7.94 (dd, J = 9.0 Hz, J = 2.6 Hz, 1 H), 8.10–8.21 (m, 4 H), 8.93 diethyl ether mixture and dried under vacuum.
(d, J = 8.8 Hz, 1 H), 9.60–9.67 (m, 3 H). 13C{1H}-NMR/DEPT135:
17: red powder, Yield: 88%. IR (cm−1): 3055 (w), 2987 (w),
δ (ppm, CD3CN) = 7.5 (prim., 4 C), 52.9 (sec., 4 C), 117.5 2946 (w), 1620 (s), 1592 (vs), 1565 (vs), 1471 (s), 1363 (s),
(tert., 1 C), 120.4 (tert., 1 C), 121.4 (tert., 1 C), 125.0 (tert., 1 C), 1289 (m), 1183 (w), 1146 (w), 1094 (w), 1036 (m), 961 (w),
125.3 (tert., 1 C), 128.4 (tert., 1 C), 128.7 (tert., 1 C), 140.5 906 (w), 764 (m), 690 (m), 654 (w), 595 (w), 511 (m), 328 (m).
(tert., 2 C), 143.5 (quat., 1 C), 144.3 (quat., 1 C), 151.0 (quat., MS: (ESI−) m/z
= 438 [M −
NEt4]−. Anal. Calcd for
1 C), 152.0 (tert., 1 C), 152.1 (tert., 1 C), 160.5 (quat., 1 C), C26H32Cl2CrN5O2·H2O: C, 53.16; H, 5.83; N, 11.92. Found: C,
161.0 (quat., 1 C), 168.0 (quat., 1 C), 168.4 (quat., 1 C). 52.48; H, 5.57; N, 11.86.
MS: (ESI−) m/z
=
490 [M
−
NEt4]−. Anal. Calcd for
19: red-brown powder, Yield: 87%. IR (cm−1): 3067 (w), 2981 (w),
C26H31Cl2CoN6O4: C, 50.25; H, 5.03; N, 13.52. Found: C, 50.05; 2924 (w), 1629 (s), 1562 (vs), 1514 (m), 1482 (s), 1366 (s), 1291 (m),
H, 4.96; N, 13.67.
1238 (w), 1178 (m), 1095 (w), 1048 (w), 1003 (w), 892 (w), 764 (w),
12: red powder, Yield: 88%. IR (cm−1): 3077 (w), 2984 (w), 695 (m), 663 (w), 510 (w), 422 (w). MS: (ESI−) m/z = 466
2924 (w), 2856 (w), 1625 (vs), 1595 (s), 1577 (s), 1486 (m), [M − NEt4]−. Anal. Calcd for C28H36Cl2CrN5O2·2H2O: C, 53.08;
1456 (m), 1404 (s), 1290 (m), 1251 (m), 1178 (m), 1092 (w), H, 6.36; N, 11.05. Found: C, 52.76; H, 6.33; N, 11.42.
This journal is © The Royal Society of Chemistry 2014
Dalton Trans., 2014, 43, 3285–3296 | 3287

View Article Online
Paper
Dalton Transactions
Synthesis of 21–23. 2.0 mmol of CoBr2·xH2O, 2.0 mmol of Autoclave experiments
the related ligand (2–4), 4.0 mmol of triethylamine and
4.42 mmol of NEt4Br were dissolved in 15 ml of DMF and
stirred under air at RT for 12 hours. DMF was removed
in vacuo and 50 ml of acetonitrile (for 21 20 ml) was added to
the residue. A precipitate was formed, which was filtered and
washed twice (for 17 thrice) with 20 ml of a 1 : 1 acetonitrile–
diethyl ether mixture. After drying under vacuum the product
was obtained. For complex 22 red-brown crystals were obtained
by slowly adding a small excess of diethyl ether into an aceto-
nitrile solution of 22 and subsequently cooling down to 2 °C.
After 2 days small red-brown crystals were formed. Using the
same procedure brown crystals for 23 could be isolated.
The copolymerization tests were conducted in a screening test-
bench consisting of eight 70 ml autoclaves (material 1.4571-
stainless austenitic steel-316Ti, 70 ml, Pmax 200 bar, Tmax
250 °C) equipped with magnetic stirring and P,T-acquisition
via multi-meters (Agilent 34970A Data Acquisition/Data Logger
Switch Unit + 34901A 20 Channel Multiplexer) and PCs. The
heating was performed with an aluminum heating block con-
trolled via the electronic contact thermometer of the magnetic
stirrers. Each configuration reported in the “reactivity” tables
(4, 5 and 6) is tested two times and, if needed, a third time;
the reproducibility was high. The autoclaves were dried at
110 °C under vacuum for a couple of hours, cooled down and
purged with argon before use. This procedure was once again
applied for two hours at 60 °C with the sole catalyst in the
autoclave. After cooling the autoclave to room temperature, the
epoxide was injected under argon via one of the autoclave
ports and the reaction mixture was pressurized with CO2. After
the reaction the autoclaves were cooled down, slowly vented
under stirring in a fume hood, opened and the remaining
solution (resp. solid) was collected with dichloromethane.
After thorough drying, the solution (resp. solid) was firstly
21: brown powder, Yield: 65%. IR (cm−1): 3086 (w),
3051 (w), 2982 (w), 2945 (w), 1626 (vs), 1598 (s), 1558 (s),
1468 (s), 1395 (s), 1285 (m), 1200 (w), 1097 (m), 1049 (w),
975 (m), 925 (w), 887 (w), 760 (m), 680 (m), 546 (m), 482 (w),
1
447 (w). H-NMR δ (ppm, CD3CN) = 1.18 (t, J = 7.2 Hz, 12 H),
3.13 (q, J = 7.2 Hz, 8 H), 7.83 (tr, J = 6.3 Hz, 2 H), 8.07–8.17 (m,
4 H), 9.00 (s, 2 H), 9.70 (d, J = 5.3 Hz, 2 H). 13C{1H}-NMR/
DEPT135: δ (ppm, CD3CN) = 7.6 (prim., 4 C), 52.9 (sec., 4 C),
123.3 (tert., 2 C), 124.9 (tert., 2 C), 125.1 (quat., 2 C), 128.4
(tert., 2 C), 140.3 (tert., 2 C), 144.1 (quat., 2 C), 153.0 (tert.,
2 C), 161.6 (quat., 2 C), 167.9 (quat., 2C). MS: (ESI−) m/z = 602
[M − NEt4]−. Anal. Calcd for C26H30Br2Cl2CoN5O2: C, 42.53; H,
4.12; N, 9.54. Found: C, 42.08; H, 4.52; N, 10.12.
1
investigated via IR and H-NMR-spectroscopy given that cyclic
monomers and polycarbonates are easily differentiated via
both spectroscopy methods. For instance, the stretching
mode27 of the CvO bond in polycarbonates can be found at
around 1750 cm−1 whereas the cyclic carbonates are typically
found at around 1800 (cis-cyclohexylene carbonate at
1802 cm−1, trans-cyclohexylene carbonate at 1818 cm−1 and
propylene carbonate at 1800 cm−1). If the solution contains a
copolymer, which was the case only for cyclohexene oxide, it
was next dissolved in a small amount of CH2Cl2 and the co-
polymer was precipitated utilizing 1M HCl/MeOH according to
the literature.28 The isolated copolymers were dried under
22: red-brown powder/crystals, Yield: 62%. IR (cm−1): 3124
(w), 3054 (w), 2987 (w), 2925 (w), 1631 (vs), 1598 (s), 1558 (vs),
1493 (s), 1478 (s), 1422 (m), 1389 (s), 1318 (s), 1293 (s), 1140
(m), 1094 (w), 1058 (m), 998 (w), 985 (w), 897 (w), 759 (m), 682
(m), 616 (w), 506 (w), 459 (w). 1H-NMR δ (ppm, CD3CN) =
1.15–1.21 (m, 12 H), 3.14 (q, J = 7.2 Hz, 8 H), 7.83–7.90 (m,
2 H), 7.95 (dd, J = 8.9 Hz, J = 2.5 Hz, 1 H), 8.14–8.20 (m, 4 H),
8.96 (d, J = 9.0 Hz, 1 H), 9.65–9.74 (m, 3 H). 13C{1H}-NMR/
DEPT135: δ (ppm, CD3CN) = 7.6 (prim., 4 C), 52.9 (sec., 4 C),
117.9 (tert., 1 C), 120.5 (tert., 1 C), 121.8 (tert., 1 C), 125.0
(tert., 1 C), 125.3 (tert., 1 C), 128.5 (tert., 1 C), 128.8 (tert., 1 C),
140.4 (tert., 1 C), 140.5 (tert., 1 C), 143.6 (quat., 1 C), 143.7
(quat., 1 C), 151.1 (quat., 1 C), 153.0 (tert., 1 C), 153.1 (tert.,
1 C), 161.2 (quat., 1 C), 161.7 (quat., 1 C), 167.5 (quat., 1 C),
167.7 (quat., 1 C). MS: (ESI−) m/z = 580 [M − NEt4]−. Anal.
Calcd for C26H31Br2CoN6O4: C, 43.96; H, 4.40; N, 11.83. Found:
C, 43.68; H, 4.41; N, 12.25.
1
vacuum and analyzed via H-NMR spectroscopy (% carbonate
linkages29), 13C-NMR-spectroscopy (tacticity) and gel per-
meation chromatography (molecular weights and polydisper-
sity indexes). However, if only the cyclic monomer was
obtained, the effective cyclic carbonate amount was calculated
by subtracting the amount of catalyst used.
Results and discussion
Synthetic procedures
23: brown powder/crystals, Yield: 74%. IR (cm−1): 3077 (w),
2984 (w), 2923 (w), 2858 (w), 1625 (s), 1594 (vs), 1485 (s), 1456
(m), 1404 (s), 1290 (w), 1251 (w), 1178 (w), 1093 (w), 1003 (m), The synthesis of the bpb-metal complexes is shown in
1
889 (w), 764 (m), 680 (m), 588 (w), 509 (w), 406 (w). H-NMR δ Scheme 2. The different bpb ligands (1–4) are obtained by con-
(ppm, CD3CN) = 1.15–1.20 (m, 12 H), 2.29 (s, 6 H), 3.13 (q, J = densation of 1,2-diaminobenzene with two equivalents of
7.3 Hz, 8 H), 7.72–7.90 (m, 2H), 8.02–8.28 (m, 4 H), 8.68 (s, 2-pyridine-carboxylic acid in pyridine under activation with tri-
2 H), 9.56–9.70 (m, 2 H). 13C{1H}-NMR/DEPT135: δ (ppm, phenyl phosphite according to Yamazaki19a and Vagg.19b The
CD3CN) = 7.6 (prim., 4 C), 20.0 (prim., 2 C), 53.0 (sec., 4 C), bpb-metal complexes are formed in a following step via reac-
124.3 (tert., 2 C), 124.4 (tert., 2 C), 127.7 (tert., 2 C), 127.7 tion of the different ligands with one equivalent of the
(quat., 2 C), 132.2 (quat., 2 C), 140.0 (tert., 2 C), 142.0 (quat., hydrated metal salts (MX2 for M = Co, X = OAc, Cl, Br and MX3
2 C), 152.7 (tert., 2 C), 163.1 (quat., 2 C). MS: (ESI−) m/z = 561 for M = Fe, Cr, X = Cl) and a slight excess of a tetraethyl
[M − NEt4]. Anal. Calcd for C28H36CoN5O2·H2O: C, 47.27; ammonium salt [NEt4]X·H2O (X = OAc, Cl and Br) in DMF. In
H, 5.38; N, 9.84. Found: C, 46.28; H, 5.34; N, 9.71.
the case of the metal halides two equivalents of triethylamine
3288 | Dalton Trans., 2014, 43, 3285–3296
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
Fig. 1 ORTEP drawings (50% probability) of 6. For the sake of clarity,
hydrogen atoms have been removed and only one cobalt complex is
shown.
Scheme 2 Synthesis of the bpb–metal complexes.
are required for the deprotonation of the ligand’s N–H,
whereas using cobalt acetate, no additional base is required.
The formation of the bpb complexes is generally straight-
forward and, for some of the complexes, a literature-known
procedure (see the Experimental section for details). However,
the synthesis of a chromium bpb-complex with a nitro-
substituent was revealed to be more complicated than
expected, yielding complex mixtures.
Fig. 2 ORTEP drawings (50% probability) of 23. For the sake of clarity,
hydrogen atoms have been omitted.
The new complexes 6–8, 11, 12, 15–17, 19, and 21–23 were
also characterized via NMR-spectroscopy, elemental analysis,
IR-spectroscopy and MS-ESI, the analytic data being in good
agreement with the values found in the literature for the
related complexes. All the complexes were obtained in fair to
excellent yields ranging from 60 to 90%. These compounds
are air-stable in the solid state and can be stored for
months without degradation. In addition, it was possible,
for the first time, to structurally characterize the ionic cobalt
complexes 6, 22 and 23 via X-ray diffractometry on single
crystals.
Fig. 3 ORTEP drawings (50% probability) of 22 showing the two mole-
Molecular structure determinations of 6, 22 and 23
cules engaged in a hydrogen bond network with 1.5 water molecules.
Suitable crystals for structure determination were obtained by
slowly adding a small excess of diethyl ether into an aceto-
nitrile solution and subsequently cooling down to 2 °C. After 2 of 6, 22 and 23 are shown in Fig. 1–3. The crystallographic
data, selected bond lengths and angles of 6, 22 and 23 are
brown crystals for 23 were formed. The thermal ellipsoid plots listed in Tables 1–4, respectively.
days small green crystals for 6, red-brown crystals for 22 and
This journal is © The Royal Society of Chemistry 2014
Dalton Trans., 2014, 43, 3285–3296 | 3289

View Article Online
Paper
Dalton Transactions
Cobalt(III)-bpb anions and the counter-ion, tetraethyl
ammonium, display no short contacts as can be representa-
tively seen in the crystal structure of complex 6 ([Co(bpb-Cl2)-
OAc2][NEt4]) (Fig. 1).
Table 1 Selected bond lengths (Å) and angles (°) for complex 6 (two
discrete complexes present in the independent unit of the elementary
cell, only one listed in this table)
Complex 6/bonds-angles
The molecular structures of compounds 6 and 23 display,
in the independent unit of the elementary cell, two discrete
complexes with water molecules engaged in hydrogen bond
networks. For complex 6 the network involves one water mole-
cule and the carbonyl moieties of acetate and amide groups
(O(7)–O(13): 2.807 Å; O(13)–O(10): 2.822 Å) whereas in com-
pound 23, the network includes 1.5 molecules water, one
bromide and the carbonyls of two amide groups per indepen-
dent unit (O(6)–O(5): 2.739 Å; O(3)–O(5): 2.757 Å; O(1)–O(5):
2.816 Å) (Fig. 2). For complex 22 a distinct single complex with
no intermolecular short contacts was found in the indepen-
dent unit of the elementary cell during the structural study
(Fig. 3, see ESI† for more details).
Cobalt atoms and the N4 ligand plane are almost coplanar
in the case of the cobalt bromide complexes 22 and 23
(maximal distortion from the mean N4-plane: 0.004 Å, distance
of Co from the mean planes ranging from 0.005 to 0.007 Å), in
the same way as can be noticed in related octahedrally co-
ordinated cobalt porphyrin derivatives.30
Co(2)–N(6)
Co(2)–N(7)
Co(2)–N(8)
Co(2)–N(5)
Co(2)–O(11)
Co(2)–O(9)
Co(2)⋯O(10)
Co(2)⋯O(12)
O(7)⋯H(131)
N(8)–Co(2)–O(9)
N(6)–Co(2)–O(9)
N(7)–Co(2)–O(11)
N(5)–Co(2)–O(11)
O(11)–Co(2)–O(9)
N(7)–Co(2)–N(8)
N(6)–Co(2)–N(7)
N(6)–Co(2)–N(5)
N(8)–Co(2)–N(5)
1.884(2)
1.8891(19)
1.989(2)
2.0011(19)
1.9249(17)
1.9249(17)
3.168
3.133
2.022
90.55(8)
88.16(8)
88.19(8)
91.21(8)
175.28(7)
82.36(8)
83.96(9)
81.93(8)
111.75(8)
All three tetraethyl-ammonium cobalt complexes 6, 22 and
23 display similar structural patterns including, on the one
hand, an octahedral geometry around the negative charged
metal center with two monodentate, apical ligands (acetate for
6, bromo for 22 and 23) and, on the other hand, the
tetradentate N4-bpb-ligand occupying the equatorial plane.
In the cobalt acetate complex 6 one independent molecule,
involved in the above-mentioned hydrogen bond network, dis-
plays coplanar cobalt and N4 ligand as in 22 and 23 whereas
the other one shows a slightly more distorted geometry
Table 2 Crystallographic data for 6, 22 and 23
6: 2{[Co(bpbCl2)OAc][NEt4]}·H2O
Crystal data
22: [Co(bpbNO2)Br2][NEt4]
23: 2{[Co(bpbMe2)Br2][NEt4]}·1.5H2O
Empirical formula
C60H74Cl4Co2N10O13
1402.95
C26H31Br2CoN6O4
710.32
C56H75.5Br4Co2N10O5.25
1409.25
Molecular mass
Crystal color
Green
Red-brown
0.147 × 0.114 × 0.023
Monoclinic
P21/c
10.2730(5)
11.9909(6)
23.3174(12)
90
Brown
0.059 × 0.049 × 0.033
Monoclinic
P21/c
17.8240(15)
16.6669(14)
20.3735(17)
90
Crystal size (mm3)
0.07 × 0.05 × 0.04
Monoclinic
P2(1)/c (no. 14)
21.8010(15)
15.0617(10)
21.8954(15)
90
Crystal system
Space group
a (Å)
b (Å)
c (Å)
α (°)
β (°)
119.0960(10)
90
6282.3(7)
101.6220(10)
90
2813.4(2)
4
1.677
3.494
101.5020(10)
90
5930.8(9)
4
1.578
3.310
2866
γ (°)
V (Å3)
Z
8
Dcalc (g cm−3
)
1.483
0.770
2920
µ (mm−1
F (0 0 0)
)
1432
Wavelength (Å)
T (K)
Θ-range (°)
Index ranges
0.71073
100(2)
1.72 to 28.49
−28 ≤ h ≤ 28, −20 ≤ k ≤ 19,
−28 ≤ l ≤ 28
0.71073
100(2)
1.783 to 28.562
−13 ≤ h ≤ 13, −15 ≤ k ≤ 15,
−30 ≤ l ≤ 31
71.073
100(2)
1.592 to 28.546
−23 ≤ h ≤ 23, −21 ≤ k ≤ 22,
−27 ≤ l ≤ 27
Solution and refinement
Number of reflections measured
Number of independent reflections
GOF
110 733
14 943
1.019
50 757
6730
1.038
0.0273
106 542
14 101
1.016
R1[I2Θ(I)]
0.0423
0.0390
wR2[I2Θ(I)]
0.0888
0.850 and −0.686
0.0690
0.846 and −0.518
0.0620
0.651 and −0.506
Largest e-max, e-min (e Å−3
)
3290 | Dalton Trans., 2014, 43, 3285–3296
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
from 1.975(2) (complex 23) to 2.001(2) (complex 6) and
“strong” bonds for the amide-N-atoms with values from
1.8763(17) (complex 22) to 1.8886(19) (complex 6). Considering
the N–Co–N angles found in the three complexes, the angles
involving the two pyridines are wider than the theoretical
value with data ranging from 109.7 to 111.7°, and the other
angles fluctuate from 81.9 to 84.5°.
The angles involving cobalt and the two axial monodentate
bond ligands (acetate for 6, bromo for 22 and 23) are between
175 and 178°, displaying an almost symmetric coordination of
the axial ligands. For compound 6 the distances between the
axial oxygen atoms and the cobalt atom are Co(2)–O(9): 1.9250(18)
and Co(2)–O(11): 1.9249(18), typical for monodentate
bound carboxylates (non-coordinated oxygens: Co(2)⋯O(10):
3.168 and Co(2)-O(12): 3.133). These values are very similar to
the data we obtained for the cobalt-acetate complex without
substitution at the ligand backbone.18 The distances between
axial bromides and the cobalt atom in compound 22 (Co–
Br(1): 2.4143(4) and Co–Br(2): 2.4029(4)) and compound 23
(Co(2)–Br(3): 2.4136(5) and Co(2)–Br(4): 2.4008(5)) are compar-
able, regardless of the substitution of the ligand linker.
Table 3 Selected bond lengths (Å) and angles (°) for complex 22
Complex 22/bonds-angles
Co(1)–N(1)
Co(1)–N(2)
Co(1)–N(3)
Co(1)–N(4)
Co(1)–Br(1)
Co(1)–Br(2)
N(2)–Co(1)–Br(2)
N(4)–Co(1)–Br(2)
N(3)–Co(1)–Br(1)
N(1)–Co(1)–Br(1)
Br(1)–Co(1)–Br(2)
N(2)–Co(1)–N(1)
N(2)–Co(1)–N(3)
N(3)–Co(1)–N(4)
N(1)–Co(1)–N(4)
1.8763(17)
1.8766(17)
1.9772(17)
1.9840(17)
2.4143(4)
2.4029(4)
90.81(6)
89.82(5)
89.61(5)
90.41(6)
1.78727(15)
84.50(8)
82.81(8)
109.74(7)
82.95(7)
Table 4 Selected bond lengths (Å) and angles (°) for complex 23 (two
discrete complexes present in the independent unit of the elementary
cell, only one listed in the table)
Complex 23/bonds-angles
Co(2)–N(5)
Co(2)–N(6)
Co(2)–N(7)
Co(2)–N(8)
Co(2)–Br(3)
Co(2)–Br(4)
N(5)–Co(2)–Br(3)
N(7)–Co(2)–Br(3)
N(6)–Co(2)–Br(4)
N(8)–Co(2)–Br(4)
Br(3)–Co(2)–Br(4)
N(5)–Co(2)–N(6)
N(6)–Co(2)–N(7)
N(7)–Co(2)–N(8)
N(8)–Co(2)–N(5)
1.975(2)
1.878(2)
1.882(2)
1.990(2)
2.4136(5)
2.4008(5)
87.87(7)
91.91(8)
91.23(7)
89.33(7)
175.17(2)
83.15(10)
84.33(10)
82.63(11)
109.89(10)
Catalytic screening of the metal bpb complexes in the
coupling of epoxides with carbon dioxide
Inspired by the promising results of the screening tests18 run
with the cobalt and iron complexes 5, 9, 13 and 20, we per-
formed similar tests with the new derivatives. The catalytic
screening was carried out in 70 ml SS316Ti autoclaves with a
standard reaction time of 20 h under “solvent-free” conditions.
The results are summarized in Table 5 for the cobalt-acetate
complexes 5–8, in Table 6 for the different metal-chloride com-
plexes 9–19 and in Table 7 for the cobalt-bromide complexes
20–24. Additionally several commercially available epoxides
(maximal distortion from the mean N4-plane: 0.02 Å, distance were examined with one of the efficient catalysts in this range:
of Co from the mean planes ranging from 0.042 Å; for more compound 10 (Table 8). The reaction parameters (cat.%, temp.
information see ESI†). The bond lengths between cobalt- and and P) used in the screening were chosen according to the
nitrogen bpb-atoms found in 6, 22 and 23 can be divided into related literature and quickly checked in a screening with
“weak” bonds for the pyridine-N-atoms with values ranging complex 10 and propylene oxide (see ESI†). Last but not least,
Table 5 Coupling reaction with cobalt-acetate-complexesa
Selectivity of the
couplingc
CO3-content (%)
of the polymerd
Conversionb
Mn (g mol−1 e
)
Mw/Mn
e
Entry
Catalyst
Epoxide
1
2
3
4
5 (Co/LH2/OAc)
6 (Co/LCl2/OAc)
7 (Co/LNO2/OAc)
8 (Co/LMe2/OAc)
CHO
CHO
CHO
CHO
64
70
35
83
100/0
100/0
100/0
100/0
100
100
100
100
9600
7600
10 100
8600
1.15
1.28
1.27
1.27
5f
6
7
5 (Co/LH2/OAc)
6 (Co/LCl2/OAc)
7 (Co/LNO2/OAc)
8 (Co/LMe2/OAc)
PO
PO
PO
PO
62
50
32
70
0/100
0/100
0/100
0/100
8
a Standard reaction conditions: 10 ml of epoxide, 20 h, 0.5 mol% catalyst, 80 °C, 50 bar for CHO and 35 bar of CO2 for PO (∼7 g CO2).
b Conversion = n(monomer units in isolated product)/n(epoxide) × 100. c Mass of isolated long chain polymers/mass of isolated cyclic monomer.
d Evaluated via 1H NMR. e Evaluated via gel permeation chromatography. f Ref. 18.
This journal is © The Royal Society of Chemistry 2014
Dalton Trans., 2014, 43, 3285–3296 | 3291

View Article Online
Paper
Dalton Transactions
Table 6 Coupling reaction with metal–chloride-complexesa
Selectivity
CO3-content (%)
of the polymerd
Conversionb
of the couplingc
Mn (g mol−1 e
)
Mw/Mn
e
Entry
Catalyst
Epoxide
1
2
3
4
5
6
7
8
9 (Co/LH2/Cl2)
CHO
CHO
CHO
CHO
CHO
CHO
CHO
CHO
CHO
CHO
CHO
PO
PO
PO
PO
PO
PO
PO
PO
PO
67
78
83
Traces
31
25
27
33
7
100/0
100/0
100/0
100
100
100
6500
8800
7900
1.32
1.18
1.33
10 (Co/LCl2/Cl2)
11 (Co/LNO2/Cl2)
12 Co/LMe2/Cl2)
13 (Fe/LH2/Cl2)
14 (Fe/LCl2/Cl2)
15 (Fe/LNO2/Cl2)
16 (Fe/LMe2/Cl2)
17 (Cr/LH2/Cl2)
18 (Cr/LCl2/Cl2)
19 (Cr/LMe2/Cl2)
9 (Co/LH2/Cl2)
10 (Co/LCl2/Cl2)
11 (Co/LNO2/Cl2)
12 Co/LMe2/Cl2)
13 (Fe/LH2/Cl2)
14 (Fe/LCl2/Cl2)
15 (Fe/LNO2/Cl2)
16 (Fe/LMe2/Cl2)
17 (Cr/LH2/Cl2)
18 (Cr/LCl2/Cl2)
19 (Cr/LMe2/Cl2)
100/0
100/0
100/0
100/0
100/0
100/0
—
0/100
0/100
0/100
97
91
100
100–
38
21
—
—
—
—
—
—
—
—
—
—
—
—
2600
2000
2700
2200
600
550
—
—
—
—
—
—
—
—
—
—
—
—
1.14
1.06
1.15
1.18
1.37
1.19
—
—
—
—
—
—
—
—
—
—
—
—
9
10
11
12f
13
14
15
16f
17
18
19
20
21
22
9
Traces
76
86
92
Traces
78
83
92
91
90
0/100
0/100
0/100
0/100
0/100
0/100
0/100
PO
PO
69
96
a Standard reaction conditions: 10 ml of epoxide, 20 h, 0.5 mol% catalyst, 80 °C, 50 bar for CHO and 35 bar of CO2 for PO (∼7 g CO2).
b Conversion = n(monomer units in isolated product)/n(epoxide) × 100. c Mass of isolated long chain polymers/mass of isolated cyclic monomer.
d Evaluated via 1H NMR. e Evaluated via gel permeation chromatography. f Ref. 18.
Table 7 Coupling reaction (PO/CO2) with cobalt–bromide-complexesa
Table 8 Epoxide–CO2-coupling reaction catalyzed by 10a
Catalyst
loading
(mol%)
Selectivity Entry
Epoxide
Product
p (bar)
Conversionb
76
of the
Entry
Catalyst
Conversionb
couplingc
1
50
1
2
3
4
5
6
7
8*
9**
10
11
12
20d (Co/LH2/Br2)
20 (Co/LH2/Br2)
20 (Co/LH2/Br2)
21 (Co/LCl2/Br2)
21 (Co/LCl2/Br2)
21 (Co/LCl2/Br2)
22 (Co/LNO2/Br2)
22 (Co/LNO2/Br2)
22 (Co/LNO2/Br2)
23 (Co/LMe2/Br2)
23 (Co/LMe2/Br2)
23 (Co/LMe2/Br2)
0.5
0.1
0.05
0.5
0.1
0.05
0.5
0.1
0.05
0.5
0.1
93
93
72
91
91
82
93
91
76
93
75
64
0/100
0/100
0/100
0/100
0/100
0/100
0/100
0/100
0/100
0/100
0/100
0/100
2
3
4
35
50
50
78
60
60
0.05
a Standard reaction conditions: 10 ml of epoxide, 20 h, *22 h, **25 h,
80 °C, 35 bar of CO2 for PO (∼7 g CO2). b Conversion = n(monomer
units in isolated product)/n(epoxide) × 100. c Mass of isolated long
chain polymers/mass of isolated cyclic monomer. d Ref. 18.
5
6
50
50
96
93
an exploratory recycle study was accomplished for catalyst 23
using propylene oxide as the substrate (Fig. 4).
7
8
50
50
99
0
In the case of the cobalt-acetate-complexes (Table 5) the
results show that cyclohexene oxide yields pure polycarbonates
with 100 per cent carbonate linkages and conversions ranging
from 35 (entry 3) to 83% (entry 4). For propylene oxide, only
the cyclic monomer is formed with yields up to 70% (entry 8).
Interestingly the same cobalt(III)/acetate combination with
N2O2 salen ligand and tetragonal pyramidal coordination geo-
a Standard reaction conditions: 10 ml of epoxide, 20 h, 0.2 mol%
catalyst, 80 °C. b Conversion = n(monomer units in isolated product)/
metry yields slightly different results as reported by Coates and n(epoxide) × 100.
3292 | Dalton Trans., 2014, 43, 3285–3296
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
99% yield, with higher molecular weights and slightly better
PDI (Mn: 14 500 g mol−1, Mw/Mn: 1.13).34 The reaction needed
however a cocatalyst like, e.g., 4-dimethylamino-pyridine (reac-
tion conditions: 80 °C, 50 atm pressure CO2, 0.2 mol% catalyst,
0.1 mol% DMAP, 24 h). Switching the axial ligand from acetate
to chloride leads to an increased reactivity with, as an
example, a maximal conversion of PO of 92% (Table 6, entry
14) instead of 32% (Table 5, entry 7). This increased reactivity
can be related to the presence of the ammonium cation in the
ionic complex and an increased reactivity of a concomitantly
formed tetraethyl ammonium halide compared to NEt4OAc
(vide infra for the mechanism proposal). However the reactivity
trend for both epoxides is reversed, the conversion decreasing
now with rising electron-donating capability of the ligand:
Me < H < Cl < NO2.
Fig. 4 Recycle experiment with catalyst 23. Standard reaction con-
ditions: 10 ml of epoxide, 20 h, 0.5 mol% catalyst, 80 °C, 35 bar of CO2.
Conversion = n(monomer units in isolated product)/n(epoxide) × 100.
Surprisingly, with complex 12 (entries 4 and 12) only traces
of product were found for both epoxides. The solubility test in
both epoxides showed that this complex displays a poor solubi-
lity in these epoxides which is most likely the explanation for
the low catalytic activity.
co-workers12a,b (polypropylene-carbonate obtained with 30%
yield at RT, under ca. 54 bar for 2 hours, catalyst: 0.2 mol%
and polycyclohexene carbonate obtained with 27% yield, PDI
and Mn comparable, run at RT under ca. 54 bar, for 20 hours,
catalyst: 0.2 mol%). Considering the influence of the bpb-
ligand structure on the conversion of the epoxides, a positive
effect was found if the aromatic linker was substituted with
electron-donating groups (Me) and a negative effect if an elec-
tron-withdrawing group (NO2) was used instead. Only a small
influence on the reactivity was noticed with the chlorine sub-
stituents probably due to “contradictory” features: on the one
hand an electron withdrawing (−I) effect and, on the other
hand, a positive mesomeric effect (+M). These results suggest
that, for the acetato complexes, the conversion increases with
rising electron-donating capability of the ligand: Me > Cl ∼ H >
NO2. Considering now the molecular weight and PDIs of the
isolated polymers, ranging from 7600 to 10 100 g mol−1 and
from 1.15 to 1.28, it seems that the substitution had a fairly
small influence on the chain lengths. Generally the gel per-
meation chromatography elugrams show a bimodal distri-
bution with an overall narrow distribution. This behaviour is
most likely due to the presence of water traces in the crystal-
line compound and the connected formation of cyclohexane-
1,2-diol. This behaviour is found in numerous copolymeriza-
tion reactions producing aliphatic polycarbonates.31 All the
copolymers isolated with this catalytic system displayed a stat-
istical arrangement of the carbonate linkages (atacticity) as
stated by the 13C-NMR spectra (CDCl3, see ESI† for spectra).
For instance the copolymer obtained with catalyst 7 displays a
strong signal at ca. 153.7 ppm for the isotactic and a broad but
weaker signal at ca. 153.3–153.1 ppm for the syndiotactic
fragments.32,33
Taking into account now the catalytic screening of the iron
chloride complexes the results reveal that switching the central
atom from cobalt to iron has only a small influence on the
reactivity towards propylene oxide. A substitution of the aro-
matic linker has also a general positive effect, leading nearly to
the same yields as for the cobalt chloride complexes (e.g. entry
18, NO2-substituted ligand; 92% yield). The effect of the substi-
tution is however not so obvious as in the former case:
complex 16 shows, despite containing electron-donating
groups, a high activity for propylene oxide, forming cyclic car-
bonate with 91% yield (entry 19). To put things in perspective,
this N4 bpb-system delivers higher yield of propylene carbon-
ate than the salan-iron-chloride system of Rieger and co-
workers15 (57%, at 80 °C, 1 mol% catalyst, 15 bar and 2 hours),
and needs however a higher CO2-pressure and longer reaction
time. Changing the test substrate to cyclohexene oxide leads to
a lower reactivity with yields reaching only 33 per cent (entry 5)
and smaller molecular weights of the isolated copolymer with
values from 2000 to 2700 g mol−1 compared to the cobalt
chloride complexes. Despite a low reactivity, these iron com-
plexes are promising if one considers that most of the reported
iron-based molecular catalysts were used in the formation of
cyclic carbonates.15,35 Our system is however not as efficient as
the other homogeneous iron-based copolymerisation catalyst
documented by Williams36 and co-workers.
Considering the chromium chloride complexes, the change
of the central metal had different influences for propylene-
and cyclohexene oxide: On the one hand the conversion is
increased up to 96% for propylene oxide (entry 22, Me-substi-
tuted ligand) which shows also that, for the chromium chlo-
ride complexes, an electron donating group has a positive effect
on the conversion. On the other hand, a very low reactivity was
found for cyclohexene oxide, with conversions culminating in
9% (entry 10). The incorporation of CO2 as well as the mole-
cular weights decreased dramatically, with a maximal carbon-
ate linkage value of only 38% (entry 9). Comparing chromium-,
The cobalt chloride bpb complexes (Table 6, entries 1–4 and
12–15), like the cobalt acetate complexes, form with cyclo-
hexene oxide selectively pure polycarbonates with molecular
weights in a narrow range, from 6500 to 8800, and with pro-
pylene oxide only the cyclic carbonate. These results can be
related to the cobalt-porphyrin-chloride system reported by
Sugimoto et al. which leads to polycyclohexene carbonate in
This journal is © The Royal Society of Chemistry 2014
Dalton Trans., 2014, 43, 3285–3296 | 3293

View Article Online
Paper
Dalton Transactions
cobalt- and iron-bpb-chlorides, a reactivity trend can be formed merely for cyclohexene oxide (entry 1) whereas for the
roughly established for both epoxides. For propylene oxide all other epoxides only the cyclic monomers were obtained. The
three metals showed high yields with the following gradation highest yields of cyclic carbonates were attained with mono-
of reactivity: Cr > Co ∼ Fe whereas for cyclohexene oxide the substituted epoxides displaying an electron-withdrawing group
difference between the metals was more significant and the (entries 5–7) as illustrated by the high conversions reached
reactivity trend was as follows: Co >> Fe > Cr.
with glycidyl ethers and epichlorhydrine. This trend was tenta-
Unfortunately, it was not possible to enlarge the bpb-metal tively explained in the literature by the ability of some electron-
acetato series and find similar reactivity trends for iron(III)- withdrawing groups to stabilize intermediates and facilitate
and chromium(III)-bpb-acetates. The synthesis of the substi- the intramolecular cyclization.38
tuted-bpb iron acetates leads surprisingly to low yields while
Styrene oxide, however, delivered the cyclic carbonate in
the un-substituted ligand worked properly18 (90% yield). The average yields (entry 3) (see NMR data in ESI†). The somewhat
maximum yield was reached with the nitro-substituted bpb- lower reactivity of SO compared to PO and CHO as well as the
iron(III)-acetate (below 30%) that led, in the catalysis, to a very preferred formation of the cyclic carbonate was already docu-
poor conversion to carbonates. The related synthesis with com- mented in some studies.39,40 The reasons depend on the
mercially available basic chromium(III) acetate ([Cr3(μ3-O)- nature of the catalytic system used, involving either inappropri-
(OAc)6(H2O)3]Cl) led to no conversion whereas freshly made ate steric or electronic parameters of the phenyl group or a
Cr(OAc)2, with subsequent oxidation, gave complexes in very lower solubility of the catalyst in the styrene oxide–CO2
low yields.
mixture. In our case, complex 10 is soluble; the low reactivity
As expected, changing the nature of the axial halide has a of SO with this system is under further study and will be evalu-
significant effect on the catalytic activity. For instance cobalt ated performing a screening with the other substituted bpb-
bromide bpb complexes display a very low catalytic activity with catalysts.
cyclohexene oxide, giving only traces of polycyclohexene car-
The presence of substituents on both sides of the epoxide,
bonate. This low reactivity towards cyclohexene oxide is due to without a constrained bicyclic structure, is disadvantageous as
a poor solubility of the complexes in the CO2–CHO mixture as shown with 2,3-dimethyl-2,3-epoxybutane (entry 8), where no
observed at the end of the test, after having opened the auto- conversion was noticed. Leaving aside purely steric consider-
claves (close to no coloration of the reaction mixtures + ations, “monocyclic” poly-substituted epoxides lack the sup-
unchanged catalysts). For comparison, the results of the plementary drive provided by the release of the strain energy
screening tests run with propylene oxide are listed in Table 7. contained in a bicyclic structure like, e.g., in cyclohexene
The substitution of the axial ligands by bromides leads to an oxide, and hence display very low conversions.41
increase of the catalytic activity with conversions up to 93%
To evaluate the robustness of the catalyst system, we carried
(entry 1), generally higher than those obtained with the cobalt- out a typical recycling experiment with PO and catalyst 23
chloride complexes. The cobalt-bpb-bromide complexes are (Fig. 4; reaction conditions: 10 ml of epoxide, 20 h, 0.5 mol%
highly efficient as shown with tests run with a catalyst loading catalyst, 80 °C, 35 bar of CO2, five consecutive runs). After the
as low as 0.05 mol% (entries 3, 6, 9 and 12). The high successful conversion of the epoxide, the cyclic carbonate was
efficiency of cobalt bromide complexes is well documented separated by vacuum distillation and the catalyst directly used
with e.g. bromo-cobalt salen complexes12a which displayed a for the next cycle. The minimal loss of the catalyst observed
high reactivity, producing however polypropylene carbonate in during the test was compensated for by adjusting the amount
high yields at room temperature. Slight differences in the cata- of propylene oxide in order to keep the epoxide-to-catalyst ratio
lytic activity were found, indicating the positive influence on equal (1 : 200). According to TGA data, catalyst 23 remains
1
the reactivity of electron-withdrawing substituents. This behav- stable up to 250 °C. Examination of the re-used catalyst via H
ior was also found for cobalt-bromide complexes newly investi- NMR and IR spectroscopy showed that the catalyst remained
gated by Gosh et al. with a similar variation of the ligand- unchanged as long as the distillation of propylene carbonate
backbone.37
was performed below 150 °C (for TGA of catalyst 23 and
Considering more generally the axial ligands/leaving groups spectra see ESI†). These preliminary results are promising and
present in the cobalt bpb complexes the following reactivity show that bpb catalysts can be reused easily without loss of
trend could be found for propylene oxide: Br > Cl > OAc.
In order to evaluate the limits of these new bpb complexes,
activity in the coupling reaction.
Mechanism proposal, influence of the substituents on reac-
we performed a somewhat broader catalytic screening with tivity. By varying minimally the structure of the ligand frame-
catalyst 10 using a range of commercially available epoxides. work with electron-donating or -withdrawing substituents, a
The results of the screening are listed in Table 8, the reactivity perceptible effect on the catalytic activity could be noticed and
trends being for the most part in agreement with the litera- correlated with the nature of metals and axial ligands used in
ture.9,10 The experiments were performed at 80 °C and a cata- the complexes. According to numerous mechanisms found in
lyst concentration of 0.2 mol%. The catalyst showed a high the literature,7–12a a general mechanism matching the findings
catalytic activity with yields up to 99% (entry 7) forming with of the study can be proposed (Fig. 3). The first step in the for-
each tested epoxide, except 2,3-dimethyl-2,3-epoxybutane, an mation of an organic carbonate involves basically the acti-
organic carbonate. A copolymer as a single product was vation of the epoxide and its ring opening to form a reactive
3294 | Dalton Trans., 2014, 43, 3285–3296
This journal is © The Royal Society of Chemistry 2014

View Article Online
Dalton Transactions
Paper
substituents can be neglected). Chromium(III) can be con-
sidered as a Lewis acid (LA) a tad harder than cobalt(III),46
while chloride can be seen as a Lewis base (LB) somewhat
harder than acetate. The Cr(III)–chloride combination demon-
strates a standard “hard acid–hard base” match. Electron-
donating substituents would force electrons in the ligand
causing the metal acidic center to be “less” hard; this would
“destabilize” the chromium chloride complex enhancing its
reactivity. A parallel can be drawn to the cobalt(III) acetate com-
plexes which display the same trend, the Co(III)–acetate pair
being however a tad softer than the Cr(III)–chloride one. Along
these lines, the higher reactivity of the Co(III)–chloride–bpb
complexes suggests also that hardness of Co(III) and acetate
match better (higher stability of the complex) than in the case
of Co(III) and chloride. Considering now the contradicting case
of the bpb-cobalt-chloride complexes, the acid–base HSAB
match is not as optimal as for the Cr(III)–chloride pair, the
Co(III) ion being softer. Electron-withdrawing substituents
would cause the metal center to be harder, allowing a more
reactive LA–LB match.
Scheme 3 Tentative mechanism for a metal–bpb catalysed coupling
reaction.
metal-alkoxide, the actual active species in the catalytic cycle.
This happens most likely in the case of the ionic bpb com-
plexes with the concomitant help of the “integrated” cocatalyst
in the nucleophilic attack on the epoxide. The ammonium
cation would play a kind of dual role: on the one hand, in the
formation of a coordinative unsaturated complex, able to coor-
dinate/activate the epoxide. On the other hand, the so-formed
ammonium halides or -acetate can play a role in a nucleophilic
attack on the epoxide.42 Although the mechanism depicted in
Scheme 3 implies a mono-molecular activation, a mechanism
involving two metallic centers, as found in some related salen
systems, cannot be discarded yet.43,44 The second step of the
reaction would be the formation of a carbonate linkage via the
CO2 insertion into the reactive metal–alkoxide bond. This
intermediate can either form a cyclic monomer via intramole-
cular rearrangement or a polycarbonate through further alter-
nating insertions of epoxide and carbon dioxide molecules
(the formation of cyclic carbonate via backbiting of the copoly-
mer is left aside in the scheme). The nucleophilicity of the
“liberated” halide also plays an important role in the opening
of the epoxide. The remaining axial ligand might also play a
role via an effective “trans effect”,10,12 either in the epoxide
coordination/activation step or in influencing the stability of
the metal–alkoxide bond.
For the cobalt-halide complexes a positive influence of the
electron-withdrawing groups was found in agreement with the
work of Ghosh and co-workers,37 which can be likely explained
with a better activation of the epoxide, the metal center being
“more” electrophilic.
The chromium(III)-chloride-complexes display on the other
hand a reverse influence of the substitution, with an increase
of the catalytic activity using electron-donating substituents.
Without getting deeper into frontier molecular orbital theories
and molecular modelling, the “Hard Soft Acid Base” prin-
ciple45 might explain the observed trends (owing to the good
solubility of the catalysts, solubility effects depending on the
Unfortunately in the case of iron(III) a substituent/reactivity
correlation is not so apparent, weakening to some extent
this LA–LB approach. Both types of substitutions had however
a positive effect on the reactivity compared to the “bare”
ligand.
Conclusions
We have synthesized new iron, chromium and cobalt com-
plexes displaying ligands with a N,N-bis(2-pyridine-carbox-
amide)-1,2-benzene framework displaying either electron
donating- or electron withdrawing substituents. These ionic
complexes encompass tetraalkylammonium counter-ions as
potential cocatalysts. Some of the new complexes: cobalt/
acetate/bpb-Cl2 complex 6, cobalt/bromo/bpb-NO2 complex 22
and cobalt/bromo/bpb-Me2 23 could be for the first time struc-
turally characterized via X-ray diffraction. The complexes were
tested towards the coupling of epoxides and carbon dioxide
and compared with the results obtained with the “parent com-
pounds”.18 The screening showed that some of the ionic cata-
lysts with substituted linkers actually work more efficiently
than those with standard linkers, leading in some cases selec-
tively to pure alternating copolymers in the case of cyclohexene
oxide and to high amounts of monomeric cyclic carbonate
with terminal epoxides (e.g. PO). A broader catalytic screening
suggests in agreement with the literature that terminal epo-
xides displaying electron-withdrawing functional groups are
favored with this sort of catalysts. In addition a recycling
screening indicates that bpb catalysts can be reused easily
without loss of activity in the coupling reaction. Based on the
data gathered in this study, we could also observe some reactiv-
ity trends relative to the substituents located at the aromatic
linker, the nature of the axial ligand and of the metal and pro-
posed a tentative mechanism.
This journal is © The Royal Society of Chemistry 2014
Dalton Trans., 2014, 43, 3285–3296 | 3295

View Article Online
Paper
Dalton Transactions
22 X.-G. Zhou, J.-L. Zuo, L. Yang, R.-N. Wei and R. Li, J. Mol.
Catal. A: Chem., 2007, 266, 284.
Acknowledgements
We thank the Helmholtz Research School Energy-Related 23 Y. Kim, Y. W. Choi, S. H. Kim, D. N. Lee and C. Kim, Acta.
Catalysis for financial support.
Cryst. E, 2006, 62, m2715.
24 Bruker, Bruker AXS Inc., Madison, Wisconsin, USA, 2007;
G. M. Sheldrick, Acta Crystallogr., Sect. A: Fundam. Crystal-
logr., 2008, 64, 112; L. Zsolnai, XPMA, University of Heidel-
berg, Germany, 1996.
Notes and references
25 C. R. Groom and F. H. Allen, Wiley Interdiscip. Rev.:
Comput. Mol. Sci., 2011, 1, 368.
1 W. Leitner, Appl. Organomet. Chem., 2000, 14, 809.
2 http://www.novomer.com,
http://www.empowermaterials.
26 MERCURY 3.0 (Build RC5); http://www.ccdc.cam.ac.uk/
mercury
27 D. J. Darensbourg, J. L. Rodgers, R. M. Mackiewicz and
A. L. Phelps, Catal. Today, 2004, 98, 485.
28 D. J. Darensbourg and S. B. Fitch, Inorg. Chem., 2008, 47,
11868.
com/
3 M. Zhang, B. Liu, L. Chen and A. Yu, Macromol. Rapid
Commun., 2002, 23, 881.
4 J. H. Clements, Ind. Eng. Chem. Res., 2003, 42, 663.
5 S. Inoue, H. Koinuma and T. Tsuruta, J. Polym. Sci., Part B:
Polym. Lett., 1969, 7, 287.
29 W. Zhang, B. Y. Liu, C.-Y. Tian, L. Zhang and W. Yan,
J. Polym. Sci., Part A: Polym. Chem., 2006, 44, 6243.
30 O. Q. Munro, S. C. Shabalala and N. Brown, Inorg. Chem.,
2001, 40, 3303.
6 S. Inoue, H. Sugimoto and H. Ohtsuka, J. Polym. Sci., Part
A: Polym. Chem., 2004, 42, 5561.
7 G. W. Coates and D. R. Moore, Angew. Chem., Int. Ed., 2004,
43, 2.
31 C. K. Williams, F. Jutz, A. Buchard, M. R. Kember and
S. B. Fredriksen, J. Am. Chem. Soc., 2011, 133, 17395.
32 K. Nozaki, K. Nakano and T. Hiyama, Macromolecules,
2001, 34, 6325.
33 G. W. Coates, M. Cheng, N. A. Darling and E. B. Lobkovsky,
Chem. Commun., 2000, 2007.
34 K. Kuroda and H. Sugimoto, Macromolecules, 2008, 41, 312.
35 M. A. Fuchs, T. A. Zevaco, E. Ember, O. Walter, I. Held,
M. Döring and E. Dinjus, Dalton Trans., 2013, 42, 5322.
36 C. K. Williams, A. Buchard, M. R. Kember and
K. G. Sandeman, Chem. Commun., 2011, 47, 212.
37 A. Ghosh, P. Ramidi, N. Gerasimchuk, Y. Gartia and
C. M. Felton, Dalton Trans., 2013, 42, 13151.
38 G.-P. Wu, S.-H. Wei, W.-M. Ren, X.-B. Lu, T.-Q. Xu and
D. J. Darensbourg, J. Am. Chem. Soc., 2011, 133, 15191.
39 G.-P. Wu, S.-H. Wei, W.-M. Ren, X.-B. Lu, B. Li, Y.-P. Zu and
D. J. Darensbourg, Energy Environ. Sci., 2011, 4, 5084.
40 N. D. Harrold, Y. Li and M. H. Chisholm, Macromolecules,
2013, 46, 692.
8 D. J. Darensbourg, Chem. Rev., 2007, 107, 2388.
9 C. K. Williams, M. R. Kember and A. Buchard, Chem.
Commun., 2011, 47, 141.
10 P. P. Pescarmona and M. Taherimehr, Catal. Sci. Technol.,
2012, 2, 2169.
11 X.-B. Lu, W.-M. Ren and G.-P. Wu, Acc. Chem. Res., 2012,
45, 1721.
12 (a) G. W. Coates, C. T. Cohen and T. Chu, J. Am. Chem. Soc.,
2005, 127, 10869; (b) C. T. Cohen, C. M. Thomas,
K. L. Peretti, E. B. Lobkovsky and G. W. Coates, Dalton
Trans., 2006, 237.
13 S. Inoue and N. Takeda, Makromol. Chem., 1978, 179, 1377.
14 D. J. Darensbourg and S. B. Fitch, Inorg. Chem., 2007, 46,
5474.
15 B. Rieger, J. B. Dengler, M. W. Lehenmeier, S. Klaus,
C. E. An-derson and E. Herdtweck, Eur. J. Inorg. Chem.,
2011, 336.
16 K. D. Janda and T. S. Reger, J. Am. Chem. Soc., 2000, 122,
6929.
41 K. Morokuma, Z. Liu and M. Torrent, Organometallics,
2002, 21, 1056.
42 X.-B. Lu, X.-J. Feng and R. He, Appl. Catal., A, 2002, 234, 25.
43 X.-B. Lu and D. J. Darensbourg, Chem. Soc. Rev., 2012, 41,
1462.
44 R. L. Paddock and S. T. Nguyen, J. Am. Chem. Soc., 2001,
123, 11498.
45 R. G. Pearson, J. Chem. Educ., 1968, 45, 581.
46 R. G. Pearson, Inorg. Chem., 1988, 27, 734.
17 K. Yu and C. W. Jones, Organometallics, 2003, 22, 2571.
18 M. Adolph, T. A. Zevaco, O. Walter, E. Dinjus and
M. Döring, Polyhedron, 2012, 48, 92.
19 (a) N. Yamazaki and F. Higashi, Tetrahedron, 1974, 30,
1323; (b) R. S. Vagg, D. J. Barnes, R. L. Chapman and
E. C. Watton, J. Chem. Eng. Data, 1978, 23, 349.
20 R. N. Mukherjee and M. Ray, Polyhedron, 1992, 11, 2929.
21 R. N. Mukherjee, M. Ray, J. F. Richardson and
R. M. Buchanan, J. Chem. Soc., Dalton Trans., 1993, 2451.
3296 | Dalton Trans., 2014, 43, 3285–3296
This journal is © The Royal Society of Chemistry 2014