Studies of electronic effects of modified pyridine-imine ligands utilized in cobalt-catalyzed meta-selective Diels-Alder reactions

Article abstract of DOI:10.1016/j.jorganchem.2013.09.020

The regioselectivity of cobalt-catalyzed Diels-Alder reactions can be controlled by the choice of ligands on the cobalt center. Ligands of the pyridine-imine type favor the formation of the 1,3-disubstitution pattern on the dihydroaromatic product. The investigation was aimed to elucidate the factors for controlling the regioselectivities and reactivity induced by electronic effects of the pyridine-imine ligands in the cobalt-catalyzed Diels-Alder reaction. For that, electron-withdrawing as well as electron donating substituents were introduced in the 4-position on the pyridine moiety and on the 4′-position of the aniline derivative used in the imine subunit of the ligands. In close synergy DFT calculations and the comparison with experimental results proved that electronic variation of the substituents at both positions have a negligible influence on the regioselectivity. However, the kinetic data for the cobalt-catalyzed Diels-Alder reactions revealed that there are great differences in the lengths of the induction periods when different cobalt pyridine-imine complexes are applied. These results could be elucidated by conductivity experiments showing that ionic homoleptic complexes Co(L) ₂²⁺/CoBr₄^2- are in equilibrium with their corresponding neutral heteroleptic complexes of type Co(L)Br₂ in solution. The equilibrium position depends on the electronic characteristics of the pyridine-imine ligands (L), thereby influencing the length of the induction period.

Full text of DOI:10.1016/j.jorganchem.2013.09.020

Journal of Organometallic Chemistry 749 (2014) 219e223
Contents lists available at ScienceDirect
Journal of Organometallic Chemistry
journal homepage: www.elsevier.com/locate/jorganchem
Studies of electronic effects of modified pyridine-imine ligands
utilized in cobalt-catalyzed meta-selective DielseAlder reactions
,
a
,
,
, 1
Gerhard Hilt a , Judith Janikowski 1, Martin Schwarzer a 1, Olaf Burghaus a
**
,
*
Dimitri Sakow b 2, Martin Bröring b , Marcel Drüschler 1, Benedikt Huber a
,
,
,
a
,
, 1
Bernhard Roling a 1, Klaus Harms a 1, Gernot Frenking a
,
,
,
*
a Fachbereich Chemie, Philipps-Universität Marburg, Hans-Meerwein-Str., 35043 Marburg, Germany
b Fachbereich Chemie, Institut für Anorganische und Analytische Chemie, TU Braunschweig, Hagenring 30, 38106 Braunschweig, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
The regioselectivity of cobalt-catalyzed DielseAlder reactions can be controlled by the choice of ligands
on the cobalt center. Ligands of the pyridine-imine type favor the formation of the 1,3-disubstitution
pattern on the dihydroaromatic product. The investigation was aimed to elucidate the factors for con-
trolling the regioselectivities and reactivity induced by electronic effects of the pyridine-imine ligands in
the cobalt-catalyzed DielseAlder reaction. For that, electron-withdrawing as well as electron donating
substituents were introduced in the 4-position on the pyridine moiety and on the 40-position of the
aniline derivative used in the imine subunit of the ligands. In close synergy DFT calculations and the
comparison with experimental results proved that electronic variation of the substituents at both po-
sitions have a negligible influence on the regioselectivity. However, the kinetic data for the cobalt-
catalyzed DielseAlder reactions revealed that there are great differences in the lengths of the induc-
tion periods when different cobalt pyridine-imine complexes are applied. These results could be eluci-
dated by conductivity experiments showing that ionic homoleptic complexes Co(L)22þ/CoBr42ꢀ are in
equilibrium with their corresponding neutral heteroleptic complexes of type Co(L)Br2 in solution. The
equilibrium position depends on the electronic characteristics of the pyridine-imine ligands (L), thereby
influencing the length of the induction period.
Received 19 July 2013
Received in revised form
6 September 2013
Accepted 15 September 2013
Keywords:
Cobalt
Cycloaddition
Pyridine-imines
Regioselectivity
Quantum-mechanical investigation
Conductivity measurements
Ó 2013 Elsevier B.V. All rights reserved.
1. Introduction
2 in excellent yields which can be easily oxidized by DDQ (¼2,3-
dichloro-5,6-dicyano-benzoquinone) to the corresponding arene
The regiodivergent synthesis of carbocyclic six-membered ring
systems via atom economic cobalt-catalyzed DielseAlder reactions
is an excellent example of a transformation where the regiose-
lectivity is efficiently controlled by the choice of the ligand system
[1] (Scheme 1).
Thereby, the highly selective synthesis of disubstituted dihy-
droaromatic compounds, such as 1, can be accomplished when a
bidentate phosphine ligand, such as dppe, is applied, leading
eventually to arenes of type 3 [2]. On the other hand, the applica-
tion of pyridine-imine-type ligands in the cobalt-catalyzed Dielse
Alder reaction leads to the highly selective formation of regioisomer
4 [3]. In recent applications also silyl- and boron-functionalized
alkynes could be applied in the regiodiverse DielseAlder reactions
with similarly high regioselectivities leading to very interesting
functionalized building blocks for follow-up transformations [4].
In a detailed quantum mechanical investigation the mechanism
of the cobalt-catalyzed DielseAlder reaction was evaluated and the
regiodivergent behavior of the Co(dppe)þ-complex [5] leading to
the para-substituted product 1, and of the Co(pyridine-imine)þ-
complex [6] predominantly generating the meta-substituted
product 2 could be explained [7]. A total of eight different reaction
pathways were found for the stepwise rather than concerted
mechanism yielding the para- or the meta-product, respectively.
The calculated ratios of the primary cycloaddition products 1 and 2
were in excellent agreement with the experimentally observed
regioselectivities.
* Corresponding authors. Fax: þ49 6421 2825677.
** Corresponding author. Fax: þ49 531 391 5387.
E-mail addresses: Hilt@chemie.uni-marburg.de (G. Hilt), m.broering@tu-bs.de
(M. Bröring), frenking@chemie.uni-marburg.de (G. Frenking).
As quintessence from the analysis of the lowest lying transition
states for both reaction pathways we came to the conclusion that
different steric hindrances of the ligands caused the change in
1
Fax: þ49 6421 2825677.
2
Fax: þ49 531 391 5387.
0022-328X/$ e see front matter Ó 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jorganchem.2013.09.020

220
G. Hilt et al. / Journal of Organometallic Chemistry 749 (2014) 219e223
Ph
2.2. Synthesis of the ligands and cobalt complexes
Ph
Co(ligand)Br2
+
+
(Fe), Zn, ZnI2, CH2Cl2
25 °C, 16 h
Me
Ph
The ligands 6ae6d were generated by condensation of 2-
formylpyridine with aniline derivatives. The cobalt-complexes
were synthesized by addition of 1.0 equiv. of the ligand to a solu-
tion of cobalt(II) bromide in THF. The mixture was stirred overnight
and the solvent was removed. The complexes were stored at
ambient temperatures without special preventive measures and
used as precatalysts without further purification.
Me
Me
Me
2
1
DDQ
ligands: dppe
pyridine-imine
=
=
Ph2P
PPh2
Ph
+
N
N Mes
Me
Ph
4
3
2.3. Representative procedure for the cobalt-catalyzed DielseAlder
reaction
Scheme 1. Regiodivergent synthesis of dihydroaromatic compounds 1 and 2 by cobalt-
catalyzed DielseAlder reaction and follow-up oxidation to the arenes 3 and 4.
regioselectivity. Accordingly, electronic modifications in a remote
position of the pyridine-imine ligand should not influence the
regioselectivity of the Co(pyridine-imine)-catalyzed cycloaddition
process. To test this hypothesis we set out to start this investigation
utilizing theoretical methods to evaluate those ligands which
cannot be synthesized in a few steps. Thereby, in silico evaluations
for those hypothetical ligands should be obtained.
Also, the cobalt-catalyzed reaction starts from a cobalt(II) pre-
cursor and a heterogeneous electron transfer from zinc powder
generates the proposed catalytically active cobalt(I) species
resulting in an induction period. The redox process should be
dependent on the electronic characteristics of the Co(pyridine-
imine)Br2 complexes resulting in different length of the induction
periods dependent from the electron donating/abstracting ability
of the ligand on cobalt. Accordingly, electronic modifications of the
ligands in remote positions should result in nearly identical ther-
modynamic and kinetic profiles, almost identical yields and
regioselectivities, but in different length of induction periods.
Anhydrous zinc iodide (64 mg, 0.2 mmol, 20 mol%), zinc dust
(13 mg, 0.2 mmol, 20 mol%), iron powder (11 mg, 0.2 mmol,
20 mol%) and Co(5c)Br2 (45 mg, 0.1 mmol, 10 mol%) were sus-
pended under argon in 1.0 mL dichloromethane. Then isoprene
(1.0 mmol) and phenylacetylene (1.0 mmol) were added and the
mixture was stirred at room temperature. The conversion of
the starting materials was monitored by GC and GCeMS analysis.
The reaction mixture was filtered over a short pad of silica gel
(pentane) and the solvent was removed under reduced pressure.
The residue was dissolved in toluene (10 mL) and stirred with DDQ
(1.1 equiv.) for 1 h. The resulting suspension was washed with 10%
NaOH/10% Na2S2O3 solution, extracted with diethyl ether, dried
over MgSO4 and the solvent was removed under reduced pressure.
The product was purified by flash column chromatography on
silica gel (eluent: pentane). The analytical data are in accordance
with the literature [2,3].
2.4. Kinetic investigations
2. Experimental
The DielseAlder reaction was performed as before and samples
were taken by syringe under argon in certain time intervals. The
progress of the reaction was determined by GC analysis via inte-
gration of the peaks.
All reagents were used as purchased from commercial suppliers
without further purification. All reactions were carried out using
standard Schlenk techniques under dry argon. Solvents were purified
and dried according to conventional methods prior to use. 1H and 13
C
3. Results and discussion
1
NMR spectra were recorded with an AVANCE 300 (300 MHz for H
and 75 MHz for 13C) or an AVANCE 500 (500 MHz for 1H and 125 MHz
for 13C). The kinetic measurements were analysed utilizing GC ana-
lyses which were recorded with a Shimadzu GC-2010 Plus Gas
Chromatograph. For chromatography the separations were carried
out on MachereyeNagel Silica 60 (0.040e0.063 mm, 230e400 mesh
ASTM). For TLC Merck TLC plates (Silica 60, F254) were used.
In order to investigate the electronic influence of a substituent
without causing steric hindrance the substituents (R1 and R2) were
placed on the 4- and 40-position in para-relation to the nitrogen
atoms of the pyridine-imine ligand (Fig. 1). Thereby, very little
steric hindrance will occur and the electronic modifications will
have a strong influence on the electron density at the reactive sites
within the cobalt complex via the trans-influence. For this purpose
the two ligand systems 5aee and 6aee were selected with
electron-donating as well as electron-withdrawing substituents on
position 4 and 40.
2.1. Computational details
Geometry optimizations of the molecules without symmetry
constraints (C1) were carried out using the Gaussian03 [8] opti-
mizer (Berny-Algorithm) [9] together with TurboMole 5.0 [10]
energies and gradients at the gradient corrected DFT level of the-
ory using Becke’s exchange functional [11] in conjunction with
Perdew’s correlation functional [12] (BP86) and Ahlrich’s def2-
SV(P) basis-sets [13]. The resolution-of-the-identity (RI) approxi-
mation [14] was applied using auxiliary basis functions [15]. This
level of theory is denoted as RI-BP86/def2-SV(P). Stationary points
were characterized as minima (number of imaginary frequencies
i ¼ 0) or transition states (i ¼ 1) by calculating the Hessian matrix
analytically (AOFORCE) [16]. The solvent-effect on the calculated
energies was estimated with COSMO [17] as single point calcula-
tions on the optimized structures. This level of theory is denoted as
COSMO//RI-BP86/def2-SV(P). The latter calculations were carried
For quantum mechanical investigations the following modifi-
cations of the pyridine moiety were selected: substituent R1 was
modified with the electron donating methoxy e (5a), the dime-
thylamino e (5b) and the electron withdrawing trifluoromethyl e
R1
N
N
N
N
R2
5a: R1 = OMe
5b
6a: R2 = OMe
6b
: R1 = NMe2
: R2 = NMe2
5c
1
(= ): R2 = H
6d
5c
6c
: R = H
: R1 = CF3
: R2 = CF3
5d
5e: R1 = F
6e: R2 = F
out with a dielectric constant
methane [18].
3
¼ 8.93 for the solvent dichloro-
Fig. 1. Modification of electronic parameters of the pyridine-imine ligands 5 and 6
(pyridine-imine) for the meta-selective cobalt-catalyzed DielseAlder reaction.

G. Hilt et al. / Journal of Organometallic Chemistry 749 (2014) 219e223
221
Ph
Ph
+
1. Co(6a-e)Br2
+
Me
Ph
Fe, Zn, ZnI2, CH2Cl2
25 °C, up to 24 h
2. DDQ, toluene
Me
Me
4
3
Scheme 2. Regioselective synthesis of aromatic compounds
catalyzed DielseAlder reactions/DDQ oxidation.
3 and 4 by cobalt-
(5d), the fluoro-substituent (5e) and the unmodified ligand R1 ¼ H
(5c). These modifications are of the highest interest because the
first carbonecarbon bond formation in the catalytic cycle of the
DielseAlder reaction takes place in trans-position to the pyridine
moiety of the ligand of type 5. Electronic modifications at this po-
sition should have the strongest influence on the reaction in terms
of reactivity and selectivity. We also modified, in a second set of
quantum mechanical calculations, the electronic nature of the
substituents R2 in the aniline-type moiety of the pyridine-imine
ligand in the same manner: the ligands 6a and 6b with the elec-
tron donating methoxy-and dimethylamino-substituents as well as
the electron deficient ligands 6d and 6e with the trifluoromethyl e
and fluoro-substituent were selected while 5c/6c again served as a
reference compound.
The synthesis of the ligands 6ae6e is trivial as the required 4-
substituted aniline derivatives are commercially available, and
imine formation with pyridine-2-carbaldehyde proceeds immedi-
ately and quantitatively upon mixing the starting materials. For
these ligands the corresponding cobalt complexes were synthe-
sized and the DielseAlder reactions were performed according to
Scheme 2. The results for the regioselectivity of these reactions are
given in Table 1.
On the other hand, the heterocyclic precursors of pyridine-
imine ligands of type 5 with modifications in the 4-position of
the pyridine subunit are not commercially available and must be
generated in multi-step reaction sequences. Therefore, we decided
not to synthesize these ligands beforehand but await the outcome
of the quantum mechanical investigation and then address the
synthesis of the most promising ligand of type 5 hopefully reaping
the rewards derived from the calculations and cutting directly to
the key ligands.
Fig. 2. Kinetic profiles of the meta-selective cobalt-catalyzed DielseAlder reaction
utilizing cobalt complexes of type Co(6aed)Br2 determined by GC analysis.
calculated values do not indicate a significant change in the
regioselectivity of the meta-selective cobalt-catalyzed DielseAlder
reaction through the variation of the substituents R1 and R2
(Table 1). The energy schemes of the meta-selective reaction
pathways for the ligands 5aee and ligands 6aed are merely iden-
tical and only very small differences can be found, as shown in
Fig. S3 (see, SI).
The results compiled in Table 1 indicate that the calculated
regioselectivities for 6ae6d correctly predict that the formation of
the meta isomer is favoured over the para isomer but the absolute
values for the meta:para ratio are smaller than the experimental
values. However, theory and experiment agree that there are only
small changes in the meta:para ratio when the electron withdrawing
substituents in 6d and 6e are replaced by electron donating groups in
6a and 6b. Even the greatest difference between the ligands 6b and
6d (entry 7/8) with a calculated selectivity increase from 66:34 to
81:19 did not result in any significant change of regioselectivity in
the experimentally observed values within their margins of error.
Moreover, the calculated changes in regioselectivity for ligands of
type 5 (entries 2e5) are even smaller than for those of type 6.
Because the calculations predict a similar negligible change for the
regioselectivity for varying the substituents in the 4-position of the
pyridine moiety of the pyridine-imine ligand, we decided not to
undertake the multiple step syntheses of pyridine-imine ligands
with modifications in the 4-position of the pyridine moiety.
The reaction profiles for the different cobalt complexes of type 5
and 6 are very similar for the DielseAlder reaction (see Fig. S3, SI).
Accordingly, very little difference should be detectable in their
catalytic activity and the overall reaction kinetics.
The results of the quantum mechanical investigations can be
summarized as follows:
Our initial hypothesis that the electronic influence on the
regioselectivity is negligible proved to be correct. We found that the
Table 1
Comparison of the quantum mechanical calculations with the experimental results
of the meta-selective cobalt-catalyzed DielseAlder reactions.
No Ligand Calc.
Regioselectivitya
Calc. Regioselectivity Regioselectivity
(Solv.)b
(Exp.)
DDEc meta:parad
DDEc meta:parad
meta:para
1
2
3
4
5
6
7
8
9
5c/6c
5a
5b
5d
5e
0.5
0.6
0.6
1.0
0.5
0.3
0.0
0.6
0.4
70:30 (75:25) 0.4
66:34 (74:26) 94:6
73:27 (77:23) 0.5
73:27 (77:23) 0.6
84:16 (85:15) 0.6
70:30 (75:25) 0.3
62:38 (73:27) 0.3
50:50 (65:35) 0.2
73:27 (77:23) 0.7
66:34 (74:26) 0.3
70:30 (77:23) nd
73:27 (78:22) nd
73:27 (78:22) nd
62:38 (74:26) nd
62:38 (73:27) 94:6
58:42 (66:34) 97:3
77:23 (81:19) 96:4
62:38 (74:26) 95:5
3.1. Kinetic effects of the electronic modifications
The quantum mechanical investigation and the experimental
results suggested that electronic modifications on the ligands of
type 5 and 6 have little effect on the reaction barriers. However, up
to now we concentrated only upon the cobalt complexes in their
proposed catalytically active form; the cobalt(I) species. The pre-
cursor complexes are activated in situ by a reduction of the Co(II)
species to their corresponding Co(I) complexes, followed by halide
abstraction by the Lewis acid (ZnI2). For this process an induction
period is generally observed. Electronic modifications of the
pyridine-imine ligands should result in different induction periods
caused by the varied reduction potential of the Co(II) catalyst
6a
6b
6d
6e
a
b
c
TURBOMOLE 5.80, RI-BP86/def2-SV(P) including Zero Point Energy correction.
TURBOMOLE 5.80, COSMO//RI-BP86/def2-SV(P).
Energy values in kcal/mol. The energy difference of the lowest activation bar-
riers. For detailed information please check Supporting information.
d
The first value is calculated only from the lowest activation energies yielding
either the meta or the para product. The values in parentheses give the meta:para
ratio calculated by considering the activation barriers of all eight reaction pathways.
nd ¼ not determined.

222
G. Hilt et al. / Journal of Organometallic Chemistry 749 (2014) 219e223
precursors. Indeed, different induction periods for the activation of
the cobalt complexes of type Co(6)Br2 were detected. The kinetic
profile of different cobalt complexes (6ae6d) are shown in Fig. 2.
The shortest induction period was detected for the DielseAlder
reaction utilizing the electron-donating methoxy-substituent in
ligand 6a, followed by the unmodified ligand 6c and the electron-
withdrawing ligand 6d. This order indicates that the electron
transfer from zinc powder to the cobalt complex [Co(6)
Br2 þ eꢀ / Co(6)Br þ Brꢀ] is not the rate limiting step in the in-
duction period, because ease of reduction should be in opposite
order following the redox potentials of the complexes. In fact, the
order indicates that the abstraction of the halide from the inter-
mediate CoI(6)Br complex to the catalytically active Co(6)þ complex
by the Lewis acid (ZnI2) is the rate determining step of the induc-
tion period. This is also plausible as the more electron-withdrawing
the ligand the stronger the halide is bound to the cobalt atom. Once
the active catalyst is formed the reaction is complete within 1e3 h
in each of the investigated cases. On the other hand, the induction
period for the DielseAlder reaction utilizing the strong electron-
donating dimethylamine-substituted ligand 6b was reproducibly
found to be very long (10e14 h), in sharp contrast to our expecta-
tions following the arguments given for the other electron-
donating methoxy-substituted ligand 6a. A rationale for this dras-
tically different behaviour of the complex Co(6b)Br2 could be the
presence of an pre-equilibrium of cobalt complexes in solution. We
envisaged that the complexes Co(6a)Br2 and Co(6b)Br2 are present
in different forms when dissolved in dichloromethane (Scheme 3).
Complex Co(6a)Br2 could favour the neutral heteroleptic form
rather than the ionic homoleptic complexes [CoBr4]2ꢀ and
[Co(6a)2]2þ while the situation for cobalt complex with ligand 6b
was proposed to be quite the opposite with the equilibrium on the
Fig. 3. Broadband conductivity spectra of pure dichloromethane and of the cobalt
complexes with ligands 6a and 6b, respectively. At intermediate frequencies, con-
ductivity plateaus caused by the ionic conductivity of the solutions are detected. Note
that the conductivity of pure dichloromethane results from ionic impurities.
For comparison, we estimated the maximum conductivity of the
dichloromethane solution with the given concentrations of the
complex Co(6b)Br2 assuming the complete formation of the ionic
species [Co(6b)2]2þ and [CoBr4]2ꢀ. The value for the diffusion
constants of the ions involved maximum conductivity was esti-
mated based on the StokeseEinstein relation and for the maximum
conductivity via the NernsteEinstein equation [19]. The estimated
value for the maximum conductivity
s
of a mixture of complexes of
type [Co(6b)2]2þ/[CoBr4]2ꢀ is
s
¼ (4.0
ꢁ
0.5)$10ꢀ6 S cmꢀ1
.
site of the homoleptic complexes [CoBr4]2ꢀ and [Co(6b)2]2þ
.
Accordingly, the methoxy-substituted complex Co(6a)Br2 is the
predominant species from which the catalytically active species can
be formed directly and only a small fraction exists as ionic species
[Co(6a)2]2þ and [CoBr4]2ꢀ. In contrast, the significantly higher
conductivity observed for the dimethylamino-substituted complex
Co(6b)Br2 indicates that the two ionic species are formed to a high
extend and that the neutral Co(6b)Br2 complex is present only in
low concentration in solution at best. One must not forget that a
contact ion pair [Co(6b)2]2þ $ [CoBr4]2ꢀ does not contribute to the
conductivity of the solution and therefore the actual concentration
of the complex Co(6b)Br2 as direct catalyst precursor might well be
reduced to a minimum thus explaining the long induction period in
the preparative DielseAlder reaction. Once the active Co(6b)þ-
species is formed the desired DielseAlder product is formed in a
short period of time with a similar rate of conversion and compa-
rable regioselectivities as found for the other cobalt complexes.
The proposed pre-equilibrium is most likely solvent dependent
and in principle a solvent-induced shift of the pre-equilibrium from
[Co(6b)2]2þ $ [CoBr4]2- towards the desired pre-catalyst Co(6b)Br2
would be desirable. Unfortunately, we experienced in the past, that
the cobalt-catalysed DielseAlder reactions utilising pyridine-imine
type-ligands are solvent-dependent. The reactions are best per-
formed in dichloromethane while in other solvents, such as
toluene, tetrahydrofuran or acetone, the reactions proceeded
sluggishly. Therefore, the aspects of influencing the pre-
equilibrium by other solvents would be interesting, but out of
scope of this investigation.
As the steric influences of the pyridine-imine ligands are
virtually identical, this assumption must be rationalized in terms of
the electronic ligand donor strengths which increases from 6a to
6b. The increase in donor strength may thus reduce the amount of
available, neutral, heteroleptic and redoxeactive complexes of
Co(6b)Br2, therefore overcompensating the effect of the shift in the
redox potential seen for 6a, 6c, and 6d.
The easiest test to determine the proposed equilibrium of the
complexes Co(6a)Br2 and Co(6b)Br2 and their ionic homoleptic
counterparts in solution are measurements of the conductivity of
the solution. The conductivity measurements were performed in
dichloromethane (Fig. 3) and a significant difference of Co(6a)Br2
and Co(6b)Br2 was observed. In these measurements the cobalt
complex Co(6b)Br2 has
a
much higher conductivity
(4.41$10ꢀ6 S cmꢀ1) compared to Co(6a)Br2 (0.76$10ꢀ6 S cmꢀ1).
2+
N
N
N
N
OMe
OMe
[CoBr4]2-
+
Co
N
N
OMe
Co
Br
Br
6a
Co( )Br2
2+
N
N
NMe2
NMe2
[CoBr4]2-
+
Co
N
N
NMe2
4. Conclusion
Co
N
N
Br
Br
The results of the quantum mechanical investigation of different
pyridine-imine ligands of type 6 applied in cobalt-catalyzed Dielse
Alder reaction are in relatively good agreement with the
Co(6b)Br2
Scheme 3. Proposal for the equilibrium of the cobalt complexes in solution.

G. Hilt et al. / Journal of Organometallic Chemistry 749 (2014) 219e223
223
experimental results indicating the correct regioselectivities
References
although not with the needed high precision. However, the calcu-
lations and experimental data are in line for ligands of type 6 which
result in almost identical regioselectivity of the meta-selective Di-
elseAlder reaction. Accordingly, electronic modifications in the li-
gands do not influence the regioselectivity significantly. The
quantum mechanical investigation of different pyridine-imine li-
gands of type 5 were undertaken to predict changes in the regio-
selectivity. In this case even smaller changes in the energy of the
transition states were found resulting in almost identical pre-
dictions for the regioselectivities. These results prompted us not to
undergo the multi-step synthesis of such ligands of type 5. The
thermodynamic data for the reaction pathways and the expected
kinetic outcome suggested that the cobalt catalysts of type Co(6)Br2
should have performed very similarly in the cobalt-catalyzed meta-
selective DielseAlder reaction. However, the investigation of the
kinetic data of ligands of type 6 in the DielseAlder reaction led to
two results. First, we identified the halide abstraction step by the
Lewis acid to be rate determining for the activation of the complex
and second, that the unexpected long induction period for the co-
balt complex [Co(6b)Br2] could only be explained by a pre-
equilibrium diminishing the concentration of the catalyst precur-
sor complex [Co(6b)Br2] in solution. Conductivity measurements of
the dichloromethane solution revealed that complex [Co(6b)Br2] is
essentially not present in solution and that the equilibrium is
shifted towards the homoleptic complexes [Co(6b)2]2þ and
[CoBr4]2ꢀ which are not direct catalyst precursors for the catalyti-
cally active cobalt species. The electronic modifications led to an
altered length of the induction periods for the activation of the
cobalt complexes caused by different positions of the pre-equilibria
of the cobalt complexes of type Co(6)Br2 species in solution.
Accordingly, the main message to learn from this contribution is
that catalysts with different electronically modified ligands which
are activated in situ exhibit induction periods which depends not
only of the mode of activation but could also depend on pre-
equilibria. The evaluation of such ligand screenings should also
include the length of the induction periods and in the next phase
the catalytic performance of the catalysts.
[1] For general reviews concerning cobalt-catalyzed carbonecarbon bond for-
mations see: (a) G. Hilt, Eur. J. Org. Chem. (2012) 4441;
(b) G. Dominguez, Perez-Castells, J. Chem. Soc. Rev. 40 (2011) 3430;
(c) C. Gosmini, A. Moncomble, Isr. J. Chem. 50 (2010) 568;
(d) G. Cahiez, A. Moyeux, Chem. Rev. 110 (2010) 1435;
(e) F. Hebrard, P. Kalck, Chem. Rev. 109 (2009) 4272;
(f) D. Leboeuf, V. Gandon, M. Malacria, in: S. Ma (Ed.), Handbook of Cyclization
Reactions, vol. 1, Wiley-VCH, Weinheim, 2009, p. 367;
(g) B.R. Galan, T. Rovis, Angew. Chem. Int. Ed. 48 (2009) 2830;
(h) I. Omae, Appl. Organomet. Chem. 22 (2008) 149;
(i) W. Hess, J. Treutwein, G. Hilt, Synthesis (2008) 3537;
(j) I. Omae, Appl. Organomet. Chem. 21 (2007) 318;
(k) S. Kotha, E. Brahmachary, K. Lahiri, Eur. J. Org. Chem. (2005) 4741;
(l) M. Malacria, C. Aubert, J.L. Renaud, in: M. Lautens, B.M. Trost (Eds.),
Science of Synthesis: Houben-Weyl Methods of Molecular Transformations,
vol. 1, Thieme, Stuttgart, 2001, p. 439;
(m) S. Saito, Y. Yamamoto, Chem. Rev. 100 (2000) 2901.
[2] (a) A.-L. Auvinet, J.P.A. Harrity, G. Hilt, J. Org. Chem. 75 (2010) 3893;
(b) G. Hilt, C. Hengst, J. Org. Chem. 72 (2007) 7337;
(c) G. Hilt, C. Hengst, Synlett (2006) 3247;
(d) G. Hilt, W. Hess, F. Schmidt, Eur. J. Org. Chem. (2005) 2526;
(e) G. Hilt, S. Lüers, K.I. Smolko, Org. Lett. 7 (2005) 251;
(f) G. Hilt, S. Lüers, K. Harms, J. Org. Chem. 69 (2004) 624;
(g) G. Hilt, K.I. Smolko, Angew. Chem. Int. Ed. 42 (2003) 2795;
(h) G. Hilt, K.I. Smolko, B.V. Lotsch, Synlett (2002) 1081;
(i) G. Hilt, S. Lüers, K. Polborn, Isr. J. Chem. 41 (2001) 317;
(j) G. Hilt, F.-X. du Mesnil, Tetrahedron Lett. 41 (2000) 6757.
[3] (a) F. Erver, G. Hilt, Org. Lett. 14 (2012) 1884;
(b) M. Arndt, G. Hilt, A.F. Khlebnikov, S.I. Kozhushkov, A. de Meijere, Eur. J. Org.
Chem. (2012) 3112;
(c) G. Hilt, M. Danz, Synthesis (2008) 2257;
(d) G. Hilt, J. Janikowski, W. Hess, Angew. Chem. Int. Ed. 45 (2006) 5204.
[4] (a) J.R. Kuttner, S. Warratz, G. Hilt, Synthesis 44 (2012) 1293;
(b) G. Hilt, J. Janikowski, Org. Lett. 11 (2009) 773.
[5] dppe ¼ 1,2-bis(diphenylphosphino)ethane.
[6] Pyridine-imine ¼ abbreviation for ligands such as N-(pyridin-2-ylmethylene)
aniline.
[7] P. Mörschel, J. Janikowski, G. Hilt, G. Frenking, J. Am. Chem. Soc. 130 (2008)
8952.
[8] M. Frisch, et al., Gaussian03, Revision E.01, Gaussian, Inc, Wallingford CT,
2004.
[9] H.B. Schlegel, J. Comput. Chem. 3 (1982) 214.
[10] R. Ahlrichs, M. Bär, M. Häser, H. Horn, C. Kölmel, Chem. Phys. Lett. 162 (1989)
165.
[11] A.D. Becke, Phys. Rev. A 38 (1988) 3098.
[12] J.P. Perdew, Phys. Rev. B 33 (1986) 8822.
[13] F. Weigend, R. Ahlrichs, Phys. Chem. Chem. Phys. 7 (2005) 3297.
[14] R. Ahlrichs, Phys. Chem. Chem. Phys. 6 (2004) 5119.
[15] (a) K. Eichkorn, O. Treutler, M. Häser, R. Ahlrichs, Chem. Phys. Lett. 242 (1995)
652;
Acknowledgments
(b) K. Eichkorn, F. Weigend, O. Treutler, R. Ahlrichs, Theor. Chem. Acc. 97
(1997) 119;
We thank the Deutsche Forschungsgemeinschaft for financial
support.
(c) F. Weigend, Phys. Chem. Chem. Phys. 8 (2006) 1057.
[16] (a) P. Deglmann, F. Furche, J. Chem. Phys. 117 (2002) 9535;
(b) P. Deglmann, F. Furche, R. Ahlrichs, Chem. Phys. Lett. 362 (2002) 511;
(c) P. Deglmann, K. May, F. Furche, R. Ahlrichs, Chem. Phys. Lett. 384 (2004)
103.
Appendix A. Supplementary data
[17] A. Klamt, G. Schüürmann, J. Chem. Soc. Perkin Trans. 2 (5) (1993) 799.
[18] Used radii/Å;: r(H) ¼ 1.4040; r(C) ¼ 1.9890; r(N) ¼ 1.8135; r(O) ¼ 1.7784;
r(F) ¼ 1.7199; r(Co) ¼ 2.2230.
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.jorganchem.2013.09.020.
[19] For the calculation of the expectable maximum conductivity, see SI.