Influence of the metal (Al, Cr, and Co) and substituents of the porphyrin in controlling reactions involved in copolymerization of propylene oxide and carbon dioxide by porphyrin metal(III) complexes. 3. Cobalt chemistry

Article abstract of DOI:10.1021/ic400068y

A series of cobalt(III) complexes LCoX, where L = 5,10,15,20- tetraphenylporphyrin (TPP), 5,10,15,20-tetrakis(pentafluorophenyl)porphyrin (TFPP), and 2,3,7,8,12,13,17,18-octaethylporphyirn (OEP) and X = Cl or acetate, has been investigated for homopolymerization of propylene oxide (PO) and copolymerization of PO and CO₂ to yield polypropylene oxide (PPO) and polypropylene carbonate (PPC) or propylene carbonate (PC), respectively. These reactions were carried out both with and without the presence of a cocatalyst, namely, 4-dimethylaminopyridine (DMAP) or PPN⁺Cl^- (bis(triphenylphosphine)iminium chloride). The PO/CO₂ copolymerization process is notably faster than PO homopolymerization. With ionic PPN⁺Cl^- cocatalyst the TPPCoOAc catalyst system grows two chains per Co center and the presence of excess [Cl^-] facilitates formation of PC by two different backbiting mechanisms during copolymerization. Formation of PPC is dependent on both [Cl^-] and the CO₂ pressure employed (1-50 bar). TPPCoCl and PO react to form TPPCo(II) and ClCH₂CH(Me)OH, while with DMAP, TPPCoCl yields TPPCo(DMAP)₂⁺Cl^-. The reactions and their polymers and other products have been monitored by various methods including react-IR, FT-IR, GPC, ESI, MALDI TOF, EXAFS, and NMR (¹H, ¹³C{¹H}) spectroscopy. Notable differences are seen in these reactions with previous studies of (porphyrin)M(III) complexes (M = Al, Cr) and of the (salen)M(III) complexes where M = Cr, Co.

Full text of DOI:10.1021/ic400068y

Article
pubs.acs.org/IC
Influence of the Metal (Al, Cr, and Co) and Substituents of the
Porphyrin in Controlling Reactions Involved in Copolymerization of
Propylene Oxide and Carbon Dioxide by Porphyrin Metal(III)
Complexes. 3. Cobalt Chemistry
Chandrani Chatterjee,† Malcolm H. Chisholm,*,† Adnan El-Khaldy,† Ruaraidh D. McIntosh,†
Jeffrey T. Miller,‡ and Tianpin Wu‡
†Department of Chemistry, Biochemistry, The Ohio State University, Columbus, Ohio 43210, United States
‡Chemical Sciences and Engineering Division, Argonne National Laboratory, 9700 South Cass Avenue, Building 200, Argonne,
Illinois 60429-4837, United States
S
* Supporting Information
ABSTRACT: A series of cobalt(III) complexes LCoX, where
L = 5,10,15,20-tetraphenylporphyrin (TPP), 5,10,15,20-
tetrakis(pentafluorophenyl)porphyrin (TFPP), and
2,3,7,8,12,13,17,18-octaethylporphyirn (OEP) and X = Cl or
acetate, has been investigated for homopolymerization of
propylene oxide (PO) and copolymerization of PO and CO2
to yield polypropylene oxide (PPO) and polypropylene
carbonate (PPC) or propylene carbonate (PC), respectively.
These reactions were carried out both with and without the
presence of a cocatalyst, namely, 4-dimethylaminopyridine
(DMAP) or PPN+Cl− (bis(triphenylphosphine)iminium chloride). The PO/CO2 copolymerization process is notably faster than
PO homopolymerization. With ionic PPN+Cl− cocatalyst the TPPCoOAc catalyst system grows two chains per Co center and
the presence of excess [Cl−] facilitates formation of PC by two different backbiting mechanisms during copolymerization.
Formation of PPC is dependent on both [Cl−] and the CO2 pressure employed (1−50 bar). TPPCoCl and PO react to form
+
TPPCo(II) and ClCH2CH(Me)OH, while with DMAP, TPPCoCl yields TPPCo(DMAP)2 Cl−. The reactions and their
polymers and other products have been monitored by various methods including react-IR, FT-IR, GPC, ESI, MALDI TOF,
EXAFS, and NMR (1H, 13C{1H}) spectroscopy. Notable differences are seen in these reactions with previous studies of
(porphyrin)M(III) complexes (M = Al, Cr) and of the (salen)M(III) complexes where M = Cr, Co.
and thus PPC must be formed under kinetic conditions.10 This
provides quite a challenge to the chemist since backbiting
reactions such as those depicted in Scheme 2 can lead to
formation of PC by degradation of a PPC chain.
In addition, in considering the reactions shown in Scheme 1,
we note that insertion of CO2 into the alkoxide bond must
occur efficiently; furthermore, since this is typically a reversible
reaction the metal−alkylcarbonate must be kinetically more
reactive than the metal−alkoxide bond in ring opening or
enchaining the epoxide for PPC formation.
INTRODUCTION
■
Use of CO2 in the production of polycarbonates, as shown in
eq 1, in contrast to the use of diphenylcarbonate, phosgene, or
one of its derivatives,1 is an attractive alternative based on atom
efficiency, cost, and environmental considerations.
There are now many reports and some excellent reviews
dealing with the use of coordination metal complexes that will
catalytically combine CO2 and epoxides to the respective cyclic
carbonate or the polycarbonate.2−9 Reactions involving CO2
and propylene oxide (PO) are particularly sensitive to the
nature of the metal and reaction conditions. As shown in
Scheme 1 there are three competing pathways leading to
polypropylene oxide (PPO), the cyclic carbonate, propylene
carbonate (PC), and the polymer, polypropylene carbonate
(PPC). In reactions A and B, PC is the thermodynamic product
Despite these rather stringent requirements, catalytic PPC
formation has been achieved with turnover frequencies (TOFs)
of ∼103 h−1.11−17 In addition, in the ring opening of PO by the
carbonate this has been achieved with both regio- and
stereoselectivity.5,6 13C{1H} NMR spectroscopy of the
carbonate signal is typically employed in determination of
Received: January 10, 2013
Published: April 4, 2013
© 2013 American Chemical Society
4547
dx.doi.org/10.1021/ic400068y | Inorg. Chem. 2013, 52, 4547−4553

Inorganic Chemistry
Article
phyrin aluminum chloride as an initiator, TPPAlCl.19,20 This
was a very poor catalyst system in terms of TOFs. (Salen)AlCl,
however, has been shown to be active in this coupling of PO
and CO2 to give PC but not PPC.21−23
Scheme 1. Competing Reactions, A, B, and C Involved in the
PO/CO2 Copolymerization Processes
a
We were attracted to use of the planar porphyrin ligands and
wished to compare the reactivity of the trivalent metal
complexes LMX, where L is one of the porphyrins shown in
Figure 1 and the metal, M, is either Al, Cr, or Co in order to
address the question what makes one preferable over the other.
In this paper we describe our studies of cobalt chemistry and
make a comparison to our earlier work with aluminum and
chromium.24,25
RESULTS AND DISCUSSION
■
Homopolymerization of PO. In the absence of a
cocatalyst TPPCoCl and PO show little if any affinity toward
either homopolymerization or in the presence of both PO and
CO2 to yield PC or PPC. However, a reaction between
TPPCoCl and PO does occur leading to formation of
TPPCo(II). Both TPPCoCl and TPPCo(II) are paramagnetic
materials in solution, and the latter is notably less soluble in
chlorinated solvents CH2Cl2 or CHCl3 from which it
precipitates as a fine powder. From this reaction we detected
formation of the organic molecule 1-chloro-2-propanol
(HOCHMeCH2Cl), which is presumably formed by decom-
position of the initially formed TPPCoOCHMeCH2Cl. The
related TPPAlOCHMeCH2Cl was previously identified in the
reaction between TPPAlCl and PO in CDCl3.26 Interestingly,
we found no evidence of formation of either an aldehyde or a
ketone in these reactions. Either of these molecules should be
easily detected by the presence of the strongly infrared active
ν(CO) group, but no such IR band was found in the product
of the reaction between TPPCoCl and PO. Decomposition of a
metal−alkoxide via β-hydrogen elimination is well docu-
mented,27 and we initially suspected that formation of Co(II)
might have occurred by a kinetically labile TPPCoH.
a
A yields PPC, B yields PC, and C yields PPO.
a
Scheme 2. Possible Backbiting Reactions To Produce PC
a
P and P′ represent growing polymer chains.
When the reaction between TPPCoCl and PO was carried
out in the presence of 1,3-cyclohexadiene, we observed
PPC microstructure at the triad sensitivity level. See Chart 1 for
the regiosequences.
1
formation of benzene by H NMR spectroscopy. This suggests
that formation of the alcohol may well occur in CH2Cl2 and
CHCl3 via a radical pathway in which the alkoxide radical
abstracts a hydrogen atom from the solvent. Irrespective of the
mechanism of this reaction it is readily apparent that
TPPCoOCHMeCH2Cl is notably more susceptible to reduc-
tion than are the related chromium and aluminum analogs.
Indeed, all our attempts to isolate the initially formed
TPPCoOCHMeCH2Cl resulted in obtaining TPPCo(II) as
determined by EXAFS, vide infra.
Chart 1. Possible Regiosequences in PPC
Even in the presence of added DMAP or PPN+Cl− the
(porphyrin)CoCl systems fail to homopolymerize PO. Addition
of PPN+Cl− leads to formation of the paramagnetic
−
PPN+TPPCoCl2 species, while DMAP reacts to form a
diamagnetic complex TPPCo(DMAP)2 Cl− in CDCl3 or
+
CD2Cl2. This is the sole product in the presence of two or
+
more equivalents of DMAP, and the cation TPPCo(DMAP)2
1
was characterized by EXAFS and H NMR spectroscopy. With
1 equiv of DMAP the equilibria involving TPPCoCl and
+
To date of all the coordinate catalysts employed those
involving the Schiff base salen ligands have been the most
efficient, and of these cobalt(III) has been the best with
chromium(III) in second place.5,15,16,18
The first report of the copolymerization of CO2 and PO by a
trivalent metal was by Inoue, who employed tetraphenylpor-
TPPCo(DMAP)2 Cl− are presumably present, along with
TPPCoCl(DMAP), though the latter compound was not
1
detected by H NMR spectroscopy.
Rather interestingly TPPCoOAc in the presence of either
DMAP or PPN+Cl− will homopolymerize rac-PO to give
regioregular PPO with isotactic enrichment (see Figure SI 1,
4548
dx.doi.org/10.1021/ic400068y | Inorg. Chem. 2013, 52, 4547−4553

Inorganic Chemistry
Article
Figure 1. (porphyrin)Co(III) catalysts employed in this study.
Supporting Information). The reaction is notably slower than
that of the chromium system, and of the three (porphyrin)-
Co(III) systems employed in this work TPP is the most active:
TOFs observed are 4 h−1 for OEP, 1 h−1 for TFPP, and 24 h−1
for TPP systems. The activity is somewhat comparable to that
of the TPPAl(III) system studied earlier. This finding indicates
that the initially formed TPPCoOCHMeCH2OAc and the
subsequent alkoxide attached to the cobalt center in the
growing chain TPPCoOCHMeCH2OP, where OP = (OCH-
MeCH2)nOAc, are less prone to decomposition with formation
of Co(II) than the chloro−alkoxide TPPCoOCHMeCH2Cl.
Although it is not necessarily fair to draw conclusions based
on the chemistry of (porphyrin)CoCl, we are inclined to the
view that certain (salen)CoX initiators may be prone to
reduction to Co(II).5,15,28 We note, for example, that Coates in
his studies of the copolymerization of PO and CO2 found that,
under equivalent reaction conditions, certain Co−X initiators
gave varying molecular weights of PPC and yet had very similar
PDI values (PDI = polydispersity index).16 A plausible
explanation is that some were prone to early decomposition
to inactive (salen)Co(II) in their reactions with PO.
(50 bar) react catalytically to form exclusively PC as noted by
earlier co-workers.30,31
A high pressure of CO2 (50 bar) favors PPC formation.
Reactions employing TPPCoOAc, PO, and 0.5 equiv of DMAP,
PPN+Cl−, or PPN+OAc− with CO2 (50 bar) are active in
formation of near regioregular PPC with HT junctions, see
Figure 2. MALDI TOF spectroscopic analysis of the PPC is
Finally, we address the question of whether or not the
TPPCo(III) system is capable of growing two chains in the
presence of an anionic cocatalyst initiator.
When TPPCoOAc and PO are allowed to react in the
presence of PPN+Cl− the molecular weight of the polymer
decreases as the PPN+Cl− concentration is increased from 0.5
to 1.0 equiv in a manner expected for two growing chains (Mn
= 4200 Da for 0.5 equiv and Mn = 3000 Da for 1.0 equiv of
PPN+Cl−). In addition, in the polymerization of PO by
TPPCoOAc we observe by ESI MS and FT-IR spectroscopy
polymer chains terminated in acetate together with some OH
end groups (see Figure SI 2, Supporting Information). The
latter arise from adventitious H2O. When to this living system
PPN+Cl− is introduced we observe chloride end groups in
addition to OAc and OH, as shown in Figure SI 3, Supporting
Information, and a bimodal polymer molecular weight by GPC.
These results parallel those we found for the TPPCr(III)
system25 and are consistent with proposals by Darensbourg and
Lu and Coates and their respective co-workers who invoke two
growing chains with (salen)Co(III) catalyst systems.7,9,16
Copolymerization of PO and CO2. TPPCoCl in the
presence of 1.0 equiv of PPN+Cl− does not provide an active
system for formation of PPC. PC is formed along with some
decomposition to Co(II). Formation of PC probably reflects
the favorable backbiting of the (TPP)CoO2COCHMeCH2Cl
moiety with regeneration of the Co−Cl bond as noted
elsewhere for related chromium and aluminum systems.25,29
In addition, at higher temperatures, TPPCoCl, PO, and CO2
Figure 2. 13C{1H} of the carbonate in PPC showing near-perfect
formation of the regio regular HT junction. HT′ has either an ether-
rich carbonate or an end group.
shown in Figure SI 4, Supporting Information. Interestingly, as
the CO2 pressure is reduced PC formation increases relative to
PPC until at 1 atm CO2 PC is predominantly formed.
In the chemistry of (salen)Co(III) complexes the need for
relatively high pressures of CO2 was noted early on by Coates
and co-workers.15,16 Our interpretation of these observations is
that higher pressures of CO2 are required to drive the
equilibrium between the Co−alkoxide and the Co−alkylcar-
bonate in favor of the latter. Unless this equilibrium favors the
alkylcarbonate the metal−alkoxide bond will be capable of
enchaining PO to give ether-rich PPC or backbiting the PPC
chain to give PC. Since the homopolymerization of PO is
relatively slow for Co(III) systems we do not see ether-rich
PPC but rather formation of PC. It is, however, also worth
noting that cobalt is like aluminum in favoring the ring opening
of PO by the alkylcarbonate bond relative to the cobalt−
alkoxide bond.24 This is in stark contrast to the chemistry of
chromium(III) where enchainment of PO is [Cr]−OR >
[Cr]−O2COR.25
4549
dx.doi.org/10.1021/ic400068y | Inorg. Chem. 2013, 52, 4547−4553

Inorganic Chemistry
Article
Comparative data for the three porphyrin systems are given
in Table 1 along with some relative data for the chemistry of
CH2Cl2. In this sample 2 axial Co−N bonds were observed, in
addition to 4 Co−N bonds from the TPP. The Co−N region
was fit to be 6 Co−N. Assuming that the 4 Co−N from the
TPP do not change (XRD shows that the bonds are slightly
larger when coordinated to an axial ligand) the bond distance of
the 2 new axial Co−N was ∼2.04 Å or about 0.12 Å longer than
the Co−N TPP bond distances. The energy of the Co K-edge
XANES pre-edge was 7.7095 keV, consistent with Co(III) (see
Figure SI 9, Supporting Information).
Table 1. TOFs Observed with (porphyrin)M(III)X
a
Catalysts
b
equiv of
TOF
%
c
d
d
entry
catalysts
PPN+Cl−
(h−1
20
)
HT
Mn
PDI
1.30
1
2
3
4
5
6
7
8
TPPCoCl
TPPCoCl
TPPCoOAc
OEPCoOAc
TFPPCoOAc
TPPAlCl
0.5
1.0
0.5
0.5
0.5
0.5
0.5
0.5
96 4900
e
28
TPPCo(H2O)2 Cl−. We are not aware of any structure of a
+
140
12
96 8500
86 3100
1.30
1.20
five-coordinate TPPCoCl, though TPPCoCl(H2O) is known
where H2O occupies the position trans to the Co−Cl bond.
Given the preparation of TPPCoCl reported33 and employed in
this study it is likely that TPPCoCl(H2O) is the correct
formation. We were not able to obtain single crystals of our
porphyrin cobalt chloride suitable for single-crystal X-ray
studies, so we again resorted to the use of EXAFS on a
powdery microcrystalline sample (see Figure SI 8, Supporting
Information). As we describe below, this sample did not prove
to have a Co−Cl bond but rather was found to be Co(III) with
four Co−N bonds in the equatorial plane assignable to the
porphyrin ligand and two longer axial bonds which could be
assignable to the light atoms O or N. The fit of all 6 Co−N
bonds was 2.00 Å. Assuming that the 4 Co−N bonds from TPP
remained unchanged, the axial Co−N bond distance was
estimated to be 2.16 Å. Given the preparation and that analysis
clearly was consistent with TPPCoCl(H2O)x we believe that
e
4.8
73
75 12 000
81 19 000
1.30
1.20
TFPPAlCl
TPPCrCl
120
1600
f
22 000,
8 100
1.10, 1.20
9
TPPCrCl
1.0
150
85 13 000
1.30
a
Reaction conditions: 0.85 mL of PO (12 mmol.), 50 bar of CO2, 0.02
mmol of catalyst, 0.01 mmol of additives when used, reaction time 5 h
b
for entries 4 and 5 and 2.5 h for the rest, 25 °C. Calculated from 1H
NMR analysis of reaction aliquots, TOFs calculated as (mmol of PO
c
consumed)/(mmol of catalyst·reaction time). Calculated from 13C
d
NMR spectra of pure polymers. GPC analysis of pure copolymers
after 100% PO consumption using polystyrene standard. Calculated
maximum molecular weight 34.8 kDa (for 1 polymer chain growing
per metal center). Observed lower molecular weights are attributed to
e
chain transfer due to the presence of adventitious water. PC was
f
produced as the major product. Ether-rich copolymer was produced
+
this sample contained the pseudo-octahedral TPPCo(H2O)2
along with a distinctly bimodal Mn distribution.
cation, with a chloride counteranion. The presence of chloride
was confirmed by elemental analysis (see Experimental Section
for details).
Concluding Remarks. It is now possible to compare the
reactivities of the three different metals under comparable
conditions and offer some reasonable speculation concerning
these intriguing reactions involving PO and CO2 that can yield
three different products as shown in Scheme 1.
aluminum and chromium. Once again we see that TPP is the
preferred porphyrin ligand, and the reactivity of the metals is in
the order Cr > Co ≈ Al.
In the presence of >1.0 equiv of PPN+Cl−, PC formation
competes with PPC, and at 5.0 equiv of PPN+Cl− only PC is
formed. We also observed by in situ react IR spectroscopy that
addition of PPN+Cl− to a living system growing PPC chains
causes decomposition of PPC and formation of PC. This
behavior is similar but somewhat slower than what we observe
for chromium25 and we believe is due to displacement of the
carbonate-bound growing chain which facilitates backbiting.
Similar proposals of Cl−-promoted chain transfer leading to PC
formation have been invoked for (salen)Co(III) systems in
their reactions with CO2.9,29,32
1. The role of the cocatalyst be it DMAP, PPN+Cl−, or
imidazole, as was often employed in the early works of
Inoue, can be understood in terms of activating the trans
M−X bond. A good donor will lengthen the M−X bond
and also by being a Lewis base increase the polarity of
the M−X bond. This will enhance CO2 insertion into the
metal−alkoxide bond, as was nicely seen in the
equilibrium involving [Al]−OR + 13CO2 = [Al]−
O213CR for the various porphyrins upon addition of
DMAP.24 The longer M−O bond in the six-coordinate
complex will also allow the PO molecule, which is to be
enchained (ring opened), to more readily approach the
metal center. It could also be argued that the trans effect
promotes dissociation of the alkylcarbonate. We find no
evidence for free polycarbonate chains during formation
of PPC as our evidence involving Cl− displacement
favors formation of PC by backbiting. We do note,
however, that Darensbourg and Lu and co-workers found
a tetrahedral quaternarized HN+ group which was bound
to the salen ligand attached to Co(III) further labilized
the alkylcarbonate by N+−H···O carbonate bonding.7,9
This hydrogen bonding would lengthen the Co−O bond
of the carbonate and thus facilitate PO activation at the
electrophilic Co(III) metal center.
EXAFS Studies. TPPCo(II). The sample examined was
prepared from reaction between TPPCoCl and PO in the
presence of 1.0 equiv of DMAP. Addition of PO to this catalyst
system led to formation of TPPCo(II). The bond distance of
the Co−N axial ligand was estimated to be 2.02 Å. The
equatorial four Co−N bond distances in TPPCo(II) are 1.94 Å.
No evidence of the Co−Cl bond was found. The axial Co−N
could also be a Co−O since both scatter X-ray identically. The
Co oxidation state was determined from the energy of the pre-
edge peak and was 7.7085 keV, identical to that of Co(II)
reference compounds; see the Experimental Section on EXAFS
for the pre-edge energy position of Co(I), Co(II), and Co(III)
reference compounds. EXAFS and XANES spectra of these
reference compounds are shown in Figures SI 5 and 6,
Supporting Information, and those of TPPCo(II) and TPPCo-
(III) samples are shown in Figures SI 7 and 8, Supporting
Information.
+
TPPCo(DMAP)2 Cl−. Microcrystalline sample was prepared
2. In our earlier gas-phase studies of the binding of PO to
from reaction between TPPCoCl and DMAP (2.5 equiv) in
the TPPM+ centers34 aluminum bound one PO molecule
4550
dx.doi.org/10.1021/ic400068y | Inorg. Chem. 2013, 52, 4547−4553

Inorganic Chemistry
the most strongly as determined by Ar atom bombard-
Article
than a metal alkoxide. Why this reactivity order is
reversed for M = Cr is not clear, but it should be noted
that the Cr−O2COR group still enchains PO faster than
M−O2COR where M = Al and Co.
ment. However, chromium and to a lesser extent cobalt-
bound PO in a cooperative manner, thus favoring the six-
+
coordinate TPPM(PO)2 cation. This is not surprising.
Aluminum with z = 13 is the cation with the least
number of electrons to provide shielding of the positive
charge, while both Cr(III) and Co(III) enjoy a significant
ligand field stabilization for the six-coordinate pseudo-
octahedral ion. We see this in this work with our
5. The reactivity of any metal center is greatly influenced by
its attendant ligands in coordinate catalysis. Thus, it is
not necessarily surprising to find that salen metal(III)-
complexes for Cr(III) and Co(III) have similar
reactivities in their reactions with CO2 and epoxides
with cobalt being slightly more reactive. This is clearly
not so for the porphyrin systems.37,36 Just why this is so
is not clear.
6. The enchainment of PO in formation of PPO at a single
metal center is in many ways similar to growth of a
polyolefin chain in olefin polymerization.38 The sub-
strate, PO or alkene, is activated by coordination to an
electrophilic metal center. The polar metal−alkyl or
metal−alkoxide bond undergoes a 1,2-addition to the
activated substrate to regenerate the metal−alkyl or
metal−alkoxide bond. We can envisage the same for the
metal−alkylcarbonate, though in addition to a 1,2-
addition akin to the alkoxide, the alkylcarbonate could
react via a six-membered transition state. These
possibilities are shown in Scheme 3.
+
characterization of TPPCo(DMAP)2 and TPPCo-
+
(H2O)2 . It is also surely this preference for six
coordination that favors formation of two growing chains
for Cr and Co but not for Al.
3. In comparing the ring opening of PO we see that the
reactivity order is Cr > Al ≈ Co. We propose that this
reflects the fact that the Cr−O bond has the greatest
polarity Crδ+−Oδ+ for both the alkoxide and the
alkylcarbonate. In comparing Cr(III) and Co(III), we
note the well-accepted trend that as one proceeds from
left to right across the d-block transition elements the
electronegativity of the metals increase. [The group 11
metals, Cu, Ag, and Au, do not react with H+ to give H2.]
Thus, Co(III) is a better oxidizing agent than Cr(III), as
was elegantly demonstrated by Taube on the now classic
example of inner-sphere electron transfer CrII + CoIII−Cl
→ CrIII−Cl + CoII.35 In this work we see the reduction of
TPPCoCl to TPPCo in its reaction with PO. Similar
formation of TPPCrII would be highly unlikely, and no
reduction of the Cr(III) porphyrin complexes was
observed in their reactions with PO.25 In general, as
one proceeds across the periodic table the electro-
negativity of the elements increases, and so while
Ti3+(aq) is a reasonable reducing agent, Fe3+(aq) is a
mild oxidizing agent. Similarly, for the lighter element
aluminum the M(III)−O bond is likely to be less polar
than that of Cr(III), and thus, aluminum and cobalt are
rather similar in their reactivities. For both Cr and Co,
the reactivity order is TPP > TFPP ∼ OEP and the latter
is virtually inactive, but for Al, the activity order is TFPP
> TPP > OEP. This order appears to reflect upon the
electrophilicity of the metal centers and in the absence of
PPN+Cl− the TFPPMCl complexes where M = Al or Cr
react with PO to give regioirregular PPO. The poorer
reactivity of the OEP complexes probably arises from the
poorer Lewis acidity of the metal centers. The interplay
of these electronic effects is also seen in the recent report
by Rieger and co-workers on the reactivity of para
substituents on TPPCoCl with PO and CO2.36
Scheme 3. Enchainment of Coordinated PO by (a) M−
a
Alkoxide Bond and (b) M−Alkylcarbonate Bond
a
P and P′ represent growing polymer chains.
7. We note that at present no M(IV)- or M(I)-containing
complexes have been reported to copolymerize epoxides
and CO2. We speculate that transition metals in the +4
oxidation state may have too covalent M−OR bonds, i.e.,
their bonds are not sufficiently polar, while M(+) metals
may lack in electrophilicity and form M−μ-OR bonds
where the oxygen lone pairs are otherwise engaged in
interactions with the metal ions and are thus not available
to attack the substrate. Only a limited number of main
group and transition metal ions in the +2 and +3
oxidation states have been found to be active and for the
latter only transition metals of the first row. We also note
that for the M(2+) ions of zinc and magnesium only
relatively low coordinate and binuclear metal com-
plexes39−42 have proved active where the larger, softer
5-coordinate Ca2+ in the complex TpbutCa(OR)(PO)
binds PO but does not ring open it.43 This does not
mean that Ca2+ ions cannot activate PO in ring opening
as the early Union Carbide heterogeneous catalyst for
formation of polyethylene oxide and PPO employed
calcium.44 Thus, aside from the obvious requirement of
steric access to the metal center in order for substrate
activation it appears that a rather specific charge/radius
may be required for any Mn+ ion, leading to activation of
In addition, in comparing the reactivity of Cr(III) and
Co(III) we note that the d6 ion is more likely to react via
dissociative substitution when compared to the “half-
filled t2g”−d3 ion. Since the reaction between the growing
chain and PO occurs within the same hemisphere the
uptake of PO at the metal center by an associative
mechanism should be easier for Cr(III), which again
would lead to its greater reactivity.
4. The reactivity order with respect to ring opening of PO,
M−O2CR > M−OR for M = Al and Co, again
emphasizes the similarity in these metals in this system.
We speculate that this is due to the more covalent nature
of the metal−alkoxide bond which leads to its low
activity. The M−O bond in the alkyl carbonate will be
longer, more polar, and more like a metal carboxylate
4551
dx.doi.org/10.1021/ic400068y | Inorg. Chem. 2013, 52, 4547−4553

Inorganic Chemistry
PO along with an apparently polar adjacent metal−OR
Article
three polystyrene standards covering the molecular weight range from
1320 to 3.15 × 106 Daltons.
bond.
Mass Spectrometry. Matrix-assisted laser desorption/ionization
time-of-flight mass spectroscopy (MALDI-TOF MS) was performed
on a Bruker Microflex mass spectrometer provided by a grant from the
Ohio BioProducts Innovation Center. The spectrometer was operated
in a linear, positive ion mode with N2 laser. Laser power was used at a
threshold level required to generate a signal. The accelerating voltage
was 28 kV. Equal amounts of the dithranol matrix and sample were
dissolved in dichloromethane (approximately 2 mg/0.5 mL). A small
amount of this solution was spotted on the target plate and allowed to
dry at room temperature before use. Electrospray ionization mass
spectroscopy (ESI MS) was carried out in positive ion mode on a
Bruker MicrOTOF mass spectrometer provided by a grant from the
Ohio BioProducts Innovation Center.
FTIR Spectroscopy. IR spectra were recorded in dichloromethane
solution using a Perkin-Elmer Spectrum GX spectrometer at room
temperature. For polymer samples films were prepared by slow
evaporation of CH2Cl2 solution of the polymer on a circular NaCl
plate of 2 mm thickness and 25 mm diameter. An ASI react-IR 1000
Reaction Analysis System was used to monitor the PO and CO2
copolymerization reactions at atmospheric pressure of CO2.
Microanalysis. Microanalysis was performed by Atlantic Microlab
Inc
Hopefully the present findings described herein will not only
contribute to the rapidly developing body of work on the
copolymerization of epoxides and CO2 by coordinate catalysis
but will also stimulate further thought and research on these
reactions which could become technologically transformative in
production of polycarbonates by a greener, atom-efficient
process involving utilization of CO2.
EXPERIMENTAL SECTION
■
Reagents and Methods. All syntheses and solvent manipulations
were carried out under a dry nitrogen atmosphere using standard
Schlenk-line and drybox techniques. Dimethylformamide was dried
and distilled over calcium hydride and degassed before use. All other
solvents were dried and degassed by standard methods. Deuterated
solvents were stored over 4 Å molecular sieves for 24 h prior to use.
Propylene oxide (Sigma Aldrich) was distilled from calcium hydride.
All three porphyrins 5,10,15,20-tetraphenyl-21H,23H-porphine
(TPPH2), 5,10,15,20-tetrakis(pentafluorophenyl)-21H,23H-porphine
(TFPPH2), and 2,3,7,8,12,13,17,18-octaethyl-21H,23H-porphine
(OEPH2) (Frontier Scientific), anhydrous CoCl2 (Aldrich), and 99%
CO2 (OSU gas stores) were used as received. Bis-
(triphenylphosphine)iminium chloride((Ph3P)2N+Cl−) (Sigma Al-
drich) was dried under vacuum and stored inside a drybox before
use. 4-Dimethylaminopyridine (DMAP) (Aldrich) was sublimed and
stored inside a drybox before use. PPN+OAc− was prepared from
PPN+Cl− following a literature procedure.45
SQUID. Direct current magnetic susceptibility data were collected
at the University of Florida using a Quantum Design MPMS-XL
SQUID susceptometer equipped with a 7 T magnet and operating in
the 1.8−300 K range. Magnetization versus field data were fit using the
program MAGNET.46 Both TPPCoCl and TPPCoOAc were found to
be diamagnetic in the solid state, as shown in Figure SI 10, Supporting
Information.
NMR Spectroscopy. 1H and 13C{1H} NMR experiments were
carried out with a Bruker DPX-400 (5 mm broad band probe) and a
Bruker DRX-500 (5 mm broad band probe) spectrometer. All
chemical shifts are in ppm relative to the solvent chloroform-d at 7.24
Synthesis of Catalysts. The series of (porphyrin)Co(III)X
catalysts, where porphyrin = TPP, TFPP, OEP and X = chloride
and acetate, were prepared according to previously reported
procedures31,47−49 and characterized by 1H NMR, ESI-MS, and
elemental analysis. Anal. Calcd for TPPCoCl(H2O): C, 72.88; H, 4.17;
N, 7.73; Cl, 4.89. Found: C, 72.72; H, 4.10; N, 7.66; Cl, 5.15. Anal.
Calcd for OEPCoOAc: C, 70.14; H, 7.28; N, 8.61. Found: C, 70.05; H,
7.38; N, 7.53. Anal. Calcd for TFPPCoOAc: C, 50.66; H, 1.02; N,
5.14. Found: C, 50.34; H, 0.89; N, 5.30.
1
ppm for H NMR and 77.23 ppm for 13C{1H} NMR spectroscopy.
EXAFS Study. X-ray absorption spectroscopy data collection and
analysis: Co K-edge (7.7090 keV) X-ray absorption measurements
were conducted on the bending magnet beamline of the Materials
Research Collaborative Access Team (MRCAT, 10-ID) at the
Advanced Photon Source (APS), Argonne National Laboratory.
Ionization chambers were optimized for maximum current with linear
response (ca. 1010 photons detected per second) using N2 (10%
absorption) in the incident X-ray detector and a mixture of ca. 30% Ar
in N2 (60% absorption) in the transmission X-ray detector. A third
detector in the series simultaneously collected a Co foil reference
spectrum with each measurement for energy calibration. The X-ray
beam was 0.5 × 1.5 mm, and data was collected in transmission
geometry in 10 min at step scan mode.
The energy of the pre-edge in the Co K-edge XANES was used to
determine the oxidation state of the samples and calibrated from the
position of standards (ClCo(I)(PPh3)3 (7.7081 keV), Co(II)-
(Ac)2·2H2O (7.7086 keV), Co(II)(AcAc)2·H2O (7.7085 keV),
Co(II)Cl2 (7.7086 keV), Co(II)F2 (7.7086 keV), Co(III)(AcAc)3
(7.7094 keV), and LiCoO2 (7.7093 keV) all obtained from Sigma-
Aldrich). EXAFS fits were determined from experimental phase shift,
and backscattering amplitudes were obtained from the Co(NH3)6Cl3
(6 Co−N at 1.94 Å) and ClCo(PPh3)3 (1 Co−Cl at 2.25 and 3 Co−P
at 2.33 Å; i.e., 4 Co-X at an average bond distance of 2.31 Å) for Co−
Cl scattering. Standard procedures based on WINXAS 3.1 software
were used to fit XAS data. EXAFS coordination parameters were
obtained by a least-squares fit in q and r space of the first-shell nearest
neighbor, k2-weighted Fourier transform data.
+
1
(TPP)Co(DMAP)2 Cl−. H NMR (CDCl3, δ, ppm): 0.25 (CH,
doublet (DMAP)); 2.90 (CH3, singlet (DMAP)); 4.05 (CH, doublet
(DMAP)); 7.7 (12H, multiplet (phenyl)); 7.9 (8H, multiplet
(phenyl)); 9.0 (8H, singlet (pyrrole)).
ASSOCIATED CONTENT
* Supporting Information
■
S
13C{1H} NMR, MALDI TOF MS, ESI MS and FT-IR spectra
of polymers, EXAFS and SQUID data of the TPPCo(III)
complexes. This material is available free of charge via the
Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*E-mail: chisholm@chemistry.ohio-state.edu.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank the Department of Energy, Office of Basic Energy
Science, Chemistry Division for support of this work. Use of
the Advanced Photon Source was supported by the U.S.
Department of Energy, Office of Basic Energy Sciences, under
contract DE-AC02-06CH11357. MRCAT operations are
supported by the Department of Energy and the MRCAT
member institutions. Funding for J.T.M. and T.W. were
provided by the Chemical Sciences, Geosciences, and
Gel Permeation Chromatography. Gel permeation chromato-
graphic (GPC) analyses were carried out at 40 °C using a Waters
Breeze system equipped with a Waters 2414 refractive index detector
and a set of two columns, Waters Styragel HR-2 and HR-4 (7.8 × 300
mm). THF was used as the mobile phase at 1.0 mL/min. Samples
were filtered before analysis. Sample concentration was 1.0%, and
injection volume was 100 μL. The calibration curve was made with
4552
dx.doi.org/10.1021/ic400068y | Inorg. Chem. 2013, 52, 4547−4553

Inorganic Chemistry
Article
(31) Jin, L.; Jing, H.; Chang, T.; Bu, X.; Wang, L.; Liu, Z. J. Mol. Catal
A: Chem. 2007, 261, 262−266.
Biosciences Division, U.S. Department of Energy, under
contract DE-AC0-06CH11357. We also thank Dr. George
Christou and Annaliese Thuijs, at the Departments of
Chemistry and Physics, University of Florida, for the SQUID
studies. Dr. Adnan El-Khaldy was on leave from Alabama A&M
University.
(32) Darensbourg, D. J.; Wei, S.-H. Macromolecules 2012, 45, 5916−
5922.
(33) Sakurai, T.; Yamamoto, K.; Naito, H.; Nakamoto, N. Bull. Chem.
Soc. Jpn. 1976, 49, 3042−3046.
(34) Chen, P.; Chisholm, M. H.; Gallucci, J. C.; Zhang, X.; Zhou, Z.
Inorg. Chem. 2005, 44, 2588−2595.
(35) Laylin, J. K. Nobel Laureates in Chemistry, 1901−1992; Chemical
Heritage Foundation: Philadelphia, PA, 1993; pp 660−661.
(36) Anderson, C. E.; Vagin, S. I.; Xia, W.; Jin, H.; Rieger, B.
Macromolecules 2012, 45, 6840−6849.
(37) Sugimoto, H.; Kuroda, K. Macromolecules 2008, 41, 312−317.
(38) Kawai, K.; Fujita, T. In Metal Catalysts in Olefin Polymerization;
Guan, Z., Ed.; Topics in Organometallic Chemistry; Springer: Berlin,
Heidelberg, 2009; pp 3−46.
(39) Moore, D. R.; Cheng, M.; Lobkovsky, E. B.; Coates, G. W.
Angew. Chem., Int. Ed. 2002, 41, 2599−2602.
(40) Moore, D. R.; Cheng, M.; Lobkovsky, E. B.; Coates, G. W. J.
Am. Chem. Soc. 2003, 125, 11911−11924.
(41) Kember, M. R.; Williams, C. K. J. Am. Chem. Soc. 2012, 134,
15676−15679.
(42) Jutz, F.; Buchard, A.; Kember, M. R.; Fredriksen, S. B.; Williams,
C. K. J. Am. Chem. Soc. 2011, 133, 17395−17405.
(43) Chisholm, M. H.; Gallucci, J. C.; Phomphrai, K. Inorg. Chem.
2004, 43, 6717−6725.
(44) Chisholm, M. H.; Zhou, Z. J. Mater. Chem. 2004, 14, 3081−
3092.
REFERENCES
■
(1) LeGrand, D. G.; Bendler, J. T. Handbook of Polycarbonate Science
and Technology; CRC Press: Boca Raton, FL, 2000; pp 7−26.
(2) Darensbourg, D. J.; Holtcamp, M. W. Coord. Chem. Rev. 1996,
153, 155−174.
(3) Coates, G. W.; Moore, D. R. Angew. Chem., Int. Ed. 2004, 43,
6618−6639.
(4) Chisholm, M. H.; Zhou, Z. J. Mater. Chem. 2004, 14, 3081.
(5) Darensbourg, D. J. Chem. Rev. 2007, 107, 2388−2410.
(6) Kember, M. R.; Buchard, A.; Williams, C. K. Chem. Commun.
2011, 47, 141.
(7) Lu, X.-B.; Darensbourg, D. J. Chem. Soc. Rev. 2012, 41, 1462.
(8) Klaus, S.; Lehenmeier, M. W.; Anderson, C. E.; Rieger, B. Coord.
Chem. Rev. 2011, 255, 1460−1479.
(9) Lu, X.-B.; Ren, W.-M.; Wu, G.-P. Acc. Chem. Res. 2012, 45, 1721−
1735.
(10) Darensbourg, D. J.; Yarbrough, J. C.; Ortiz, C.; Fang, C. C. J.
Am. Chem. Soc. 2003, 125, 7586−7591.
(11) S, S.; Min, J. K.; Seong, J. E.; Na, S. J.; Lee, B. Y. Angew. Chem.,
Int. Ed. 2008, 47, 7306−7309.
(12) Na, S. J.; S, S.; Cyriac, A.; Kim, B. E.; Yoo, J.; Kang, Y. K.; Han,
S. J.; Lee, C.; Lee, B. Y. Inorg. Chem. 2009, 48, 10455−10465.
(13) Paddock, R. L.; Nguyen, S. T. Macromolecules 2005, 38, 6251−
6253.
(45) Darensbourg, D. J.; Pala, M.; Waller, J. Organometallics 1983, 2,
1285−1291.
(46) Davidson, E. R. MAGNET; Indiana University: Bloomington,
IN, 1999.
(47) Paolesse, R.; Licoccia, S.; Bandoli, G.; Dolmella, A.; Boschi, T.
Inorg. Chem. 1994, 33, 1171−1176.
(14) Ren, W.-M.; Zhang, X.; Liu, Y.; Li, J.-F.; Wang, H.; Lu, X.-B.
(48) Adler, A. D.; Longo, F. R.; Kampas, F.; Kim, J. J. Inorg. Nucl.
Chem. 1970, 32, 2443−2445.
Macromolecules 2010, 43, 1396−1402.
(15) Qin, Z.; Thomas, C. M.; Lee, S.; Coates, G. W. Angew. Chem.,
Int. Ed. 2003, 42, 5484−5487.
(49) Lauher, J. W.; Ibers, J. A. J. Am. Chem. Soc. 1974, 96, 4447−
4452.
(16) Cohen, C. T.; Chu, T.; Coates, G. W. J. Am. Chem. Soc. 2005,
127, 10869−10878.
(17) Darensbourg, D. In Synthetic Biodegradable Polymers; Rieger, B.,
Kunkel, A., Coates, G. W., Reichardt, R., Dinjus, E., Zevaco, T. A.,
̈
Eds.; Advances in Polymer Science; Springer: Berlin/Heidelberg,
2012; Vol. 245, pp 1−27.
(18) Cohen, C. T.; Coates, G. W. J. Polym. Sci., Part A: Polym. Chem.
2006, 44, 5182−5191.
(19) Inoue, S.; Koinuma, H.; Tsuruta, T. J. Polym. Sci., Polym. Lett.
1969, 7, 287−292.
(20) Inoue, S.; Koinuma, H.; Tsuruta, T. Makromol. Chem. 1969,
130, 210−220.
(21) Lu, X.-B.; Zhang, Y.-J.; Jin, K.; Luo, L.-M.; Wang, H. J. Catal.
2004, 227, 537−541.
(22) Lu, X.-B.; He, R.; Bai, C.-X. J. Mol. Catal A: Chem. 2002, 186,
1−11.
(23) Lu, X.-B.; Zhang, Y.-J.; Liang, B.; Li, X.; Wang, H. J. Mol. Catal
A: Chem. 2004, 210, 31−34.
(24) Chatterjee, C.; Chisholm, M. H. Inorg. Chem. 2011, 50, 4481−
4492.
(25) Chatterjee, C.; Chisholm, M. H. Inorg. Chem. 2012, 51, 12041−
12052.
(26) Chisholm, M. H.; Zhou, Z. J. Am. Chem. Soc. 2004, 126, 11030−
11039.
(27) Bradley, D. C. Alkoxo and Aryloxo Derivatives of Metals;
Academic Press, 2001; pp 27−30.
(28) Shen, Y.-M.; Duan, W.-L.; Shi, M. J. Org. Chem. 2003, 68,
1559−1562.
(29) Tian, D.; Liu, B.; Gan, Q.; Li, H.; Darensbourg, D. J. ACS Catal.
2012, 2, 2029−2035.
(30) Paddock, R. L.; Hiyama, Y.; McKay, J. M.; Nguyen, S. T.
Tetrahedron Lett. 2004, 45, 2023−2026.
4553
dx.doi.org/10.1021/ic400068y | Inorg. Chem. 2013, 52, 4547−4553