Zinc and cobalt complexes based on tripodal ligands: Synthesis, structure and reactivity toward lactide

Article abstract of DOI:10.1039/c3dt52629a

The coordination chemistry of a series of pro-ligands ([L ¹]-[L⁶]) with cobalt and zinc derivatives has been studied. All complexes have been characterized by multinuclear NMR, elemental analysis, and by single-crystal X-ray diffraction studies. Polymerization of rac-lactide takes place at 130°C in the presence of cobalt and zinc complexes to yield polymers under solvent free conditions with controlled molecular masses and narrow polydispersities.

Full text of DOI:10.1039/c3dt52629a

View Article Online
View Journal
Dalton
Transactions
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: C. THOMAS, E.
Brulé, M. J. Tschan, J. Guo, T. Roisnel, R. Legay, B. Rigaud, G. Durieux, M. Rager and S. K. Raman, Dalton
Trans., 2014, DOI: 10.1039/C3DT52629A.
This is an Accepted Manuscript, which has been through the
Royal Society of Chemistry peer review process and has been
accepted for publication.
Accepted Manuscripts are published online shortly after
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
Information for Authors.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the Ethical guidelines still
apply. In no event shall the Royal Society of Chemistry be held
responsible for any errors or omissions in this Accepted Manuscript
or any consequences arising from the use of any information it
contains.
www.rsc.org/dalton

Page 1 of 39
Dalton Transactions
View Article Online
DOI: 10.1039/C3DT52629A
For Table of Contents entry only
Zinc and Cobalt Complexes Based on Tripodal Ligands:
Synthesis, Structure and Reactivity toward Lactide
Mathieu J.-L. Tschan, Jia Guo, Sumesh K. Raman, Emilie Brulé, Thierry Roisnel, Marie-
Noelle Rager, Rémi Legay, Guillaume Durieux, Baptiste Rigaud, Christophe M. Thomasa*
Abstract. Polymerization of lactide takes place in the presence of cobalt and zinc chloro
complexes to yield polymers under solvent free conditions.
1

Dalton Transactions
Page 2 of 39
View Article Online
DOI: 10.1039/C3DT52629A
Zinc and Cobalt Complexes Based on Tripodal Ligands:
Synthesis, Structure and Reactivity toward Lactide
Mathieu J.-L. Tschan,a Jia Guo,a Sumesh K. Raman,a Emilie Brulé,a,c Thierry Roisnel,b Marie-
Noelle Rager,d Rémi Legay,e Guillaume Durieux,a Baptiste Rigaud,e Christophe M.
Thomasa,*
a Chimie ParisTech, UMR CNRS 7223, 75005 Paris, France. Fax: +33(0)143260061; Tel:
+33(0)144276721; E-mail: christophe.thomas@chimie-paristech.fr
b Centre de Diffractométrie X, UMR CNRS 6226, Université de Rennes 1, 35042 Rennes,
France
c Université Pierre et Marie Curie, 4 Place Jussieu, 75005 Paris, France
d Chimie ParisTech, NMR Facility, 11 rue Pierre et Marie Curie, 75005 Paris, France
e Université de Caen, ENSICAEN, UMR 6507, 6 Boulevard Maréchal Juin, 14050 Caen
Abstract. The coordination chemistry of a series of pro-ligands ([L1]-[L6]) with cobalt
and zinc derivatives has been studied. All complexes have been characterized by multinuclear
NMR, elemental analysis, and by single-crystal X-ray diffraction studies. Polymerization of
rac-lactide takes place at 130 °C in the presence of cobalt and zinc complexes to yield
polymers under solvent free conditions with controlled molecular masses and narrow
polydispersities.
Keywords. Catalysis, Polylactide, Ring-Opening Polymerization, Zinc, Cobalt, Tripod
1

Page 3 of 39
Dalton Transactions
View Article Online
DOI: 10.1039/C3DT52629A
Introduction
Biodegradable polymers have recently gained great attention as a replacement for
synthetic (petroleum-based) materials.1 Among the new biodegradable polymers that have
been developed during the last decade,2 polylactides (PLAs) are of particular interest.3 The
availability of monomer feedstocks from renewable resources has given PLAs an increasing
prominence in the market place.4 Methods for the synthesis of these polymers now rely on the
controlled ring-opening polymerization (ROP) of lactide by well-defined metal initiators,
which produce high-molecular-weight polymers with interesting mechanical properties.5
The coordination-insertion mechanism of lactide polymerization proceeds via coordination of
the monomer to the metal centre, followed by insertion into the metal-alkoxide species
through the acyl-oxygen bond with retention of configuration. This results in a new metal-
alkoxide species capable of further insertion. In a biomimetic spirit,6 a mechanistic analogy
may be drawn between this process and the hydrolysis reactions performed by a large class of
zinc metalloenzymes.
The ubiquity of zinc in metalloenzyme chemistry has been related to its flexible
coordination chemistry, substitutional lability, Lewis acidity, combined with intermediate
polarizability.7 Much of the importance of zinc enzymes originates from their amidase activity
involving the cleavage of amide bonds.8 For instance, the metallo β-lactamases scaffold is
designed to accommodate one or two zinc ions able to activate a nucleophilic hydroxide for
the hydrolysis of the β-lactam ring. Upon substitution with Co2+, the resulting enzyme is a
more efficient catalyst than the native enzyme for the hydrolysis of some β-lactams
suggesting that the role of the metal centre is principally to provide the nucleophilic
hydroxide, rather than to act as a Lewis acid to polarize the carbonyl group.7b
In our search for new polymerization catalysts, we focused our efforts on investigating the
catalytic activity of mononuclear zinc and cobalt complexes supported by tripod ligands.9 The
2

Dalton Transactions
Page 4 of 39
View Article Online
coordination chemistry of tripod ligands has been intensively studied in recent years. Of
DOI: 10.1039/C3DT52629A
particular interest are pyrazolylborate ligands to which Vahrenkamp and Parkin have
contributed synthetic, model complex, and complex reactivity studies.10 Another important
class of tripod ligands are tris(picolyl)amine derivatives which can easily be obtained with
different donor arms. Depending on the nature and number of donor groups, such compounds
can accommodate a broad range of coordination geometries as multidentate ligands for
various metal centres. Some of the resulting complexes can be used as efficient catalysts in a
variety of processes. Early reports on tripodal ligands were focused on the coordination
properties of monoanionic phenoxy based ligands of the type [L2NO]– but there is no example
of multidentate ligand with soft phosphorus donor atoms.11 Therefore we decided to
investigate the coordination of tripodal containing [L2NO] and [L3N] donor environments and
the catalytic behavior of the corresponding zinc and cobalt chloride complexes.
Although efficient single site initiators have been reported for the controlled ROP of
lactide, a great challenge remains in performing catalytic polymerization of lactide, thus
optimizing productivity of the metal-based species and minimizing contamination of polymer
with metal residues. Inoue first reported the use of metalloporphyrins in the immortal
polymerization of lactones using transfer agents such as alcohols.12 This initial discovery
inspired the development of other catalytic systems.13 As already mentioned, model studies of
different zinc enzymes have been used to determine how the coordination environment
around the zinc centre affects the activity of the model. To this end, researchers have
constructed active zinc enzyme mimics. In particular, tripodal zinc complexes have been
widely investigated as models. For instance, it has been shown that, in the presence of
methanol or ethanol, a tris(pyrazolyl)borate zinc complex acts as a catalyst for the ring-
opening of β-butyrolactone, resulting in the formation of the corresponding esters of β-
hydroxybutyric acid. In this regard, we envisaged that our [L2NO]M(Cl)n and [L3N]M(Cl)n
3

Page 5 of 39
Dalton Transactions
View Article Online
chloride complexes would have the potential to be labile to alcohol exchange; that is, a rapid
DOI: 10.1039/C3DT52629A
exchange would occur between the metal species and alcohol. Since the growing species of
lactide polymerization is an alkoxide, alcohol would then behave as a reversible chain transfer
agent. Herein we report the synthesis, characterization, and LA polymerization reactivity of
different zinc(II) and cobalt(II) complexes.
Synthesis and structural characterization of the pro-ligands. The tripodal pro-ligands
described in this study were prepared in a straightforward manner, by reductive amination,
following a one-pot procedure using commercially available or easily accessible starting
materials (Figure 1).11c,14,15 The new compounds [L2]H, [L3]H, L4, and L5 were obtained in
good to excellent yields. [L3]H was prepared by two successive reductive amination reactions
from picolyl amine, introducing first the phenolic moiety followed by the phosphine arm. L5
is obtained in a one-step reaction between diphenylphosphinobenzaldehyde and an
ammonium salt. All compounds were obtained as colourless or yellowish powders, and their
identity was confirmed by NMR and high-resolution mass spectrometry or elemental analysis.
Such ligand architecture allows for steric and electronic variations on the different donor
fragments (Figure 1).
4

Dalton Transactions
Page 6 of 39
R4
View Article Online
DOI: 10.1039/C3DT52629A
R4
R1
R2
R1
R2
NaBH(OAc)3
+
HN
N
R3
R3
CHO
X
X
[L1]H: X = OH, R1 = R2 = 2-pyridyl, R3 = R4 = tBu
[L2]H: X = OH, R1 = R2 = 2-pyridyl, R3 = C(Ph)3, R4 = Me
[L3]H: X = OH, R1 = 2-(diphenylphosphino)benzyl, R2 = 2-pyridyl, R3 = R4 = tBu
L4: X = PPh2, R1 = 2-(diphenylphosphino)benzyl, R2 = 2-pyridyl, R3 = R4 = H
L5: X = PPh2, R1 = R2 = 2-(diphenylphosphino)benzyl, R3 = R4 = H
L6: X = PPh2, R1 = R2 = 2-pyridyl, R3 = R4 = H
Ph Ph
Ph
OH
OH
N
OH
N
N
N
N
N
Ph2P
N
N
[L1]H (N3OtBu)
[L3]H (N2PO)
[L2]H (N3OCPh3
)
Yield: 85%
Yield: 75%
Yield: 75%
PPh2
PPh2
N
PPh2
N
N
N
N
Ph2P
Ph2P
PPh2
N
L6 (N3P)
Yield: 85%
L5 (NP3)
Yield: 61%
L4 (N2P2)
Yield: 90%
Figure 1. Synthesis of [Ln]
The crystal structures of compounds [L1]H, [L2]H, L4 L5 and L5·HCl were determined by
,
,
X-ray diffraction studies (Figures 2-6). A summary of the relevant crystallographic data and
refinement parameters for [L1]H, [L2]H, L4 L5and the protonated ligand L5·HCl can be found
,
in Table 1. Interestingly, the molecular structure of the ammonium cation [L5H]+ (Figure 6)
revealed the steric properties of pro-ligand L5 in a structural motif akin to that involving metal
coordination, due to the interaction of all coordinating atoms, the nitrogen and the three
phosphorus atoms with the proton H1. These interactions are evidenced when comparing with
the molecular structure of L5 (Figure 5) in which the tripod minimizes the steric interactions
5

Page 7 of 39
Dalton Transactions
View Article Online
between all atoms. Considering the plane containing the three methylenic carbons, the
DOI: 10.1039/C3DT52629A
nitrogen atom is above the plane and the three phosphorus atoms located on the other side; the
lone pair of nitrogen pointing out from the NP3 tetrahedra. In [L5H]+, the proton is inside the
NP3 tetrahedra, showing the cavity that could create the ligand L5 around a metal centre upon
coordination.
Table 1. Summary of X-ray data collection and refinement
[L1]H
[L2]H
L4
L5
L5·HCl
2
Empirical
C27H35N3O
C39H35N3O C44H38N2P2
C57H48NP3 C60H52Cl10NP3 C45H40ClN3OZn
formula
Formula weight
417.58
100(2)
0.71073
monoclinic
P21/n
561.7
100(2)
656.7
150(2)
839.9
150(2)
0.71073
triclinic
P-1
1234.44
150(2)
739.62
150(2)
T (K)
Wavelength (Å)
Crystal system
Space group
a (Å)
0.71073
triclinic
P-1
0.71073
monoclinic
P21/c
0.71073
triclinic
P-1
0.71073
monoclinic
P21/n
10.052(2)
90
10.9042(6)
112.417(2)
13.0721(6)
91.203(2)
14.7332(7)
104.786(2)
1860.58(16)
2
14.0718(9) 15.6653(16)
90 98.534(4)
13.5580(8) 18.0030(18)
98.664(2) 112.551(4)
19.6384(9) 18.1172(19)
10.8501(19)
93.544(6)
13.843(2)
95.702(6)
20.392(3)
96.184(7)
3021.8(8)
2
9.1363(15)
90
α (°)
b (Å)
24.616(6)
110.971(12)
10.268(2)
90
22.025(5)
91.470(11)
19.679(4)
90
β (°)
c (Å)
γ (°)
V (Å3)
90
3704.0(4)
4
92.618(4)
4637.0(8)
2
2372.4(9)
4
3958.6(14)
4
Z
Dc (g.cm-3)
µ (mm-1)
F(000)
1.169
1.003
1.178
0.150
1384
1.203
0.167
1768
1.357
1.241
0.071
0.060
0.579
0.725
904
596
1268
1544
Crystal size
(mm)
0.55 x 0.5 x 0.55 x 0.32
0.27 x 0.18
x 0.11
0.56 x 0.39
x 0.25
0.57 x 0.51 x
0.42
0.1 x 0.1 x 0.1
0.43
x 0.18
Crystal color
colourless
colourless
colourless
colourless
colourless
colourless
42068
Collected
reflections
Unique
17275
34702
30098
65501
44988
5268
8390
8432
20787
13664
8975
reflections
Rint
0.069
282
0.0481
391
0.0504
433
0.0637
1099
0.0554
671
0.2219
407
Variable
parameters
Goodness-of-fit
1.013
1.093
1.055
1.038
1.052
0.912
R1, wR2 (I>2σ)a
0.0672,
0.1496
0.0524,
0.1518
0.0451,
0.122
0.0488,
0.1132
0.0631,
0.1481
0.0986, 0.2197
6

Dalton Transactions
Page 8 of 39
Δρmax/min (e.Å-3)
View Article Online
DOI: 10.1039/C3DT52629A
0.28, -0.262
0.243, -
0.235
0.32, -0.259 0.356, -0.34 1.563, -1.534
0.798, -1.086
a R1 = ∑ | |Fo| - |Fc| | / ∑ |Fo|; wR2 = {∑ [w(Fo2 - Fc2)2] / ∑ [w(Fo ) ]}1/2
2 2
Figure 2. Molecular structure of pro-ligand [L1]H (hydrogen atoms and solvent molecules are
omitted for clarity)
Figure 3. Molecular structure of pro-ligand [L2]H (hydrogen atoms and solvent molecules are
omitted for clarity)
7

Page 9 of 39
Dalton Transactions
View Article Online
DOI: 10.1039/C3DT52629A
Figure 4. Molecular structure of pro-ligand L4 (hydrogen atoms are omitted for clarity)
Figure 5. Molecular structure of tripodal pro-ligand L5 (hydrogen atoms are omitted for
clarity)
8

Dalton Transactions
Page 10 of 39
View Article Online
DOI: 10.1039/C3DT52629A
Figure 6. Molecular structure of tripodal ammonium salt L5·HCl (hydrogen atoms and
chloride anion are omitted for clarity)
Synthesis and structural characterisation of the cobalt(II) and zinc(II) complexes. The
coordination chemistry of the ligands was studied towards zinc and cobalt dichloride. [L2NO]
and [L3N] ligands were used to synthesize a series of twelve new complexes 1-12, such as
[L2NO]MCl and [L3N]MCl2 (M = Zn, Co). Reactions of the neutral pro-ligands L4, L5, and L6
with one equiv. of ZnCl2 or CoCl2 in tetrahydrofuran at 50ºC afford cleanly the bischloro-zinc
and -cobalt complexes 4-6 and 10-12 (Figure 7). These compounds were isolated in good
yields and are readily soluble in chlorinated solvents. These complexes were characterized by
elemental analysis, NMR spectroscopy and molar conductivity, UV/Vis measurements, X-ray
diffraction studies for some of them. The results of molar conductivity measurements suggest
that all zinc complexes are non-electrolytes in DMF (see Supporting Information). In
1
accordance with previous studies and H NMR data (vide infra), these observations are
indicative of the formation of (di)chloride species, in which one of the pyridyl/phosphino
moieties may remain uncoordinated. By contrast the molar conductivity values of cobalt(II)
dichlorides indicate that one of the coordinated chloride has been replaced by DMF molecule.
This assignment of cationic structures was supported by UV/Vis analyses of 7, 9 and 11.
9

Page 11 of 39
Dalton Transactions
View Article Online
DOI: 10.1039/C3DT52629A
Figure 7. Complexes 1-12 (R = Ph).
The monochloro zinc and cobalt complexes 1-3 and 7-9 were synthesized via a salt
metathesis route (Scheme 1). The salt metathesis reactions are two-step reactions. First a
solution of pro-ligand is combined with sodium hydride in order to form an anionic
phenoxide. In a second step, the preformed sodium-phenoxide is added to a metal chloride
and the synthesis of the metal complex is accompanied by the formation of sodium chloride.
As an example, the synthetic scheme for the preparation of 1 is shown Scheme 1. The
1
coordination of the ligand L1 to the zinc metal centre is confirmed by H NMR analysis
(Figure 8). Firstly, the signal of the pyridine α-proton is shifted from 8.4 ppm to 9.4 ppm.
Secondly, an AB pattern for benzylic CH2 protons confirms the non-equivalence of the
benzylic protons of each picolyl arm upon coordination (C-C bond rotation is now
10

Dalton Transactions
Page 12 of 39
View Article Online
impossible). Finally, the benzylic protons of the phenolic arm remain equivalent and appear as
DOI: 10.1039/C3DT52629A
1
a singlet in the H NMR spectrum, suggesting the formation of a monomeric species with
rigid Cs symmetry on the NMR timescale. As further supported by the solid state structure of
2 (vide infra), these observations are consistent with both pyridyl groups coordinated onto the
metal centre. The same features are also observed for L2 and L3 coordinated zinc complexes
(see Experimental Section).
Scheme 1. Synthesis of 1.
1
In the H NMR of complex 5 (Figure 9), all the signals, especially the benzylic CH2 are
broaden singlets, suggesting a fluxional structure, probably due to the weak interaction
between the zinc and the phosphorus atoms; the small coordination induced shifts Δδ 31P
31
31
(complex-free ligand) are 4.2, 0 and –2.5 (δ P(L5) = –16.3 ppm; δ P(5) = –11.1, –16.3, –
18.8 ppm), and supports weak interactions of two of the phosphorus atoms with the zinc metal
centre. Variable-temperature NMR spectroscopy of complex 5 in CDCl3 revealed
decoalescence of the resonances upon lowering the temperature, eventually resulting in the
observation of individual resonances for two NCH2-phosphine groups at –30°C. These
observations are also consistent with only two of the phosphine groups coordinated onto the
metal centre and fast exchange of the coordinated and free phosphine groups on the NMR
timescale at relatively high temperature. The enhanced steric flexibility within L5, originating
from the presence of less donating phosphino groups, and the corresponding more opened
11

Page 13 of 39
Dalton Transactions
View Article Online
coordination sphere at the metal centre might be responsible for the formation of DaOmI: 1i0x.1t0u3r9e/C3oDfT52629A
31
monomeric (5) and dimeric/oligomeric species.11 In P NMR spectra, one shielded and one
deshielded signals compared to the free ligand signal are observed (see experimental section).
The same situation, that is a coordinated pyridyl/phosphino group and a free
pyridyl/phosphino group, can be envisioned for the other metal complexes. Alternatively, we
cannot rule out that all σ-donor groups are coordinated to the metal centre in 1-12, either in a
symmetric or dissymmetric fashion. In the case of nitrogen containing pro-ligand
homologues, observation of exchange mechanism between two pyridine arms on the zinc
atom was reported for a (N4)ZnCl2 system (N4 = N-[(2-pyridyl)methyl]-2,2’-
dipyridylamine).16 In addition, as already observed by Kol in zinc complexes supported by
aminophenolate ligands bearing 2-methoxyethyl arms, one arm can be tightly coordinated and
the other one can be weakly coordinated onto the metal centre.17
Figure 8. 1H NMR spectra of pro-ligand [L1]H (up) and 1 (down) (CDCl3, 25°C, 300 MHz).
12

Dalton Transactions
Page 14 of 39
View Article Online
DOI: 10.1039/C3DT52629A
Figure 9. Variable temperature 1H NMR experiment for (L5)ZnCl2 (5); benzylic region (400
MHz, from 40°C to –30°C, CDCl3).
The crystal structure of complex 2 was determined by X-ray diffraction studies. Complex 2
is monomeric in the solid state and features a five-coordinated zinc centre in a distorted
bipyramidal trigonal geometry (Figure 10). Furthermore, the geometric parameter τ defined
by Addison and Reedijk is 0.80, which implies that the coordination geometry is best
described as a distorted bipyramidal trigonal.18 Consistent with the NMR observations in
solution for complex 1 (vide supra), the two pyridyl groups per molecule are coordinated,
together with the N and O atoms of the ligand, forming one six-membered metallacycle and
two five-membered metallacycles. Therefore, though potentially tridentate in solution, the
tripodal aminophenolate is actually tetradentate. The coordination sphere is completed by the
chloro ligand. The Zn-O and Zn-N bond distances fall in normal ranges for similar
complexes.11 The Zn(1)–N(11) (2.123(6) A˚) and Zn(1)–N(1) (2.141(6)A˚) bond distances are
13

Page 15 of 39
Dalton Transactions
View Article Online
comparable, likely reflecting a relatively strong σ-donor effect of the pyridyl groups. The six-
DOI: 10.1039/C3DT52629A
membered metallacycle adopts a boat conformation, as observed in other trisaminophenolate
zinc complexes.
Figure 10. Molecular structure of 2 (hydrogen atoms and solvent molecule are omitted for
clarity).
Paramagnetic NMR spectroscopy of the cobalt complexes. The NMR of paramagnetic
species has been extensively studied over the past decades.19,20,21,22 In diamagnetic molecules,
the chemical shift arises from the orbital shift linked to the shield constant. In paramagnetic
molecules, a hyperfine shift is added modifying the observed chemical shift. Even though the
presence of a paramagnetic centre in the molecule modifies the NMR spectra, some
information can still be extracted from NMR experiments.23 Despite the acceleration of
relaxation processes, the integration of NMR signals is still linked to the number of nuclei
involved in the resonance.24 Moreover, by setting the parameters to the correct values, the
recording of 2D NMR spectra is possible, allowing seeing the correlations between nuclei
spin magnetic moments.
14

Dalton Transactions
Page 16 of 39
View Article Online
DOI: 10.1039/C3DT52629A
Table 1. Chemical shift, spin-lattice relaxation time, half-height width and integration of the
1H NMR spectrum of 12 in CDCl3
α
β
β'
m + b
13.6
26.1
30
b’
c
a
γ
p
o
δ (ppm)
131.5 64.3 43.0
13.2
3.6
-3.3
-5.6
-7.5 -12.5
T1 (ms)
0.4
800
2
14.4
75
2
7.6
60
2
35.2 38.5 10.2 21.2
49
110
2
1.3
240
4
ν (Hz)
½
40
1
25
1
45
1
65
2
Integration
5
As expected, the cobalt(II) complex (L6)CoCl2 12 is paramagnetic, displaying broad
signals in its 1H and 31P NMR spectra.25 The 1H NMR spectrum displays ten resonances over
the fourteen expected for a CS symmetrical complex (Figure 11). The most down-fielded
resonances at 131.5, 64.3 and 43 ppm can be easily attributed to protons on the pyridine cores
as often attributed in similar iron(II) complexes.26 The COSY experiment with the {43.0 ppm;
-5.6 ppm} correlation gives the last signal for the pyridines cycles (Figure 1S). The attribution
of the different signals is allowed by the correlation between relaxation times and distance to
the paramagnetic centre. The broadest peak at 131.5 ppm corresponds to the nucleus in α
position. The two signals at 64.3 ppm and 43.0 ppm can be attributed to the β and β’ protons,
respectively. The phenyl groups of the triarylphosphine are easily assigned with the peak
integration. Those aromatic cycles are the only possibility to have four or two equivalent
protons. The protons in ortho position are the closest to the metal centre and thus are the most
shielded and shifted to -12.5 ppm. In addition the meta protons resonate at 13.6 ppm and the
two para protons at -7.5 ppm. The relative position of these signals is different from the
position of the pyridine protons signals, which can be explained by the difference of the
coordination mode. The direct bonding through the nitrogen donor atom increases the residual
electronic spin density on the pyridine atoms yielding a stronger Fermi contact interaction.
Finally, for the last aromatic ring, 4 non-equivalent protons are identified. The proton a in
ortho position is assigned to the -3.3 ppm signal, the broadest and quickest relaxing. The other
15

Page 17 of 39
Dalton Transactions
View Article Online
signals are attributed following the same pattern as the two phenyl rings. As prDoOtoI: n10s.10o3n9/Ca3lDlT52629A
aromatic cycles are subjected to strong paramagnetic shift, all donor atoms are coordinated to
the metal centre. On this spectrum the benzylic protons are not detected either due to their fast
relaxation at the NMR timescale or due to an excessive peak width.
The description of the previous complex enables a faster interpretation of other cobalt(II)
1
complexes with the same framework. In particular, the H NMR spectrum of 7 displays the
same paramagnetic pattern (Figure 12). The pyridine protons signals are observed in the same
range as in 12. The main difference concerns the two aromatic protons of the phenolate ring
which resonate at 46.3 ppm and 29.3 ppm. Like previously, the most down-fielded signal is
attributed to the b position. The greater paramagnetic shift of these protons is explained by the
increased electronic spin density on the ring as there are fewer rings on which it can be
delocalized. Even if they are in the diamagnetic region, the tBu protons are less shielded than
in the free protonated ligand. The protons in ortho position appear as a broad signal at 3.45
ppm and the other ones as a sharper signal at 3.75 ppm.
16

Dalton Transactions
Page 18 of 39
View Article Online
DOI: 10.1039/C3DT52629A
Figure 11. 1H NMR spectrum of complex 12 (CDCl3, 25°C, 300 MHz).
Figure 12. 1H NMR spectrum of complex 7 (CDCl3, 25°C, 300 MHz).
As a d7 ion, the cobalt centre can either be in a low spin configuration (S=1/2) or a high
spin configuration (S=3/2). A change of the spin configuration in a given geometry has a
strong influence on the ionic radius.27 The magnetic moment can be directly measured by
SQUID experiments or by the NMR Evan’s method.28,29 Therefore, the magnetic moment of
complex 12 was measured with both techniques and gave, respectively, μeff=4.0 μB and
μeff=3.9 μB. As the complex is not in the same state for these two methods, the results can
differ due to different accessible configurations in the liquid state. Nevertheless, both methods
indicate a cobalt in its high spin configuration.30 The magnetic moment of 7, determined using
17

Page 19 of 39
Dalton Transactions
7
View Article Online
the Evans’ method in solution, was found to be 4.8 µB, consistent with a d metal center
DOI: 10.1039/C3DT52629A
having three unpaired electrons.
Reactivity towards rac-lactide. A range of catalytic species have been employed to
mediate the polymerization of LA.31 However, in spite of great efforts in this field, there is
still demand for highly active, nontoxic, and stable complexes that can be easily handled. In
this regard, ROP catalysts have not been widely investigated in bulk polymerizations,
although there is significant interest in using metal-based catalysts that are active in bulk
monomer. Indeed, bulk polymerizations are industrially attractive and compliant with the
principles of green chemistry.32 For instance, Feijen et al. demonstrated that a chiral Schiff-
base (SALEN type) aluminium complex based on (R,R)-cyclohexanediamine is an
isoselective initiator for the polymerization of racemic lactide at 130°C.33 Interestingly, this
catalytic system proceeded under solvent free conditions, and gave high-yield reactions,
controlled molecular weights and narrow polydispersities. Unfortunately, this aluminium-
based initiator was not sufficiently active to be industrially viable, as two days were necessary
to convert 200 eq. of lactide.33 Also, Hillmyer reported a detailed study of analogous Zn, Mg,
and Co dinuclear complexes, showing that the dizinc complex is the most active of the series;
in melt lactide at 140ºC, more than 4000 eq of lactide were converted by the dinuclear zinc
complex to produce atactic polylactide with low molecular weight control (Mn 57 kg.mol-1;
Mw/Mn 2.6).34 Only one other example of cobalt-based [Co(OAr)3][Na(THF)6] complexes was
reported for melt lactide polymerization with moderate activity and control.35
As noted earlier, a great challenge in the production of PLA is the catalytic polymerization
of lactide, thus optimizing productivity of the metal-based species and minimizing
contamination of polymer with metal residues. In this regard, we were interested in
investigating the catalytic behaviour of our chloride complexes for lactide polymerization in
18

Dalton Transactions
Page 20 of 39
View Article Online
combination with an alcohol under solvent free conditions. Representative resultsDoObI:s10e.r10v3e9d/C3iDnT52629A
these polymerizations using benzyl alcohol as a reversible transfer agent are collected in
Table 2. All complexes36 were tested for ring-opening polymerization of rac-lactide, at 130°C
in solvent free conditions. As control experiments the homoleptic zinc and cobalt dichloride
precursors were tested in the presence or absence of alcohol for the ring-opening
polymerization of rac-lactide. By contrast to previous observations,37 we observed activity in
the case of zinc dichloride in the presence or absence of alcohol at 130°C, even if the
polymerization rate and control were low (Table 2, entry 1). However, CoCl2 is inactive for
the ROP of lactide in bulk at 130ºC (Entry 2). Generally, in the absence of alcohol, the Zn
complexes (Table 2, entries 4, 7, 9 and 11) and the Co complexes (Table 2, entry 17) gave
also modest reactivities and controls.
Table 2. Polymerization of Lactide using Co and Zn complexesa
Mn
(g.mol-1)d
Time
(h)
Conv.
(%)b
Mn,calc
(g.mol-1) c
TOF
(h-1)
d
Entry Initiator [LA]/[I] [BnOH]/[I]
Mw/Mn
1000
1000
1000
1000
1000
5000
1000
1000
1000
1000
1000
1000
1000
1000
1000
1000
1000
1000
10
10
10
-
3
3
11
0
NDe
ND
ND
ND
ND
37
0
1
2
ZnCl2
CoCl2
ND
3
4
1
2
3
13
28
75
44
27
83
5
1800
40300
10800
31700
38800
11900
ND
1100
41000
10500
17000
8200
11300
ND
1.06
2.02
2.00
1.71
1.32
1.16
ND
43
0.25
0.25
0.25
2
1120
3000
8800
135
415
16
5
2
10
10
-
6
2
7
3
8
3
10
-
2
9
4
3
10
11
12
13
14
15
16
17
18
4
10
-
3
66
12
77
13
45
26
10
7
9300
17300
11000
ND
6700
20800
12300
ND
1.05
1.22
1.38
ND
218
120
770
43
5
1
5
10
10
10
10
10
-
1
7
3
8
3
6400
3700
ND
5000
2300
ND
1.21
1.07
ND
150
86
9
3
10
11
11
3
33
3
ND
ND
ND
23
10
3
60
8600
7300
1.08
200
19

Page 21 of 39
Dalton Transactions
a
b
1
View Article Online
Polymerization was performed in melt rac-lactide at 130°C. Conversion was determined by H NMR
DOI: 10.1039/C3DT52629A
(integration of the methyl resonances of LA ad PLA). c Mn,calc = 144.00 × ([LA]/[I]) × Conv. / ([BnOH]/[I]). d Mn
and Mw/Mn of polymer determined by SEC-RI in THF at RT using polystyrene standards and Mn corrected by the
Mark-Houwink parameter (0.58). e not determined.
Notably, the combination based on zinc complex 1 and 10 equiv. of benzyl alcohol
readily polymerized 130 equiv. of rac-LA in 3h at 130ꢀ°C, leading to PLA with an
experimentally determined Mn value in close agreement with the calculated one (Entry 3).
Interestingly, for most of the systems tested, we observed that in the presence of an excess of
benzyl alcohol (10 equiv.), polymers with narrow polydispersities and controlled molecular
weights, which could be predicted from the monomer-to-alcohol ratio, were obtained.38
Depending on the ligand framework, the zinc initiator is not always more active than its Co
analog. For instance, the cobalt complex 7 (Entry 13), which bears ligand L1 with a tert-butyl
group in the ortho position of the phenolate, is as active as the Zn analog 1 (Entry 3). When
the ortho-substituent is a bulkier CPh3 substituent, the corresponding zinc complex 2 is much
more active and exhibits the highest activity observed in this study with a TOF up to 8800 h-1
(Entry 6). However, introduction a bulkier ortho-substituent on the N3O ligand also decreases
the control of polymerization (Mn shorter than the expected one and PDI broadened).
Remarkably, as opposed to the chloro metal complexes that were previously evaluated in
similar conditions, the zinc complex 2 is significantly more efficient.37 Indeed, these data lie
in the same range as the highest activities obtained with unstable catalytic systems, the major
difference being that the present zinc catalytic system is much easier to handle.37
The phosphorus atom is a soft donor and has a weak interaction with the zinc metal centre.
Therefore, by replacing a pyridine arm by a phosphine arm, the metal centre is more
accessible and more electrophilic. This is evidenced in the ROP of LA by an increased
polymerization rate for 3 compared to 1 (Entries 3 and 8). Also, the replacement of a pyridyl
group in 4 by a phosphine group in 5 increases the activity of the initiators from a TOF of 218
to 770 h-1 respectively (Entries 10 and 12). We assume that, since the phosphine ligand L5 is
20

Dalton Transactions
Page 22 of 39
4
View Article Online
DOI: 10.1039/C3DT52629A
more labile than the ligand L , the environment around the zinc metal centre is less hindered,
resulting in better activity. Similarly to what was obtained in the case of N3O ligands, PDI are
slightly broadened (from 1.05 to 1.4) as the activity increases. In the case of neutral NmPn
ligand derivatives, the zinc is bonded to two chloride atoms, meaning that in the presence of
benzyl alcohol, the propagation of two polymer chains can occur. However the molecular
weights of the polymer obtained for the zinc dichloride complexes 4 and 5 suggest that only
one polymeric chain is formed per metal atom.
Among the cobalt derivatives, complexes 8 and 11 exhibit good activity and excellent
control for ROP of lactide at 130ºC (Entries 14 and 18). As already observed for zinc
complexes, the cobalt complex 8 (Entry 14), which bears ligand L2 with a trityl group in the
ortho position of the phenolate, is much more active than the Co analog 7 bearing tert-butyl-
substituted ligand L1 (Entry 13). The poor reactivity of complex 10 (Entry 16) might arise
from the low solubility of the corresponding pre-catalyst. Overall, such performances
compare well with those reported for some related, efficient cobalt and zinc chloro systems.5
Finally, a somehow disappointing feature of these systems remains the absence of
stereocontrol achieved in the ROP of racemic lactide, whatever the substituents on the ligands
and the nature of the donor fragments.
Conclusion.
We have reported a new series of zinc and cobalt chloride complexes supported by readily
available tripodal ligands. The corresponding versatile systems enable the production of large
quantities of polymer with small amounts of catalyst, optimizing productivity and thus
minimizing the contamination of polymer with metal residues. These organometallic
catalysts, which are easily synthesized, are thermally robust, thus allowing their use in bulk
conditions. These results suggest a number of new avenues for biodegradable polymers.
21

Page 23 of 39
Dalton Transactions
View Article Online
Indeed the discovery of efficient and productive organometallic catalysts for theDsOyIn: 1t0h.1e0s39is/Co3DfT52629A
biodegradable polymers is a crucial requirement for a sustainable growth of the chemical
industry.
Acknowledgements. CNRS, ENSCP, ANR (grant ANR-10-PDOC-010-01), Fondation
Pierre-Gilles de Gennes, and French Ministry of Research and Higher Education are thanked
for financial support of this work. The authors would like to thank Purac for a generous loan
of rac-lactide. CMT is grateful to the Institut Universitaire de France.
Experimental Section.
General considerations. All reactions were carried out under an argon atmosphere using
Schlenk techniques or a glove box (<1 ppm O2, <2 ppm H2O). Tetrahydrofuran, pentane and
toluene were dried with solvent purification system (PureSolv, Innovative technology, Inc.)
and degassed by freeze pump thaw cycles prior to use. Anhydrous dichloroethane, etheral
solution of HCl (1M), 2-picolylamine, 3,5-di-tert-butyl-4-hydroxybenzaldehyde, sodium
triacetoxyborohydride (stored in a glove box) were purchased from Aldrich and used as
received. Anhydrous cobalt dichloride (CoCl2) and zinc dichloride (ZnCl2) were purchased
from Strem chemicals. 2-diphenylphosphinobenzaldehyde was prepared by previously
described method.39 Deuterated solvents from Eurisotop were dried and degassed by freeze
pump thaw cycles (deuterated chloroform was distilled over phosphorus pentoxide, deuterated
benzene was distilled over sodium/benzophenone). NMR spectra were recorded on a Bruker
Avance 300, Avance 400 or AM-500 MHz spectrometer (at the Ecole Nationale Supérieure
de Chimie de Paris and at University of Caen). Chemical shifts are reported in ppm and
1
coupling constants J in hertz. NMR spectra were calibrated using residual H and 13C
31
resonances of deuterated solvents. External calibration for P was made with 85% H3PO4 as
22

Dalton Transactions
Page 24 of 39
13
1
1 Vie1w3Article Online
DOI: 10.1039/C3DT52629A
standard. Assignment of signals was made from C{1H}{31P}, 2D H-1H COSY, H- C
HMQC and HMBC NMR experiments. The paramagnetic spectrum sequence was as
followed: a π/6 pulse followed by a 550 ms acquisition and a 2 s delay for a spectrum width
of 220 ppm with 64 k points. After the FFT, a 0.3 Hz filter is applied. For the T1
measurements, the inversion-recovery sequence was used: π pulse, variable delay τ, π/2 pulse
and acquisition. The paramagnetic COSY 45 used an acquisition time of 32 ms and a delay of
1 s with TD1 = 512, TD2 = 2048 with an accumulation of 64 scans. For the paramagnetic
complexes, numbers in brackets are the peak half-width in Hz and the number of protons for
this signal. X-ray diffraction analyses were done at the Laboratoire de Diffractométrie X,
UMR 6226, CNRS-Université de Rennes 1 (France). Elemental analyses (C, H, N) were
performed at the Elemental Analysis Service of the London Metropolitan University (UK)
and are the average of two independent determinations. High-resolution mass spectra were
performed at Institut Parisien de Chimie Moléculaire, UMR CNRS 7201, University Pierre et
Marie Curie (France). Size exclusion chromatography (SEC) of polymers was performed in
THF at 35 °C using an Agilent 1260 Infinity Series GPC (ResiPore 3 μm, 300 x 7.5 mm, 1.0
ml/min, UV (250 nm) and RI (PL-GPC 220) detectors). The number average molecular
masses (Mn) and polydispersity index (Mw/Mn) of the polymers were calculated with reference
to a universal calibration vs. polystyrene standards (limits Mw = 200 to 400,000 g/mol).
SQUID measurements were recorded on a MPMS-XL7 apparatus. UV-visible spectra were
recorded on a Safas-Monaco spectrometer from 250 nm to 1000 nm with 2 nm steps. Molar
conductivities of freshly prepared 10-3 mol/L solutions of cobalt complexes in DMF were
measured on a CDM 210 conductivity meter.
Figure 13: [L1]H and [L2]H with position numbering.
23

Page 25 of 39
Dalton Transactions
View Article Online
DOI: 10.1039/C3DT52629A
General procedure: Synthesis of N-(2-hydroxybenzyl)-N,N-bis(2-pyridylmethyl)amine
proligands (N3O)H. Into a solution of di-(2-picolyl)amine (1 g, 5 mmol) and appropriate
salicylaldehyde (5.5 mmol) in 1,2-dichloroethane (12.5 ml), sodium triacetoxyborohydride
(1.7 g, 8 mmol) was added. The mixture was stirred overnight at room temperature, and water
(15 ml) was poured into the solution to stir for another 30 min. The organic phase was
separated from aqueous phase and dried over magnesium sulfate (MgSO4). The solvent was
removed under vacuum and final product was purified by recristillisation in cold methanol.
N-(3,5-ditert-butyl-2-hydroxybenzyl)-N,N-bis(2-pyridylmethyl)amine (N3OtBu)H ([L1]H):
White solid with 85 % yield. 1H NMR (300 MHz, 25 °C, CDCl3): δ = 8.47 (d, J = 4.2 Hz, 2H,
H1), 7.53 (t, J =4.2 Hz, 2H, H2), 7.29 (d, J = 4.8 Hz, 2H, H4), 7.13 (d, J = 2.4 Hz, 1H, H9),
7.08 (m, 2H, H3), 6.79 (d, J = 2.4 Hz, 1H, H11), 3.80 (s, 4H, H6), 3.73 (s, 2H, H7), 1.38 (s,
9H, H17), 1.19 (s, 9H, H14); 13C NMR (75 MHz, 25 °C, CDCl3): δ = 158.1 (C5), 153.8
(C13), 149.0 (C1), 140.4 (C8), 136.6 (C4), 135.6 (C12), 124.5 (C9), 123.6 (C2), 123.1 (C11),
122.2 (C3), 121.7 (C10), 56.5 (C6), 58.3 (C7), 34.7 (C16), 34.1 (C15), 31.7 (C17), 29.7 (C14)
N-(3-trityl-5-methyl-2-hydroxybenzyl)-N,N-bis(2-pyridylmethyl)amine
(N3OCPh3)H
1
([L2]H): White solid with 75 % yield. H NMR (300 MHz, 25 °C, CDCl3): δ =8.32 (d, J = 4
Hz, 2H, H1), 7.42 (t-d, J1= 4 Hz, J2 = 1.8 Hz, 2H, H2), 7.30-6.95 (m, 17H, H3, H14), 6.85 (d,
J = 9 Hz, 2H, H4), 6.82 (d, J = 1.8 Hz, 1H, H13), 6.74 (d, J = 1.8 Hz, 1H, H11), 3.66 (s, 2H,
H7), 3.59 (s, 4H, H6), 2.08 (s, 3H, H16); 13C NMR (75 MHz, 25 °C, CDCl3): δ = 157.5 (C5),
24

Dalton Transactions
Page 26 of 39
View Article Online
153.8 (C9), 148.7 (C1), 146.2 (C4), 136.8 (C14), 133.8 (C8), 131.3 (C14), 131.0 (C10), 129.6
DOI: 10.1039/C3DT52629A
(C12), 126.9 (C14), 126.5 (C13), 125.2 (C14), 123.8 (C2), 122.3 (C11), 122.1 (C3), 63.3
(C15), 58.6 (C7), 57.9 (C6), 20.9 (C16); HRMS exact mass calcd for [M+H]+ 601.3342,
found 601.3330.
Figure 14: [L3]H with position numbering.
2
1
HO
17
16
13
3
N
8
12
11
7
4
5
6
N
9
25
10
14
18
19
24
Ph2P
20
15
21
22
23
L3
N-(3,5-ditert-butyl-2-hydroxybenzyl)-N-(2-(diphenylphosphino)benzyl)-N-bis(2-
pyridylmethyl)amine (N2PO)H ([L3]H): Two steps of reductive amination. A solution of
3,5-ditert-butyl-4-hydroxybenzaldehyde (2.38 g, 10.1 mmol) and 2-(aminomethyl)pyridine (1
g, 9.2 mmol) in methanol (100 ml) was stirred under reflux for 2h. The solution was cooled
down to room temperature and sodium borohydride (1.74 g, 46 mmol) was added in small
portions. After NaBH4 was completely dissolved, the solvent was removed under vacuum.
Dichloromethane (100 ml) and water (100 ml) were added to crude product and the mixture
was stirred for 1h. The organic phase was washed two times more with water and dried over
magnesium sulfate (MgSO4). After removal of solvent, the product was further dried under
vacuum for 24h. It was then dissolved in 1,2-dichloroethane (23 ml) along with 2-
diphenylphosphino-benzaldehyde (2.93 g, 10.1 mmol) and sodium triacetoxyborohydride (3.2
g, 14.7 mmol) was then added. The mixture was stirred overnight at room temperature, and
water (15 ml) was poured into the solution to stir for another 30 min. The organic phase was
separated from aqueous phase and dried over magnesium sulfate (MgSO4). The solvent was
removed under vacuum and final product was purified by recrystallization in cold methanol to
25

Page 27 of 39
Dalton Transactions
1
View Article Online
DOI: 10.1039/C3DT52629A
give a white powder (4.1 g, 6.9 mmol, 75 %). H NMR (400 MHz, 25 °C, CDCl3): δ = 10.67
(s, 1H, OH), 8.51 (d, J = 3.6 Hz, 1H, H1), 7.62 (t-d, J1 = 5.7 Hz, J2 = 1.2 Hz, 1H, H2), 7.57
(m, 1H, H22), 7.27-7.38 (m, 8H) 7.17-7.26 (m, 5H), 7.10-7.16 (m, 2H), 6.80-6.87 (m, 2H),
3.93 (d, J = 2 Hz, 2H, H18), 3.76 (s, 2H, H7), 3.74 (s, 2H, H6), 1.45 (s, 9H, H16), 1.26 (s, 9H,
13
H14); C NMR (100 MHz, 25 °C, CDCl3): δ = 158.1 (C5), 153.8 (C13), 149.0 (C1), 142.3
(d, JP-C = 23 Hz, C19), 140.5 (C8), 136.8 (d, JP-C = 12 Hz, C25), 136.6 (C3), 136.5 (C20),
135.5 (C12), 134.0 (d, Jp-c = 20 Hz, C25) 133.6 (C21), 130.5 (d, JP-C = 4 Hz, C24), 129.3
(C23), 128.8 (C25), 128.6 (d, JP-C = 7 Hz, C25), 127.7 (C22), 124.4 (C11), 123.9 (C10), 123.0
(C9), 122.2 (C4), 121.9 (C2), 59.7 (C6), 58.1 (C7), 56.0 (d, JP-C = 22 Hz, C18), 35.0 (C17),
34.1 (C15), 31.8 (C16), 29.7 (C14); 31P{1H} (120 MHz, 25°C, CDCl3): δ = –16.2. HRMS
exact mass calcd for [M+H]+ 562.2853, found 562.2870.
Figure 15: L4 and L5 with position numbering.
16
17
15
14
4
PPh2
5
P
13
3
12
11
N
8
N
PPh2
1
2
6
7
1
N
7
9
P
2
9
8
10
Ph2P
3
4
10
6
11
5
L4
L5
Synthesis of N,N-bis(2-(diphenylphosphino)benzyl)-1-(pyridin-2-yl)methanamine (N2P2,
L4): A solution of 2-pyridinemethylamine (0.279 g, 2.60 mmol) and 2-diphenylphosphino-
benzaldehyde (1.50 g, 5.17 mmol) in degassed dichloroethane (10 mL) was stirred few
minutes. Then, sodium triacetoxyborohydride (1.42 g, 6.72 mmol) was added. The
heterogeneous mixture was stirred for 6h at room temperature, untill full conversion of the
26

Dalton Transactions
Page 28 of 39
31
View Article Online
DOI: 10.1039/C3DT52629A
aldehyde (the disappearance of the aldehyde can be monitored by P NMR). When the
reaction was completed, the solvent was evaporated and the product was extracted with
ethylacetate (AcOEt). The organic phase was washed with an aqueous saturated solution of
NaHCO3. and the organic phase was separated and dried over magnesium sulfate (MgSO4).
The solvent was evaporated to give the crude product as a white solid foam. The solid was
washed several times with pentane and dried. L4 was obtained pure as a white powder (1.53 g,
2.33 mmol, 90 %). Single crystals for X-ray diffraction structure analysis were obtained by
1
slow evaporation of a concentrated solution of L4 in acetone. H NMR (500 MHz, 25 °C,
CDCl3): δ = 8.34 (m, 1H, H5), 7.70 (m, 2H, H9), 7.53 (td, J = 7.71, J =1.9 Hz, 1H, H3), 7.38
(brd, J = 7.71 Hz, 1H, H2), 7.29 (m, 14H, H17, H16, H10), 7.18 (m, 8H, H15), 7.08 (m, 3H,
4
13
H11, H4), 6.79 (m, 2H, H12), 3.83 (d, JP-H = 2.1 Hz, 4H, H7), 3.65 (s, 2H, H6); C NMR
(125 MHz, 25 °C, CDCl3): δ = 159.9 (C1), 148.7 (C5), 143.8 (d, J P-C = 22.0 Hz, C8), 136.8
(d, J = 11.0 Hz, C14), 136.2 (C3), 136.1 (d, J P-C = 14.5 Hz, C13), 133.9 (d, J P-C = 19.6 Hz,
C15), 133.3 (C12), 128.82 (C9), 128.80 (C10), 128.6 (C17), 128.5 (d, J P-C = 6.9 Hz, C16),
31
126.9 (C11), 122.7 (C2), 121.7 (C4), 59.9 (C6), 56.1 (d, J P-C = 23.4 Hz, C7); P{1H} (200
MHz, 25°C, CDCl3): δ = –16.3. Anal. Calcd (%) for C44H38N2P2: C, 80.47; H, 5.83; N, 4.27.
Found: C, 80.37; H, 5.72; N, 4.18.
Synthesis of tris(2-(diphenylphosphino)benzyl)amine (NP3, L5): A suspension of
ammonium benzoate (0.154 g, 1.11 mmol) in a solution of 2-diphenylphosphinobenzaldehyde
(1.00 g, 3.45 mmol) in degassed dichloroethane (10 mL) was stirred for few minutes. Then,
sodium triacetoxyborohydride (1.06 g, 5 mmol) was added. The heterogeneous mixture was
stirred for 72h at room temperature, untill full conversion of the aldehyde (the disappearance
31
of the aldehyde can be monitored by P NMR). When reaction was completed, the solvent
was evaporated and the product was extracted with ethylacetate. The organic phase was
27

Page 29 of 39
Dalton Transactions
View Article Online
washed with an aqueous solution of NaOH (1 M) three times. The organicDOpI:h1a0.s1e039w/Ca3DsT52629A
separated and dried over magnesium sulfate (MgSO4). The solvent was evaporated to give the
crude mixture as a solid foam. The solid can be purified by chromatography over silica gel
column (gradient of elution from hexane to hexane/diethylether 2:1 v/v, Rf ~ 0.5) and
obtained as a white foam (0.6 g, 0.71 mmol, 61 %). If there is no other amine by-products, L5
can also be purified by precipitation of the ammonium chloride salt (c.f. synthesis of L5·HCl)
followed by deprotonation (a solution of L5·HCl in dichloromethane is vigorously stirred
with aqueous NaOH (1M)). Single crystals for X-ray diffraction structure analysis were
obtained by slow evaporation of a concentrated solution of L5 in CH2Cl2. 1H NMR (300 MHz,
25°C, D6-acetone): δ = 7.77 (m, 3H, H3), 7.30 (m, 21H, H-Ar), 7.12 (m, 15H, H-Ar) 6.72
4
13
(m, 3H, H6), 3.78 (d, JP-H = 2.8 Hz, 6H, H1); C{1H} NMR (125 MHz, 25°C, CDCl3): δ =
144.1 (d, JP-C = 22.5 Hz, C2), 136.9 (d, J P-C = 11.2 Hz, C8), 135.9 (d, J P-C = 14.5 Hz, C7),
133.9 (d, JP-C = 19.5 Hz, C9), 133.0 (C6), 128.8 (C3), 128.7 (d, JP-C = 5.1 Hz, C4), 128.5
31
(C11), 128.4 (d, JP-C = 6.9 Hz, C10), 126.7 (C5), 55.7 (d, JP-C = 23.6 Hz, C1); P{1H} (200
MHz, 25°C, D6-acetone): δ = –16.3. Anal. Calcd (%) for C57H48NP3: C, 81.51; H, 5.76; N,
1.67. Found: C, 81.44; H, 5.81; N, 1.55.
Synthesis of tris(2-(diphenylphosphino)benzyl)ammonium chloride (by protonation of
L5 with HCl) (L5·HCl): L5 was dissolved in diethyl ether and an etheral solution of HCl
(1M) was added dropwise until no further precipitate was formed. The solid was filtered off
over celite and recovered by dissolution in dichloromethane (CH2Cl2). The solvent was
evaporated to dryness and L5·HCl was obtained as a white solid quantitatively. Single
crystals for X-ray diffraction structure analysis were obtained by slow evaporation of a
concentrated solution of L5·HCl in CHCl3. 1H NMR (300 MHz, 25 °C, CDCl3): δ = 8.72 (brs,
1H, NH), 7.69 (d, J = 7.1 Hz, 4H, H-Ar), 7.45–7.00 (brm, 38H, H-Ar), 5.12 (brs, 6H, N-CH2);
28