Zwitterionic Cobalt Complexes with Bis(diphenylphosphino)(N-thioether)amine Assembling Ligands: Structural, EPR, Magnetic, and Computational Studies

Article abstract of DOI:10.1021/acs.inorgchem.5b02889

The coordination of two heterofunctional P,P,S ligands of the N-functionalized DPPA-type bearing an alkylthioether or arylthioether N-substituent, (Ph₂P)₂N(CH₂)₃SMe (1) and (Ph₂P)₂N(p-C

Full text of DOI:10.1021/acs.inorgchem.5b02889

Article
pubs.acs.org/IC
Zwitterionic Cobalt Complexes with
Bis(diphenylphosphino)(N‑thioether)amine Assembling Ligands:
Structural, EPR, Magnetic, and Computational Studies
†
†
‡,§
‡
‡
Christophe Fliedel, Vitor Rosa, Bertrand Vileno, Nathalie Parizel, Sylvie Choua,
∥
⊥
‡
,†
Christophe Gourlaouen, Patrick Rosa, Philippe Turek, and Pierre Braunstein*
†
‡
∥
Laboratoire de Chimie de Coordination, Laboratoire POMAM, and Laboratoire de Chimie Quantique, Institut de Chimie (UMR
177 CNRS), Universite de Strasbourg, 4 rue Blaise Pascal, CS 90032, 67081 Strasbourg, France
French EPR Federation of Research (REseau NAtional de Rpe interDisciplinaire, RENARD), Fed
7081 Strasbourg, France
7
§
́
́ ́
eration IR-RPE CNRS 3443,
6
⊥
CNRS, Universite
́
de Bordeaux, ICMCB, UPR9048, F-33600 Pessac, France
*
S Supporting Information
ABSTRACT: The coordination of two heterofunctional P,P,S
ligands of the N-functionalized DPPA-type bearing an
alkylthioether or arylthioether N-substituent, (Ph P) N-
2
2
(
CH ) SMe (1) and (Ph P) N(p-C H )SMe (2), respectively,
2 3 2 2 6 4
toward cobalt dichloride was investigated to examine the
influence of the linker between the PNP nitrogen and the S
atoms. The complexes [CoCl (1)] (3) and [CoCl (2)] (4)
2
2
2
2
have been isolated, and 3 was shown by X-ray diffraction to be
a unique dinuclear, zwitterion containing one CoCl moiety bis-
chelated by two ligands 1 and one CoCl3 fragment
coordinated by the S atom of a thioether function. The FT-
IR, UV−vis, and EPR spectroscopic features of 3 were
analyzed as the superposition of those of constitutive
fragments identified by a retrosynthetic-type analysis. A similar approach provided insight into the nature of 4 for which no
X-ray diffraction data could be obtained. A comparison between the spectroscopic features of 4 and of its constitutive fragments,
[
CoCl(2) ]PF (7) and [H2′] [CoCl ] (8) (2′ = NH (p-C H )SMe), and between those of 4 and 3 suggested that 4 could
2
6
2
4
2
6
4
either have a zwitterionic structure, similar to that of 3, or contain a tetrahedral dicationic bis-chelated Co center associated with
a CoCl dianion. Magnetic and EPR studies and theoretical calculations were performed. Doublet spin states were found for the
4
pentacoordinated complexes [CoCl(1) ]PF (5) and 7 and anisotropic quadruplet spin states for the tetrahedral complexes
2
6
[
CoCl (H1′)] (6) (1′ = NH (CH ) SMe) and 8. A very similar behavior was observed for 3 and 4, consisting in the
3
2
2 3
juxtaposition of noninteracting doublet and quadruplet spin states. Antiferromagnetic interactions explain the formation of
dimers for 6 and of layers for 8. The EPR signatures of 3 and 4 correspond to the superposition of low-spin nuclei in 5 and 7 and
high-spin nuclei in 6 and 8, respectively. From DFT calculations, the solid-state structure of 4 appears best described as
zwitterionic, with a low-spin state for the Co1 atom.
INTRODUCTION
successfully applied to the catalytic oligo- or polymerization of
ethylene.
■
6
,7
The profuse chemistry of short-bite ligands, particularly the
diphosphines bis(diphenylphosphino)methane (DPPM) and
bis(diphenylphosphino)amine (DPPA) and their derivatives,
magnified by the diversity of their coordination modes
The study of the electronic structure of first-row transition
metal ions in relation with their stereochemistry is a very
relevant field in coordination and biological chemistry, and it is
essential to analyze how stereochemistry, electronic structure,
(
monodentate, bridging, chelating), instigated a longstanding
8
,9
1
and magnetic properties correlate. We became interested in
the coordination chemistry of N-thioether-functionalized
DPPA-type ligands, such as (Ph P) N(CH ) SMe (1) and
interest across the scientific community. The versatility of
DPPA-type ligands can be amplified by the introduction of
various N-substituents and empowers an extensive number of
applications, such as the formation of mono- or polynuclear
2
2
2 3
(
Ph P) N(p-C H )SMe (2), with an alkyl or an aryl spacer,
2 2 6 4
2
respectively (Chart 1, top) because of (i) the versatility of the
complexes or the anchoring of coordination compounds into
3
4
mesoporous materials or on metallic surfaces. Metal
complexes with such short-bite ligands are involved in various
catalytic reactions, and their Ni and Cr derivatives have been
Received: December 14, 2015
5
©
XXXX American Chemical Society
A
DOI: 10.1021/acs.inorgchem.5b02889
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Chart 1
coordination spheres (coordination numbers and geometries)
13
associated with diverse chemical and physical properties. One
should of course remember that the structures of the metal
complexes in solution and in the solid state may not be
identical, which justifies the use of various complementary
methods for their investigation.
Following these observations and recent findings on the
14
magnetic properties of polynuclear Co(II) complexes, we
investigated the coordination chemistry of ligands 1 and 2 in
cobalt complexes, a metal not much represented in the DPPA-
type chemistry landscape. Unexpectedly, dinuclear structures
were obtained, as established by X-ray diffraction in the case of
the zwitterionic complex [CoCl (1)] (3), which contains one
2
2
CoCl moiety bis-chelated by two ligands 1 and one CoCl3
fragment coordinated by the S atom of a thioether function. In
the absence of X-ray diffraction data for [CoCl (2)] (4) and in
2
2
order to better understand its occurrence and properties, we
compared its spectroscopic features with those of potential
constitutive building blocks. This approach was first validated
by performing a similar analysis with 3. The comparisons were
based on FT-IR, UV−vis, and EPR spectroscopic data,
magnetic measurements, and theoretical calculations.
thioether function, which can act as an intra- or intermolecular
donor toward a metal center or be used to deposit/anchor
complexes on metal surfaces, and (ii) the possibility to evaluate
the differences of reactivity induced by a change of the linker
that connects the PNP moiety and the thioether donor group,
for example, when going from a flexible propyl to a rigid
phenyl. Previous studies have shown that (i) bis-chelated
dicationic Pd(II) complexes with ligands 1 and 2 could be used
to assess molecular anchoring on Au surfaces (Janus micro-
RESULTS AND DISCUSSION
■
1. Synthesis of the Co(II) Complexes and Solid-State
Molecular Structures. To evaluate the influence of the linker
that connects the S-donor and the PNP nitrogen atom of
ligands 1 and 2 on the nature of their cobalt(II) complexes,
these ligands were initially reacted with an equimolar amount of
4
spheres), (ii) the N-functionalization influences the behavior
of Ni(II) complexes containing such tritopic P,P,S ligands in
6a
catalytic ethylene oligomerization, and (iii) a combination of
dicationic Ni(II) complexes containing two chelating ligands 1
or 2 and Zn metal allows the activation of both C −Cl bonds
anhydrous CoCl in dichloromethane. In both cases, dark green
2
sp3
of CH Cl under mild conditions and affords mixed phosphine,
solutions were obtained and the same workup was applied.
After filtration of the reaction mixture, evaporation to dryness
of the filtrate, and washing of the residue, a green solid was
isolated. Crystallization from a dichloromethane/pentane
mixture led to dichroic green/red crystals of complex 3 (ligand
1) and green polycrystalline [CoCl (2)] (4) (ligand 2) in 55%
2
2
10
phosphonium ylide species.
Further indications that subtle changes in ligand design can
strongly affect the nature of the resulting coordination
complexes were also provided by the reactions between FeCl2
and ligand 1 or 2 (1:1 molar ratio) since they led, under strictly
similar reaction conditions, to an infinite coordination polymer
2
2
and 83% yields, respectively (Scheme 1).
11
31
or a dinuclear complex, respectively. It was concluded that
electronic factors accounted for the unusual unsymmetrical
chelation of the iron center (Chart 1, bottom), raising the
question of a possible lability of phosphorus donors in solution,
which is relevant to catalyst behavior. Whereas studies on the
coordination chemistry of DPPA-type ligands have mostly
Whereas P NMR spectroscopy is commonly used to assess
the coordination behavior of N-functionalized DPPA-type
1
ligands, this was not possible here owing to the paramagnetic
nature of the Co(II) center. The precise determination of the
coordination sphere around the metal center thus required X-
ray diffraction studies on single crystals, complemented by
physical and theoretical investigations (see below). While the
molecular structure of complex 3 in 3·2CH Cl was
5
,12
focused on metals of groups 6, 10, and 11, Ru and Rh, those
on Fe and Co remain scarce, although these earth-abundant
metals are of considerable current interest and display versatile
2
2
unambiguously established by X-ray diffraction studies (Figure
a
Scheme 1. Reactions of Ligands 1 and 2 with Anhydrous CoCl2
a
Conditions: (i) CH Cl , room temperature, 12 h, (ii) crystallization from CH Cl /pentane.
2
2
2
2
B
DOI: 10.1021/acs.inorgchem.5b02889
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
), no suitable single crystals could be obtained for complex 4
Article
1
P4 angles of 71.69(6)° and 71.33(7)°, respectively. The second
Co center (Co2) is in a distorted tetrahedral coordination
environment containing three similarly spaced Cl ligands [Co−
Cl bond lengths in the range 2.236(2)−2.246(2) Å], whereas
the Co2−S2 bond of 2.396(2) Å is significantly longer,
consistent with the larger covalent radius of S vs Cl. The angles
involving S2 span a broader range [between Cl3−Co2−S2
despite many attempts.
9
7.74(7)° and Cl4−Co2−S2 108.49(8)°] than those involving
only chlorides and the cobalt center [between Cl4−Co2−Cl2
1
12.79(9)° and Cl3−Co2−Cl2 117.92(8)°].
The structure of complex 4 could not be determined by
single-crystal X-ray diffraction analysis, but elemental analysis
performed on a microcrystalline powder was in agreement with
the 1:1 ligand/metal ratio used, as in 3 (see Experimental
Section). Interestingly, in the ESI-MS spectra (MeCN/CH Cl
2
2
+
solutions) of 3 and 4, both the [M − H − CoCl ] (m/z 1005.1
4
+
(
1
3) and 1073.1 (4)) and the [M − CoCl ] fragments (m/z
3
040.1 (3) and 1108.1 (4)), were detected as major peaks,
Figure 1. View of the molecular structure of the zwitterionic complex
however with a 4:1 and 2:1 ratio in intensity, respectively.
We expected to gain more information about the nature of 4
from a comparison of its far-infrared (far-FT-IR) spectrum in
the region corresponding to the ν(Co−Cl) vibrations (600−
3
in 3·2CH Cl . Solvent molecules and hydrogen atoms were omitted
2 2
for clarity. Ellipsoids are represented at the 50% probability level.
Selected bond lengths and angles are reported in Table 1.
−1
1
00 cm ) with that of complex 3. However, the spectra
The dinuclear complex 3 contains one CoCl moiety bis-
recorded were not conclusive (see below).
chelated by two ligands 1 and one CoCl fragment coordinated
3
At this stage, two hypotheses concerning the structure of
complex 4 could be reasonably envisaged: (i) a structure similar
to that of complex 3, with a CoCl moiety bis-chelated by two
by the S atom of one thioether function (Figure 1). The 1:1
ligand/metal molar ratio used in the synthesis is retained in the
structure of the complex, as confirmed by the elemental analysis
on ground and dried crystals (see Experimental Section). To
the best of our knowledge, there are only 3 examples of
structurally characterized cobalt complexes containing one or
two N-substituted DPPA-type ligands, and none of them
ligands 2 and a CoCl fragment coordinated by a thioether
3
2a
donor (Chart 2, left), or (ii) an ion pair with a tetrahedral
dicationic bis-chelated Co center associated with a CoCl4
dianion (Chart 2, right). The latter situation, i.e., the formation
of a ML /MCl (L = chelating ligand) formula isomer of a
2
4
exhibits the observed [CoCl(P,P) ] or the anticipated
2
15
MCl L complex, has been recently observed in a Ni(II)
[
CoCl (P,P)] motif (Figure S25 in ESI). Panda et al.
2
2
complex supported by the monosulfido ligand 1·S, in which a
reported the formation of a mixed-valent dinuclear complex
[
of a N-aryl DPPA-type ligand with CoCl2.
II
III
dicationic Ni(II) center was bis-chelated by two (P,PS)
Co Cl (P,P) ]/[Co Cl (NH -R)], resulting from the reaction
2 2 3 2
15a
18
ligands, while a NiCl dianion balanced the charge.
The
4
Since the molecular structure of 4 was not directly accessible
by X-ray crystallography or indirectly by comparison of its
spectroscopic data (FT-IR, UV−vis, EPR see below) with those
of zwitterion 3, we targeted the formation of mononuclear
complexes/fragments potential building blocks of 4 to access
their spectroscopic fingerprints, compare them with those of 4,
and see whether the latter could be analyzed as being the
“simple” superposition of those of the constitutive fragments.
First applied to the structurally characterized complex 3 for
validation, this strategy led us to synthesize the complex
[
CoCl (RSR′)] motif is also rare, with only three structures
3
16
II
available in the CSD database; two of them are Co (μ -Cl)
dinuclear complexes, and one contains a Co center (Figure
2
III
S26 in ESI). The two chemically different Co centers in 3
exhibit different coordination geometries and are formally part
of a cationic (Co1) and an anionic fragment (Co2). The
pentacoordination around Co1 corresponds to a slightly
distorted trigonal bipyramidal geometry, with P1, P3, and Cl1
being coplanar and the apical positions being occupied by P2
and P4. The longest basal P−Co bond is P1−Co1 [2.283(2)
Å], whereas the Co1−P3 [2.221(2) Å] and Co1−Cl [2.232(2)
Å] bond lengths are similar; likewise, those between the metal
center and the apical ligands are similar [Co1−P2 2.248(2) Å
and Co1−P4 2.233(2) Å)]. The distortion away from a regular
trigonal bipyramidal geometry, as quantified by a continuous
[CoCl(1)
bis-chelated CoCl unit and the zwitterionic complex
[CoCl (H1′)] (6) (1′ = NH (CH SMe), which displays
the anionic S-CoCl moiety (Scheme 2).
]PF (5) containing the anticipated cationic (P,P)-
2
6
)
2 3
3
2
3
Complex 5 was obtained in nearly quantitative yield by
17
symmetry measure S(D ) = 2.11 (see Table S10 in ESI), is
reaction of anhydrous CoCl with ligand 1 in a 1:2 metal/ligand
3h
2
induced by the PNP chelate, with P1−Co1−P2 and P3−Co1−
ratio in the presence of excess KPF and was isolated as red
6
Chart 2. Proposed Zwitterionic (4, left) and Ion Pair (4′, right) Forms for Complex [CoCl (2)]2
2
C
DOI: 10.1021/acs.inorgchem.5b02889
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Scheme 2. Synthesis and Structure of the Zwitterionic Complex 3 and Its Representative [CoCl(1) ]PF (5) and [CoCl (H1′)]
2
6
3
(6) “Fragments”
+
crystals. Elemental and MS analyses were in agreement with its
crystal structure (Figure 2). The structural parameters around
found at m/z 1040.2 and attributed to the [M − PF ]
6
fragment. The far-FT-IR spectrum of 5 was also recorded
and compared with that of the zwitterionic complex 3 (see
below).
Gratifyingly, the reaction of 5 with 2 equiv of anhydrous
CoCl2 also afforded the zwitterionic complex 3 with
concomitant formation of “CoClPF ” (not isolated).
6
The complex [CoCl (H1′)] (6) (1′ = NH (CH ) SMe) was
3
2
2 3
obtained stepwise by (i) addition of a HCl aqueous solution to
a CH Cl solution of 3-(methylthio)propylamine 1′, the amine
2
2
precursor to ligand 1, leading to the corresponding ammonium
chloride (not isolated), and (ii) addition of CoCl , which
2
captured a chloride ion and completed its coordination sphere
with the thioether donor function of the ligand. The structure
of the zwitterionic complex 6 was confirmed by single-crystal X-
ray diffraction studies and, as expected, revealed structural
parameters very similar to those of Co2 in complex 3·2CH Cl
2
2
(
Figure 3 and Table 1). The anionic cobalt(II) center has a
distorted tetrahedral coordination geometry with Co−Cl bond
lengths ranging from 2.233(8) (Cl2) to 2.264(8) Å (Cl3), close
to those found in 3, and a Co1−S1 bond of 2.365(9) Å, shorter
than in complex 3 [Co2−S2 2.396(2) Å].
Figure 2. View of the molecular structure of the cationic complex in 5·
CH Cl . Counter ion, solvent molecule, and hydrogen atoms omitted
2
2
for clarity. Ellipsoids are represented at the 50% probability level.
Selected bond lengths and angles are reported in Table 1.
Complex 6 was also characterized by EA, ESI-MS (major
+
peak m/z 234.9 attributed to [M − Cl] ), and FT-IR
spectroscopy (see below).
the Co center (Co1) are, as expected, very close to those found
Our strategy consisting in the independent synthesis of
mononuclear cobalt(II) complexes (5 and 6) with ligand 1
relevant to complex 3 being validated, it was applied to
complexes based on ligand 2. We thus prepared and isolated
the different moieties (7 and 8) related to complex 4 (Scheme
for Co1 in complex 3·2CH Cl (selected bond lengths and
2
2
angles are reported in Table 1). The cobalt(II) center is bis-
chelated by two ligands 1 and adopts again a distorted trigonal
bipyramidal coordination geometry, as quantified by a S(D ) =
3
h
2.47 measure (Table S10), in which the base contains two
3
).
Similarly to complex 5, the cationic bis(P,P)-chelated
phosphorus atoms from different ligands (P1, P3), with Co−P
bond lengths of 2.240(1) (P1) and 2.264(1) Å (P3) and one
chlorine [Co−Cl1 2.227(1) Å], while two phosphorus donors
occupy the apical positions and exhibit Co−P bond lengths
close to those for the basal ones [Co1−P2 and Co1−P4
complex [CoCl(2)
between anhydrous CoCl
]PF
(7) was synthesized by reaction
and ligand 2, in a 1:2 metal/ligand
2
6
2
ratio, in the presence of excess KPF . Complex 7 was isolated
6
2
.239(2) and 2.241(2) Å, respectively]. The cationic charge on
after recrystallization as green crystals and characterized by EA,
+
the cobalt center is balanced by an isolated PF anion. The
ESI-MS (m/z 1108.2 [M − PF ] ), and FT-IR spectroscopy.
6
6
major peak recorded in the ESI-MS spectrum of complex 5 was
Moreover, its molecular structure was determined in the solid-
D
DOI: 10.1021/acs.inorgchem.5b02889
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Table 1. Selected Bond Lengths (Angstroms) and Angles
state by X-ray diffraction analysis (Figure 4). As expected, the
(
degrees) in the Solid-State Structures of 3·2CH Cl , 5·
cationic cobalt(II) center has a distorted trigonal bipyramidal
2
2
CH Cl , 6, 7, and 8
coordination geometry, as quantified by a S(D ) = 2.18
2
2
3h
measure (Table S10), in which the base contains two
phosphorus atoms from two ligands (P1, P3) and is completed
by one chlorine atom (Cl1), while the apical positions are
occupied by the two other P atoms (P2, P4) from the chelates.
Interestingly, the characteristic structural parameters, i.e., bond
lengths and angles around the cationic cobalt center, are very
similar to those found in the analogous complex 5 with ligand 1
3
·2CH Cl
5·CH Cl
2
7
2
2
2
Co1−P1
Co1−P2
Co1−P3
Co1−P4
Co1−Cl1
Co2−S2
Co2−Cl2
Co2−Cl3
2.283(2)
2.248(2)
2.221(2)
2.233(2)
2.232(2)
2.396(2)
2.246(2)
2.236(2)
2.242(2)
2.240(1)
2.239(2)
2.264(1)
2.241(2)
2.227(1)
2.232(1)
2.223(1)
2.245(9)
2.223(1)
2.224(1)
(
Table 1). As was observed for 5, addition of 2 equiv of
anhydrous CoCl to a CH Cl solution of complex 7 led to the
2
2
2
formation of complex 4 (evidenced by FT-IR and EA).
Attempts to synthesize independently the anionic moiety
potentially present in the zwitterionic complex 4 by a similar
Co2−Cl4
P1−Co1−P2
P3−Co1−P4
P1−Co1−Cl1
P3−Co1−Cl1
S2−Co2−Cl2
Cl2−Co2−Cl3
P1−N1−P2
P3−N2−P4
71.69(6)
71.33(7)
119.5(7)
128.0(7)
100.4 (7)
117.9 (8)
102.6(3)
99.0(3)
71.65(5)
71.32(5)
134.8(6)
118.8(6)
71.93(3)
72.20(3)
procedure to that used to access the S-CoCl fragment of 6 led
3
127.6(4)
123.9(4)
surprisingly to a different outcome. Reaction of the amine
MeS(p-C H )NH (2′), precursor to ligand 2, with aqueous
6
4
2
HCl, followed by treatment with CoCl , led to an intense blue
2
powder (further identified as complex 8) along with some
101.7(2)
101.7(2)
99.4(2)
unreacted CoCl . While the analytical results performed on this
100.2(2)
2
product were not consistent with the expected [CoCl (H2′)]
6
3
formulation, its ESI-MS spectrum only revealed the peak
corresponding to the ammonium chloride salt 2′·HCl. The far-
FT-IR spectrum of the isolated compound was also different
Co1−S1
Co1−Cl1
Co1−Cl2
Co1−Cl3
2.365(9)
2.251(8)
2.233(8)
2.264(8)
from that of the zwitterionic complex [CoCl (H1′)] (6) (see
3
below). Fortunately, suitable crystals for X-ray diffraction
S1−Co1−Cl1
S1−Co1−Cl2
S1−Co1−Cl3
Cl1−Co1−Cl2
Cl1−Co1−Cl3
Cl2−Co1−Cl3
C1−S1−C4
108.3(3)
analysis were obtained and revealed that in the solid state the
114.1(3)
96.6(3)
113.5(3)
113.0(3)
110.3(3)
101.2(2)
8
structure of complex 8 is composed of a CoCl dianion and two
4
H2′ ammonium cations (Figure 5). This observation is in
accordance with the fact that unreacted CoCl was isolated and
2
suggests that an ion-pair formulation for complex 4, i.e.,
[Co(2) ][CoCl ], could be also reasonably envisaged instead of
2
4
a zwitterionic structure.
Attempts were made to synthesize trinuclear cobalt(II)
Co1−Cl4
Co1−Cl1
Co1−Cl2
Co1−Cl3
Cl1−Co1−Cl4
Cl2−Co1−Cl4
Cl3−Co1−Cl4
Cl1−Co1−Cl2
Cl1−Co1−Cl3
Cl2−Co1−Cl3
C1−S1−C7
C8−S2−C14
2.272(6)
complexes in which both thioether functions of ligand 1 or 2
2.271(5)
2.257(6)
2.293(5)
−
would be coordinated to a CoCl or [CoCl ] moiety by
2
3
reacting (i) complex 3 (or 4) with 1 equiv CoCl2 or
[
HNEt ][CoCl ] and (ii) complex 5 (or 6) with 2 equiv
105.7(2)
3 3
110.7(2)
106.8(2)
115.3(2)
108.6(2)
109.4(2)
103.0(1)
104.6(1)
[HNEt ][CoCl ] (generated in situ), respectively. Unfortu-
3 3
nately, these reactions did not lead to the expected compounds,
but few crystals of the mixed-valent Co(II)/Co(III) polynuclear
complexes [CoCl (1) CoCl ] (9) and [CoCl (2) ]-
2
2
3
n
2
2
(H2′)[CoCl ] (10) (Scheme 4) were isolated, and their crystal
4
structures are shown in the ESI (Figures S27 and S28).
Since a rational synthesis could not be devised nor full
analytical data obtained for complexes 9 and 10, they will not
be discussed further. However, we note similarities between the
coordination polymer 9 and 3 (the pentacoordinated Co(II)
and the tetracoordinated Co(II) centers of 3 being replaced by
a hexacoordinated Co(III) and a pentacoordinated Co(II)
center, respectively) and between the cationic complexes in 10
and 7 (the pentacoordinated Co(II) center of the latter being
replaced by a hexacoordinated Co(III) center).
2
. Far-FT-IR and UV−vis Spectroscopic Data. With the
aim to validate our approach consisting in the synthesis of
molecular synthons (here compounds 5−8) to clarify the
structure of an unknown (complex 4) and more complex
architecture (complexes 3 and 4), we studied their vibrational
spectra in the far-infrared region corresponding to the ν(Co−
Cl) vibrations and their absorption electronic spectra in
solution.
Figure 3. View of the molecular structure of [CoCl (H1′)] (6) (1′ =
3
NH (CH ) SMe). Ellipsoids are represented at the 50% probability
2
2 3
level. Selected bond lengths and angles are reported in Table 1.
E
DOI: 10.1021/acs.inorgchem.5b02889
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Scheme 3. Synthesis and Proposed Structure of Complex 4 and the Relevant “Fragments” [CoCl(2) ]PF (7) and
2
6
[
H2′] [CoCl ] (8)
2
4
Scheme 4. Polynuclear Complexes [CoCl (1) CoCl ] (9)
2
2
3 n
and [CoCl (2) ](H2′)[CoCl ] (10) (see Figures S27 and
2
2
4
S28)
Figure 4. View of the molecular structure of the cationic complex in 7.
Counterion and hydrogen atoms omitted for clarity. Ellipsoids are
represented at the 50% probability level. Selected bond lengths and
angles are reported in Table 1.
and 296 cm− with weak shoulders (Figure S5 in ESI). Both
1
compounds also presented weak absorptions in the 430−360
−1
cm region. The complexity of these spectra encouraged us to
compare the spectrum of each compound with those of their
suggested fragments (5 and 6 for 3 and 7 and 8 for 4, Schemes
2
and 3, respectively).
As expected, the far-IR spectra of complexes 5 and 6, when
superimposed to that of complex 3, showed significant
similarities (Figure 6). The spectrum of 6 exhibits three
absorptions at 320, 305, and 289 cm− assigned to the distorted
1
Figure 5. View of the molecular structure of 8. Ellipsoids are
represented at the 50% probability level. Selected bond lengths and
angles are reported in Table 1.
While the far-FT-IR spectrum of complex 3 exhibited four
absorptions of medium intensity at ν = 530 (well separated
−1
from the other), 510, 490, and 480 cm and one intense band
at ν = 320 cm− with weaker shoulders (Figure S1 in ESI), the
spectrum of complex 4 exhibited three large and distinct bands
1
Figure 6. Superimposition of the IR spectra of complexes 3, 5, and 6
in the 600−100 cm region.
−1
−1
at ν = 557, 526, and 495 cm and two intense bands at ν = 307
F
DOI: 10.1021/acs.inorgchem.5b02889
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
tetrahedral S-CoCl moiety, and their pattern is comparable to
region, the spectrum of 6 (CoCl fragment) also exhibits broad
3
3
that found for 3, however shifted to lower wavenumbers
and poorly resolved absorption bands, with the most intense at
−1
(
Figures S1 (3) and S3 (6) in ESI). Interestingly, none of these
289 cm . Therefore, it is not possible to unambiguously
conclude on the basis of FT-IR data about the presence of a
CoCl or a CoCl fragment in complex 4.
absorption bands was detected in the spectrum of 5. In
contrast, various strong vibration bands between 550 and 480
4
3
−1
cm are present in both spectra of 3 and 5 (Figures S1 (3) and
S2 (5) in ESI).
We then recorded the UV−vis spectra of all the isolated
compounds with ligand 1. The spectra of the dinuclear
zwitterionic complex 3 and of the mononuclear cationic
A comparison between the far-FT-IR spectra of complexes 4
and 7 also revealed close similarities with four absorptions at
complex 5 (both in CHCl ) presented the same characteristic
3
−
1
5
54, 526, 495, and 423 cm for 4 and 556, 526, 492, and 425
absorption bands at 350, 375, and 470 nm (Figure 8).
However, the spectrum of 3 exhibits a further absorption at 585
−1
cm for 7 (Figures 7, and S5 (4) and S6 (7) in ESI). However,
nm (in CHCl and THF) that is not present in the spectrum of
3
5
, but a similar absorption (λ 590 nm) was recorded in the
spectrum of the zwitterion 6 (in THF because of insolubility in
4
CHCl , Figure 8 insert), which is characteristic of a T (P) ←
3
1
4
19
A2 transition split by spin−orbit coupling. These observa-
tions led us to conclude that the spectrum of 3 can be viewed as
+
the superposition of the spectra of complexes 5 (Co Cl(P,P)
2
−
fragment) and 6 (SCo Cl fragment), and this encouraged us
3
to study further the UV−vis spectra of complexes 7 and 8,
potential mononuclear building blocks of complex 4.
Interestingly, in the case of the complexes supported by
ligand 2, we initially observed that 4 exhibited a different
spectrum in CHCl and in THF, in contrast to 3, which has
3
similar spectra in both solvents. While the spectrum of 4 in
CHCl closely resembles those of complexes 7 (in CHCl ) and
3
3
3
(Figure S9 in ESI), with a very intense absorption band below
3
30 nm, a large shoulder until 375 nm, and two of lower
Figure 7. Superimposition of the IR spectra of complexes 4, 7, and 8
in the 600−100 cm region.
−1
intensity at 460 and 690 nm, its spectrum in THF is clearly
different (Figure 9). The strong absorptions below 330 nm
remain present in the spectrum of 4 in THF, but instead of a
shoulder, there is a strong, sharp absorption at 370 nm. The
most important differences are found at higher wavelengths,
complex 4 exhibits strong and broad absorption bands centered
−1
at 295 cm , in contrast to 7. The spectrum of 8 exhibits three
very intense, well-separated absorptions at 322, 294, and 272
−1
cm (Figure S7 in ESI, the splitting of the T mode expected
since the absorption at 460 nm for 4 and 7 in CHCl is not
3
2
−
for a perfectly tetrahedral [CoC14] dianion results from a
lowering of the local symmetry). Compared with those of 4, the
different shape of the bands could be due to intermolecular
nonclassical H−Cl bonds or packing effects. In this spectral
present in the spectrum of 4 in THF, while an absorption band
of low intensity appeared at 585 nm, which is also present in
the spectrum of complex 8 in THF (Figure 9 insert).
Noteworthy, the UV−vis spectrum reported for the
−
4
−4
−4
Figure 8. UV−vis spectra of compounds 3 (2 × 10 M in CHCl ), 5 (4 × 10 M in CHCl ), and 6 (3 × 10 M in THF) and zoom in the 525−
3
3
6
50 nm region, highlighting the fact that the spectrum of the dinuclear zwitterionic complex 3 is nearly the sum of those from the mononuclear
complexes 5 and 6.
G
DOI: 10.1021/acs.inorgchem.5b02889
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 9. UV−vis spectra of compounds 4 (4 × 10 M in THF and 2 × 10−4 M in CHCl ), 7 (2 × 10 M in CHCl ), and 8 (4 × 10 M in THF)
−
4
−4
−4
3
3
and zoom in the 475−775 nm region.
−
59
[
CoCl (THF)] anion exhibits the similar two characteristic
due to the hyperfine interactions with the Co nucleus (I = 7/
2), and the overall signal shape is well reproduced (Table 2).
The X-band EPR spectra of powders of 6 and 8 recorded at 5
K (Figure 10) correspond to Co(II) in a HS quartet spin state.
This is usually recognized by poorly resolved resonances,
broadened by a combination of large g anisotropy, sizable spin−
orbit couplings, and admixture of excited state character into
3
−
1
−1
absorption bands at 587 and 693 cm (vs 585 and 690 cm
for 4 and 8), which were ascribed to the d−d transition from
the ground state A to the T (P) state. Altogether, the fact
that complex 8 exhibits different spectra in CHCl and THF
and that the spectra of complexes 4 and 8 (in THF) present
4
4
20
2
1
3
exactly the same pattern in the 550−750 nm region as the
−
25
[
CoCl (THF)] anion is in favor of a zwitterionic form (vs ion
the magnetic ground state. Two strategies may be considered
3
pair, Chart 2) for the dinuclear complex 4, whose weakly
to obtain visually satisfactory simulations of HS Co(II) EPR
spectra. Both rely on an effective S = 1/2 ground state. The
positions of the resonances are then matched either by
considering an axial g tensor, hence introducing rhombicity of
coordinated CoCl fragment would be displaced by THF. The
3
UV−vis spectrum of a pseudotetrahedral [CoCl(μ-Cl)(Me -
2
cAAC)] complex, formally presenting a LCoCl fragment, also
2
3
26
exhibits a similar pattern, with two strong absorption bands
the zero-field slitting (ZFS), or by considering an axial ZFS
−1
21
27
around 580 and 650 cm . The spectra of complexes 4 and 8
and by tuning the anisotropy of the g tensor. It is worth
in THF present exactly the same pattern
mentioning that mathematically one can often achieve a
The UV−vis studies revealed clearly that the spectrum of the
bis-cobalt complex 3 combines the different absorption bands
observed for its constitutive mononuclear fragments 5 and 6,
independently of the solvent used. With complexes supported
by ligand 2, we observed that the spectrum recorded for
complex 4 is solvent dependent and different from that of
complex 3, while both couples of complexes 5/7 and 6/8
exhibit the same characteristic absorptions (Figures S9−11 in
ESI). These elements do not allow us to conclude on the exact
structure of complex 4: in a noncoordinating solvent, such as
similarly satisfying visual simulation by using a M = 3/2
S
ground state and varying the rhombicity in both g and ZFS,
although this provides a wholly indeterminate set of solutions.
It is noteworthy that the EPR signal extends well beyond the g
= 2 region at low temperature, so that more transitions than
2
8
expected may be observed. In many instances, these
“additional” lines appear at fields above 0.5 T, so that partly
out-of-range transitions may induce artifacts into the
simulations. Simulations of the spectra for 6 and 8 (Table 3)
are proposed, while we are fully aware of all these difficulties.
These simulations will serve as references for the spectral
simulation of the bis-Co complexes described hereafter. The
narrower peak-to-peak line width in the signal of 8 may express
an isotropic tetrahedral surrounding of the Co nucleus in a
+
chloroform, it behaves like the cationic bis-chelate [Co Cl-
(
P,P) ] complex 7, while in coordinating THF, it presents
2
undoubtedly the absorptions recorded for complex 8.
. EPR Measurements. The X-band EPR spectra of
3
powder samples of complexes 5 and 7 recorded at 5 K are
shown in Figure 10, together with their simulations. As
reported in the literature, five-coordinate low-spin (LS)
cobalt(II) complexes have been found with soft phosphorus
CoCl environment, while it was anisotropic tetrahedral in 6
4
(CoCl S).
3
The X-band EPR spectra of the powder samples of
complexes 3 and 4 show both the LS and the HS features
2
2
donor ligands. The fit parameters are consistent with a
(Figures S12 and 10, respectively) with g (HS) = 5.83 and
eff
23
trigonal bipyramidal structure (Table 2). A low-spin (LS)
Co(II) S = 1/2 doublet is suggested to have weaker anisotropy
than the corresponding S = 3/2 high-spin (HS) species.
Moreover, pulsed EPR indicates longer relaxation times, which
are compatible with the usually well-separated energy levels
g (LS) = 2.17 for 3 and g (HS) = 5.24 and g (LS) = 2.11 for
eff
eff
eff
4. The anisotropic EPR spectral parameters deduced from
simulations are summarized in Table 4.
Numerical simulations satisfactorily accounted for a 1:1 ratio
of LS and HS species for the binuclear species. Unfortunately,
grain size effects were observed for a polycrystalline powder of
24
encountered for the LS state. The spectral width is mostly
H
DOI: 10.1021/acs.inorgchem.5b02889
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 10. EPR spectra of powder samples of complexes 4−8 and of a frozen solution of complex 3 recorded at 5 K: experimental (upper trace) and
simulation (lower trace).
Table 2. g-Tensor and Hyperfine Tensor Components (hfcc’s, A) Deduced by Numerical Simulations of the EPR Spectra of
Powder Samples of 5 and 7 recorded at 5 K
a
S
g1
g2
g3
A1 (MHz)
A2 (MHz)
A3 (MHz)
lwpp (mT)
7
5
1/2
1/2
2.057
2.040
2.288
2.350
19
19
120
140
5.5 G+L
3.5 G+L
2.087
19
a
lwpp corresponds to the peak-to-peak line width of the EPR signal with Gaussian (G) and Lorentzian (L) line shape broadening.
Table 3. EPR Parameters Deduced by Numerical
Simulations of the EPR Spectra of Powder Samples of 6 and
8
Co hfcc. We did not try to attenuate these effects upon finer
grinding, being more concerned with the sample stability
(structural and chemical). To remove the underlying ambiguity,
we studied sample 3 in frozen solution (Figure 10). This study
actually demonstrated the similarity of the LS Co EPR signals
for complexes 3 and 6. EPR studies cannot firmly conclude the
possible magnetic exchange between the two Co centers.
Accounting for the estimation of such exchange coupling from
SQUID measurements (see below), the main conclusion is that
the two remote Co magnetic centers are very weakly correlated
(J < 1 cm ). At the energy scale of EPR, this is always within
the strong exchange limit with J ≫ A.
4. Magnetic Measurements. We confronted the EPR
results with the magnetic data on polycrystalline samples.
ZFS D (cm−1)
lwpp (mT)
a
S
g⊥
g∥
8
6
3/2
3/2
2.240
2.033
2.340
2.120
>1.5
>1.5
18 G+L
65 G+L
a
lwpp corresponds to the peak-to-peak line width of the EPR signal for
Gaussian (G) and Lorentzian (L) line shape broadening.
−1
3
, which resulted in an orientation-dependent EPR signal.
Although the low-spin region of polycrystalline 3 clearly shows
the hyperfine interaction with the I = 7/2 Co (Figure S12 in
ESI), grain size effects prevent a proper assessment of the LS
5
9
I
DOI: 10.1021/acs.inorgchem.5b02889
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Table 4. EPR Parameters Deduced from Simulations of the Experimental Spectra of a Frozen Solution of 3 and of a
Polycrystalline Powder 4 Recorded at 5 K
S = 1/2
S = 3/2
a
D (cm−1
a
g1 = g2
g3
A = A (MHz)
A3 (MHz)
lwpp (mT)
%
g⊥
g∥
)
lwpp (mT)
%
1
2
4
3
2.057
2.058
2.288
2.410
19
50
120
180
31 L
54
50
2.240
2.300
2.340
2.433
>1.5
>1.5
58 L
46
50
25 G+L
48 G+L
a
lwpp corresponds to the peak-to-peak line width of the EPR signal with Gaussian (G) and Lorentzian (L) line shape broadening.
Measurements were performed over the 2−300 K range under
or 10 kOe. Figure 11 illustrates the evolution of the χ T
1
M
Figure 12. Variation of the χ T product with temperature measured
M
under 1 kOe for compounds 6 (circles) and 8 (triangles), with the
simulated curve (red line), for 2S = 3/2 spins in AF interaction, with g
−1
−1
=
2.147, J = −0.4 cm , and D = +7 cm .
Figure 11. Variation of the χ T product with temperature measured
M
under 10 kOe for compounds 5 (circles) and 7 (triangles).
mol− for compound 6. The latter shows straightforward
1
product with temperature for 5 and 7. Compound 5 shows
Curie−Weiss behavior above 50 K, with parameter values of
3
3
−1
Curie behavior above 20 K, with a Curie value of 0.482(1) cm ·
2.161(2) cm ·K·mol and −3.0(3) K (Figure S15 in ESI), in
−1
K·mol , while compound 7 shows Curie−Weiss behavior with
agreement with largely uncorrelated tetrahedral Co(II) ions in a
3
−1
4
31−35
0
.5663(3) cm ·K·mol and −1.7(1) K (Figure S13 in ESI).
A
2
ground state,
́
with a spin-only Lande factor g =
Only a slight decrease of the χ T products is observed at low
2.147(2), significantly larger than the free-electron value. The
M
temperatures, in agreement with the low Weiss temperatures,
showing that the Co(II) ions are mostly uncorrelated.
Isothermal magnetizations tend toward saturation values close
latter and the high TIP values are expected because of the
4
mixing of the ground state with the low-lying T state issued
2
4
36
from the same F atomic term. Below 50 K, a significant
to 1 μ (Figure S14 in ESI), thus supporting a low-spin S = 1/2
decrease of the χ T product is observed down to 2 K, in
B
M
ground state. Curves at 5 and 8 K are superimposed on the
Brillouin curve for g = 2.119(1) and 2.345(1) for compounds 5
and 7, respectively. Compounds 5 and 7 show distorted
agreement with the observed negative Curie−Weiss temper-
ature. Typical nesting behavior is observed in low-temperature
isothermal magnetization measurements (Figure S16 in ESI),
showing the presence of antiferromagnetic interactions or/and
single-ion magnetic anisotropy. Simulations taking into account
only single-ion anisotropy proved unsatisfactory to simulate the
magnetic behavior observed, all the more so since rhombicity
on g and/or the zero-field splitting cannot be reliably extracted
from a powder measurement. The crystalline structure shows
close contacts between chlorine and cobalt atoms of
neighboring complexes, forming effectively pairs of complexes
in the structure (Figure S17 in ESI). Using a simple dinuclear
model with single-ion anisotropy allows for an acceptable
trigonal bipyramidal CoP Cl coordination geometries with C2
4
symmetry at most. For low-spin Co(II) in TBPY-ML5
2
geometry, the resulting E′ ground state is expected to have
́
its orbital momentum quenched but with a Lande g value quite
above the free electron value by spin−orbit mixing. Indeed, the
effective g values found for compounds 5 and 7, either from
isothermal magnetizations or the Curie parameter (2.267(5)
and 2.457(1), respectively) are in the range found by EPR for
II
these compounds and for other low-spin Co P Cl com-
4
2
2,23,29
plexes.
Intriguingly, isothermal magnetizations at 1.8 K
deviate significantly from the Brillouin curve and cannot be
fitted with another Brillouin with reasonable values, which may
be symptomatic of the onset of some intermolecular magnetic
interactions between Co(II) centers that may not show zero-
field splitting but are expected to have strongly anisotropic g
tensors.
simulation of both χ T and isothermal magnetization curves
M
−1
with the following parameters: g = 2.147, J = −0.4 cm , D = +7
−1
37
−1
cm . The interaction parameter agrees with the −0.8 cm
value deduced from the Curie−Weiss temperature using the
38
mean-field approach. Easy-axis single-ion anisotropies (D <
0) led to a worse agreement at low temperature for the χMT
curve. The resulting effective g factor value is above the average
value of 2.062 found by EPR on a powder sample of 8. Some
discrepancy between single-ion anisotropy values had to be
expected given the inherent uncertainty for their determination
from powder values.
Figure 12 shows the evolution of the χ T product with
M
temperature for 6 and 8. For both compounds, temperature-
independent paramagnetic (TIP) contributions have to be
7
taken into account, a common feature for tetrahedral d Co(II)
3
0
−6
3
compounds, with a reasonable TIP value of 450 × 10 cm ·
J
DOI: 10.1021/acs.inorgchem.5b02889
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
For compound 8 (Figure 12), even with a strong TIP
contribution of 600 × 10 cm ·mol , as observed in other
one high-spin Co(II) with tetrahedral coordination, the single-
ion properties vary somewhat, in agreement with the
−6
3
−1
2
−
30,36
salts containing a tetrahedral CoCl4 anion,
the χ T value
differences seen in the powder EPR spectra of compounds 3,
M
−6
3
still increases at high temperatures. A TIP value of 1500 × 10
5, and 6. For 4, the χ T value at high temperature of 3.1 cm ·K·
M
3
−1
−1
cm ·mol would be required to account for the high-
mol is quite close to the sum of the corresponding values for
4
3
−1
temperature behavior, a value that would imply a T excited
compounds 7 and 6 at 2.7 cm ·K·mol , indicating that here
also the preferred formulation would be one low-spin Co(II)
with trigonal bipyramidal coordination and one high-spin
Co(II) with tetrahedral coordination.
2
3
9
state lying at unphysical low energy. Nevertheless, a Curie−
Weiss law gives a satisfactory fit above 50 K (Figure S15 in
ESI), with values of 4.04 cm ·K·mol and −10.1 K, depending
on the TIP contribution used. These values support again
3
−1
For both compounds a strong decrease of the χ T product is
M
4
tetrahedral Co(II) ions with a A ground state but with an
observed below 50 K but without any corresponding peak in
the susceptibility curve, which indicates that any antiferromag-
netic interaction is of low magnitude. Indeed, the structure of
compound 3 does not evidence obvious interaction pathways.
2
unusually high spin-only Lande
temperature behavior parallels that observed for compound 6
Figure 12), with no maximum observable at low temperature
́
factor g = 2.9. The low-
(
on the susceptibility curve but with an isothermal magnet-
ization that does not seem to be close to saturation even at 1.8
K and 70 kOe (Figure S16 in ESI). Close examination of the
The decrease observed is thus caused mainly by the zero-field
4
splitting of the A ground state of the tetrahedrally coordinated
2
Co(II) ions. Though the problem is largely overparametrized,
considering axially anisotropic g tensors for both Co(II) ions,
the zero-field splitting, and a small intramolecular interaction
between both Co(II) ions, reasonable sets of parameters can be
2−
crystal packing of 8 shows that the CoCl4 ions form layers
parallel to the ac plane, which are separated along the b axis by
the ammonium moieties (Figure S18 in ESI). The arrangement
of the anions within the layers is close to a 2D honeycomb
lattice, with close Cl···Cl contacts of about 4 Å. The anomalous
magnetic behavior observed may thus be accounted for by
correlation effects within those layers between the strongly
41
found that simulate the overall behavior of both χ T and
M
isothermal magnetization curves (Figures 13 and S20−S21 in
ESI). We used the g factors found by EPR for the
pentacoordinated low-spin Co(II) and adjusted accordingly
the g factor for the tetrahedral Co(II) with an arbitrary ratio g⊥/
g∥ fixed at 1.1 for both compounds (this ratio corresponds to
that observed by EPR for the high-spin Co(II) in compound
4). In both cases, using easy-axis zero-field splitting led to worse
simulations. The parameters found for compound 3 were J =
40
anisotropic tetrahedral Co(II). Its correct modeling would
require further anisotropic measurements on single crystals that
are outside the scope of this study.
Figure 13 reports the χ T product evolution with temper-
M
ature for 3 and 4. It is readily seen that both compounds show
−
1
−
0.3 cm , S = 1/2, g = 2.135, and g = 2.402 and S = 3/2, g
⊥ ∥ ⊥
−1
=
2.401, g = 2.641, and D = +15 cm and for compound 4 J =
∥
−1
−
=
0.4 cm , S = 1/2, g = 2.057, and g = 2.288 and S = 3/2, g
⊥ ∥ ⊥
−1
2.31, g = 2.541, and D = +10 cm .
∥
5
. Theoretical DFT Calculations. DFT calculations using
the ADF package were performed to determine the geometries
and electronic structures of the complexes. Geometry
optimizations were performed starting from the X-ray
structures. In all cases we found that spin contamination was
negligible. For comparison, we first computed mononuclear
complexes, viewed as building blocks of the dinuclear
complexes. Several conformations for 5 and 7 were considered,
but only one has been optimized where the phenyl rings are
roughly parallel to the P−N−P planes for 7, and the positions
of the aliphatic chains for 5 are consistent with the X-ray
structure. As established experimentally, the Co1 atom in 5 and
Figure 13. Variation of the χ T product with temperature measured
M
7
has an S = 1/2 ground state, and therefore, the structures
under 1 kOe for compounds 3 (black circles) and 4 (black triangles),
and simulated curves (blue and red curves, respectively) for the
interacting {1/2;3/2} spin system (see text for parameters).
were optimized as a doublet. The calculated geometries for 5
and 7 are similar to the arrangement of the ligands found in the
X-ray structures. Relevant computed geometrical parameters
are given in Table S2 (ESI) and compared to experimental
values. The resulting bond lengths and angles are similar to
those found experimentally. The largest deviation concerns the
computed Co−P bond distances, which are ca. 0.01−0.1 Å
longer than the observed values. The size of the molecules
prevented optimization with larger basis sets or higher theory
level methods to obtain more accurate values. The coordination
sphere around the Co1 atom is distorted trigonal bipyramidal,
with S(D ) of 2.25 and 2.32, respectively, quite close to
very similar characteristics with Curie−Weiss behavior above 20
3
−1
K (Figure S19 in ESI) and parameters of 3.408(3) cm ·K·mol
3
−1
and −4.1(2) K and 3.147(2) cm ·K·mol and −3.0(1) K,
respectively. For both compounds, TIP contributions were
taken into account with a reasonable value of 350 × 10 cm ·
mol for 4, while 3 showed some slight contamination with
ferromagnetic impurities, which could not be deconvoluted
from the TIP contribution.
−6
3
−1
As indicated above, only for compound 3 is the solid-state
structure clearly established, and one could anticipate magnetic
properties for 3 close to the sum of those of compounds 5 and
3h
experimental values (2.47 and 2.18) (Table S10). The
arrangement of the phenyl moieties and carbon chains is
quite well reproduced for 7 and 5, respectively. As shown in
Figure 14b for 5, the spin density is exclusively centered on the
cobalt cation, with very small contributions from the chlorine
6
. A comparison of the respective χ T products readily shows
M
that although this is the case, in agreement with one low-spin
Co(II) in trigonal bipyramidal coordination environment and
K
DOI: 10.1021/acs.inorgchem.5b02889
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
43
reported in the literature. According to calculated tetrahedral
geometries and EPR data, the Co(II) ions adopt a high-spin
configuration S = 3/2. The computed g tensors obtained for
both compounds are almost axial and reproduce quite
satisfactory the EPR data.
For 3 complex, EPR and SQUID data prompted us to assign
a low-spin state S = 1/2 to Co1 and a high-spin state S = 3/2 to
Co2. The computed values for 3 reproduced both environ-
ments of Co1 and Co2 atoms and indicated a reasonable
agreement with the X-ray diffraction data. For the Co1−P and
Co2−S bond distances, the same overestimation was observed
as that mentioned above for mononuclear complexes.
Compared to the related mononuclear complexes, the MO
diagram for 3 nearly reproduces the addition of the individual
MO diagrams. The HOMOs are almost identical in shape with
the HOMOs of 5 and 6. The computed spin density
distribution presented in Figure 15 indicates that the unpaired
Figure 14. (a) ADF view drawing of the Kohn−Sham singly occupied
molecular orbitals (SOMOs) of 5 (top) and 7 (bottom). (b) Spin
density plots for 5 (top) and 7 (bottom).
and phosphorus atoms. This spin density is mostly generated
by the molecular orbital (SOMO) given in Figure 14a, which is
π-antibonding between the Co1 3d orbitals and the chlorine
and phosphorus atoms. The situation is almost identical for 7,
with a spin density again mostly localized on the cation.
However, the associated SOMO shows some delocalization
through the π system of the phenyl rings. Finally, the different
substituents on nitrogen have a slight influence on the
electronic structure. The EPR parameters were also calculated
to evaluate the g tensors through a single point at the B3LYP
level of DFT with the ORCA package on the ADF-optimized
geometries. The calculated values are listed in Table S3 (ESI).
The simulation of the spectra indicated a more rhombic g
tensor for 5 than for 7, which is axial with g > g . These results
∥
⊥
are similar to those reported on triphosphine complexes and
support the assignment of a trigonal bipyramidal structure with
Figure 15. Spin density plots calculated for 3.
42
2
2
a mixed d /dxy ground state. This is consistent with the
spin is located on both Co(II) ions. On first inspection, powder
EPR spectra and SQUID experiments indicate that both
binuclear complexes 3 and 4 correspond to the superposition of
the fragments developed above and EPR data were simulated
with two distinct g tensors corresponding to two paramagnetic
centers. Thus, for both dinuclear compounds, we started our
computation on the whole molecule by replacing one Co atom
x −y
spin density map calculated for the low-spin mononuclear
complexes localized on the trigonal equatorial plane. The
computed g tensors obtained for both compounds are almost
axial and reproduce quite satisfactory the EPR data for 7. A
more important discrepancy is observed for 5 that is attributed
to the absence of the surrounding medium, i.e., calculations
were performed on the gas phase.
The ground state geometries for 6 and 8 were computed
using the experimentally determined S = 3/2 ground states.
Geometrical parameters calculated for both 6 and 8 indicated a
four-coordinated metal center in a tetrahedral environment,
which is in good agreement with the experimental data.
However, for complex 6, the calculated distance between Co2
and the nitrogen atoms was divided by approximately two: the
zwitterionic structure induces significant Coulombic attraction.
The Co2−S and Co2−Cl distances are slightly overestimated
and present the same behavior as that described above for 5
and 7. For 6 and 8, the metal 3d orbitals contribute to the
frontier orbitals with a high participation of the Cl and S atoms.
The three highest occupied molecular orbitals (SOMO) for 6
are depicted in Figure S22 (ESI). The EPR spectrum of 8 is
dominated by two broad features, whereas for 6 only a very
broad signal is observed. The spectral features are similar to the
EPR spectra of some substituted tetrahedral complexes
10
with a diamagnetic Zn(II) center (d , S = 0) to reach
individual g tensors. The g tensor extracted from EPR
simulation for 3 indicates g ,g > g for Co1 and supports a
x
y
z
low-spin ground state identical to 5. This is in agreement with
the computed g tensor obtained for the low-spin Co1 which
reproduces quite satisfactory the EPR data. For the high-spin
Co2, a more important discrepancy was observed between the
calculated and the experimental g values.
As no X-ray structure is available for 4, two hypotheses could
be formulated (Chart 2), one compatible with a zwitterionic
form similar to 3 and the other with a Co1 center associated
with a CoCl dianion. For the first hypothesis, we used the data
4
from the crystal structure of the mononuclear Co complex 7 as
a fragment to build the dinuclear complex with a similar
environment for both Co atoms, as in 3 (Figure 1).
As shown in Table S2 (ESI), the calculated geometrical
parameters reproduce well the experimental values obtained for
3. The shorter distance calculated between the two metal
L
DOI: 10.1021/acs.inorgchem.5b02889
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
centers compared to 4 (∼1.3 Å) is consistent with the flexibility
of the alkyl chains in 3, which lead to a curved zwitterionic
form. As expected, the SOMOs are localized over the cobalt
ions with a small participation of the Cl, S, and P atoms as
observed for 3 (Figure S23 in ESI). The computed spin density
distribution for the zwitterionic form (Figure S24 in ESI)
indicates that the unpaired spin is localized on both Co atoms.
Single-point analysis on the optimized structure using the
method described above for 3 provides nearly axial g tensors
very similar to those for the parent complex 7 for the low-spin
Co1 and in agreement with the experimental powder EPR
spectra. The results for the high-spin Co2 are more nuanced
with an important discrepancy.
investigated. A combination of solution and solid-state data
was used to facilitate comparisons between the molecular
fragments and the dinuclear complexes and also with literature
data. However, not every technique could be used both in
solution and in the solid state. After validation with 3, this
“lego-type” approach was applied to complex 4, obtained by
reaction of ligand 2 with CoCl in a 1:1 molar ratio, for which
2
the structure could not be determined by single-crystal X-ray
diffraction. We thus synthesized and characterized the
complexes [CoCl(2) ]PF (7) and [H2′] [CoCl ] (8) (2′ =
2
6
2
4
NH (p-C H )SMe) as models of the elementary bricks
2
6
4
constitutive of 4. Two possibilities were envisaged for the
structure of 4, a zwitterionic structure similar to that of complex
3, or an ion pair containing a tetrahedral dicationic bis-chelated
The second structural hypothesis, thereafter noted 4′, was a
complex where the Co1 center is associated with a CoCl4
dianion. We only optimized the dication without any
counterion for the low- and high-spin states configurations.
The calculated geometrical parameters are given in Table S2
Co center associated with a CoCl dianion. It was concluded
4
from the spectroscopic data and DFT calculations that the
former was more likely. We can anticipate that the approach
developed in this work will be useful in other polynuclear
systems to evaluate the degree of intramolecular cooperativity
between building blocks and in complex systems for which
structural information cannot always be obtained by single-
crystal X-ray diffraction.
(ESI) for the Co1 low-spin state. In the absence of the Cl atom,
the tetracoordinated Co1 could adopt a trigonal pyramidal
structure because the steric bulk of the P-phenyl rings prevents
a planar structure. The energy difference between the high- and
the low-spin states for the anionic form is very small and in
favor of the doublet by 0.5 eV. We also compared the energies
of the 4′ low-spin state and 4, which is lower by 10.6 eV. Note
that the energy for the anionic form 4′ was assessed by addition
EXPERIMENTAL SECTION
General Procedures. All operations were carried out using
standard Schlenk techniques under inert atmosphere. Solvents were
purified and dried under nitrogen by conventional methods. IR spectra
were recorded in the region 4000−100 cm on a Nicolet 6700 FT-IR
spectrometer (ATR mode, SMART ORBIT accessory, Diamond
■
of the energies of the two fragments 4′ and CoCl . Such a
4
difference could have been expected as the gas-phase approach
disfavors charged system: nothing stabilizes the cation−dianion
pair. Therefore, the solid-state structure of 4 is probably best
described as a zwitterionic form with a low-spin state for the
Co1 atom.
−1
crystal). Elemental analyses were performed by the “Service de
microanalyses”, Universite de Strasbourg, and by the “Service Central
́
d’Analyse”, USR-59/CNRS, Solaize. Electrospray mass spectra (ESI-
MS) were recorded on a microTOF (BrukerDaltonics, Bremen,
Germany) instrument using nitrogen as drying agent and nebulizing
gas, and MALDI-TOF analyses were carried out on a BrukerAutiflexII
TOF/TOF (BrukerDaltonics, Bremen, Germany) using dithranol
(1.8.9 trihydroxyanthracene) as a matrix. All other reagents were used
as received from commercial suppliers. Ligands 1 and 2 were prepared
according to literature methods.
CONCLUSION
■
This work has illustrated the nontrivial coordination chemistry
of Co(II) complexes with two tritopic P,P,S ligands of the N-
functionalized DPPA type and provided further examples of the
versatile coordination geometries displayed by this metal
center. As expected, the different nature of the N-substituent
in the assembling ligands (Ph P) N(CH ) SMe (1) and
Synthesis of the Complexes. We first describe the complexes
containing ligand 1 and then those with ligand 2.
2
2
2 3
Complex [Co Cl (1) ] (3). Synthesis Resulting from the Reaction
2
4
2
(
Ph P) N(p-C H )SMe (2), an alkylthioether or an arylth-
2 2 6 4
of CoCl with Ligand 1. To a suspension of anhydrous CoCl (0.082
2
2
ioether group, respectively, has no significant impact on the
P,P-chelating ability of these diphosphine ligands, which is their
main characteristics. However, the higher flexibility of the
alkylthioether N-substituent and its more electron-rich S atom
facilitates coordination of a second metal center. Thus, the
dinuclear complex 3, which was obtained by reaction of ligand
g, 0.63 mmol) in CH Cl (10 mL) was added a solution of ligand 1
2
2
(
0.300 g, 0.63 mmol) in CH Cl (20 mL). The solution quickly turned
2 2
to green and was stirred at room temperature for 12 h. The solution
was filtered through a Celite pad to ensure removal of any unreacted
starting material, and the solvent was removed under reduced pressure.
The resulting dark green solid was washed with diethyl ether and
pentane and recrystallized from a mixture of CH Cl /pentane (1:5),
1
with CoCl in a 1:1 molar ratio, has an original zwitterionic
2
2
2
yielding complex 3 (0.210 g, 55%). Yield was based on the crystalline
material; however, the FTIR spectra of the green powder (precipitated
or isolated after drying of the mother liquor) were similar to those of
crystals of 3. Red/green crystals suitable for single-crystal X-ray
diffraction were grown from a mixture of CH Cl /pentane. Anal. Calcd
structure consisting of a cationic CoCl center, bis-chelated by
two PNP ligands, and one anionic CoCl moiety linked to the
3
N-thioether function. The reaction of ligand 2 with CoCl in a
2
1:1 molar ratio led also to a dinuclear complex, 4. To gain
2
2
better insight into the structures and properties of these
complexes, a retro-synthetic approach was developed, consist-
ing in the synthesis of mononuclear fragments as similar as
possible to those constituting the dinuclear complexes, in order
to compare their spectroscopic signatures (FT-IR, UV−vis,
EPR) with those of the dinuclear complex. First, this approach
was successfully applied to 3, whose spectroscopic character-
istics were analyzed as the superposition of those of the
corresponding mononuclear fragments. Magnetic and EPR
studies and theoretical calculations were performed in order to
gain a more complete understanding of the systems
for C H Cl Co N P S (1206.78): C, 55.74; H, 4.84; N, 2.32.
Found: C, 55.33; H, 4.83; N, 2.17. FTIR (νmax(solid)/cm ) 1770w,
56 58 4 2 2 4 2
−1
1480w, 1433s, 1310w, 1184w, 1091s, 1069sh, 1037sh, 997m, 879m,
−1
7
34s, 694vs. Far-FTIR (νmax(solid)/cm ) 531s, 511s, 502s, 490s,
4
82s, 319vs. MS (ESI): m/z (ranked by decreasing intensity) = 1005.1
+
+
[
M − H − CoCl ] , 1040.1 [M − CoCl ] .
4
3
Synthesis Starting from Complex 5. Solid anhydrous CoCl
2
(
5
0.0175 g, 0.135 mmol) was added to a deep red solution of complex
(0.080 g, 0.067 mmol) in CH Cl (20 mL). The resulting suspension
2
2
was stirred overnight, filtered through a Celite pad to remove the
“CoCl(PF )” formed, and concentrated to a quarter of its original
6
volume under reduced pressure. Addition of pentane (50 mL) led to
M
DOI: 10.1021/acs.inorgchem.5b02889
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
the precipitation of a green powder, isolated by filtration, dried under
reduced pressure, and identified as complex 3 by FT-IR and EA (0.043
g, 53% based on complex 5). FT-IR data are superimposable to those
described above. Anal. Calcd: C, 55.74; H, 4.84; N, 2.32. Found: C,
X-ray Data Collection, Structure Solution, and Refinement
for All Compounds. Suitable crystals for the X-ray analysis of all
compounds were obtained as described above. The intensity data were
4
4
collected on a Kappa CCD diffractometer (graphite monochromated
Mo Kα radiation, λ = 0.71073 Å) at 173(2) K. Crystallographic and
experimental details for the structures are summarized in Table S1
(ESI). The structures were solved by direct methods (SHELXS-97)
5
5.46; H, 4.77; N, 2.29.
Complex [CoCl(1) ]PF (5). A solution of ligand 1 (0.500 g, 1.06
2
6
mmol) in CH Cl (20 mL) was added to a mixture of anhydrous
2
2
2
CoCl (0.069 g, 0.53 mmol) and KPF (0.097 g, 0.53 mmol) in
and refined by full-matrix least-squares procedures (based on F ,
2
6
4
5
CH Cl (10 mL). The solution turned to brown and was stirred for 12
SHELXL-97) with anisotropic thermal parameters for all the non-
hydrogen atoms. The hydrogen atoms were introduced into the
geometrically calculated positions (SHELXS-97 procedures) and
refined riding on the corresponding parent atoms. For some
2
2
h. The solution was filtered through a Celite pad to ensure removal of
any unreacted starting material, and the solvent was removed under
reduced pressure. The resulting red/brown solid was washed with
diethyl ether and pentane and recrystallized from a mixture of
CH Cl /pentane, affording red crystals of complex 5 (0.590 g, 94%).
4
6
compounds a MULTISCAN correction was applied.
Magnetic Measurements. Measurements were performed on
microcrystalline samples enclosed in 30 μm polyethylene bags.
Weights were accurately determined with a Mettler Toledo MX5
microbalance. The bags diamagnetic contribution was subtracted from
the measured magnetic moments with a diamagnetic correction
combined to a small Curie tail determined previously on a massive
sample. Measurements were performed on Quantum Design MPMS-
2
2
Anal. Calcd for C H ClCoF N P S ·3CH Cl (1441.25): C, 49.17;
56
58
6
2
5
2
2
2
H, 4.48; N, 1.94. Found: C, 48.78; H, 4.47; N, 1.87. FTIR
−
1
(
7
4
νmax(solid)/cm ) 1434m, 1312w, 1187w, 1091s, 999w, 834vs,
−1
34s, 693s. Far-FTIR (νmax(solid)/cm ) 556vs, 530vs, 513vs, 501vs,
+
90vs, 418m. MS (ESI): m/z 1040.2 [M − PF ] .
6
Complex [CoCl (H1′)] (6) (1′ = NH (CH ) SMe). An aqueous HCl
3
2
2 3
5
XL with a dc probe and MPMS-7 with a RSO probe SQUID
solution (0.40 mL of 37% HCl in H O, corresponding to 4.75 mmol of
2
magnetometers; data were corrected when necessary with locally
HCl) was added to a solution of 3-(methylthio)propylamine (0.53 mL,
4
7
developed procedures using multipole fitting. Sample diamagnetic
contributions were approximated using Pascal constants. Magnetic
simulations were calculated with a locally modified version of the
0
.500 g, 4.75 mmol) in CH Cl . A white precipitate was
2 2
instantaneously formed. Anhydrous CoCl (0.617 g, 4.75 mmol) was
2
then added to the suspension, which turned blue, and the mixture was
stirred for 2 h. The volatiles were then removed under reduced
pressure, and the solid was washed with diethyl ether (2 × 20 mL).
Complex 6 was isolated as a blue powder (1.11 g, 86%). Anal. Calcd
for C H Cl CoNS (271.50): C, 17.70; H, 4.45; N, 5.16. Found: C,
4
8
Magpack software.
EPR measurements. X-band (Larmor frequency ≈ 9.3 GHz) EPR
spectra were recorded with a continuous-wave EMXplus spectrometer
(
BrukerBiospin GmbH, Germany) equipped with a high-sensitivity
4
12
3
−1
resonator (4119HS-W1, Bruker). The spectrometer was tuned so the
settings (modulation coils, incident microwave power) did not distort
the EPR signal. Measurements were carried out on powdered samples
held at 4−5 K. Simulations were completed with the EasySpin free
software (http://www.easyspin.org).
1
1
9
2
7.66; H, 4.75; N, 5.29. FTIR (νmax(solid)/cm ) 3102s, 2941m,
572m, 1486s, 1443m, 1244w, 1217w, 1114w, 1097m, 1040w, 978m,
−1
32m, 827w, 773w. Far-FTIR (νmax(solid)/cm ) 489w, 459w, 433w,
+
89vs MS (ESI): m/z 234.9 [M − Cl] .
Complex [CoCl (2)] (4). The same procedure as for complex 3 was
2
2
Computational Details. Ground state electronic structures have
used with anhydrous CoCl (0.051 g, 0.39 mmol) and ligand 2 (0.200
2
been calculated by DFT methods using the Amsterdam Density
g, 0.39 mmol) and afforded complex 4 as green crystalline solid (0.208
4
9
Functional (ADF2010.02). Geometry optimization was based on the
g, 83%). Anal. Calcd for C H Cl Co N P S (1274.81): C, 58.41; H,
62
54
4
2
2 4 2
5
0,51
hybrid B3LYP exchange-correlation functional.
A standard triple-ζ
4
.27; N, 2.20. Found: C, 58.77; H, 4.54; N, 2.35. FTIR (νmax(solid)/
−1
Slater basis set was used for all atoms. All calculations were performed
cm ) 1586w, 1491m, 1433s, 1299w, 1271w, 1217m, 1158w, 1093s,
within the unrestricted formalism. Relativistic effects were not
1
026w, 1010w, 998w, 934m, 894s, 811w, 736s, 691vs. Far-FTIR
5
2
−1
included. The ORCA package was used to calculate EPR properties
through a single point on ADF-optimized geometries. B3LYP
(
νmax(solid)/cm ) 554m, 526s, 495s, 423w, 295s. MS (ESI): m/z
ranked by decreasing intensity) = 1073.1 [M − H − CoCl ] , 1108.1
+
(
4
53
+
exchange-correlation functional was applied with TZVP basis for
[M − CoCl ] .
3
the metal centers and the remaining atoms. Representations of the
molecular structures and orbitals were made using ADFview v06. The
obtained wave functions were analyzed with the DGRID package to
Synthesis Starting from Complex 7. The same procedure as for
complex 3 was followed, with CoCl (0.0186 g, 0.143 mmol) and
2
complex 7 (0.090 g, 0.072 mmol) in CH Cl (20 mL), affording
2
2
54
compute maps of spin density.
complex 4 (0.057 g, 62%). FT-IR data are superimposable to those
described above. Anal. Calcd: C, 58.41; H, 4.27; N, 2.20. Found: C,
5
8.38; H, 4.49; N, 2.30.
ASSOCIATED CONTENT
■
Complex [CoCl(2) ]PF (7). The same procedure as for complex 5
2
6
*
S
Supporting Information
was used with anhydrous CoCl (0.064 g, 0.49 mmol), KPF (0.091 g,
2
6
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.inorg-
chem.5b02889.
0
.49 mmol), and ligand 2 (0.500 g, 0.99 mmol) and afforded green
crystals of complex 7 (0.580 g, 94%). Anal. Calcd for
C H ClCoF N P S ·H O (1272.50): C, 58.52; H, 4.44; N, 2.20.
6
2
54
6
2
5
2
2
−1
Found: C, 58.60; H, 4.43; N, 1.87. FTIR (νmax(solid)/cm 1490m,
434m, 1217m, 1094m, 937m, 898m, 836vs, 736m, 694s. Far-FTIR
X-ray data collection and structure refinement for all
compounds, CCDC numbers 1440248−1440252 (3·
2(CH Cl ), 4·CH Cl , 5, 7, and 8) and 1451018−
1
−1
(νmax(solid)/cm ) 556s, 526vs, 514sh, 492vs, 443m. MS (ESI): m/z
+
2
2
2
2
1
108.2 [M − PF ] .
6
1
451019 (9 and 10·CH Cl ); far-FT-IR spectra and
Complex [H2′] [CoCl ] (8) (2′ = NH (p-C H )SMe). The same
2 2
2
4
2
6 4
procedure as for complex 6 was used with anhydrous CoCl (0.466 g,
3
corresponding to 3.59 mmol of HCl), and 4-(methylthio)aniline (0.45
mL, 0.500 g, 3.59 mmol). Unreacted CoCl was removed by water
washing before additional Et O washing, and complex 8 was isolated as
a blue powder (0.760 g, 44%, based on CoCl used). Anal. Calcd for
C H Cl CoN S .4H O (553.26): C, 30.39; H, 5.10; N, 5.06. Found:
C, 30.26; H, 5.17; N, 4.92. FTIR (νmax(solid)/cm ) 3444m, 3372m,
UV−vis spectra of compounds 3−8; molar susceptibility
plots vs T and isothermal magnetization plots for
compounds 3−8; representation of the magnetic pair
compound 5; representation of the magnetic layers in
compound 8; relevant computed geometrical parameters
for complexes 3−8; ADF view drawings of the Kohn−
Sham highest occupied molecular orbitals (SOMOs) of
complexes 6 and 4; spin density plots calculated for 4;
principal g values for complexes 5−8 and Zn(II) model
complexes of 3 and 4 (PDF)
2
.59 mmol), aqueous HCl solution (0.29 mL of 37% HCl in H O,
2
2
2
2
1
4
20
4
2
2
2
−1
2
1
2
989w, 2947w, 1611m, 1453s, 1402m, 1309w, 1184s, 1082w, 1034m,
−1
006s, 794s, 710w. Far-FTIR (νmax(solid)/cm ) 322m, 294s, 272s,
+
26m, 168m, 125vs MS (ESI): m/z 140.1 [CH S(C H )NH ] .
3
6
4
3
N
DOI: 10.1021/acs.inorgchem.5b02889
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
X-ray crystallographic data in CIF format (CIF)
Article
Jordan, R. F. Organometallics 2008, 27, 2107−2117. (e) Dennett, J. N.
L.; Gillon, A. L.; Heslop, K.; Hyett, D. J.; Fleming, J. S.; Lloyd-Jones,
C. E.; Orpen, A. G.; Pringle, P. G.; Wass, D. F.; Scutt, J. N.;
Weatherhead, R. H. Organometallics 2004, 23, 6077−6079. (f) Cooley,
N. A.; Green, S. M.; Wass, D. F.; Heslop, K.; Orpen, A. G.; Pringle, P.
G. Organometallics 2001, 20, 4769−4771. (g) Wass, D. F. (BP
Chemicals Ltd.) WO01/10876, 2001. (h) Hofmann, P.; Perez-Moya,
L. A.; Steigelmann, O.; Riede, J. Organometallics 1992, 11, 1167−1176.
i) Hofmann, P.; Perez-Moya, L. A.; Krause, M. E.; Kumberger, O.;
Muller, G. Z. Z. Naturforsch., B: J. Chem. Sci. 1990, 45, 897−908.
7) For general recent reviews on Cr-based catalysts see, e.g.,
a) Dagorne, S.; Fliedel, C. Top. Organomet. Chem. 2013, 41, 125−
71. (b) McGuinness, D. S. Chem. Rev. 2011, 111, 2321−2341.
c) Agapie, T. Coord. Chem. Rev. 2011, 255, 861−880. (d) van
AUTHOR INFORMATION
Corresponding Author
E-mail: braunstein@unistra.fr.
■
*
Notes
The authors declare no competing financial interest.
(
̈
ACKNOWLEDGMENTS
We are grateful to the CNRS, the Minister
Paris), the French EPR Federation of Research (REseau
■
(
(
̀
e de la Recherche
(
1
(
́ ́
NAtional de Rpe interDisciplinaire, RENARD, Federation IR-
RPE CNRS 3443), the DFH/UFA (International Research
Training Group 532-GRK532, Ph.D. grant to C.F.), and the
Leeuwen, P. W. N. M.; Clement, N. D.; Tschan. Coord. Chem. Rev.
2011, 255, 1499−1517. For other examples of Cr-based catalysts see,
e.g., (e) Reddy Aluri, B.; Peulecke, N.; Peitz, S.; Spannenberg, A.;
Fundaca
SFRH/BPD/73253/2010 to C.F. and SFRH/BPD/44262/
008 to V.R.) for financial and scientific support. We thank Dr.
̃
̂
̧ o para a Ciencia e Tecnologia (FCT) (fellowships
Mu
̈
ller, B. H.; Schulz, S.; Drexler, H.-J.; Heller, D.; Al-Hazmi, M. H.;
hl, A.; Muller, W.; Rosenthal, U. Dalton Trans. 2010,
2
Mosa, F. M.; Wo
3
̈
̈
9, 7911−7920. (Cr and Ni complexes): (f) Zhang, J.; Braunstein, P.;
L. Karmazin and Miss. C. Bailly (Service de radiocrystallogra-
phie, Institut de Chimie, Unistra) and Dr. R. Pattacini for the
Hor, T. S. A. Organometallics 2008, 27, 4277−4279. (g) Weng, Z.;
Teo, S.; Hor, T. S. A. Dalton Trans. 2007, 38, 3493−3498.
́
X-ray diffraction studies, Melanie Boucher for technical
(
8) Lippard, S. J.; Berg, J. M. Principles of Bioinorganic Chemistry;
assistance, and Dr. D. Specklin for the graphical Figures 1−5.
We thank the high-performance computing center of the
University of Strasbourg for computation facilities.
University Science Books: Mill Valley, CA, 1994; Chapter 12.
9) (a) Raphael, A. L.; Gray, H. B. J. Am. Chem. Soc. 1991, 113,
038−1040. (b) Geno, M. K.; Halpern, J. J. Am. Chem. Soc. 1987, 109,
238−1240. (c) Perutz, M. F.; Fermi, G.; Luisi, B.; Shaanan, B.;
Liddington, R. C. Acc. Chem. Res. 1987, 20, 309−321. (d) Halpern, J.
Science 1985, 227, 869−875. (e) Garret, T. P. J.; Glingeleffer, D. J.;
Guss, J. M.; Rogers, S. J.; Freeman, H. C. J. Biol. Chem. 1984, 259,
822−2825.
10) Ghisolfi, A.; Condello, F.; Fliedel, C.; Rosa, V.; Braunstein, P.
Organometallics 2015, 34, 2255−2260.
11) Fliedel, C.; Rosa, V.; Falceto, A.; Rosa, P.; Alvarez, S.;
Braunstein, P. Inorg. Chem. 2015, 54, 6547−6559.
12) For recent examples see, e.g., (a) Ok, F.; Aydemir, M.; Durap, F.
(
1
1
REFERENCES
■
(
1) For selected reviews see, e.g., (a) Hierso, J.-C.; Amardeil, R.;
Bentabet, E.; Broussier, R.; Gautheron, B.; Meunier, P.; Kalck, P.
Coord. Chem. Rev. 2003, 236, 143−206. (b) Appleby, T.; Woollins, J.
D. Coord. Chem. Rev. 2002, 235, 121−140. (c) Bhattacharyya, P.;
Woollins, J. D. Polyhedron 1995, 14, 3367−3388. Other specific
examples are cited along the manuscript.
2
(
(
(
́
2) (a) Rosa, V.; Fliedel, C.; Ghisolfi, A.; Pattacini, R.; Aviles, T.;
Braunstein, P. Dalton Trans. 2013, 42, 12109−12119. (b) Fliedel, C.;
(
Pattacini, R.; Braunstein, P. J. Cluster Sci. 2010, 21, 397−415.
Appl. Organomet. Chem. 2014, 28, 38−43. (b) Naik, S.; Mague, J. T.;
Balakrishna, M. S. Inorg. Chim. Acta 2013, 407, 139−144. (c) Sarcher,
C.; Lebedkin, S.; Kappes, M. M.; Fuhr, O.; Roesky, P. W. J. Organomet.
Chem. 2014, 751, 343−350. (d) Aydemir, M.; Meric, N.; Kayan, C.;
Ok, F.; Baysal, A. Inorg. Chim. Acta 2013, 398, 1−10.
(
c) Ganesamoorthy, C.; Balakrishna, M. S.; Mague, J. T. Inorg. Chem.
2
́
009, 48, 3768−3782. (d) Rodriguez-Zubiri, M.; Gallo, V.; Rose, J.;
Welter, R.; Braunstein, P. Chem. Commun. 2008, 64−66. (e) Gallo, V.;
Mastrorilli, P.; Nobile, C. F.; Braunstein, P.; Englert, U. Dalton Trans.
2
006, 2342−2349.
3) (a) Lin, B.; Liu, Z.; Liu, M.; Pan, C.; Ding, J.; Wu, H.; Cheng, J.
Catal. Commun. 2007, 8, 2150−2152. (b) Posset, T.; Rominger, F.;
Blumel, J. Chem. Mater. 2005, 17, 586−595. (c) Schweyer-Tihay, F.;
Braunstein, P.; Estournes, C.; Guille, J. L.; Lebeau, B.; Paillaud, J.-L.;
Richard-Plouet, M.; Rose, J. Chem. Mater. 2003, 15, 57−62.
d) Braunstein, P.; Kormann, H.-P.; Meyer-Zaika, W.; Pugin, R.;
Schmid, G. Chem. - Eur. J. 2000, 6, 4637−4646.
4) Fliedel, C.; Faramarzi, V.; Rosa, V.; Doudin, B.; Braunstein, P.
Chem. - Eur. J. 2014, 20, 1263−1266.
5) For recent examples see, e.g., (a) Aydemir, M.; Baysal, A.; Sahin,
(
13) (a) Catalysis without Precious Metals; Bullock, R. M., Ed.; Wiley-
(
VCH: Weinheim, 2010. (b) Biochemistry and Inorganic Chemistry:
Influence on Structure and Reactivity; Swart, M., Costas, M., Eds.; John
Wiley & Sons, Ltd.: Oxford, U.K., 2015. (c) Poli, R.; Allan, L. E. N.;
Shaver, M. P. Prog. Polym. Sci. 2014, 39, 1827−1845. (d) Pedersen, K.
̈
̀
́
S.; Bendix, J.; Cler
e) Grohmann, A. Dalton Trans. 2010, 39, 1432−1440. (f) Debuigne,
A.; Poli, R.; Jer me, C.; Jer me, R.; Detrembleur, C. Prog. Polym. Sci.
009, 34, 211−239.
14) (a) Ghisolfi, A.; Monakhov, K.; Yu; Pattacini, R.; Braunstein, P.;
Lopez, X.; de Graaf, C.; Speldrich, M.; van Leusen, J.; Schilder, H.;
Ko
́
ac, R. Chem. Commun. 2014, 50, 4396−4415.
(
(
́
o
̂
́ ̂
o
(
2
(
(
́
̈
̈ ̈
mgum, B.; Ozkar, S. Inorg. Chim. Acta 2011, 378, 10−18.
E.; Gu
̈
gerler, P. Dalton Trans. 2014, 43, 7847−7859. (b) Pattacini, R.;
(
b) Ganesamoorthy, C.; Balakrishna, M. S.; Mague, J. T. J. Organomet.
Teo, P.; Zhang, J.; Lan, Y.; Powell, A. K.; Nehrkorn, J.; Waldmann, O.;
Chem. 2009, 694, 3390−3394. (c) Ganesamoorthy, C.; Balakrishna, M.
S.; Mague, J. T.; Tuononen, H. M. Inorg. Chem. 2008, 47, 7035−7047.
Hor, T. S. A.; Braunstein, P. Dalton Trans. 2011, 40, 10526−10534.
15) (a) Naktode, K.; Kottalanka, R. K.; Panda, T. K. J. Coord. Chem.
(
(
d) Paolillo, R.; Gallo, V.; Mastrorilli, P.; Nobile, C. F.; Rose,
́
J.;
m, B.;
̈ ̈
Biricik, N.; Durap, F.; Ozdemir, I.; Gurbuz, N.; Ang, W. H.; Dyson, P.
2
014, 67, 3042−3053. (b) Zheng-Zhi, Z.; Ao, Yu; Hua-Ping, X.; Ru-Ji,
W.; Hong-Gen, W. J. Organomet. Chem. 1994, 470, 223−229.
c) Werner, H.; Lippert, F.; Peters, K.; von Schnering, H. G. Chem.
Ber. 1992, 125, 347−352.
16) (a) Soobramoney, L.; Bala, M. D.; Friedrich, H. B. Dalton Trans.
2014, 43, 15968−15978. (b) Slepukhin, P. A.; Kim, D. M.; Charushin,
V. N. Russ. J. Coord. Chem. 2010, 36, 615−617. (c) Baumann, G.;
Marzilli, L. G.; Nix, C. L.; Rubin, B. Inorg. Chim. Acta 1983, 77, L35−
L37.
̈ ̈
Braunstein, P. Organometallics 2008, 27, 741−746. (e) Gumgu
̈
(
J. Appl. Organomet. Chem. 2007, 21, 711−715. (f) Mandal, S. K.;
Gowda, G. A. N.; Krishnamurthy, S. S.; Zheng, C.; Li, S.; Hosmane, N.
S. J. Organomet. Chem. 2003, 676, 22−37.
(
(
6) For examples of Ni-based catalysts see, e.g., (a) Ghisolfi, A.;
Fliedel, C.; Rosa, V.; Monakhov, K. Yu.; Braunstein, P. Organometallics
014, 33, 2523−2534. (b) Schultz, M.; Eisentrager, F.; Regius, C.;
2
̈
Rominger, F.; Hanno-Igels, P.; Jakob, P.; Gruber, I.; Hofmann, P.
Organometallics 2012, 31, 207−224. (c) Song, K.; Gao, H.; Liu, F.;
Pan, J.; Guo, L.; Zai, S.; Wu, Q. Eur. J. Inorg. Chem. 2009, 2009, 3016−
(17) Alvarez, S.; Llunell, M. Dalton Trans 2000, 3288. See also:
Alvarez, S. Chem. Rev. 2015, 115, 13447−13483 and references cited
therein..
3
024. (d) Lavanant, L.; Rodrigues, A.-S.; Kirillov, E.; Carpentier, J.-F.;
O
DOI: 10.1021/acs.inorgchem.5b02889
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
(
18) Ghisolfi, A.; Fliedel, C.; Rosa, V.; Pattacini, R.; Thibon, A.;
Monakhov, K. Yu.; Yu; Braunstein, P. Chem. - Asian J. 2013, 8, 1795−
805.
19) Cotton, F. A.; Wilkinson, G.; Murillo, C.; Bochmann, A. M.
Advanced Inorganic Chemistry, 6th ed.; John Wiley & Sons: New York,
999; pp 817−822.
20) Li, J.; Merkel, S.; Henn, J.; Meindl, K.; Do
W.; Ghadwal, R. S.; Stalke, D. Inorg. Chem. 2010, 49, 775−777.
21) Mondal, K. C.; Samuel, P. P.; Roesky, H. W.; Carl, E.; Herbst-
(44) Bruker-Nonius. Kappa CCD Reference Manual; Nonius BV: The
Netherlands, 1998.
(45) Sheldrick, G. M. Acta Crystallogr., Sect. A: Found. Crystallogr.
1
(
2008, 64, 112−122.
(46) (a) Spek, L. J. Appl. Crystallogr. 2003, 36, 7−13. (b) Blessing, B.
Acta Crystallogr., Sect. A: Found. Crystallogr. 1995, 51, 33−38.
(47) Palamarciuc, T.; Oberg, J. C.; El Hallak, F.; Hirjibehedin, C. F.;
1
(
̈
ring, A.; Roesky, H.
Serri, M.; Heutz, S.; Let
690−9695.
́
48) Borras-Almenar, J. J.; Clemente-Juan, J. M.; Coronado, E.;
́
ard, J.-F.; Rosa, P. J. Mater. Chem. 2012, 22,
9
(
(
Irmer, R.; Stalke, D.; Schwederski, B.; Kaim, W.; Ungur, L.; Chibotaru,
Tsukerblat, B. S. Inorg. Chem. 1999, 38, 6081−6088; and J. Comput.
L. F.; Hermann, M.; Frenking, G. J. Am. Chem. Soc. 2014, 136, 1770−
Chem. 2001, 22, 985−991.
1
773.
(
49) ADF2010.02; Theoretical Chemistry, Vrije Universiteit:
(
22) (a) Farona, M. F.; White, J. F. Inorg. Chem. 1971, 10, 1080−
081. (b) Dyer, G.; Meek, D. W. J. Am. Chem. Soc. 1967, 89, 3983−
987.
23) Nishida, Y.; Kida, S. Bull. Chem. Soc. Jpn. 1975, 48, 1045−1046.
24) Pietrzyk, P.; Srebro, M.; Radon, M.; Sojka, Z.; Michalak, A. J.
Phys. Chem. A 2011, 115, 2316−2324.
25) Marts, A. R.; Greer, S. M.; Whitehead, D. R.; Woodruff, T. M.;
Amsterdam, The Netherlands, SCM, 2010.
1
3
(
(
(
50) Becke, A. D. J. Chem. Phys. 1993, 98, 5648−5652.
51) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B: Condens. Matter
Mater. Phys. 1988, 37, 785−789.
52) (a) Neese, F. ORCA An Ab Initio, DFT and Semiempirical
Electronic Structure Package, version 2.9.1; Universitat Bonn: Bonn,
Germany, 2012. (b) Neese, F. J. Chem. Phys. 2007, 127, 164112.
53) (a) Schafer, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys. 1992, 97,
571−2577. (b) Schafer, A.; Huber, C.; Ahlrichs, R. J. Chem. Phys.
994, 100, 5829−5835.
54) Kohout, M.DGrid, version 4.6; Radebeul, 2011.
(
́
(
̈
(
Breece, R. M.; Shim, S. W.; Oseback, S. N.; Papish, E. T.; Jacobsen, F.
(
̈
E.; Cohen, S. M.; Tierney, D. L. Appl. Magn. Reson. 2011, 40, 501−
2
1
̈
5
11.
(
(
26) Bennett, B. Curr. Top. Biophys. 2002, 26, 49−57.
27) Myers, W. K.; Duesler, E. N.; Tierney, D. L. Inorg. Chem. 2008,
(
4
7, 6701−6710.
28) Jacobsen, F. E.; Breece, R. M.; Myers, W. K.; Tierney, D. L.;
Cohen, S. M. Inorg. Chem. 2006, 45, 7306−7315.
29) Nishida, Y.; Shimohori, H. Bull. Chem. Soc. Jpn. 1973, 46, 2406−
410.
30) Figgis, B. N.; Hitchman, M. A. Ligand Field Theory and Its
Applications; Wiley-VCH.: New York, 2000; pp 228−263.
31) Banci, L.; Bencini, A.; Benelli, C.; Gatteschi, D.; Zanchini, C.
(
(
2
(
(
Struct. Bonding (Berlin) 1982, 52, 37−86.
(
32) Cossee, P.; van Arkel, A. E. J. Phys. Chem. Solids 1960, 15, 1−6.
(
33) DeW, W.; Horrocks, D. A., Jr.; Burlone. J. Am. Chem. Soc. 1976,
9
(
1
(
2
(
8, 6512−6516.
34) Dyrek, K.; Dyrek, M.; Wesełucha-Birczyn
985, 4, 169−172.
35) Qu, T.; Kuang, X.-Y.; Li, Y.-F.; Chai, R.-P. Chem. Phys. Lett.
011, 504, 170−174.
36) Boca, R. Theoretical foundations of Molecular Magnetism;
́
ska, A. Polyhedron
̌
Elsevier: Lausanne (Switzerland), 1999; Chapters 5 and 9.
2
z
2
x
2
̂
̂
̂
̂
̂
̂
̂
̂
y
(
37) HamiltonianH = βg ∑ S·H − 2JS S + ∑ DS + E(S − S )
1
2
z2JS(S + 1)
(
(
38) θ =
39) It must be noted that an uncorrected ferromagnetic impurity
3
k
would give a similar contribution than a very high TIP, but
magnetization measurements at 100 K did not evidence such an
impurity.
(
40) Though no maximum is observed on the susceptibility plot, a
similar behavior was observed for the isothermal magnetization of a
D Heisenberg honeycomb lattice model, see: Curely, J.; Lloret, F.;
Julve, M. Phys. Rev. B: Condens. Matter Mater. Phys. 1998, 58, 11465−
1483.
41) Only visual agreement was sought for. Since parameters are
2
́
1
(
likely to be heavily correlated, the parameters are certainly neither
unique nor optimal.
(
42) (a) Drago, R. S.; Stahlbush, J. R.; Kitko, D. J.; Breese, J. J. Am.
Chem. Soc. 1980, 102, 1884−1889. (b) Stelzer, O.; Sheldrick, W. S.;
Subramanian, J. J. Chem. Soc., Dalton Trans. 1977, 966−970.
(
43) (a) Krzystek, J.; Ozarowski, A.; Zvyagin, S. A.; Telser, J. Inorg.
Chem. 2012, 51, 4954−4964. (b) Krzystek, J.; Zvyagin, S. A.;
Ozarowski, A.; Fiedler, A. T.; Brunold, T. C.; Telser, J. J. Am. Chem.
Soc. 2004, 126, 2148−2155. (c) Romerosa, A.; Saraiba-Bello, C.;
Serrano-Ruiz, M.; Caneschi, A.; Mckee, V.; Peruzzini, M.; Zanobini, F.
Dalton Trans. 2003, 3233−3239. (d) Makinen, M. W.; Kuo, L. C.;
Yim, M. B.; Wells, G. B.; Fukuyama, J. M.; Kim, J. E. J. Am. Chem. Soc.
1
985, 107, 5245−5255.
P
DOI: 10.1021/acs.inorgchem.5b02889
Inorg. Chem. XXXX, XXX, XXX−XXX