Synthesis of cobalt(II)-α-diimines complexes and their activity as mediators in organometallic mediated radical polymerization of vinyl acetate

Article abstract of DOI:10.1016/j.ica.2017.11.041

This study describes the synthesis and characterization of three α-diimine-cobalt complexes of the type [CoCl₂(R-DAB)] (R-DAB = R–N[dbnd]CH–CH[dbnd]N–R; R = Mes, Dipp and Dipp^?) and their application as mediators for the cobalt-media

Full text of DOI:10.1016/j.ica.2017.11.041

Accepted Manuscript
Research paper
Synthesis of cobalt(II)-α-diimines complexes and their activity as mediators in
organometallic mediated radical polymerization of vinyl acetate
Beatriz A. Riga, Marina D. Neves, Antonio E.H. Machado, Diesley M.S.
Araújo, Jhonathan R. Souza, Otaciro R. Nascimento, Vinícius T. Santana, Carla
C.S. Cavalheiro, Valdemiro P. Carvalho-Jr, Beatriz E. Goi
PII:
S0020-1693(17)31116-7
https://doi.org/10.1016/j.ica.2017.11.041
ICA 18011
DOI:
Reference:
Inorganica Chimica Acta
To appear in:
Received Date:
Revised Date:
Accepted Date:
17 July 2017
11 October 2017
22 November 2017
Please cite this article as: B.A. Riga, M.D. Neves, A.E.H. Machado, D.M.S. Araújo, J.R. Souza, O.R. Nascimento,
V.T. Santana, C.C.S. Cavalheiro, V.P. Carvalho-Jr, B.E. Goi, Synthesis of cobalt(II)-α-diimines complexes and
their activity as mediators in organometallic mediated radical polymerization of vinyl acetate, Inorganica Chimica
Acta (2017), doi: https://doi.org/10.1016/j.ica.2017.11.041
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Synthesis of cobalt(II)-α-diimines complexes and their activity as mediators
in organometallic mediated radical polymerization of vinyl acetate
Beatriz A. Riga1, Marina D. Neves1, Antonio E. H. Machado2, Diesley M. S. Araújo2,
Jhonathan R. Souza2, Otaciro R. Nascimento3, Vinícius T. Santana3, Carla C. S. Cavalheiro4,
Valdemiro P. Carvalho-Jr1 and Beatriz E. Goi1*
1 Faculdade de Ciências e Tecnologia, UNESP Universidade Estadual Paulista, CEP 19060-
900, Presidente Prudente, SP, Brazil.
2 Instituto de Química, UFU Universidade Federal de Uberlândia, CEP 38400-902,
Uberlândia, MG, Brazil.
3 Instituto de Física de São Carlos, USP Universidade de São Paulo, CEP 13563-120, São
Carlos, SP, Brazil.
4 Instituto de Química de São Carlos, USP Universidade de São Paulo, CEP 13566-590, São
Carlos, SP, Brazil.
Correspondence to: Beatriz G. Goi (e-mail: beatriz_goi@fct.unesp.br)
1

Abstract
This study describes the synthesis and characterization of three α-diimine-cobalt complexes of
the type [CoCl2(R-DAB)] (R-DAB = R-N=CH-CH=N-R; R =Mes, Dipp and Dipp*) and their
application as mediators for the cobalt-mediated radical polymerization (CMRP) of vinyl
acetate (VAc) using 2,2′-Azobis(2-methylpropionitrile) (AIBN) as initiator. The complexes
were characterised by elemental analyses, FTIR, UV-vis, electron paramagnetic resonance
spectroscopy, cyclic voltammetry, and computational methods. The systematic variation of
the reaction conditions, such as [initiator]/[Co] and [monomer]/[Co] molar ratios at 65 °C,
affected the polymerization rates and the molecular weights, reaching a certain level of
control. The VAc polymerization was initiated by AIBN in the presence of DMSO using the
α-diimine-cobalt complexes as mediators with [DMSO]/[Co] = 1 at 65 °C. The control over
the polymer produced was not improved after the addition of DMSO. Kinetics studies and
computational investigations support a tailorable cobalt complex reactivity mainly altered by
steric factors of the α-diimine ligands.
Keywords: α-diimine-cobalt complexes, CMRP, vinyl acetate, radical polymerization
2

1. INTRODUCTION
Since vinyl acetate (VAc) radical polymerization mediated by cobalt (CMRP)
was first reported by Jérôme et al.1, it became more than an answer to those who
believed CMRP was efficient only for acrylic monomers; it became the most
promising technique to achieve a highly efficient living radical polymerization (LRP)
of this monomer.2 This methodology belongs to a specific class of controlled
polymerization named organometallic-mediated radical polymerization (OMRP).3
Effective control over VAc radical polymerization is desirable due to the fact that
poly(vinyl acetate) (PVAc) can be exclusively produced by the radical mechanism4
and, as additional motivation, there is a wide range of applications for PVAc, from
ordinary applications as additives to paints or adhesive emulsions, to a large variety of
conceivable new materials such as amphiphilic structures and nanofibrous material
applied to hard tissue engineering.2,4-8
These reasons are sufficient to motivate the exhaustive efforts of many researchers to
find an ideal controlling system. It has long been reported by Jérôme and co-workers
that [Co(acac)2] can successfully mediate VAc polymerization, offering great results in
terms of narrow molecular weight distribution, high monomer conversion, and linear
dependence between molecular weight and conversion.9-11 Even before that study,
Wayland and co-workers had already reported cobalt porphyrinates that displayed only
moderate control over the CMRP of VAc.12,13 Recently, Peng et al. have announced the
use of [Co(Salen*)] as a suitable controlling agent to promote the radical
polymerization of VAc with large monomer conversion and high living character.14
However, cobalt(II)-α-diimine complexes for controlled radical polymerization of
3

vinyl acetate have not been reported yet. A general mechanism illustrating the role
played by cobalt complexes in CMRP reaction is presented in Scheme 1.
Scheme 1. General mechanism of cobalt-mediated radical polymerization.
Thereby, to contribute to the advancement of research in CMRP of VAc, we
herein report, for the first time, vinyl acetate radical polymerization mediated by
cobalt(II)-α-diimines. α-Diimine ligands are molecules that have been extensively
studied by Koten and Vrieze.15 This type of ligand presents a 1,4-diaza-1,3-butadiene
skeleton (N=C–C=N) that has aroused much interest owing to its versatile coordination
behaviour. When these ligands pertain to the class of 1,4-disubstituted 1,4-diaza-1,3-
butadienes (RN=CH–CH=NR), they are frequently represented by the label R–DAB,
where R is the functional group attached to the azomethine groups.
These ligands are bidentate, very rigid, and can easily have their backbone
changed to modify the electronic and steric effects at the metal centre.16-20 Moreover,
diazabutadiene (or α-diimine) ligands are able to stabilize metal centres in formally
low oxidation states (such as cobalt II or III); this is a direct consequence of the
presence of a low energy ligand-based π* orbital able to accept electron density from
the metal, establishing a π back-bonding.21,22 Fe(II) and Ni(II) complexes bearing α-
diimine ligands have already been reported as catalysts for atom transfer radical
polymerization of styrene and acrylic monomers,23-24 OMRP of VAc,25 and even α-
4

diimine-Co(II) complexes were applied as catalysts on polymerizations of other
monomers such as ethylene19 and 1,3-butadiene.20
This study describes the synthesis and characterization of α-diimine Co(II) complexes
and their application as controlling agents for CMRP of VAc initiated by AIBN (2,2′-
Azobis(2-methylpropionitrile)). The influences of the ligand structure, focusing the steric
hindrance at the cobalt centre, were assessed by modification of the substituent on the 1 and 4
ligand positions. Different conditions of reaction time, concentrations of VAc and AIBN as
initiator, as well as the use of DMSO as additive were investigated.
5

2. RESULTS AND DISCUSSION
2.1. Synthesis and characterization
α
-Diimines are excellent ligands for four-coordinate cobalt complexes due to
their ease of preparation and their versatility to modification. They are readily prepared
in high yield from the condensation of 2 equiv of the appropriate amine with 1 equiv of
glyoxal.26-31 The [Co(R-DAB)] complexes used in this study are shown in Scheme 2
(Mes = 2,4,6-trimethylphenyl; Dipp = 2,6-diisopropylphenyl; Dipp* = 2,6-
(diphenylmethyl)-4-methylphenyl). The complexes 2a-c were prepared in good yields
from the metallation of the corresponding R-DAB ligands with CoCl2 under mild
1
conditions by stirring and heating at 40 °C (Scheme 3). The FTIR and H and 13C
NMR spectroscopic characteristics of the ligands 1a-c have been described in previous
reports.32-34 The complexes 2a-c were characterized by elemental analyses, cyclic
voltammetry, FTIR, UV–Vis, EPR spectroscopy, and computational methods.
The experimental section shows the results of the elemental analyses (C, H, and
N); the results agree with those calculated for the suggested formula. Single crystals of
the compounds could not be isolated from any organic solution, thus no definite
structures can be described. However, the analytical and spectroscopic data enable us
to predict possible structures. The FTIR spectra of ligands 1a-c indicate absence of
peaks attributable to carbonyl and amino groups, and the appearance of new peaks
1
between 1624-1612 cm− is assigned to
ν
(C=N) stretching vibration (Table 1). These
peaks confirm the formation of the corresponding azomethine group. The latter bands
1
are also observed in the FTIR spectra of complexes 2a-c in the range 1630-1590 cm− .
6

These bands in the complexes are shifted to lower wavenumbers when compared with
the ligands, suggesting the participation of azomethine nitrogen atoms in the
1
coordination. Additional bands between 260 and 290 cm− are also observed in the
spectra of all complexes assigned to
ν
(Co N) stretching vibration, which confirms the
−
coordination of azomethine nitrogen atoms to the Co centre. The bands in the region of
1
300-345 cm− are assigned to
ν
(Co Cl) stretching vibrations. This assignment is quite
−
reasonable and in agreement with previous assignments for related tetrahedral
complexes.16-19
The electronic spectra of the free ligands and cobalt(II) complexes were
recorded in CH2Cl2. The absorption spectra of 1a-c exhibit absorption peaks at the
regions of 259 nm and 297-365 nm (Table 1). The bands at higher energy arise from n-
π* transitions and are associated with the azomethine chromophore, whereas the bands
at lower energy are attributable to π-π* transitions within the phenyl rings.35 All
compounds display an intense band at approximately 230 nm; this excitation is
attributed to the π-π* transitions arising from the azomethine group.16,17 The absorption
spectra of 2a-c exhibit excitations between 260-326 nm, which are assignable to
ligand-to-cobalt(II) charge transfer. Two transitions that give rise to broad bands
between 502 and 631 nm are also observed: The absorption at higher energy occurs
due to cobalt(II)-to-ligand charge transfer and the absorption at lower energy is
attributed to metal-ligand bonding-to-antibonding transition, the latter transition is
provided by the mixing between the metal dπ and R-DAB ligand pπ* orbitals (π back-
bonding).36
7

Scheme 2. Structures of cobalt(II)-α-diimines complexes used to mediate vinyl
acetate radical polymerization in this work.
Scheme 3. Synthesis of complexes 2a-c.
8

Table 1. Infrared and electronic absorption data for R-DAB ligands and
[CoCl2(R-DAB)] complexes.
1
FTIR (cm−−− )
UV-Vis (nm)
Ligand/Complex
Ligand
1612
Complex
Ligand
Complex
234 ( *)
1a
1b
1c
/
/
/
2a
2b
2c
1593 (C=N) 236 (
ν
π
-
π
*)
*)
*)
π-π
326 (LMCT)
502 (MLCT)
343
312
280
ν
ν
(Co-Cl) 259 (n-
π
(Co-Cl) 365 (
(Co-N)
(C=N) 231 (
π
-π
ν
592(d
231 (
π
-
π
*)
1624
1624
1600
ν
π
-
π
*)
*)
*)
π-π
*)
323 (LMCT)
406 (MLCT)
310
300
279
ν
(Co-Cl) 256 (n-
π
ν
(Co-Cl) 357 (
(Co-N)
(C=N) 231 (
π
-π
ν
631 (d
231 (
π
-
π
*)
1630
ν
π
-
π
*)
*)
*)
π-π
*)
260 (LMCT)
521 (MLCT)
312
297
279
ν
(Co-Cl) 259 (n-
(Co-Cl) 297 (
(Co-N)
π
ν
π π
-
ν
626 (d
π
-π*)
2.2. EPR spectra
EPR spectra of complexes 2a-c were recorded at T = 5 K from powder samples (not
shown) and from the frozen CH2Cl2 solutions of the samples that are shown in Fig. 1. All
samples presented a broad signal characteristic of CoII ions with a high spin S=3/2
configuration in distorted tetrahedral symmetry. Powder and frozen solution spectra differed
only in the linewidth values because magnetic dipolar interactions among neighbours
9

broadened the powder EPR spectra. EasySpin,37 a package of programs working under
Matlab,38 was used to simulate the experimental spectra. The spin Hamiltonian describing
these systems is
H0 = µBB · (g · S) + S ·D · S ,
(1)
where µB is the Bohr magneton, B is the applied magnetic field, g is the giromagnetic tensor
and D is the zero field splitting expressed by the two components D and E (axial and rhombic
terms, respectively). The ratio E/D expresses the distortion of the symmetry and is usually
limited to the interval 0 ≤ E/D ≤ 1/3, where 0 indicates perfect axial symmetry and 1/3
indicates the maximum distortion in the z-direction (|Dzz| or |D33|) and minimum in the x-
direction (|D11|). The spectra are characteristic of transitions in the lower |3/2, 1/2> doublet
populated at low temperatures. Simulations of the frozen solution spectra were first performed
considering effective g’ values for an S’=1/2 effective spin in order to estimate the E/D
values18 and are shown in Fig. 1. Simulations of the spectra at T = 5 K using Equation 1 for
CoII ions with a high spin S=3/2 configuration and distorted tetrahedral symmetry were used
to estimate the minimum value of D parameters.
The EPR parameters obtained from these simulations are described in Table 2. The
best simulations gave g values and D values in agreement with the references18,39-41 despite the
uncertainty imposed by the broad linewidths. E/D ratios are similar for the three complexes
studied and indicates a rhombic distortion in the range 22%-35% in comparison with the
maximum distortion (E/D=1/3).
2.3. Theoretical calculations
The optimized structures of the pseudo-tetrahedral Co(II)-α-diimine complexes
are presented in Fig. 2. Relevant bond distances and angles of the studied complexes are
10

presented in Table 3. The atoms follow the numbering shown in Fig. 4. As expected, all these
complexes form distorted tetrahedrals. The Co-Cl distances are longer than Co-N distances,
and the solvation of these complexes tends to increase some geometrical parameters,
enlarging the distortion of the tetrahedral structure. The data in Table 2 suggests that the
dihedral angle (N-C-C-N) of the five-membered ring involving the Co(II) and the nitrogen
atoms is strongly influenced by the amount and size of the substituents present in the phenyl
groups. The parameters g1, g2, and g3, obtained theoretically, are consistent with what is
expected of species submitted to a field of rhombic symmetry18,40 and compatible with
pseudo-tetrahedral Co(II) complexes that display the general formula CoL2X2.39 Regarding
the parameter E/D, although smaller than the values estimated from experimental parameters,
the theoretical values also suggest a high rhombic distortion, similarly to the experimental
values. It should be emphasized that, in an environment with maximum rhombicity, the ratio
E/D is limited to 0.33.42 Analysis of the geometric parameters estimated for these compounds,
either isolated or solvated (Table 3), shows that they are distorted tetrahedrals in the vicinity
of Co (II). It can also be noticed that the two substituent groups anchored to the aromatic
structures attached to these rings play no determining role on the distortion and general
geometry of these complexes. However, although in a small extension, tension is reduced for
the solvated structures especially between cobalt and chlorides.
11

Figure. 1. Frozen CH2Cl2 solution EPR spectra ( — ) and simulations (- - -) of 2a (A), 2b (B)
e 2c (C).
12

Table 2. EPR parameters obtained from EPR spectra and using DFT single-point
calculations.
Linewidtha H Strainb
Dc
> 6
Sample
2a
g1
g2
g3
E/D
5.3543e 4.0012 e 2.0707 e
52
71
46
(0, 66, 59)
(0, 42, 72)
(0, 68, 74)
0.096
0.026
0.073
0.029
0.118
REVPBEd 2.0571
2.1119
2.0804
-4.01
> 6
5.4162 e 4.3435 e 2.2440 e
2b
2.0569
2.1129
2.0807
-4.07
> 6
REVPBE
2c
5.6826 e 3.9754 e 2.0509 e
2.0819
2.3787
2.0784
-20.74
0.038
REVPBE
b
c
a linewidth in mT; anisotropic broadening in mT for the directions of g1, g2 and g3; D in
d
cm-1; the GGA functional REVPBE was used to calculate g1, g2, g3, and D and E/D were
estimated using the diagonal terms of the ZFS tensor,42 calculated using the functional
BP8643; e effective g values.
It should be emphasized that the angles Cl-Co-Cl shown in Table 3 for the non-
solvated complexes, of approximately 137o for the complexes 2a and 2b, correspond to a
deviation of nearly 23% with respect to the expected for a tetrahedral compound based on
Co(II), Fig. 3, consistent with the rhombic distortion in the range of 22%-35% in comparison
with the maximum distortion42. For complex 2c, the lowest value found for the Cl-Co-Cl
angle (129o) is due to the angular tension caused by the groups diphenylmethyl, Fig. 2,
leading to a minor deviation (16%). In contrast, for the angle N-Co-N, the estimated
deviations with respect to the expected in a non-distorted tetrahedron (103o) are larger: 32%
for 2a and 2b, and 29% for 2c, respectively.
13

Complex 2a
Complex 2b
Complex 2c
Figure 2. Views of the optimized structures of complexes 2a, 2b, and 2c.
14

Figure. 3. Representation of the tetrahedral compound Co(NH2)2Cl2: the angles Cl-Co-Cl and
N-Co-N are, respectively, 110.96o and 103.31o. This structure was optimized under conditions
similar to those used in the structural optimization of the other complexes.
Figure 4. Representation of the central part of the complexes showing the numbered atoms.
Colours of the atoms: white for hydrogen, grey for carbon, blue for nitrogen, light blue for
cobalt, and green for chlorine.
15

Table 3. Relevant bond distances and angles related to the studied Co(II) complexes in
the gas phase. The values in parenthesis are related to the complex in dichloromethane.
Complexes
Geometric parameter
2a
2b
2c
d(Co-Cl(3)), Å
2.2249 (2.2620)
2.2250 (2.2622)
2.1020 (2.1140)
2.1015 (2.1143)
1.4243 (1.4192)
1.4244 (1.4189)
137.252 (119.082)
78.284 (79.791)
111.853 (120.084)
111.646 (120.240)
-9.222 (-12.251)
2.2235 (2.2600)
2.2235 (2.2600)
2.1034 (2.1196)
2.1036 (2.1197)
1.4272 (1.4212)
1.4271 (1.4212)
2.2170 (2.2630)
2.2170 (2.2588)
2.1190 (2.1167)
2.1191 (2.1080)
1.4264 (1.4175)
1.4264 (1.4204)
d(Co-Cl(2)), Å
d(Co-N(5)), Å
d(Co-N(4)), Å
d(C(18)-N(5)), Å
d(C(10)-N(4)), Å
Θ(Cl(3)-Co-Cl(2)), o
Θ(N(5)-Co-N(4)), o
Θ(Cl(3)-Co-N(5)), o
Θ(Cl(2)-Co-N(4)), o
D(N(5)C(7)-C(6)N(4)), o
137.485 (118.466) 128.644 (117.538)
78.290 (79.681) 80.222 (80.353)
101.000 (106.142) 102.057 (105.497)
100.978 (106.141) 102.049 (107.122)
8.696 (11.185)
9.817 (8.439)
2.4. Cyclic Voltammetry
The electrochemical activity of the complexes 2a-c was studied by cyclic voltammetry
in scan rate of 100 mV s−1 in CH2Cl2 solution containing 0.1 M n-Bu4NPF6 as supporting
electrolyte. Under these conditions the redox potential (E1/2) for the Fc/Fc+ couple occurred at
470 mV, with ∆Ep = 208 mV; results for 2a-c are collected in Table 4. E1/2 values lie in the
range from −269 to –630 mV, indicating that the Co(II) centres are readily oxidized and
potentially suited to facile organo-cobalt(III) formation (Table 4).44 Their peak-to-peak
separations (∆Ep) are lower to the ferrocene-ferrocenium couple, indicating that these
16

CoII/CoIII couples are facile and reversible. The peak-to-peak separation, ∆Ep, is somewhat
larger than the canonical value for ideal Nernstian behavior (59 mV), indicating some
reorganization of the coordination sphere of the cobalt centre. This may reflect preferred
coordination geometries for each redox state, CoII, d7 and CoIII, d6. One-electron redox
processes in the region between –700 and –250 mV and an irreversible oxidation process
above 750 mV were observed for all complexes. The first redox process can be assigned to
the CoII/CoIII metal centre and the more positive oxidation potential is assigned to the
arylimine moieties. Overall, there is a clear shift in the CoII/CoIII redox potentials towards
more positive values as the electron-withdrawing ability of the aryl substituents is increased
(Dipp* > Mes > Dipp).
Table 4. Cyclic voltammetrya results for complexes 2a-c in the absence/presence
of DMSO.
Presence of DMSOb
Absence of DMSO
E1/2 (Eox)/mV
E1/2 (Eox)/mV
∆
Ep/mV
87
∆
Ep/mV
68
-323 (-279)
-630 (-585)
-269 (-223)
-313 (-279)
-658 (-615)
-307 (-268)
2a
2b
2c
91
86
92
77
a Conditions: CH2Cl2, n-Bu4NPF6 (supporting electrolyte, 0.1 mol L−1), [Co] = 5 mmol L−1,
scan rate = 100 mV s−1), platinum disk and wire (working and auxiliary electrode), Ag/AgCl
(reference electrode). E1/2 is the half-wave potential for the complex; ∆Ep is the cathodic-
anodic peak separation. b [DMSO]/[Co] = 1.
17

2.5. CMRP of VAc
The complexes 2a-c were used as mediators for the polymerization of VAc in
bulk at 65 °C. To begin this study, a variation of [initiator]/[Co] ratio was performed to
find the best reaction condition to control the production of radicals. The effect of the
[AIBN]/[Co] molar ratio on the VAc conversion is illustrated in Fig. 5. The conversion
values showed favourable dependence on the [AIBN]/[Co] molar ratio in the range
from 0.5 to 11.5 at 65 °C for 12 h. The conversions increase exponentially throughout
the evaluated concentration range in all cases. Co(II)-α-diimines (2a-c) mediated VAc
polymerization seem to be mainly controlled by degenerative transfer (DT) process as
well since this polymerization requires an extra equivalent of radicals, once only reach
low monomer conversion with [AIBN]/[Co] = 0.5. Mn values for 2a-b show a curve
with profile similar to those of conversion, which presents an increase in molecular
weights of polyVAc up to [AIBN]/[Co] = 11.5, whereas for 2c the Mn values reach a
maximum at 3.25 followed by a drop for higher ratios. For [AIBN]/[Co] ratio ≥ 1.5, the
CMRP mechanism becomes mostly DT, which requires the rapid exchange of the
propagating radicals by the controlling agents in order to minimize the termination
reactions. For complex 2c the increase of the ratio [AIBN]/[Co] from 3.25 decreases
the Mn values, even though an increase in the conversion values is observed. This
clearly indicates that the greater steric hindrance in the complex 2c prevents the
deactivation of the propagating radicals and favours the chain termination reactions,
which consequently results in lower molecular weights.
18

For [AIBN]/[Co] = 3.25, narrow (Mw/Mn = 1.40-1.42) molecular weight
distributions of polyVAc were obtained with 2a and 2b, whereas a broad distribution
(Mw/Mn = 2.10) was obtained with 2c. An increase of AIBN concentration resulted in
polymers even more disperse, with Đ values higher than 1.5. Therefore, considering
the good results obtained, this AIBN concentration ([AIBN]/[Co] = 3.25) was selected
for the upcoming studies.
The effect of the [VAc]/[Co] molar ratio on the conversion of isolated polymer
is presented in Table 5. Polymerization appeared to be sensitive to the
monomer/mediator ratio, with a decrease in the conversion values of polyVAc with the
increase of the [VAc]/[Co] ratio for all complexes. Furthermore, an increase in the
[VAc]/[Co] ratio resulted in broader Đ and molecular weights higher than the predicted
ones. This can be explained by the fact that higher [VAc]/[Co] ratio imply in a
decrease in the concentration of AIBN, and thus the concentration of radicals. With
declined radical concentration, the polymerization rate decreases as well as the VAc
conversion. The experimental molecular weights closer to the theoretical values with
narrowest Đ values were obtained with [VAc]/[Co] = 542, suggesting the most
satisfactory VAc concentration for the upcoming studies.
CMRP of VAc mediated by 2a
conditions ([VAc]/[AIBN]/[Co] = 542/3.25/1 at 65 °C) showed a similar length of
induction period ( 270 min) and first order kinetic plots with different slopes, as
-c as a function of time under established optimal
∼
assessed by the linear dependence of ln([M]0/[M]t) versus time (Fig. 6). The slope in
the first-order kinetic plots allowed us to calculate the kinetic constants of reactions
5
1
5
mediated by each one of the complexes (kobs = 1.96 × 10− s− for 2a kobs = 1.62 × 10−
,
1
5
1
s− for 2b, and kobs = 2.55 × 10− s− for 2c). The induction period was rationalized as
19

the time required to form organo-Co(III) complexes from the Co(II) species and
radicals.45,46 Nevertheless, the theoretical molecular weights predicted were higher than
those obtained experimentally. The molecular weight distribution curves were
monomodal throughout the polymerization (Fig. S14). A linear increase of molecular
weight with conversion (Fig. 7), coupled with Đs < 1.43, illustrate a certain control that
2a-b exert over polymerizations of VAc. In contrast, the molecular weight of polyVAc
did not increase linearly as a function of VAc conversion and presented high Đ values
when mediated by 2c. The molecular weights were high from the beginning, much
higher than the calculated values, and did not depend linearly on conversion. Shaver
and co-workers conducted a systematic variation of ligand substituents in cobalt
complexes correlating with their redox potential for a better understanding of the role
of the metal-carbon bond strengths towards improved control and tunability of
polymerisations.47 The studies pointed out that an increase in the electron density
around the metal centre (lower redox potential) favours the increase of the conversion,
but to a loss of controllability in the polymerization. When evaluating the mediating
efficiency of complexes 2a-c, better control levels were achieved with the complexes
with lower redox potential. However, it is clear that steric effects also play an
important role in determining the reactivity of complexes 2a-c. A smaller discrepancy
between observed and calculated molecular weight, and narrower molecular weight
distribution were obtained for less sterically hindered complexes. It can be proposed
that higher steric hindrance restricts the equilibrium between the mediator and the
propagating radical species.
20

100
90
80
70
60
50
40
30
20
10
0
2.10
60
2.55
50
40
30
20
10
0
2.38
1.53
1.91
2.36
1.73
1.94
1.93
1.61
2.02
1.70
1.40
1.42
1.43
2
0
1
3
4
5
6
[AIBN]/[Co]
7
8
9
10 11 12 13
Fig. 5 Dependence of conversion and Mn values on the [AIBN]/[Co] ratio for CMRP of VAc
with 2a (■), 2b (●) and 2c (▲); [VAc]/[Co] = 542 with 40 µmol of complex in bulk at 65 °C
for 12 h. The numbers correspond to the Đ values for each run.
21

Table 5. Bulk polymerization of VAc with different amounts of VAc at 65°C in 12 h.
a
Entry [CoII]/[AIBN]/[VAc] Complex Conv. % Mn,GPC Mn,th
Đb
1
2
3
4
5
6
7
8
9
1/3.25/271
1/3.25/542
1/3.25/1084
42
50
72
40
33
55
12
19
25
16800 9800
1.60
2a
2b
2c
2a
2b
2c
2a
2b
2c
10200 11700 1.73
67500 16800 2.28
14300 18700 2.12
8200
15400 1.42
54600 25700 2.10
13800 11200 2.28
3200
17700 2.10
46500 23300 2.29
aMn,th = [VAc]0/[Co]0 × MW(monomer) × conv (%); bĐ = Mw,gpc/Mn,gpc
.
22

1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
50
40
30
20
10
0
0
2
4
6
Time (h)
8
10
12
Figure 6. Dependence of ln([VAc]0/[VAc]t) conversion on the reaction time for CMRP
of VAc with 2a (■), 2b (●) and 2c (▲); [Co]/[AIBN]/[VAc] = 1/3.25/542 with 40
µmol of complex in bulk at 65 °C.
50
40
30
20
10
0
3,5
3,0
2,5
2,0
1,5
1,0
0
5
10 15 20 25 30 35 40 45 50 55 60
Conversion (%)
Figure 7. Dependence of Mn and Đ on the conversion for CMRP of VAc with 2a (■),
2b (●) and 2c (▲); [Co]/[AIBN]/[VAc] = 1/3.25/542 with 40 µmol of complex in bulk
at 65 °C; Mn(theor) (black dashed-line).
23

CMRP of vinyl monomers in the presence of DMSO mediated by organocobalt
has already been reported as a strategy to achieve better levels of control in
polymerization.48-52 These studies suggest that the coordination of DMSO to the Co
centre exerts a trans effect to facilitate the weakening of the C-Co bond in the axial
plan, favoring the exchange of radicals by the reversible termination mechanism. In
light of this results, VAc polymerization was initiated by AIBN in the presence of
DMSO using 2a-c as mediators with [DMSO]/[Co] = 1 at 65 °C. Using this ratio, a
DMSO molecule coordinates to the cobalt centre and the VAc polymerization,
previously conducted by the DT pathway, is reverted to a reversible termination (RT)
mechanism due to the blocked reaction site. All three Co(II) complexes were able to
mediate the radical polymerization of VAc, but with different rates and levels of
control, under these conditions. The conversion curves evidence a shorter induction
period as compared with CMRP of VAc in the absence of DMSO in all cases (Fig. 8).
Đs were actually broader, with more deviation between theoretical and experimental
molecular weights indicating that the VAc polymerizations in the presence of DMSO
mediated by 2a-c were not preventing termination reactions. The polymerization
control has not demonstrated to be improved, as can be verified by the molecular
weight deviation and the broader molecular weight distribution (Fig. 9).
In order to further investigate the reason why the polymerization control was not
improved, we investigated the complexes 2a-c in the presence of DMSO. The new Co species
formed in solution in the presence of DMSO ([DMSO]/[Co] = 1) were identified by cyclic
voltammetry. Experimentally determined electrochemical parameters for complexes 2a–c
with DMSO are collected in Table 4, which shows the half-wave potentials (E1/2) and peak-to-
peak separation (∆E). The ∆E for all complexes with DMSO is something lower than without
24

DMSO, indicating an increase in the redox reversibility (Table 4). The ∆E (with or without
DMSO) are comparable to the ferrocene-ferrocenium couple, indicating that these CoII/CoIII
couples are facile and reversible. Further, a computational investigation was carried out to
investigate the coordination mode and arrangement of the DMSO and VAc radical ligands in
the Co center (2a-c) (Fig. 10). As can be verified, DMSO, bonded through O, and VAc radical
ligands are cis-positioned each other around the Co centre according to a distorted
octahedron, strongly influenciated by the five-membered ring.
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
50
40
30
20
10
0
0
2
4
6
8
10
12
Time (h)
Figure 8. Dependence of ln([VAc]0/[VAc]t) and conversion on the reaction time for
CMRP of VAc with 2a (■), 2b (●) and 2c (▲); [Co]/[AIBN]/[VAc]/[DMSO] =
1/3.25/542/1 with 40 µmol of complex in bulk at 65 °C.
25

16
14
12
10
8
6
4
2
3,05
3,0
2,5
2,0
1,5
1,0
0
5
10
15
20
25
30
35
40
45
50
55
Conversion (%)
Figure 9. Dependence of Mn and Đ on the conversion for CMRP of VAc with 2a (■),
2b (●) and 2c (▲); [Co]/[AIBN]/[VAc]/[DMSO] = 1/3.25/542/1 with 40 µmol of
complex in bulk at 65 °C. Mn(theor) (black dashed-line).
Such structures have multiplicity M=3 (triplet) as a result of the incorporation of VAc
as a free radical. Different from the expected, results of the calculations involving the VAc
radical and DMSO-Co-α-diimine complexes show a geometric arrangement different from
that reported by Debuigne et al.52 These results help explain why the VAc polymerization in
the presence of DMSO was inadequately controlled by complexes 2a-c. In addition to the lack
of trans-labilizing effect in the axial plan in these complexes to facilitate the weakening of the
C-Co bond, the arrangement involving the VAc radical and DMSO-Co-α-diimine complexes
exerts greater steric hindrance in the Co centre, which hinders the trapping of the propagating
radical.
26

Complex 2a + DMSO + VAc
Complex 2b + DMSO + Vac
Complex 2c + DMSO + Vac
Fig. 10 Representation of DMSO and VAc coordinated to the complexes 2a (A), 2b (B), and
2c (C).
27

Table 6. Thermodynamic parameters determined to the cobalt complexes [Co-VAc] e
[Co-VAc-DMSO].
∆Ga (kJ/mol)
49.74
∆H (kJ/mol)
-22.59
T∆S (kJ/mol)
-72.33
Complexes
[2a-VAc]
95.83
-24.00
-119.83
-74.45
[2a-VAc-DMSO]
[2b-VAc]
37.92
-36.53
88.63
-38.04
-126.67
138.34
[2b-VAc-DMSO]
[2c-VAc]
-367.33
-74.34
-228.59
-228.64
-154.30
[2c-VAc-DMSO]
aThe thermodynamic parameters were calculated from the following chemical equations:
Co-α-diimines + VAc → [Co-VAc]
Co-α-diimines + VAc + DMSO → [Co-VAc-DMSO]
Scheme 4. Gibbs free energy diagram relating the differences between the energies of
complexes [Co-VAc] and [Co-VAc-DMSO]
28

These statements can be further confirmed by the comparison between the
thermodynamic parameters obtained to VAc bonded to the Co complexes in the presence and
absence of DMSO (Table 6, Scheme 4). According to the calculations, the variation of ∆G
between the [(2a-c)-VAc] and [(2a-c)-VAc-DMSO] species shows that the alkyl-Co
complexes obtained in the presence of DMSO are much less favoured than those in the
absence of DMSO. The most important factor contributing to achieve the more
thermodynamically favoured species is the entropy required to bring together the initial
complex and the alkyl-Co complex.
3. CONCLUSION
Three cobalt(II)-α-diimine complexes (2a-c) were prepared and studied by FTIR, UV-
Vis, cyclic voltammetry, electron paramagnetic resonance spectroscopy, and computational
methods. The DT-CMRP of VAc mediated by 2a-c as a function of time under established
optimal conditions ([VAc]/[AIBN]/[Co] = 542/3.25/1 at 65 °C) were first order in monomer,
as assessed by the linear dependence of ln([M]0/[M]t) versus time. The kinetics constant was
measured by the slope of the linear graph, ln[VAc]0/[VAc]t versus time (kobs = 1.96 × 10-5 s-1
for 2a, kobs = 1.62 × 10-5 s-1 for 2b, and kobs = 2.55 × 10-5 s-1 for 2c). The data show smaller
discrepancy between the observed and calculated molecular weights, and narrower molecular
weight distribution were obtained for less sterically hindered complexes. Computational
investigations and electrochemical studies associated with polymerization kinetics allowed us
to comprehend how steric effects may modulate the catalytic reactivity of 2a-c. The RT-
CMRP of VAc mediated by 2a-c induced by the presence of DMSO, at proportion
29