A spin-crossover Co(II) complex catalyzes the activation of sp3 C–H bonds by two-electron oxidants

Article abstract of DOI:10.1016/j.ica.2018.06.012

We recently found that a high-spin Co(II) complex with N,N'-dibenzyl-N,N'-bis(2-pyridylmethyl)-1,2-cyclohexanediamine (bbpc) was capable of catalyzing the oxidation of weak sp³ C–H bonds by iodosobenzene (PhIO) and meta-chloroperbenzoic acid (MCPBA). The bbpc compound could also be oxidized to a Co(III)-peroxo species capable of reacting with aldehydes. The ligand N-benzenesulfonyl-N,N’-bis(2-pyridylmethyl)-1,2-cyclohexanediamine (bsbpc) reacts with Co(ClO₄)₂ to yield a spin-crossover complex, [Co(bsbpc)(MeCN)₂](ClO₄)₂. The bsbpc complex catalyzes the oxidation of benzylic and allylic C–H bonds by PhIO and MCPBA to approximately the same extent as the high-spin bbpc compound. The bsbpc complex differs more substantially from the bbpc compound in that it cannot be oxidized to a stable Co(III)-peroxo complex and shows no reactivity with aldehydes.

Full text of DOI:10.1016/j.ica.2018.06.012

InorganicaChimicaActa482(2018)206–212
Contents lists available at ScienceDirect
Inorganica Chimica Acta
journal homepage: www.elsevier.com/locate/ica
Research paper
A spin-crossover Co(II) complex catalyzes the activation of sp3 C–H bonds by
two-electron oxidants
Angela Bell-Taylor, John D. Gorden, Emily E. Hardy, Christian R. Goldsmith⁎
Department of Chemistry and Biochemistry, Auburn University, Auburn, AL 36849, United States
A R T I C L E I N F O
A B S T R A C T
Keywords:
Cobalt
C-H activation
Homogeneous catalysis
We recently found that a high-spin Co(II) complex with N,N'-dibenzyl-N,N'-bis(2-pyridylmethyl)-1,2-cyclohex-
anediamine (bbpc) was capable of catalyzing the oxidation of weak sp3 C–H bonds by iodosobenzene (PhIO) and
meta-chloroperbenzoic acid (MCPBA). The bbpc compound could also be oxidized to a Co(III)-peroxo species
capable of reacting with aldehydes. The ligand N-benzenesulfonyl-N,N’-bis(2-pyridylmethyl)-1,2-cyclohex-
anediamine (bsbpc) reacts with Co(ClO4)2 to yield a spin-crossover complex, [Co(bsbpc)(MeCN)2](ClO4)2. The
bsbpc complex catalyzes the oxidation of benzylic and allylic C–H bonds by PhIO and MCPBA to approximately
the same extent as the high-spin bbpc compound. The bsbpc complex differs more substantially from the bbpc
compound in that it cannot be oxidized to a stable Co(III)-peroxo complex and shows no reactivity with alde-
hydes.
1. Introduction
Cobalt is much better known for its ability to catalytically activate sp3
C–H bonds through organometallic coupling reactions [14–16]. We
Iron complexes have long been explored for their abilities to cata-
lyze the oxidation of C–H bonds by two-electron terminal oxidants, such
as H2O2 [1]. The C–H activation performed by a variety of non-heme
iron oxidants is strongly influenced by the spin-state of the metal
center. A series of Fe(IV)-containing complexes with the 1,4,8,11-tet-
ramethyl-1,4,8,11-tetraazacyclotetradecane (TMC) ligand, [Fe(O)
(TMC)(X)]n+, was found to activate weak C–H bonds [2]. The rate of
these reactions is dependent on the identity of the axial ligand, X, trans
to the oxo. Faster rates were associated with ligands that allowed better
access to the high-spin quintet state. A subsequent study suggested that
the inability of low-spin Fe(IV) oxo complexes to activate stronger C–H
bonds derives partly from the need to access this more active high-spin
state during the reactions [3]. High-spin Fe(IV) oxo species were later
characterized and were found to be indeed more reactive and less stable
than their low-spin counterparts [4–6]. Spin-state also influences the
reactivity of Fe(III)-OOR and Fe(III)-OOH complexes. For the Fe(III)-
OOR compounds, the low-spin state favors O–O bond cleavage, whereas
the high-spin state promotes scission of the Fe–O bond [7,8]. High-spin
Fe(III)-OOH complexes are proposed to react with weak C–H bonds
more rapidly than low-spin analogs; the low-spin complexes, however,
are better at activating stronger C–H bonds [9].
recently found that a high-spin Co(II) complex with the bulky tetra-
dentate N-donor ligand N,N'-dibenzyl-N,N'-bis(2-pyridylmethyl)-1,2-
cyclohexanediamine (bbpc, Scheme 1) could catalyze the oxidation of
9,10-dihydroanthracene, 1,4-cyclohexadiene, and xanthene by iodoso-
benzene and meta-chloroperbenzoic acid [17]. One drawback of the
catalyst is that the activity persists for fewer than 30 min. Mass spec-
trometry of the C–H activation reactions reveals that the bbpc ligand
degrades over this time; the methylene bridges between the phenyl
groups and the amines appear to be susceptible to oxidation.
In an effort to prolong catalysis, we developed a new organic ligand
that contains a sulfonyl linkage in place of the methylene between the
metal-coordinating portion of the ligand and the phenyl group: N-
benzenesulfonyl-N,N’-bis(2-pyridylmethyl)-1,2-cyclohexanediamine
(bsbpc, Scheme 1). We hypothesized that removing weak and poten-
tially oxidizable benzylic C–H bonds would eliminate pathways for
catalyst degradation and thereby extend the reactivity with external
substrates. To our mild surprise, we found that the bsbpc ligand yielded
a spin-crossover, rather than a high-spin, Co(II) complex. Despite this
substantial change to the electronic character of the metal, the com-
pound is still capable of catalyzing C–H activation. The bbpc ligand had
also been found to stabilize a Co(III)-peroxo species which was capable
of aldehyde deformylation [17]. The bsbpc ligand, conversely, does not
support this chemistry.
Non-organometallic cobalt-mediated sp3 C–H activation, con-
versely, has been observed only rarely, and most of this has been
stoichiometric rather than catalytic with respect to the metal [10–13].
⁎
Corresponding author.
E-mail address: crgoldsmith@auburn.edu (C.R. Goldsmith).
https://doi.org/10.1016/j.ica.2018.06.012
Received 24 April 2018; Received in revised form 6 June 2018; Accepted 7 June 2018
Availableonline15June2018
0020-1693/©2018ElsevierB.V.Allrightsreserved.

A. Bell-Taylor et al.
InorganicaChimicaActa482(2018)206–212
Table 1
Selected crystallographic data for bsbpc and 1
Parameter
bsbpc
[Co(bsbpc)(MeCN)2](ClO4)2·MeCN (2)
Formula
MW
Crystal system
Space group
a (Å)
b (Å)
c (Å)
a (deg)
b (deg)
C24H28N4O2S
436.58
C30H37Cl2CoN7O10S
817.57
Triclinic
¯
Orthorhombic
Pbca (#61)
16.5482(16)
11.2725(10)
23.390(2)
90
90
90
4363.2(7)
8
P (
1
) (#2)
10.6448(5)
12.2374(6)
13.2826(6)
93.521(4)
90.421(4)
95.262(4)
1719.56(14)
2
Scheme 1.
2. Experimental section
g (deg)
V (Å3)
2.1. Materials
Z
Cryst color
T
Reflns collected
Unique reflns
R1 (F, I > 2σ(I))
wR2 (F2, all data)
Colorless
180(2)
21798
3989
0.0593
Orange
180(2)
21604
7040
0.0448
0.1197
Most chemicals were purchased from Sigma-Aldrich. Acetonitrile
(MeCN) and cyclohexanecarboxaldehyde (CCA) were bought from
Acros Organics. 2-Phenylpropionaldehyde (2-PPA) and iodosobenzene
(PhIO) were obtained from TCI America. All deuterated solvents were
purchased from Cambridge Isotopes. Methanol (MeOH) and ethanol
(EtOH) were acquired from Pharmaco-Aaper and EMD Chemicals, re-
spectively. H2O2 (50% wt) was bought from Fisher Chemical and was
calibrated by titration of KMnO4 and H2SO4 in water. N,N’-Bis(2-pyr-
0.1139
R1 = Σ∣∣Fo∣ − ∣Fc∣∣/Σ∣Fo∣; wR2 = [Σw(Fo2 − Fc2)2/Σw(Fo2)2]1/2
.
applied using the program SADABS. The structure for bspbc was solved
with the ShelXL software package using direct methods and refined
using the Olex2 program. A single crystal of [Co(bsbpc)(MeCN)2](ClO4)
(1) was mounted on a CryoLoop with Krytox oil and optically aligned
on a Rigaku XtaLAB Synergy diffractometer coupled to a Rigaku Hypix
detector. Initial intensity measurements were performed using
PhotonJet source, 30 microfocused sealed tube (MoKα,
idylmethyl)-1,2-cyclohexanediamine was prepared through
a pre-
viously reported procedure [18]. PhIO was kept in a 253 K freezer when
not in use. H2O2 and MCPBA were kept in a 277 K refrigerator when not
in use. Xanthene and 9,10-dihydroanthracene (DHA) were re-
crystallized twice from EtOH before their use in reactivity assays. CCA
was distilled prior to its use.
Caution: Perchlorate salts of metal complexes, peroxides, and per-
acids are all potentially explosive and should be handled with care.
Precautionary measures include working with small quantities and low
concentrations of these reagents. Appropriate safety measures, such as
protective shields, should be employed during their syntheses and
subsequent handling.
a
a
W
λ = 0.71073 Å) with high-brilliance and high-performance focusing
multilayer optics. The diffraction images for 1 were processed and
scaled using the CrysAlisPro software. Empirical absorption correction
using spherical harmonics, implemented in SCALE3 ABSPACK scaling
algorithm was applied. The structure was solved with the ShelXT soft-
ware package and refined using XH software. Selected crystallographic
information is listed on Table 1. Atomic coordinates and additional
structural information are provided as Supporting Information. The two
reported structures have been deposited in the Cambridge Crystal-
lographic Data Centre (CCDC 1834775-1834776).
2.2. Instrumentation
1H nuclear magnetic resonance (NMR) spectra were recorded on
either a 600, 400, or 250 MHz AV Bruker NMR spectrometer. Optical
data were collected with a Varian Cary 50 spectrophotometer and were
processed and analyzed using software from the WinUV Analysis Suite.
IR data were obtained using a Shimadzu IR Prestige-21 FT-IR spectro-
photometer. A Thermo Scientific Trace GC Ultra Gas Chromatograph
and Thermo Scientific TR-1 and TG-WAXMS columns were used for gas
chromatography (GC). A Johnson Matthey magnetic susceptibility
balance (model MK I#7967) was used to measure the magnetic mo-
ments of solid samples of the metal complexes. Pascal’s constants were
used to estimate the diamagnetic correction to the susceptibility [19].
High-resolution mass spectrometry (HR-MS) data were collected at the
Mass Spectrometer Center at Auburn University on a Bruker microflex
LT MALDI-TOF mass spectrometer via direct probe analysis operated in
the positive ion mode. Crystalline samples were dried, stored under N2,
and sent to Atlantic Microlabs (Norcross, GA) for elemental analysis.
2.4. Synthesis
N-Benzenesulfonyl-N,N’-bis(2-pyridylmethyl)-1,2-cyclohex-
anediamine
(bsbpc).
N,N’-Bis(2-pyridylmethyl)-1,2-cyclohex-
anediamine (0.593 g, 100 mM) and benzenesulfonyl chloride (0.707 g,
100 mM) were dissolved in 20 mL of dry THF under N2 and stirred for 1
h at room temperature (RT). 20 mL of deionized H2O was subsequently
added, and the organic solvent was removed by rotovaporation. A sa-
turated KOH solution was added dropwise until the pH was above 9;
precipitation was observed during the basification. The product was
extracted with three portions of CH2Cl2. These extracts were combined,
dried with Na2SO4, and filtered. The CH2Cl2 was removed, and the
residue was dissolved in a minimal amount of MeCN and cooled in a
refrigerator. Slow diffusion of Et2O into the cooled solution deposited
light brownish crystals. Rinsing with cold ether removed the brown
color to yield the product as a white crystalline solid (0.646 g, 66%).
The crystals were suitable for single crystal X-ray diffraction. 1H NMR
(CDCl3, 294 K, 400 MHz): δ 8.48 (1H, d, J = 4.84 Hz), 8.20 (1H, d,
J = 4.24 Hz), 7.77 (2H, q, J = 8.75 Hz), 7.54 (4H, m), 7.38 (2H, t,
J = 7.74 Hz), 7.11 (3H, m), 4.70 (1H, q, J = 8.67 Hz), 4.37 (1H, q,
J = 8.7 Hz), 3.74 (2H, m), 3.25 (1H, q, J = 10.99 Hz), 2.42 (1H, s),
2.30 (1H, s), 2.08 (1H, m), 1.61 (2H, s), 1.43 (2H, m), 1.22 (1H, m),
1.08 (2H, quintet, J = 10.59 Hz). 13C NMR (CDCl3, 294 K, 100 MHz):
160.59, 157.96, 148.79, 148.74, 140.80, 136.66, 136.30, 132.33,
128.89, 127.20, 123.23, 122.41, 121.553, 121.51, 58.47, 52.24, 32.04,
29.73, 25.59, 24.22. IR (KBr, cm−1): 3448 (w), 3246 (m), 3059 (w),
2.3. Crystallographic Studies
A single crystal of bsbpc was mounted on a glass fiber with Krytox
oil and optically aligned on a Bruker APEX Quazar X-ray diffractometer
using a digital camera. Initial intensity measurements were performed
using an IμSX-ray,
a 30 W microfocused sealed tube (MoKα,
λ = 0.71073 Å) with high-brilliance and high-performance focusing
Quazar multilayer optics. Standard APEX software was used for de-
termination of the unit cell and data collection control. The SAINT
software was used for data integration including Lorentz and polar-
ization corrections, and semi-empirical absorption corrections were
207

A. Bell-Taylor et al.
InorganicaChimicaActa482(2018)206–212
3007 (w), 2988 (w), 2953 (w), 2930 (m), 2857 (w), 2804 (w), 2362
(w), 2337 (w), 1592 (s), 1571 (m), 1504 (m), 1477 (m), 1437 (m), 1388
(m), 1369 (m), 1327 (m), 1289 (m), 1266 (w), 1152 (m), 1088 (w),
1033(m), 998 (m), 963 (m), 916 (m), 900 (m), 880 (m), 862 (w), 827
(m), 796 (m), 760 (m), 691 (m), 598 (m), 552 (m), 465 (w). MS (ESI):
Calcd [bsbpcH]+: 437.2011; Found: 437.1924.
different UV/vis spectra in MeCN and MeOH, suggesting that the MeCN
ligands are readily exchangeable (Fig. S6). The bands in MeCN between
400 and 550 nm have relatively low intensities; the energies and in-
tensities of these bands are consistent with d-d transitions for high-spin
and six-coordinate Co(II) ions [20,21]. The complex was analyzed by
electron paramagnetic resonance (EPR) at 4 K and 77 K (Fig. 2). At 77
K, we observed a strong feature at g = 2.26, consistent with low-spin Co
(II). At 4 K, we observed the same feature as well as ones at a lower field
which are consistent with high-spin Co(II) [22]. At first glance, the
absence of the high-spin signal at 77 K is odd since raising the tem-
perature should increase the proportion of the Co(II) that is high-spin.
High-spin Co(II) is not visible at 77 K, however, due to rapid spin re-
laxation [22,23]. The rapid spin relaxation precludes an accurate ac-
count of the relative concentrations of the high-spin and low-spin
species by spin quantification. When the high-spin portion of the
spectrum is treated as an effective S’ = ½ system [24], it can be fit to
g′x = 3.2, g′y = 3.8, g′z = 5.8. Room temperature magnetic susceptibility
measurements on a powdered sample are consistent with a magnetic
dipole moment of 4.0 μB, which is at the lower end of the range an-
ticipated for a high-spin Co(II) complex [23].
Complex 1 can be crystallized from MeCN/ether mixtures (Fig. 1B,
Table 1). Each unit cell contains one [Co(bsbpc)(MeCN)2]2+ ion, two
perchlorate anions, and an MeCN solvent molecule. The coordination
around the Co(II) consists exclusively of N-donors; the O atoms from the
sulfonyl group do not bond directly to the metal center. The metrical
parameters associated with the dication are consistent with the metal
center being low-spin Co(II). Four of the Co-N bonds are markedly
shorter than the other two, with lengths ranging from 1.986 to 2.021 Å.
Cationic high-spin Co(II) complexes with near-octahedral coordination
geometries, conversely, typically have Co-N bond lengths between 2.10
and 2.25 Å [21,25,26]. The Co-N bonds to the tertiary amine and to the
MeCN molecule trans to it are much longer, at 2.440 Å and 2.135 Å,
respectively. The elongation along this N(5)-Co(1)-N(3) axis is con-
sistent with the Jahn-Teller distortion anticipated for a low-spin d7
metal center. The spin-state differs from that of the Co(II) complex with
bbpc; although this compound was not structurally characterized, a
variety of spectroscopic measurements suggest that it is entirely high-
spin both as a solid and in MeCN solutions, even when cooled to 4 K
[17].
[(N-Benzenesulfonyl-N,N’-bis(2-pyridylmethyl)-1,2-cyclohex-
anediamine)-diacetonitrilocobalt(II)] perchlorate ([Co(bspbc)
(MeCN)2](ClO4)2, 1). Co(ClO4)2∙6H2O (0.11 g, 30 mM) and bsbpc
(0.13 g, 30 mM) were placed under N2 and dissolved in 10 mL of MeCN.
After stirring for 30 min at RT, the solvent was removed by rotovap.
The residue was dissolved in a minimal volume of heated MeCN and
cooled in a refrigerator. Slow diffusion of Et2O into the saturated so-
lution yielded the product (0.150 g, 63%) as red crystals suitable for
single crystal X-ray diffraction. Solid-state magnetic susceptibility
(294 K): µeff = 4.0 µB. Optical spectroscopy (MeCN, 298 K): 434 nm
(20 M−1 cm−1), 496 nm (27 M−1 cm−1). Optical spectroscopy (MeCN,
298 K): 540 nm (shoulder, 31 M−1 cm−1). EPR (MeCN, 77 K, X-band):
g = 2.26. EPR (MeCN, 4 K, X-band): g = 2.26, g’x = 3.2, gy’ = 3.8,
g’z = 5.8. Elem. anal. Calcd for C28H34CoCl2N6O10S·0.5CH3CN: C, 43.70;
H, 4.49; N, 11.42. Found: C, 43.62; H, 4.61; N, 11.47.
2.5. Reactivity
Unless stated otherwise, all reactivity studies were done at 298 K in
MeCN under air. The yields of organic products were determined by GC;
all reported values are the average of the measurements for at least
three independent reactions. The listed errors represent one standard
deviation. The identities and yields of products were confirmed by
comparing the GC retention times with those of authentic compounds
and by comparing the peak integrations with that of a non-reactive 1,2-
dichlorobenzene internal standard.
For the C–H activation reactions, the Co(II) compound 1, non-re-
active internal standard, and hydrocarbon substrate were dissolved in
2.5 mL of MeCN, with initial concentrations of 1.0 mM and 50 mM,
respectively. 25 equiv. of either PhIO or MCPBA were added as a solid,
after which the reaction vessel was sealed. During the reaction, aliquots
were removed via syringe, diluted with ether, filtered through silica gel,
and analyzed via GC. Parallel reactions were run in CD3CN and ana-
lyzed by 1H NMR to confirm that no products were lost via this work-
up.
3.2. Reactivity
Aldehyde deformylation reactions were attempted with CCA and 2-
PPA as the substrates; these followed a previously described protocol
[17]. Neither of the anticipated organic products of CCA oxidation,
cyclohexanone and cyclohexene, were observed by GC.
Mixtures of 1 and iodosobenzene (PhIO) can activate weak C–H
bonds, such as the benzylic ones found in 9,10-dihydroanthracene
(DHA) and xanthene or the allylic ones found in 1,4-cyclohexadiene
(CHD) (Scheme 2) [27]. As was observed for [Co(bbpc)(MeCN)2]
(ClO4)2, an additional Lewis acid beyond the Co(II) is not needed to
enable this reactivity [17].
3. Results
3.1. Synthesis and Spectroscopic Characterization
The substrate DHA is converted exclusively into anthracene
(Table 2). The reactivity halts within 30 min, with peak yields of ap-
proximately 7 equiv. of anthracene per equiv. of Co(II). Oxygenated
products, such as anthrone and anthroquinone, are not observed, even
when the reactions are run under air. CHD is oxidized exclusively to
benzene when PhIO is used as the terminal oxidant. The turnover
numbers are much higher than they are for DHA, with approximately
19 equiv. of benzene produced per equiv. of catalyst. Xanthene is oxi-
dized exclusively to xanthone and appears to be about as reactive a
substrate as DHA, based on the number of turnovers. The yields of
benzene and xanthone likewise maximize within 30 min and are not
significantly impacted by exposure to air. No oxidation of either DHA,
CHD, or xanthene occurs in the absence of the metal complex [17].
Neither cyclohexane nor cyclohexene react with 1 and PhIO to ob-
servable degrees, even when 500 mM of these substrates are added.
The sulfonyl linkage between the phenyl group and the rest of the
bsbpc ligand remains intact, but this does not prevent catalyst decom-
position. The side reactivity is instead directed towards the picolyl arms
N-Benzenesulfonyl-N,N′-bis(2-pyridylmethyl)-1,2-cyclohex-
anediamine (bsbpc) was prepared in 66% yield from a reaction between
benzenesulfonyl chloride and N,N’-bis(2-pyridylmethyl)-1,2-cyclohex-
anediamine. The compound can be crystallized from mixtures of MeCN
and ether (Fig. 1A, Table 1), allowing pure bsbpc to be obtained
without chromatography. The only complication in the synthesis is the
sensitivity of the benzenesulfonyl chloride starting material to water;
this necessitates the use of rigorously dry THF. We initially pursued the
doubly substituted ligand N,N′-di(benzenesulfonyl)-N,N’-bis(2-pyr-
idylmethyl)-1,2-cyclohexanediamine, but reacting N,N′-bis(2-pyr-
idylmethyl)-1,2-cyclohexanediamine with larger excesses of benzene-
sulfonyl chloride has thus far resulted in mixtures of that target product
and bsbpc. These compounds are difficult to separate from each other,
prompting us to first focus on the more readily accessible bsbpc.
Mixing bsbpc and Co(ClO4)2 in MeCN provides [CoII(bsbpc)
(MeCN)2](ClO4)2 (1) in 63% yield (Fig. 1). The complex displays
208

A. Bell-Taylor et al.
InorganicaChimicaActa482(2018)206–212
Fig. 1. ORTEP representations of (A) bspbc and (B) [Co(bsbpc)(MeCN)2]2+. All thermal ellipsoids are drawn at 50% probability. All H atoms, solvent molecules, and
counteranions have been removed for clarity. The atoms have been relabeled from their CIF designations in order to facilitate comparison. Fuller details regarding the
structures can be found in Table 1 and the CIF file submitted as Supporting Information. The bond lengths for [Co(bsbpc)(MeCN)2]2+ are as follows: Co(1)-N(1),
2.000(2) Å; Co(1)-N(2), 2.012(2) Å; Co(1)-N(3), 2.4395(19) Å; Co(1)-N(4), 2.020(2) Å; Co(1)-N(5), 2.134(2) Å; Co(1)-N(6), 1.986(2) Å.
Substituting meta-chloroperbenzoic acid (MCPBA) for PhIO results
in similar chemistry, albeit with generally lower yields of anthracene
and benzene (Table 3). The yield of xanthone from xanthene, con-
versely, increases. The only major difference with respect to the product
distributions is that MCPBA can directly react with CHD to yield an
epoxide, but not benzene [17]. Xanthene and DHA do not react with
MCPBA at all without a catalyst. This competing reaction lowers the
yield of benzene by consuming the terminal oxidant prior to its reaction
with 1. The two reactions completely deplete the MCPBA by 30 min
when CHD is used as the substrate. H2O2 and O2 are not competent
terminal oxidants for the C–H activation catalyzed by 1. The bbpc
system displayed identical trends with respect to the terminal oxidants,
being capable of C–H activation with PhIO and MCPBA but not with
H2O2 or O2 [17].
The bsbpc ligand differs from bbpc in that it does not appear to
support aldehyde deformylation chemistry. Reacting 1 with H2O2 and a
base, such as triethylamine, does not result in the formation of a stable
Co(III)-peroxo species, as assessed by the lack of change in its UV/vis
spectrum. Mixtures of 1, H2O2, and base likewise do not react with
Fig. 2. X-Band EPR spectra of a solution of 1.0 mM 1 in MeCN at 77 K (red) and
4 K (blue). For the low-spin signal observed at both temperatures, g = 2.26. For
the high-spin signal observed at 4 K, g’x = 3.2, gy’ = 3.8, gz’ = 5.8. The solution
cyclohexanecarboxaldehyde or 2-propionphenylaldehyde, the two most
commonly employed substrates for aldehyde deformylation studies
was prepared from a crystalline sample of 1.
[17,28–31].
4. Discussion
We originally were interested in N-benzenesulfonyl-N,N’-bis(2-pyr-
idylmethyl)-1,2-cyclohexanediamine (bsbpc, Scheme 1) as a more oxi-
datively robust substitute for N,N'-dibenzyl-N,N'-bis(2-pyridylmethyl)-
1,2-cyclohexanediamine (bbpc), which we used in several studies of
Scheme 2.
high-spin Mn(II), Fe(II), and Co(II) complexes [17,32–34]. The Co(II)
complex with bbpc catalyzes the oxidation of weak sp3 C–H bonds by
two-electron oxidants, such as iodosobenzene (PhIO) and meta-chlor-
operbenzoic acid (MCPBA), and can be oxidized to a Co(III)-peroxo
species that can subsequently react with aldehydes [17]. The bsbpc and
bbpc ligands are similar in that they are both neutral and tetradentate
N-donor ligands. The bsbpc ligand differs from bbpc in that only one of
the amines is linked to a phenyl ring and that the linkage between the
two functionalities is a sulfonyl group rather than a methylene. At-
tempts to isolate a closer analog to bbpc with two sulfonyl-linked
phenyl groups are ongoing.
(Figs. S8 and S9). MS analysis of the substrate-free reactions between 1
and PhIO shows two new m/z peaks at 345.0536 and 280.9838 which
are consistent with the loss of one and two picolyl groups respectively.
Another m/z peak at 453.1954 corresponds to the addition of an O
atom to the ligand, indicating that it has been oxidized. Possible sites of
oxidation include the methylene carbons between the pyridine rings
and the diamine backbone, the amines, or the pyridines. The observed
turnover numbers are all below the experimental maximum of 25,
suggesting that complex 1 may lose its catalytic activity before the
terminal oxidant is depleted. The MS data show no clear evidence of
higher-valent cobalt complexes, such as mononuclear Co(III)-OH or Co
(IV)=O species.
The bsbpc ligand is prepared in two steps from commercially
available materials, and pure material can be obtained in moderately
high yields (∼65%) from the final reaction mixture without the need
for chromatography (Fig. 1A). A Co(II) complex, [Co(bsbpc)(MeCN)2]
209

A. Bell-Taylor et al.
InorganicaChimicaActa482(2018)206–212
Table 2
Oxidation of Hydrocarbons by PhIO Catalyzed by 1
Substrate
Atmosphere
Air
Time (min)
Products
Yield (%)a
TONb
TON (bbpc)c
9,10-Dihydroanthracene (DHA)
30
60
120
30
60
Anthracene
30( 2)
28( 2)
28( 2)
32( 4)
32( 2)
30( 4)
7.3
7.1
7.2
7.8
8
12
10
11
11
N2
Anthracene
120
7.6
1,4-Cyclohexadiene (CHD)
Air
N2
30
60
120
30
60
Benzene
Benzene
72( 4)
76( 2)
74( 2)
78( 8)
76( 6)
78( 4)
18
19
19
20
19
19
16
14
120
Xanthene
Air
N2
30
60
120
30
60
Xanthone
Xanthone
28( 2)
34( 4)
32( 2)
28( 2)
30( 2)
32( 2)
7.2
8.4
7.9
6.8
7.4
8
6
8.5
12
3.2
2.9
3.7
120
Standard reaction conditions: All reactions were run at 298 K in 2.5 mL of MeCN. The starting concentrations of the cobalt(II) catalyst (1) and the substrate in all
reactivity assays were 1.0 mM and 50 mM, respectively. 25 equiv. of PhIO, relative to the 1, were added at the beginning of the reaction. For each time point, an
aliquot of the reaction mixture was diluted with ether, filtered through silica gel, and analyzed via GC. The products were identified by comparing the retention times
with those of authentic samples of anthracene, xanthone, and benzene. The concentrations of each organic product were calibrated relative to that of an internal
standard (dichlorobenzene) with a known concentration.
aPhIO is the limiting reagent.
bTurnover number, defined as the number of moles of oxidized organic product generated per mole of 1.
cFrom analogous reactions with [Co(bbpc)(MeCN)2](ClO4)2 reported in Reference 17.
high-spin d7 measurements and the high sensitivity of these measure-
Table 3
Oxidation of Hydrocarbons by MCPBA Catalyzed by 1.
ments to sample preparation [35], we cannot provide an estimate of the
proportions of high-spin and low-spin Co(II) in solution at room tem-
perature.
Substrate Atmosphere Time
(min)
Products
Yield (%)a TONb TON
(bbpc)d
The low-spin form of 1 crystallizes readily from mixtures of MeCN
and ether (Fig. 1B). The Co-N bond lengths and the Jahn-Teller elon-
gation along the N(3)-Co(1)-N(5) axis provide strong support for the
spin-state assignment. Not surprisingly, the longest Co-N bond is to the
tertiary amine with the benzenesulfonyl substituent. With the bbpc li-
gand, both amines are connected to benzyl groups. That there is sub-
stantial steric bulk around both amines should promote longer Co-N
bonds and consequently favor high-spin Co(II) complexes. A similar
trend has been observed for Fe(II) complexes with pyridylamine li-
gands: methylating the 6-positions of the pyridines is enough to change
the spin-state of the Fe(II) from low-spin to high-spin.[36–38] That only
one of the amines in bsbpc has been derivatized may allow the other
donor atoms to approach the cobalt closely enough to support the low-
spin state. We were unable to obtain structural data for the high-spin
form of 1. The UV/vis data in MeCN are consistent with a high-spin and
six-coordinate Co(II) complex [20,21]; the d-d bands for lower-co-
ordinate Co(II) centers are normally much more intense [39].
The change to the spin-state does not significantly impact the C–H
activation reactivity (Table 2). The replacement of the methylene group
by the sulfonyl linkage does not prolong the catalysis, and neither 1 nor
[Co(bbpc)(MeCN)2]2+ promotes C–H activation beyond 30 min. De-
spite the introduction of the sulfonyl group to the ligand framework,
the catalyst still degrades. MS analysis reveals both oxidation of the
ligand and loss of one or both of the picolyl arms (Fig. S9). The me-
thylene linkages between the pyridine rings and the amines become the
new weak spots of the catalyst. The product distributions of both Co(II)
catalysts are identical, with anthracene, benzene, and xanthone being
observed as the sole organic products from DHA, CHD, and xanthene,
respectively. Both catalysts work with either PhIO or MCPBA as the
terminal oxidant but fail to oxidize hydrocarbons when either O2 or
H2O2 is used in this capacity. None of the four terminal oxidants was
found to activate benzylic or allylic sp3 C–H bonds without a catalyst at
298 K within 120 min.
DHA
CHD
Air
N2
30
30
Anthracene 20( 1)
Anthracene 18( 2)
4.8
4.7
8.8
7.2
Air
N2
30
30
Benzene
Epoxidec
Benzene
Epoxidec
51( 1)
49( 1)
66( 4)
36( 12)
13
16
13
13
Xanthene Air
N2
30
30
Xanthone
Xanthone
38( 4)
36( 6)
9.5
9.1
Standard reaction conditions: All reactions were run at 298 K in 2.5 mL of
MeCN. The starting concentrations of the cobalt(II) catalyst (1) and the sub-
strate in all reactivity assays were 1.0 mM and 50 mM, respectively. 25 equiv. of
MCPBA, relative to the 1, were added at the beginning of the reaction. For each
time point, an aliquot of the reaction mixture was diluted with ether, filtered
through silica gel, and analyzed via GC. The products were identified by
comparing the retention times with those of authentic samples of anthracene,
benzene, and 1,4-cyclohexadiene oxide. The 1,4-cyclohexadiene monoxide
standard was synthesized from the uncatalyzed reaction between CHD and
MCPBA. The concentrations of each organic product were calibrated relative to
that of an internal standard (dichlorobenzene) with a known concentration.
aMCPBA is the limiting reagent.
bTurnover number, defined as the number of moles of oxidized organic product
generated per mole of 1. c1,4-Cyclohexadiene monoxide.
dFrom analogous reactions with [Co(bbpc)(MeCN)2](ClO4)2 reported in Ref.
[17].
(ClO4)2 (1), can be prepared by mixing Co(ClO4)2 and the bsbpc ligand
in MeCN. Although the donor atoms are similar for bbpc and bsbpc, 1
displays a spin-crossover whereas the bbpc complex is entirely high-
spin, even at 4 K [17]. The room temperature magnetic susceptibility of
a solid sample of 1 is more consistent with a high-spin d7 metal center,
but the EPR spectra at 77 K and 4 K show a strong spectroscopic feature
at g = 2.26 indicative of low-spin Co(II) (Fig. 2). Due to the large
contribution of orbital angular momentum to the magnetic moments for
210

A. Bell-Taylor et al.
InorganicaChimicaActa482(2018)206–212
reminiscent of that displayed by an entirely high-spin Co(II) complex
recently reported by our laboratory, with the biggest observed differ-
ences relating to the reactivity of the substrate xanthene. The high-spin
Co(II) complex with bbpc could also be oxidized to a Co(III)-peroxo
complex capable of reacting stoichiometrically with aldehyde, but the
bsbpc complex is unable to perform this chemistry.
Acknowledgements
Scheme 3.
The described work was financially supported by Auburn University
and a grant from the American Chemical Society- Petroleum Research
Fund (grant number 55301-ND3). We thank Prof. Evert Duin for his
assistance with the acquisition and interpretation of the EPR data.
The bsbpc catalyst is not as active as the high-spin bbpc complex
when DHA is the substrate. With this substrate, the bbpc catalyst turns
over 12 times with PhIO as the terminal oxidant and 7 times with
MCPBA [17]. The activities of the two Co(II) complexes are approxi-
mately equal with CHD. Surprisingly, 1 is more efficient at oxidizing
xanthene to xanthone, and these reactions are not sensitive to O2. With
[Co(bbpc)(MeCN)2]2+ conversely, the yield of xanthone steadily in-
creases over time when exposed to air. The increasing yield of xanthone
for the bbpc catalyst was proposed to result from a free radical pro-
pagation rather than the regeneration of the cobalt catalyst [17], but
this chemistry appears to be negligible when 1 is used in its place. The
substrates that get oxidized to dehydrogenated products (DHA and
CHD) react identically under air and N2 with either 1 or [Co(bbpc)
(MeCN)2]2+ as the catalyst.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in the
online version, at https://doi.org/10.1016/j.ica.2018.06.012.
References
[1] M. Costas, M.P. Mehn, M.P. Jensen, L. Que Jr., Chem. Rev. 104 (2004) 939–986.
[2] C.V. Sastri, J. Lee, K. Oh, Y.J. Lee, J. Lee, T.A. Jackson, K. Ray, H. Hirao, W. Shin,
J.A. Halfen, J. Kim, L. Que Jr., S. Shaik, W. Nam, Proc. Natl. Acad. Sci., U.S.A. 104
(2007) 19181–19186.
There was a great deal of ambiguity with respect to the identity of
the active oxidant in the C–H activation catalyzed by [Co(bbpc)
(MeCN)2]2+, and we had previously speculated that a Co(III)-hydroxo
or Co(IV)-oxo species could be responsible [17]. Regrettably, the bsbpc
ligand did not stabilize any potential intermediates enough to allow for
their identification by MS, UV/vis, or EPR. The lack of CHD epoxidation
with PhIO as the terminal oxidant may be evidence against a Co(II)-
PhIO complex as the metal-based oxidant, for PhIO complexes with
other metal ions have been found to convert alkenes to epoxides
[40,41].
The bsbpc and bbpc cobalt compounds differ much more markedly
in their aldehyde deformylation chemistry. The bbpc compound can be
oxidized to a Co(III)-peroxo species that could convert cyclohex-
anecarboxaldehyde (CCA) to mixtures of cyclohexanone and cyclo-
hexene [17]. Conversely, we were unable to convert 1 to a Co(III)-
peroxo species. Reactions between CCA and mixtures of 1, H2O2, and
base produce neither cyclohexanone nor cyclohexene, suggesting that
the Co(III)-peroxo oxidants necessary for this reaction do not form even
transiently. Previously reported cobalt-promoted aldehyde deformyla-
tion has been found to be highly sensitive to changes in the ligand. The
five-coordinate Co(II) complex with the tetraazamacrocyclic ligand 13-
TMC (Scheme 3) has relatively long Co-N bonds, with the equatorial Co-
N bond distances averaging 2.068 Å; this compound reacts with H2O2
and base to yield a Co(III)-peroxo species that can itself react with al-
[3] K.-B. Cho, H. Chen, D. Janardanan, S.P. de Visser, S. Shaik, W. Nam, Chem.
Commun. 48 (2012) 2189–2191.
[4] J. England, Y. Guo, E.R. Farquhar, V.G. Young Jr., E. Münck, L. Que Jr., J. Am.
Chem. Soc. 132 (2010) 8635–8644.
[5] J. England, M. Martinho, E.R. Farquhar, J.R. Frisch, E.L. Bominaar, E. Münck,
L. Que Jr., Angew. Chem. Int. Ed. 48 (2009) 3622–3626.
[6] D.C. Lacy, R. Gupta, K.L. Stone, J. Greaves, J.W. Ziller, M.P. Hendrich, A.S. Borovik,
J. Am. Chem. Soc. 132 (2010) 12188–12190.
[7] N. Lehnert, R.Y.N. Ho, L. Que Jr., E.I. Solomon, J. Am. Chem. Soc. 123 (2001)
12802–12816.
[8] F. Namuswe, T. Hayashi, Y. Jiang, G.D. Kasper, A.A. Narducci Sarjeant, P. Moënne-
Loccoz, D.P. Goldberg, J. Am. Chem. Soc. 132 (2010) 157–167.
[9] L.V. Liu, S. Hong, J. Cho, W. Nam, E.I. Solomon, J. Am. Chem. Soc. 135 (2013)
3286–3299.
[10] F.F. Pfaff, S. Kundu, M. Risch, S. Pandian, F. Heims, I. Pryjomska-Ray, P. Haack,
R. Metzinger, E. Bill, H. Dau, P. Comba, K. Ray, Angew. Chem. Int. Ed. 50 (2011)
1711–1715.
[11] D.C. Lacy, Y.J. Park, J.W. Ziller, J. Yano, A.S. Borovik, J. Am. Chem. Soc. 134
(2012) 17526–17535.
[12] A.I. Nguyen, R.G. Hadt, E.I. Solomon, T.D. Tilley, Chem. Sci. 5 (2014) 2874–2878.
[13] O. Villanueva, N.M. Weldy, S.B. Blakey, C.E. MacBeth, Chem. Sci. 6 (2015)
6672–6675.
[14] Z.-Z. Zhang, Y.-Q. Han, B.-B. Zhan, S. Wang, B.-F. Shi, Angew. Chem. Int. Ed. 56
(2017) 13145–13149.
[15] W.N. Palmer, P.J. Chirik, ACS Catal. 7 (2017) 5674–5678.
[16] M. Sen, B. Emayavaramban, N. Barsu, J.R. Premkumar, B. Sundararaju, ACS Catal. 6
(2016) 2792–2796.
[17] Q. Zhang, A. Bell-Taylor, F.M. Bronston, J.D. Gorden, C.R. Goldsmith, Inorg. Chem.
56 (2017) 773–782.
[18] R.R. Fenton, R.S. Vagg, P. Jones, P.A. Williams, Inorg. Chim. Acta 128 (1987)
219–229.
[19] G.A. Bain, J.F. Berry, J. Chem. Educ. 85 (2008) 532–536.
[20] F.A. Cotton, G. Wilkinson, Advanced Inorganic Chemistry, 5 ed., John Wiley & Sons,
New York, 1988.
[21] R.J.M. Klein Gebbink, R.T. Jonas, C.R. Goldsmith, T.D.P. Stack, Inorg. Chem. 41
(2002) 4633–4641.
[22] F.S. Kennedy, H.A.O. Hill, T.A. Kaden, B.L. Vallee, Biochem. Biophys. Res. Comm.
48 (1972) 1533–1539.
[23] R.S. Drago, Physical Methods for Chemists, 2 ed., Surfside Scientific Publishers,
Gainesville, FL, 1992.
[24] O. Kahn, Molecular Magnetism, VCH, New York, NY, 1993.
[25] Q. Zhang, J.D. Gorden, R.J. Beyers, C.R. Goldsmith, Inorg. Chem. 50 (2011)
9365–9373.
[26] R.D. Shannon, Acta Cryst., A 32 (1976) 751–767.
[27] F.G. Bordwell, J.-P. Cheng, G.-Z. Ji, A.V. Satish, X. Zhang, J. Am. Chem. Soc. 113
(1991) 9790–9795.
dehydes [29,42].
A five-coordinate Co(II) complex with 12-TMC
(Scheme 3) has shorter equatorial Co-N bonds (average 1.988 Å). This
compound can be oxidized to a Co(III)-peroxo species, but it does not
react with aldehydes. Although this is admittedly a small sample size,
shorter Co-ligand bond distances thus far seem to disfavor aldehyde
deformylation. The connection between the aldehyde deformylation
reactivity and spin-state remains ambiguous. In the TMC chemistry,
both Co(II) precursors are high-spin and both Co(III)-peroxo complexes
are low-spin, and spin-crossover activity was not noted for any of these
complexes.
5. Conclusions
[28] M.S. Seo, J.Y. Kim, J. Annaraj, Y. Kim, Y.-M. Lee, S.-J. Kim, J. Kim, W. Nam, Angew.
Chem. Int. Ed. 46 (2007) 377–380.
[29] J. Cho, R. Sarangi, W. Nam, Acc. Chem. Res. 45 (2012) 1321–1330.
[30] J. Annaraj, Y. Suh, M.S. Seo, S.O. Kim, W. Nam, Chem. Commun. (2005)
4529–4531.
A neutral N-donor ligand, bsbpc, forms a spin-crossover complex
with Co(II). The Co(II) complex is capable of catalyzing the oxidation of
weak sp3 C–H bonds by PhIO and MCPBA. Overall, the catalysis is
[31] A. Zielinksa, L. Skulski, Tetrahedron Lett. 45 (2004) 1087–1089.
211

A. Bell-Taylor et al.
InorganicaChimicaActa482(2018)206–212
[32] Y. He, J.D. Gorden, C.R. Goldsmith, Inorg. Chem. 50 (2011) 12651–12660.
[33] Q. Zhang, J.D. Gorden, C.R. Goldsmith, Inorg. Chem. 52 (2013) 13546–13554.
[34] Y. He, C.R. Goldsmith, Chem. Commun. 48 (2012) 10532–10534.
[35] D. Jackson Rudd, C.R. Goldsmith, A.P. Cole, T.D.P. Stack, K.O. Hodgson,
B. Hedman, Inorg. Chem. 44 (2005) 1221–1229.
[36] K. Chen, L. Que Jr., J. Am. Chem. Soc. 123 (2001) 6327–6337.
[37] Y. Zang, J. Kim, Y. Dong, E.C. Wilkinson, E.H. Appelman, L. Que, J. Am. Chem. Soc.
119 (1997) 4197–4205.
[38] P. Gütlich, A. Hauser, H. Spiering, Angew. Chem. Int. Ed. 33 (1994) 2024–2054.
[39] R. Han, A. Looney, K. McNeill, G. Parkin, A.L. Rheingold, B.S. Haggerty, J. Inorg.
Biochem. 49 (1993) 105–121.
[40] Y. Yang, F. Diederich, J.S. Valentine, J. Am. Chem. Soc. 113 (1991) 7195–7205.
[41] W. Nam, J.S. Valentine, J. Am. Chem. Soc. 112 (1990) 4977–4979.
[42] J. Cho, R. Sarangi, H.Y. Kang, J.Y. Lee, M. Kubo, T. Ogura, E.I. Solomon, W. Nam, J.
Am. Chem. Soc. 132 (2010) 16977–16986.
212