Deciphering preferred geometries of pyridylmethylamines-based complexes: A robust strategy combining NMR, DFT and X-ray

Article abstract of DOI:10.1016/j.ica.2019.119070

The preparation of pyridylmethylamines (pma)-ZnBr₂ and -CoBr₂ complexes is described. Accurate structural informations in both solution and solid state have been obtained using an approach combining advanced NMR such as pure shift gr

Full text of DOI:10.1016/j.ica.2019.119070

InorganicaChimicaActa498(2019)119070
Contents lists available at ScienceDirect
Inorganica Chimica Acta
journal homepage: www.elsevier.com/locate/ica
Deciphering preferred geometries of pyridylmethylamines-based complexes:
A robust strategy combining NMR, DFT and X-ray
Benjamin Largea, Maissa Meddebb, José Enrique Herbert Puchetab,c, Anne Gauchera,
⁎
Marie Cordierd, Corinne Gosminid, Jonathan Farjonb,e, , Audrey Auffrantd,⁎, Damien Prima,⁎
a Institut Lavoisier de Versailles, UVSQ, CNRS, Université Paris-Saclay, 78035 Versailles, France
b Institut de Chimie Moléculaire et des Matériaux d’Orsay, University of Paris-Sud, UMR 8182, 91405 Orsay, France
c Consejo Nacional de Ciencia y Tecnología Laboratorio Nacional de Investigación y Servicio Agroalimentario y Forestal, Universidad Autónoma Chapingo, Km 38.5
Carretera México – Texcoco, Chapingo 56230, Estado de México, Mexico
d Laboratoire de Chimie Moléculaire, CNRS, École Polytechnique, IP Paris, 91128 Palaiseau, France
e CEISAM UMR-CNRS 6230, Faculté des Sciences et Techniques, 2 Rue de la Houssinière, BP 92208, 6230 Nantes, France
A R T I C L E I N F O
A B S T R A C T
Keywords:
Pyridylmethylamine
Zn
Co
Complexation
X-ray
The preparation of pyridylmethylamines (pma)-ZnBr2 and -CoBr2 complexes is described. Accurate structural
informations in both solution and solid state have been obtained using an approach combining advanced NMR
such as pure shift gradient-encoded selective refocusing (PS-GSERF) and conventional NOESY experiments, DFT
calculations and X-ray analysis. The methodology developed has allowed a clear identification and character-
ization of preferred conformations and configurations at an atomic resolution. Our study has evidenced some key
features of the overall 3D structure of pma-Zn and pma-Co complexes which shapes are set by the geometry of
the metallacycle, the configuration of the sp3 nitrogen atom, the equatorial position of the benzyl side arm as
well as the preferred spatial arrangement of the chiral side arm with respect to the metallacycle.
1. Introduction
focused emerging attention in several areas. Indeed, pma-Zn complexes
were found efficient as enantiomeric excess determination tool [26,27]
Over the last two decades, complexation properties of pyr-
idylmethylamines (pma) and their derivatives to various transition
metals have attracted continuous interest for a broad scientific com-
munity [1–10]. The pma scaffold comprises a pyridine ring substituted
by a methylamine pendant arm and belongs to the general 1,2-N,N
bidentate ligands family. Those pma ligands display two distinct ni-
trogen atoms: the N-heterocyclic and the sp3 nitrogen atoms. The
growing popularity of pma ligands arise by part from their facile
synthesis. Advantageously, their electronic and steric properties can be
customized by the easy installation of various substituents at the pyr-
idine ring, the methylene carbon atom and the sp3 nitrogen atom
(Fig. 1). This ligand modularity enables pma to complex metals such as
Cu, Re, Fe, Pd, Pt, Mg, Ti, Ni or Yb. Moreover, its ability to generate
pma-H+ complexes has also been recently demonstrated [11]. Such
compounds revealed especially valuable as catalytic systems in various
synthetic organo- and metallo-promoted transformations including
oxidative CeC bond formation, cycloaddition reactions, Friedel-Crafts
alkylation, Henry reaction, Suzuki coupling, CeH arylation [11–25].
Among the various transition metal complexes, pma-ZnII association
and in the polymerization of rac-lactide [28,29].
In addition, as a consequence of the crucial role played by ZnII ions
in several biological processes and diseases associated with the varia-
tion of its concentration [30–36], pma-based platforms have been used
as model ligands in molecular recognition, fluorescent probes and
imaging agent [1,37–42]. In this context, both the coordination mode
and the geometry of pma-ZnII complexes are key elements to study in
order to fully understand and/or predict their behaviour and reactivity.
Indeed, tiny modifications of the coordination sphere are known to
affect the reactivity of Zn species within the active sites of Zn-con-
taining enzymes, deeply impact catalytic activities and induce con-
formational modifications [43–45]. While the preparation of pma-ZnCl2
complexes and the description of their properties accompanied by solid
state data were disclosed in a few reports [28,37], the pma-ZnBr2
complexes are more scarcely described [46,47]. In this communication
we focused on the preparation of pma-ZnBr2 complexes based on N,N-
bidentate ligands. In order to gain accurate structural information in
both solution and solid state, we describe herein the first issue of an
approach combining DFT calculations and advanced NMR such as pure
⁎ Corresponding authors.
E-mail addresses: jonathan.farjon@univ-nantes.fr (J. Farjon), audrey.auffrant@polytechnique.edu (A. Auffrant), Damien.prim@uvsq.fr (D. Prim).
https://doi.org/10.1016/j.ica.2019.119070
Received 17 June 2019; Received in revised form 29 July 2019; Accepted 8 August 2019
Availableonline20August2019
0020-1693/©2019ElsevierB.V.Allrightsreserved.

B. Large, et al.
InorganicaChimicaActa498(2019)119070
modulation sites
R1
N
N
N
Br Zn
Br
R2
R3
or
Br Zn
Br
NH
NH
ZnBr2
*
N
R
S
THF, 16 h
N
N
HN
N
*
2
79%
1
58%
= Pd, Cu, Pt, Yb ...
or H+
CoBr2
THF, 16 h
LS or LR
N
this work
N
or
Br Co NH
Br
Br Co NH
Br
Me
Ph
Me
S
R
*
N
N
*
N
N
Zn
Co
Ph
3
70%
4
77%
Br
Br
Br
Br
Scheme 2. Preparation of complexes 1–4.
Fig. 1. Pma modulation sites, transition metal and proton complexes and ZnII
and CoII complexes.
independent enantiomers 1 and 2 respectively. Similarly, pma-CoBr2
complexes 3 and 4 could be isolated in 70 and 77% yield from ligands
CoBr2 and LS and LR respectively in freshly distilled THF (Scheme 2).
shift gradient-encoded selective refocusing (PS-GSERF) and standard
NOESY experiments towards well-defined neutral ZnII complexes. This
methodology has allowed clear identification and characterization of
preferred geometries at an atomic resolution. Based on the Zn com-
plexes model, we also describe the preparation of unprecedented Co
complexes from the same ligand family. Indeed, Co exhibits a com-
parable coordination sphere to Zn, is a cheap metal that have found
ongoing developments in catalysis for example [48–52]. In this context,
gaining a first set of structural information is undoubtedly of broad
interest. The geometry of Co complexes is analyzed from solid-state
data and finally compared to their Zn analogues.
2.2. NMR as a probe to decipher the complex geometries
During the complexation process several complexes might be ob-
tained arising from two possible configurations of the nitrogen neo
stereocenter and the two relative positions of the pendant nitrogen arm
(equatorial or axial) of the metallacycle. The J-resolved technique,
which is well-known to extract JH,H couplings as a source of structural
information, such as conformational preferences, has been applied to
complex 1 (LSZnBr2). However, this method was limited since all JH,H
evolve at the same time rendering spectra difficult to interpret. Among
NMR tools, the GSERF sequence is recognized as an original tool for
simplifying the analysis of JH,H couplings [56]. Thus, different 2D
GSERF maps have been recorded for 1 as well as some pure-shift ver-
sions allowing an enhancement of the spectral resolution along the
direct dimension [58,59].
2. Results and discussion
2.1. Complex synthesis
Ligands LS and LR were first prepared according to classical tech-
niques [20,53–55], starting from commercially available pyridin-2-
carboxaldehyde and the corresponding benzylic amine in the presence
of MgSO4. The imine intermediate which is quantitatively formed after
2 h stirring at room temperature is reduced with NaBH4 in MeOH af-
fording the target pma ligands in quantitative yields (Scheme 1).
Complexation of ligands LS and LR to ZnBr2 was next examined.
Both ligands were independently reacted with anhydrous ZnBr2 in
distilled THF. The clear reaction mixture was stirred for 16 h at room
temperature. In contrast to other pma-metal complexes which sponta-
neously precipitated [20], complexes 1 and 2 remained soluble and
their isolation required solvent evaporation and several washes with
Et2O. Both complexes were isolated as white solids in 58 and 79% yield.
Both 1H and 13C NMR spectra of complexes 1 and 2 display identical
characteristic chemical shifts. As confirmed by solid state data (vide
infra), LS and LR undergo a complexation process leading to two
By selecting different nuclei couple among which the nitrogen
proton (H9), the pseudo benzylic methylene protons (H6a axial and H6e
equatorial), and the pendant arm proton (H7) as shown in Fig. 2, it was
easy to extract JH9,H6 and JH9,H7 respectively (see Fig. 3). The latter JH,H
are key data directly linked to the complex geometry through the well-
3
known Karplus curve. Thus, JH9,H7 = 12.5 Hz corresponds to a dihe-
dral angle close to 180° (see Fig. 3 and DFT calculations vide infra). It
3
has been more difficult to accurately measure JH9,H6 with GSERF ex-
periment (see Fig. 3C) due to the strong coupling between the pseudo
benzylic protons H6. In fact, H6e/H6a remains a second order system
whatever the temperature from 193 to 323 K. To push over such a limit,
the pure-shift GSERF allows easily measuring an averaged coupling for
these two benzylic 1H: JH9,H6 = 7.5 Hz thanks to the vanishing of JH,H
leading to improved spectral resolution by reducing the linewidth by 3
(Fig. 3D). The latter corresponds to a dihedral angle H9eNeCeH6a
*
H6e
H6a
H2N
1)
N
N
H7
HN
DCM, MgSO4
*
Zn
8
N
H
C
N
3
2) NaBH4 ,MeOH, 2 h rt
O
H10
H9
H10
LS
LR
quant.
quant.
Fig. 2. Numbering system adopted for protons of interest according to solid and
liquid state data (vide infra).
Scheme 1. Preparation of ligands LS and LR.
2

B. Large, et al.
InorganicaChimicaActa498(2019)119070
Fig. 3. GSERF spectra (A, C) and pure-shift GSERF (B,D) on complex 1 at 193 K (see ESI). [57] In spectra A and B, H7 is excited and H7 + H9 are selected for
measuring JH9,H7; In spectra C and D, H6 are excited and H6 + H9 are selected for measuring JH9,H6
.
close to 180° and therefore close to 90° for H9eNeCeH6e. These results
confirm the most stable conformation of the metallacycle envelop
leading to the equatorial position of the pendant arm nitrogen sub-
stituent.
To better understand the impact of different parameters on the
overall geometry of the zinc complex, such as the configuration of the
nitrogen stereocenter, the equatorial or axial position of the pendant
arm nitrogen substituent and its spatial arrangement relative to H9, DFT
calculations were carried out.
computed by mean of an iterative change of the H9eNeCeH7 dihedral
angle by 10° (for full details see ESI, Fig. S7). The less constrained
conformer displays an anti-conformation of H7 and H9 hydrogen atoms
characterized by a dihedral angle of 174° as shown in Fig. 5. All other
potential conformations display higher energies ranging from 13.2 to
43.1 kJ.mol−1
.
This computed analysis confirms the selective complexation of li-
gand LS and the formation of complex 1 (LSZnBr2) as the preferred
geometry which architecture is set by the combination of three key
features including (i) the S configuration of the sp3 nitrogen atom, (ii)
the equatorial position of the benzyl side arm and (iii) the preferred
conformation of the benzylic substituent. These results are further
supported by the solid-state structure analysis (vide infra).
In order to rationalize DFT structures in comparison to NMR data,
interatomic short and long distances as well as JH,H couplings have been
considered.
2.3. DFT calculations
The structures of ligand LS and its corresponding complexes 1 have
been optimized with the B3LYP/GEN 6-311G(d,p) (H,C,N) LANL2DZ
(Zn,Br)/DFT level of theory (full details are given in the ESI) [60]. Four
possible combinations labelled 1eqNS (equatorial position of the pen-
dant arm and S configuration of the nitrogen atom), 1eqNR, 1axNS and
1axNR may arise from complexation. As shown in Fig. 4 1eqNS is the
preferred spatial arrangement of the pendant arm nitrogen substituent
over the corresponding 1axNS, 1eqNR and 1axNR structures by 1.1, 5.9
and 10.2 kJ/mol. Interestingly, inversion of the nitrogen configuration
from S to R led to a clear discrimination of both potential equatorial
complexes. 1eqNS is favoured by 10.2 kJ.mol−1 over the 1eqNR isomer.
In accordance with NMR data, computed data imply that coordination
of LS affords 1eqNS as the preferred isomer of four possible combina-
tions.
In addition, as shown by the aforementioned NMR study, the con-
formation of the pendant nitrogen arm relative to H9 is of crucial im-
portance and has also to be considered. Thus, we next examined the
preferred spatial arrangement of the benzylic side arm with respect to
the metallacycle in the most stable 1eqNS isomer. To this end, possible
conformations of the equatorial benzylic substituent have been
2.4. Comparison of NMR and DFT data
The calculation of JH,H for 1 has to take into account the structures
of lowest energy. If 1eqNS is the preferred geometry, 1eqNS and 1axNS
exhibit close computed energies and thus both have to be taken into
consideration. As the input or calculation of the weight of each struc-
ture is given by the Boltzmann distribution, other analogues arising
from 1eqNR and 1axNR might also contribute if their computed energies
are comparable with those of 1eqNS and 1axNS.
We thus compute all possible conformations for 1eqNR, 1axNR,
1eqNS and 1axNS by mean of an iterative incrementation of the cor-
responding H9eNeCeH7 dihedral angle by 10° in order to identify the
conformers of lowest energy for all combinations (for full details see
ESI) [60].
Among all lowest energy conformers for 1eqNR, 1axNR, 1eqNS and
3

B. Large, et al.
InorganicaChimicaActa498(2019)119070
Fig. 4. Comparison of computed potential geometries for complex 1 (LSZnBr2) with relative energies as compared to the most stable complex.
NSCS diastereoisomer is more stable than the NRCS one as energetically
quantified by DFT (compare 1eqNS(174) to 1axNR(59), Table 2). This
joint experimental (NMR) and theoretical studies shows that JH,H cou-
pling constants and dH,H are good structural markers to determine the
most stable conformers as well as the preferred diastereoisomer formed
upon complexation of pma to ZnII cation.
Fig. 5. Computed preferred conformation of the equatorial benzylic substituent
for complex 1.
2.5. Solid-state study
1axNS, the five lowest ones which exhibit a ΔE ≤ 10.2 kJ/mol relative
to 1eqNS have been considered. Their 3D structure, corresponding en-
ergies and dihedral angles between H9-H7 are gathered in Table 1, to-
gether with the corresponding Boltzmann probability and calculated
JH,H thanks to the DFT/Gauge Invariant Atomic Orbitals (GIAO) ap-
proach (ESI).
X-ray diffraction structure was used to confirm that the pre-
dominant structure of the zinc complex determined in solution does
correspond to the one collected from the solid-state data. To do so, we
were able to obtain single crystals suitable for X-ray analysis for each
enantiomeric complex by slow diffusion of pentane into concentrated
THF solution. Ortep plots of ZnBr2 complexes are shown in Fig. 7 and
selected bond lengths and angles that account for the 3D structure of
complexes 1 and 2 are gathered in Table 2. Both complexes are iso-
morphous with very similar cell parameters.
They adopt a tetrahedral geometry due with no substantial differ-
ence in bond lengths and angles between 1 and 2 (Table 3).
These structures are very similar to that published for LSZnCl2 [61]
the principal difference being the longer Zn-Br bonds (2.34 Å) com-
pared to the Zn-Cl ones (2.20 Å on average) which is explained by the
size of the halides. Otherwise bond lengths and angles are very similar
for the three structures (Table 3). It is noteworthy that for both 1 and 2
complexes, the benzylic fragment lies in equatorial position, as antici-
pated from the study in solution. In the same manner, the anti-con-
formation of the H7 and H9 hydrogen atoms was confirmed by this
solid-state analysis (Fig. 8).
Moreover, for these complexes the unit cell contains four molecules
(space group P 212121), and, as for the chloride complex, the only ob-
served short contacts take place between hydrogens and bromines (see
Fig. S13). Our further objective was to gain informations on the overall
3D structure of Co complexes. We were especially interested in de-
termining if the observed preferred diasteroisomer formation and
conformation evidenced for zinc complexes would be the same with
another metal.
The JH,H value between H6 and H9 in the 1eqNS conformer was
calculated as a weighted average over 4 subsequent major contributors
(see Tables 1 and S1). This value of 6.4 Hz is close to the experimental
one of 7.5 Hz. In the same manner, JH7,H9 has been calculated at
11.2 Hz close to the 12.5 Hz value found experimentally. Same calcu-
lations for 1eqNR and 1axNR considering the most stable conformer
gave an averaged value JH6,H9 = 6.8 Hz, close to the experimental one.
However, the calculated JH7,H9 = 4.0 Hz is far from the measured value,
confirming that this diastereoisomer is not formed in solution. A sig-
nificant contribution of other conformers of higher energy remains
unlikely.
Fig. 6 shows the variation between calculated and experimental JH,H
(JH6,H9 and JH7,H9) coupling constants for each conformer and evi-
dences that 1eqNS(174) and 1axNS(170) are conformers with the
highest contribution, due to their smallest calculated-versus-experi-
mental differences.
The same approach was conducted for interatomic hydrogen dis-
tances (dH,H) in complex 1: calculated values for the 1eqNS(174),
1axNS(170), 1axNS(79) and 1eqNR(173) structures have been com-
pared to those obtained from NOESY experiments (Table 2 and ESI).
Only the most relevant dH,H between flexible moieties were used to
differentiate conformers and diastereoisomers: H7-H9 and H6-H10 allow
distinguishing the most stable conformers: 1eqNS(174) and
1axNS(170), since their |d(DFT)-d(NMR)| lead to smaller values
(Table 2).
The structure of cobalt complexes 3 and 4, for which NMR studies
would be more difficult to conduct due to the paramagnetic relaxation
effect of CoII, were investigated in the solid-state (Fig. 9).
For those pma-cobalt complexes no previous X-ray structure was
reported. The structure of 3 and 4 resemble those of the Zn compounds
and are very similar to each other in terms of metrics and angles
(Table 4). They crystallize in a non centrosymmetric chiral orthorombic
This approach has also been used to determine the most stable
diastereoisomer between NSCS and NRCS. Again, the calculated/ex-
perimental differences |d(DFT)-d(NMR)| allow confirming that the
4

B. Large, et al.
InorganicaChimicaActa498(2019)119070
Table 1
Probabilities and calculated JH,H for lowest energy conformers of complex 1.
a
Structure
Conformer (dihedral angle H9-H7)
Energy relative to 1eqNS
Probability (%)
64.8
JH6e,H9 (Hz)
3.80
JH6a,H9 (Hz)
11.78
JH7-H9
11.05
JH6,H9
6.38b
JH7,H9
av
1eqNS(174)
0
11.18b
1axNS(170)
+1.1
32.7
6.69
0.51
12.05
1axNS (79)
+7.1
0.8
8.47
0.36
1.04
1eqNR(173)
+10.2
0.1
5.67
11.65
4.7
1axNR(59)
+5.9
1.6
13.20
0.37
4.02
6.78
4.02
a
Averaged coupling: (JH6e,H9 + JH6a,H9)/2.
Conformationaly weighted calculated JH,H according to probabilities.
b
space group (P212121), with cell parameters close to those of complexes
1 and 2. As observed for the zinc complexes the conformation of the
metallacycle changes with the stereochemistry of the methylene
carbon. Noteworthy the torsion angle is in this case a little bit larger in
absolute value for 4 bearing the LR ligand than for 3. This induces a
small widening of the N-Co-Br angle. As for 1 and 2, the cell unit
contains 4 molecules (space group P 212121) with most short contacts
being observed between bromine and hydrogens (Fig. 10).
zinc complex 1, structures are close but not similar with some differ-
ences (RMS = 3.3 Å). These weak discrepancies are probably due to the
effect of the solvent on the most stable conformer (in yellow Fig. 11A
and Table 1) as compared to the crystallographic structure (in purple
Fig. 11A). This effect is also seen for the Co complex 3 showing a fit
between structures of 3.4 Å (Fig. 11C).
NMR data on complex 3 were difficult to interpret despite DFT/
GIAO derived NMR data (not shown) due to the paramagnetic relaxa-
tion induced by the CoII metallic center. Alternatively, complex 3 could
be crystallized and X-ray derived structures of complex 3 (in light blue
Fig. 11B) are almost surprisingly similar to complex 1 (in purple
Fig. 11B) as showed by the RMS = 1.66A°. On the contrary, the DFT
derived structures between Zn and Co are close but with small con-
formational differences especially in the Br-Metal-Br angle and also for
the CH(Me)Ph lateral chain position (Fig. 11D). A general tendency put
into relief by the previous comparative study is the very close five
membered ring metallacycle conformation with an equatorial
The geometry of Zn and Co complexes markedly differs from the PdII
analogue due to the propensity of Zn and Co to form tetrahedral ar-
rangements instead of square planar for the Pd counterpart. The cor-
responding LRPdCl2 as well as all pyridylmethylamine-based Pd com-
plexes within this series displays
a
topology set by
-interaction between the pyridine and phenyl ring of the ligand (see
ESI) [20].
In order to validate our DFT/NMR/X-ray approach, DFT/NMR and
X-ray derived structures have been compared (see Fig. 11A). For the
a strong
π
5

B. Large, et al.
InorganicaChimicaActa498(2019)119070
Fig. 6. 1H-1H scalar coupling constant differences [ΔJH,H DFT-NMR] between calculated (DFT) and experimental (NMR) approaches, respectively by GIAO single
point calculations and GSERF experiments for the five most stable structures of complex 1.
Table 2
Table 3
Absolute difference between DFT (QM) calculated and NMR derived distances
(dH,H) by using NOESY experiments for the four most stable contributors of
complex 1 and for both possible diastereomers in the most stable conformation
labelled 1eqNS(174) and 1axNR(59).
Selection of bond lengths (Å) and angles (°) for complexes 1, 2 and already
reported LSZnCl2 for comparison.
1 (LSZnBr2)
2 (LRZnBr2)
LSZnCl2
61
Zn1-N1
Zn1-N2
Zn1-X1
Zn1-X2
N1-C5
C5-C6
C6-N2
N1-Zn1-N2
N1-Zn1-X2
Br1-Zn1-X2
N2-Zn1-X1
Zn1-N1-C6-N2
H9n-N2-C7-H7
2.037(5)
2.095(5)
2.344(1)
2.351(1)
1.338(7)
1.500(8)
1.482(8)
83.7(2)
113.6(2)
118.01(3)
109.1(2)
12.30
2.029(8)
2.096(8)
2.354(2)
2.344(1)
1.34(1)
1.49(1)
1.49(1)
83.6(3)
109.8(2)
118.00(6)
117.7(2)
−12.85
177.92
2.037(3)
2.089(3)
2.2025(12)
2.2134(11)
1.332(5)
1.511(5)
1.479(4)
83.58(12)
113.45(11)
117.13(4)
109.70(9)
13.09
1eqNR(173) 1eqNS(174) 1axNS(79) 1axNS(170) 1axNR(59)
H8-H7
0.0105
0.0240
0.0136
0.3249
0.4053
0.6396
0
0.0172
0.0110
0.0286
0.3399
0.2423
0.9045
1.8637
0.5909
0.0038
0.0110
0.0136
0.3209
0.3447
0.8995
0.6971
0.1009
0.0132
0.0494
0.0253
0.3324
0.3253
0.6406
0.7640
0.7093
H6a-H6e
H7-H10
H10-H8
H7-H6av
H9-H6av
0.0209
0.0136
0.3282
0.1753
0.6279
0.6796
0.1118
H6av-H10 1.7976
H7-H9 0.6209
−172.57
−175.74
substituent whatever the state of matter. Moreover, the lateral chain CH
(Me)Ph show some slight conformational changes depending on the
considered metal and the state of matter.
performed on silica gel 60 F254 and the plates were visualized with UV
light (254 nm) or a potassium permanganate solution (1 g with 2 g of
K2CO3 in 200 mL of water). The crude products were purified by pre-
parative thin layer chromatography on silica gel 60 PF254 or by column
chromatography using silica gel Merck 60. Known compound structures
were assigned by comparison with the literature spectroscopic data.
3. Experimental section
3.1. General procedures
3.1.1. Synthesis
All non-aqueous reactions were carried out under an atmosphere of
argon in flame- or oven-dried glassware with magnetic stirring. All
reagents and solvents obtained from commercial sources were used
without further purification unless otherwise noted. CH2Cl2 was dis-
tilled over CaH2 before use. Anhydrous THF and Et2O were used after
distillation over sodium metal. Thin layer chromatography (TLC) was
3.1.2. NMR
Routine 1H and 13C NMR spectra were recorded on a Bruker AM
360, AM 300, DPX 200 and DPX 250 spectrometers at room tempera-
ture; Chemical shifts for 1H and 13C were referenced internally ac-
cording to the residual solvent resonances and reported in ppm relative
Fig. 7. Ortep plots of 1 (LSZnBr2, left) and 2 (LRZnBr2, right) with 50% probability thermal ellipsoids. Most Hydrogen atoms were omitted for clarity.
6

B. Large, et al.
InorganicaChimicaActa498(2019)119070
Fig. 8. Side-view of 1 (LSZnBr2, left) and 2 (LRZnBr2, right) using 50% probability thermal ellipsoids. Most Hydrogen atoms were omitted for clarity as well as
bromines and five carbons of the phenyl ring.
to CDCl3 or (CD3)2CO. All coupling constants (J values) are given in
hertz (Hz). Data appear in the following order: chemical shift in ppm,
multiplicity (s, singlet; d, doublet; t, triplet; q, quadruplet; m, multi-
plet), coupling constant J, number of protons, and assignment.
Advanced NMR spectra were performed at 9.4 T on a Bruker DRX
400 spectrometer using a 1H/13C/X Triple Broad Band Inverse probe
equipped with a z field gradient coil and a standard variable-tem-
perature unit (BVT 3000). All experiments carried out at low tem-
perature (193 K), were prepared using (CD3)2CO as solvent. All 3D
GSERF spectra were obtained by recording 4096 × 32 × 24 matrices
converted by the pure shift macro [62,63] to 2D 4096 × 32 points in
the F1 dimension. No apodization was applied in each dimension prior
the double Fourier transform. Phased 2D maps were obtained using the
Quadrature Sequential Mode.
For the zinc complex, a 40 ms (400 Hz) Burp pulses (EBURP-2 for
excitation, REBURP for refocusing and a time reversal EBURP-2 for flip
back) have been used in the SERF block for selecting 1H areas. In the
band selective 1H–1H decoupled spectra a RSNOB [64] selective pulse
of 40 ms duration and bandwidth of 65 Hz. The homodecoupled spectra
were acquired with number of t2 increments (i.e., number of FID
chunks) equal to 32, the duration of FID chunk is 19.2 ms, the number
of complex data points of constructed FID in 1H dimension is 4096, the
recycling delay is 1s, and the number of scans is 4.
Table 4
Selection of bond lengths (Å) and angles (°) for complexes 3, 4.
3 (LSCoBr2)
4 (LRCoBr2)
Co1-N1
Co1-N2
Co1-Br1
Co1-Br2
N1-C5
C5-C6
C6-N2
N1-Co1-N2
N1-Co1-Br2
Br1-Co1-Br2
N2-Co1-Br1
Co1-N1-C6-N2
H9n-N2-C7-H7
2.017(3)
2.080(3)
2.3555(7)
2.3744(7)
1.340(5)
1.511(5)
1.485(5)
84.4(1)
110.6(1)
118.01(3)
109.1(2)
12.78
2.03(1)
2.071(8)
2.361(2)
2.369(2)
1.32(1)
1.52(1)
1.47(1)
83.4(3)
116.1(3)
116.17
112.2(2)
−19.92
−173.94
174.4
Windows or Mercury.
3.2. Experimental procedures
3.2.1. Reductive amination
To a stirred solution of amine (1 eq) and MgSO4 (2 eq), di-
chloromethane or THF (1 mL per 0.1 mmol) was added the carbox-
aldehyde derivative (1 eq). The mixture is stirred at room temperature
for 2 h. After filtration and concentrated under vacuum, MeOH (1 mL
per 0.1 mmol) and NaBH4 (3 eq) were added. The reaction mixture was
stirred at room temperature for 2 h. The solution was then concentrated
under vacuum and 10 mL of saturated aqueous solution of NH4Cl was
added, the aqueous phase was extracted with EtOAc (3 × 10 mL). The
combined layers were dried (MgSO4), filtered and concentrated under
vacuum. The crude product was purified by flash chromatography on
silica gel.
1H–1H exchange spectroscopy (NOESY) is routinely used to access
dynamic processes within a particular time‐scale window. Series of 2D
spectra are acquired with different mixing times tmix to generate
build‐up curves associated with NOEs that produce variations of diag-
onal‐to‐cross‐peak volume ratios. Buildup curves comprising signal in-
tensity dependence of the cross- and diagonal-peaks with the mixing
time (tmix) was fitted according to the calculation detailed in the
Supporting Information [46].
3.1.3. X-ray
3.2.2. Complexation to MBr2
X-ray crystallography data were collected at 150 K on a Bruker
Kappa APEX II diffractometer using a Mo-κ (λ = 0.71069 Å) X-ray
source and a graphite monochromator. The crystal structures were
solved using SIR 97 [64] or Shelxt [65] and refined using Shelxl-97 [66]
or Shelxl-2013. ORTEP drawings were made using ORTEP III [67] for
MBr2 (0.70 mmol) is added to a solution of LS or LR (155 mg,
0.73 mmol) in THF (4 mL). The solution is stirred for 24 h at room
temperature. The reaction mixture is concentrated under vaccum. The
solid residue is next suspended in Et2O (10 mL) and filtrated. The solid
Fig. 9. Ortep plots of 3 (LSCoBr2, left) and 4 (LRCoBr2, right) 50% probability thermal ellipsoids. Most Hydrogen atoms were omitted for clarity.
7

B. Large, et al.
InorganicaChimicaActa498(2019)119070
Fig. 10. Side-view of
3 4
(LSCoBr2, left) and
(LRCoBr2, right) using 50% probability thermal el-
lipsoids. Most Hydrogen atoms were omitted for
clarity as well as bromines and five carbons of the
phenyl ring.
is washed three times with Et2O (10 mL) and dried under vacuum to
yield the expected complex 1, 2, 3 or 4.
3.2.5. Complex 1 (yield: 58% white solid)/complex 2 (yield: 79% white
solid)
NMR 1H: (400 MHz, (CD3)2CO), δ(ppm), J(Hz): 1.62 (d, J = 6.1,
3H); 3.85 (m, 2H); 4.17 (q, J = 6, 1H); 5.25 (m, 1H); 7.43 (t, J = 7,
1H); 7.51 (t, J = 7.3, 2H); 7.63 (d, J = 7.3, 2H); 7.72 (d, J = 7.9, 1H);
7.78 (t, J = 6, 1H); 8.22 (t, J = 7.4, 1H); 8.63 (d, J = 4.8, 1H). NMR
13C: (100 MHz, (CD3)2CO), δ(ppm): 24.21; 51.69; 60.34; 124.76;
125.87; 127.98; 129.00; 130.89; 141.95; 147.98; 156.82. HRMS: cal-
culated for C14H16N2Br2ZnNa [M+Na+] = 456.8869; found:
456.8875.
3.2.3. N-[(S)-1-phenylethyl]-N-[1-(2-pyridinyl)methyl]amine LS
The crude product was purified by flash chromatography on silica
gel with pentane/AcOEt (75/25) as the eluent. Yield: quant. Yellow oil.
NMR 1H: (400 MHz, (CD3)2CO), δ(ppm), J(Hz): 1.31 (d, J = 7.5,
3H); 3.61 (s, 2H); 3.77 (q, J = 6, 1H); 7.26 (m, 2H); 7.35 (t, J = 7, 2H);
7.44 (m, J = 7.3, 3H); 7.76 (td , J = 7.7, 1.7, 1H); 8.5 (d, J = 4, 1H).
NMR 1H: (200 MHz, CDCl3), δ(ppm), J(Hz): 1.42 (d, J = 6.5, 3H); 2.65
(brs, 1H); 3.75(s, 2H); 3.83 (q, J = 6.5, 1H); 7.09–7.41 (m, 7H); 7.59
(td, J = 7.6, 1.7, 1H); 8.55 (d, J = 4.7, 1H). NMR 13C : (75 MHz,
CDCl3), δ(ppm): 23.83; 52.45; 57.42; 121.23; 121.82; 126.18; 126.57;
127.87; 135.71; 144.77; 148.67; 159.20. HRMS: calculated for
C14H17N2 [M+H+] = 213.1392; found: 213.1388. Data in ac-
cordance with literature [68].
3.2.6. Complex 3 (yield: 70% blue solid)/complex 4 (yield: 77% blue
solid)
HRMS: calculated for C14H16N2BrCo [M Br] = 349.9823; found:
349.9824.
4. Conclusions
This solid-state study of Zn- and Co-complexes evidenced some key
features regarding the conformation of the metallacycle, the config-
uration of the sp3 nitrogen atom set during the complexation process,
the pseudo equatorial position of the benzyl side arm as well as the
preferred spatial arrangement of the chiral side arm with respect to the
metallacycle. It perfectly fits with the structures of 1 and 2 evidenced in
solution thanks to an NMR study complemented for 1 by DFT calcula-
tions. The combination of various structural elucidation methods like
DFT/NMR in solution and X-ray in solid state is the key to better un-
derstand complex stability regarding its potential reactivity. Comparing
different sources of structural data is the only one way to make a given
approach more robust for investigating unknown complexations. In
3.2.4. N-[(R)-1-phenylethyl]-N-[1-(2-pyridinyl)methyl]amine LR
The crude product was purified by flash chromatography on silica
gel with pentane/AcOEt (75/25) as the eluent. Yield: quant. Yellow oil.
NMR 1H: (200 MHz, CDCl3), δ(ppm), J(Hz): 1.36 (d, J = 6.5, 3H);
2.35 (brs, 1H); 3.7 (s, 2H); 3.78 (q, J = 6.6, 1H); 6.99–7.36 (m, 7H);
7.47 (td, J = 7.6, 1.7, 1H); 8.49 (d, J = 4.5, 1H). NMR 13C: (75 MHz,
CDCl3), δ(ppm): 23.77; 52.37; 57.33; 121.13; 121.70; 126.10; 126.28;
127.78; 135.59; 144.69; 148.57; 159.10. HRMS: calculated for
C14H17N2 [M+H+] = 213.1392; found: 213.1399. Data in ac-
cordance with literature [55].
Fig. 11. Overlay of A) X-ray (purple)/DFT (yellow) for complex 1, B) of X-ray data for complexes 1 (purple) and 3 (light blue), C) X-ray (light blue) and DFT (orange)
data of complex 3 and D) DFT data comparison between complex 1 (yellow) and complex 3 (orange). RMS of different overlays are indicated as a comparison marker.
(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
8

B. Large, et al.
InorganicaChimicaActa498(2019)119070
particular, for Zn-pma complex, NMR/DFT approach remains a great
alternative for deciphering the ligand folding around the metal when
the resulting complex could not diffract and/or be crystallized.
(2002) 389–404.
[26] F.A. Scaramuzzo, E. Badetti, G. Licini, C. Zonta, Eur. J. Org. Chem. (2017)
1438–1442.
[27] L.A. Joyce, M.S. Maynor, J.M. Dragna, G.M. da Cruz, V.M. Lynch, J.W. Canary,
E.V. Anslyn, J. Am. Chem. Soc. 133 (2011) 13746–13752.
[28] S. Nayab, H. Lee, J.H. Jeong, Polyhedron 43 (2012) 55–62.
[29] C.K. Williams, L.E. Breyfogle, S.K. Choi, W. Nam, V.G. Young Jr., M.A. Hillmyer,
W.B. Tolman, J. Am. Chem. Soc. 125 (2003) 11350–11359.
[30] M.P. Cuajungco, G.J. Lees, R.R. Kydd, R.E. Tanzi, A.I. Bush, Nutr. Neurosci. 2
(1999) 191–208.
Acknowledgements
The authors acknowledge the MENRS, Centre National de la
Recherche Scientifique – France, Ecole Polytechnique, Universities of
Versailles Saint‐Quentin‐en‐Yvelines, Paris–Saclay, as part of the
“Investissements d′Avenir” program (ANR‐11‐IDEX‐0003‐02), and
CHARMMMAT labex (ANR‐11‐LABEX‐0039) for funding to J.E.H.P. J.F.
thanks his partner Sandrine Bouchet for unfailing assistance. Louis
Ricard and Sophie Bourcier are deeply thanked for discussion on X-ray
structures and HRMS analysis.
[31] S.W. Suh, K.B. Jensen, M.S. Jensen, D.S. Silva, P.J. Kesslak, G. Danscher,
C.J. Frederickson, Brain Res. 852 (2000) 274–278.
[32] X. Huang, M.P. Cuajungco, C.S. Atwood, R.D. Moir, R.E. Tanzi, A.I. Bush, J. Nutr.
130 (2000) 1488S–1492S.
[33] E. Andrasi, E. Farkas, D. Gawlik, U. Rosick, P. Bratter, J. Alzheimer’s Dis. 2 (2000)
17–26.
[34] D.W. Choi, J.Y. Koh, Annu. Rev. Neurosci. 21 (1998) 347–375.
[35] L.M.T. Canzoniero, D.M. Turetsky, D.W. Choi, J. Neurosci. 19 (1999) RC31.
[36] J.H. Weiss, S.L. Sensi, J.Y. Koh, Trends Pharm. Sci. 21 (2000) 395–401.
[37] Y.-K. Song, S.K. Kang, Y.-I. Kim, Bull. Korean Chem. Soc. 36 (2015) 374–377.
[38] M. Royzen, J.W. Canary, Polyhedron 58 (2013) 85–91.
[39] Z. Dai, X. Xu, J.W. Canary, Chem. Commun. (2002) 1414–1415.
[40] Z. Dai, J.W. Canary, New J. Chem. 31 (2007) 1708–1718.
[41] J. Liang, J.W. Canary, Chirality 23 (2011) 24–33.
[42] Y.-H. Chiu, G.J. Gabriel, J.W. Canary, Inorg. Chem. 44 (2005) 40–44.
[43] B.L. Vallee, D.S. Auld, Acc. Chem. Res. 26 (1993) 543–551.
[44] Y. Yang, Y. Hayashi, Y. Fujii, T. Nagano, Y. Kita, T. Ohshima, J. Okuda, K. Mashima,
Catal. Sci. Technol. 2 (2012) 509–513.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.ica.2019.119070.
References
[45] S.-S. Wu, X.-W. Zhang, W.-L. Dai, S.-F. Yin, W.-S. Li, W.-Q. Ren, C.-T. Au, Appl.
Catal. A 341 (2008) 106–111.
[46] J.E.H. Pucheta, J.-M. Gillet, D. Prim, J. Farjon, ChemPhysChem 17 (2016)
[1] M. Kruppa, B. König, Chem. Rev. 106 (2006) 3520–3560.
[2] D. Chandra, T. Ohji, K. Kato, T. Kimura, Langmuir 27 (2011) 11436–11443.
[3] A. Congreve, R. Kataky, M. Knell, D. Parker, H. Puschmann, K. Senanayake,
L. Wylie, New J. Chem. 27 (2003) 98–106.
1034–1045.
[47] M. Westerhausen, T. Bollwein, K. Polborn, Z. Naturforsch. 55b (2000) 51–59.
[48] S.M. Ujwaldev, N.A. Harry, M.A. Divakar, G. Anilkumar, Catal. Sci. Technol. 8
(2018) 5983–6018.
[49] A. Mukherjee, D. Milstein, ACS Catal. 8 (2018) 11435–11469.
[50] M. Moselage, J. Li, L. Ackermann, ACS Catal. 6 (2016) 498–525.
[51] C.E.I. Knappke, S. Grupe, D. Gärtner, M. Corpet, C. Gosmini, A.J. von Wangelin,
Chem. Eur. J. 20 (2014) 6828–6842.
[52] G. Cahiez, A. Moyeux, Chem. Rev. 110 (2010) 1435–1462.
[53] A. Requet, H. Yalgin, D. Prim, Tetrahedron Lett. 56 (2015) 1378–1382.
[54] V. Terrasson, D. Prim, J. Marrot, Eur. J. Inorg. Chem. (2008) 2739–2745.
[55] G. Grach, A. Dinut, S. Marque, J. Marrot, R. Gil, D. Prim, Org. Biomol. Chem. 9
(2011) 497–503.
[4] C. Hemmert, M. Renz, H. Gornitzka, B. Meunier, J. Chem. Soc., Dalton Trans.
(1999) 3989–3994.
[5] J.M. Rowland, M.M. Olmstead, P.K. Mascharak, Inorg. Chem. 39 (2000)
5326–5332.
[6] J. Chang, S. Plummer, E.S.F. Berman, D. Striplin, D. Blauch, Inorg. Chem. 43 (2004)
1735–1742.
[7] M. Van den Heuvel, T.A. Van den Berg, R.M. Kellogg, C.T. Choma, B.L. Feringa, J.
Org. Chem. 69 (2004) 250–262.
[8] T. Wang, X.-Q. Hao, X.-X. Zhang, J.-F. Gong, M.-P. Song, Dalton Trans. 40 (2011)
8964–8976.
[9] C. Hemmert, M. Renz, H. Gornitzka, S. Soulet, B. Meunier, Chem. Eur. J. 5 (1999)
1766–1774.
[56] N. Giraud, L. Beguin, J. Courtieu, D. Merlet, Angew. Chem. Int. Ed. 49 (2010)
[10] W. Baratta, G. Chelucci, S. Gladiali, K. Siega, M. Toniutti, M. Zanette, E. Zangrando,
3481–3484.
P. Rigo, Angew. Chem. Int. Ed. 44 (2005) 6214–6219.
[11] S. Marque, V. Razafimahaléo, A. Dinut, G. Grach, D. Prim, X. Moreau, R. Gil, New J.
Chem. 37 (2013) 2683–2690.
[57] 193K was found optimum temperature in order to investigate spatial arrangement.
At room temperature, signal broadening prevent the probing of spatial arrange-
ment, but is acceptable for structural elucidation.
[12] K. Suzuki, P.D. Oldenburg, L. Que Jr., Angew. Chem. Int. Ed. 47 (2008) 1887–1889.
[13] L. Zhang, Z. Zuo, X. Wan, Z. Huang, J. Am. Chem. Soc. 136 (2014) 15501–15504.
[14] P. Nareddy, L. Mantilli, L. Guénée, C. Mazet, Angew. Chem. Int. Ed. 51 (2012)
3826–3831.
[58] J.E.H. Pucheta, M. Candy, O. Colin, A. Requet, F. Bourdreux, E. Galmiche-Loire,
A. Gaucher, C. Thomassigny, D. Prim, M. Mahfoudh, E. Leclerc, J.-M. Campagne,
J. Farjon, PCCP 17 (2015) 8740–8749.
[59] J.A. Aguilar, S. Faulkner, M. Nilsson, G.A. Morris, Angew. Chem. Int. Ed. 49 (2010)
[15] J. Chen, T. Xi, Z. Lu, Org. Lett. 16 (2014) 6452–6455.
[16] C. Nájera, J. Gil-Moltó, S. Karlström, L.R. Falvello, Org. Lett. 5 (2003) 1451–1454.
[17] H. Brunner, M. Schimdt, H. Schönenberger, Inorg. Chim. Acta 123 (1986) 201–207.
[18] W.J. Kerr, M. Middleditch, A.J.B. Watson, Synlett (2011) 177–180.
[19] A. Requet, O. Colin, F. Bourdreux, S.M. Salim, S. Marque, C. Thomassigny, C. Greck,
J. Farjon, D. Prim, Magn. Res. Chem. 52 (2014) 273–278.
[20] G. Grach, G. Pieters, A. Dinut, V. Terrasson, R. Medimagh, A. Bridoux,
V. Razafimahaleo, A. Gaucher, S. Marque, J. Marrot, D. Prim, R. Gil, J. Giner Planas,
C. Vinas, I. Thomas, J.-P. Roblin, Y. Troin, Organometallics 30 (2011) 4074–4086.
[21] V. Terrasson, J. Giner Planas, C. Vinas, F. Teixidor, D. Prim, M.E. Light,
M.B. Hursthouse, Organometallics 29 (2010) 4130–4134.
3901–3903.
[60] Similar computed details can be found in the for enantiomeric ligand LR and
complex 2. In this case, calculations point, similarly to the LS/1 couple, towards the
preferential formation of one complex which geometry combines the preferred
equatorial position of the benzyl side arm, the preferred conformation of the
benzylic substituent and the R configuration of the sp3 nitrogen atom.
[61] Q.T. Nguyen, J.H. Jeong, Acta Crystallogr. E 64 (2008) m457.
[62] http://nmr.chemistry.manchester.ac.uk/?q=node/256.
[63] Ē. Kupče, J. Boyd, I.D. Campbell, J. Magn. Reson., Ser B 106 (1995) 300–303.
[64] A. Altomare, M.C. Burla, M. Camalli, G.L. Cascarano, C. Giacovazzo, A. Guagliardi,
A.G.G. Moliterni, G. Polidori, R. Spagna, J. Appl. Crystallogr. 32 (1999) 115–119.
[65] G.M. Sheldrick, Acta Cryst. A 71 (2015) 3–8.
[22] O. Colin, H. Boufroura, C. Thomassigny, S. Perato, A. Gaucher, J. Marrot, D. Prim,
Eur. J. Org. Chem. (2017) 746–752.
[66] G.M. Sheldrick, SHELXL-97, Universität Göttingen, Göttingen, Germany, 1997.
[67] L.J. Farrugia, ORTEP-3 Program, Department of Chemistry, University of Glasgow,
2001.
[23] A. Aillerie, V. Rodriguez-Ruiz, R. Carlino, F. Bourdreux, R. Guillot, S. Bezzenine-
Lafollée, R. Gil, D. Prim, J. Hannedouche, ChemCatChem 8 (2016) 2455–2460.
[24] S. Perato, B. Large, Q. Lu, A. Gaucher, D. Prim, ChemCatChem 9 (2017) 389–392.
[25] M. Westerhausen, T. Bollwein, N. Makropoulos, S. Schneiderbauer, M. Suter,
H. Nöth, P. Mayer, H. Piotrowski, K. Polborn, A. Pfitzner, Eur. J. Inorg. Chem.
[68] W. Goldeman, B. Boduszek, Phosphorus, Sulfur, Silicon Related Elem. 184 (2009)
1413–1425.
9