Controlled acrylate insertion regioselectivity in diazaphospholidine- sulfonato palladium(II) complexes

Article abstract of DOI:10.1021/om300755j

Diazaphospholidine-sulfonato Pd(II) complexes [{κ²-P,O-(N- Ar₂C₂H₄N₂P)C₆H ₄SO₃}PdMe(L)] 1-L (L = dmso, pyridine, lutidine, or μ-LiCl(solvent); 1a: Ar = Ph, 1b: Ar = 2-MeC₆H₄, 1c: Ar = 2-MeOC₆H₄, 1d: Ar = 2,4,6-Me₃C ₆H₂, 1e: Ar = 2,6-iPr₂C₆H ₃, 1f: Ar = 2,6-(p-tolyl)₂C₆H₃) were prepared and structurally characterized. The regioselectivity of methyl acrylate (MA) insertion into the Pd-Me bond is entirely inverted from >93% 1,2-insertion for bulky substituents (1d-f, yielding the insertion products [(PO)Pd{κ²-C,O-CH₂CHMeC(O)OMe], 12) to the usual electronically controlled 2,1-insertion (>95%) for the less bulky Ar = Ph (1a, yielding the insertion product [(PO)Pd{κ²-C,O- CHEtC(O)OMe], 11, and β-H elimination product methyl crotonate). DFT studies underline that this is due to a more favorable insertion transition state (2,1- favored by 12 kJ mol^-1 over 1,2- for 1a) vs destabilization of the 2,1-insertion transition state in 1d,e. By contrast, MA insertion into the novel isolated and structurally characterized hydride and deuteride complexes [{κ²-P,O-(N-Ar₂C ₂H₄N₂P)C₆H₄SO ₃}PdR(lutidine)] (Ar = 2,6-iPr₂C₆H₃; 9e: R = H, 10e: R = D) occurs 2,1-selectively. This is due to the insertion occurring from the isomer with the P-donor and the olefin in trans arrangement, rather than the insertion into the alkyl from the cis isomer in which the olefin is in proximity to the bulky diazaphospholidine. 1a-f are precursors to active catalysts for ethylene polymerization to highly linear polyethylene with M _n up to 35 000 g mol^-1. In copolymerization experiments, norbornene was incorporated in up to 6.1 mol % into the polyethylene backbone.

Full text of DOI:10.1021/om300755j

Article
pubs.acs.org/Organometallics
Controlled Acrylate Insertion Regioselectivity in Diazaphospholidine-
Sulfonato Palladium(II) Complexes
Philipp Wucher,† Philipp Roesle,† Laura Falivene,‡ Luigi Cavallo,§ Lucia Caporaso,*,‡
Inigo Gottker-Schnetmann,† and Stefan Mecking*,†
̈
†Chair of Chemical Materials Science, Department of Chemistry, University of Konstanz, Universitatsstrasse 10, 78457 Konstanz,
̈
Germany
‡Department of Chemistry, University of Salerno, Via Ponte Don Melillo, 84084 Fisciano, Salerno, Italy
§Kaust Catalysis Center, Chemical and Life Sciences and Engineering, King Abdullah University of Science and Technology, Thuwal
23955-6900, Saudi Arabia
S
* Supporting Information
ABSTRACT: Diazaphospholidine-sulfonato Pd(II) complexes [{κ2-P,O-(N-Ar2C2H4N2P)C6H4SO3}PdMe(L)] 1-L (L = dmso,
pyridine, lutidine, or μ-LiCl(solvent); 1a: Ar = Ph, 1b: Ar = 2-MeC6H4, 1c: Ar = 2-MeOC6H4, 1d: Ar = 2,4,6-Me3C6H2, 1e: Ar =
2,6-iPr2C6H3, 1f: Ar = 2,6-(p-tolyl)2C6H3) were prepared and structurally characterized. The regioselectivity of methyl acrylate
(MA) insertion into the Pd−Me bond is entirely inverted from >93% 1,2-insertion for bulky substituents (1d−f, yielding the
insertion products [(P∧O)Pd{κ2-C,O-CH2CHMeC(O)OMe], 12) to the usual electronically controlled 2,1-insertion (>95%) for
the less bulky Ar = Ph (1a, yielding the insertion product [(P∧O)Pd{κ2-C,O-CHEtC(O)OMe], 11, and β-H elimination product
methyl crotonate). DFT studies underline that this is due to a more favorable insertion transition state (2,1- favored by 12 kJ
mol−1 over 1,2- for 1a) vs destabilization of the 2,1-insertion transition state in 1d,e. By contrast, MA insertion into the novel
isolated and structurally characterized hydride and deuteride complexes [{κ2-P,O-(N-Ar2C2H4N2P)C6H4SO3}PdR(lutidine)] (Ar
= 2,6-iPr2C6H3; 9e: R = H, 10e: R = D) occurs 2,1-selectively. This is due to the insertion occurring from the isomer with the P-
donor and the olefin in trans arrangement, rather than the insertion into the alkyl from the cis isomer in which the olefin is in
proximity to the bulky diazaphospholidine. 1a−f are precursors to active catalysts for ethylene polymerization to highly linear
polyethylene with Mn up to 35 000 g mol−1. In copolymerization experiments, norbornene was incorporated in up to 6.1 mol %
into the polyethylene backbone.
toward polar moieties.2 Thus, cationic palladium and nickel α-
INTRODUCTION
The selective insertion of an olefinic substrate into a metal−
carbon bond is a decisive step in many catalytic transformations.
■
diimine complexes 3 (Figure 1) can copolymerize ethylene and
1-olefins with acrylates, yielding branched copolymers.3 Due to
In the area of polymer synthesis, this is impressively illustrated by
isotactic polypropylene. Its properties ultimately result from a
highly regio- and stereoselective insertion of propylene into the
growing chain. By comparison to the large-scale application of
catalytic polymerization of apolar olefins such as ethylene and
propylene,1 an insertion (co)polymerization of electron-
deficient monomers such as acrylates is challenging. Consid-
erable progress in this area has been achieved by the
development of d8-metal (late transition metal) complexes.
Their less oxophilic nature, in comparison to their early
transition metal counterparts, renders them more tolerant
the propensity of the cationic Pd(II) catalysts for “chain walking”
by a series of rapid β-hydride elimination and reinsertion events,
the polymers formed are highly branched, with the acrylate-
derived repeat units located at the end of branches.
In contrast, highly linear ethylene-methyl acrylate copolymers
without any chain-walking-derived microstructure are obtained
with neutral Pd(II) complexes 2-L based on anionic phosphine-
Received: August 7, 2012
© XXXX American Chemical Society
A
dx.doi.org/10.1021/om300755j | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Figure 1. Neutral diazaphospholidine-sulfonato Pd(II) complexes (left), neutral phosphine-sulfonato Pd(II) complex (center), and cationic Pd(II)
diimine complexes (right).
Scheme 1. Preparation of Diazaphospholidine-Pd(II) Complexes
sulfonato ligands [{κ2-P,O}-(2-MeOC6H4)2PC6H4SO2O]−. A
variety of polar comonomers in addition to methyl acrylate
(MA), e.g., vinyl ether, acrylonitrile, vinyl fluoride, acrylamides,
vinyl acetate, vinyl sulfones, acrylic acid, substituted norbor-
nenes, and allyl acrylate, have been effectively copolymerized
with ethylene by complexes of type 2.2i,4,5 These unique catalytic
properties can be related to the unsymmetric nature of the
anionic chelating ligand, with a soft phosphine and a hard
sulfonate oxygen donor.6 With respect to influencing the olefin
insertion step, however, the immediate environment of the metal
center is rather open in these square-planar complexes: the
sulfonate donor bears little steric bulk, and the aryl substituents
on the phosphorus donor of 2 point away from the metal center.
We have recently communicated that by incorporation of the P-
donor into a diazaphospholidine heterocycle, with N-bound aryl
moieties, the latter can be forced into closer proximity to the
metal center to invert the regioselectivity of acrylate insertion.7
We now give a full account of the insertion behavior and
catalytic properties of this class of complexes.
B
dx.doi.org/10.1021/om300755j | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Figure 2. ORTEP plot of complexes 1a-lut, 1b-lut, 1b-ClLi(acetone), 1c-lut, 1d-lut, 1d-pyr, 1e-lut, 1e-dmso, and 1e-pyr drawn with 50% probability
ellipsoids. All hydrogen atoms and cocrystallized solvent molecules are omitted for clarity.
mononuclear dimethylsulfoxide complexes 1-dmso can be
prepared by addition of equimolar amounts of AgBF4 and
DMSO to a methylene chloride solution of the chloride-bridged
complexes 1-ClLi(solvent). Addition of the stronger coordinat-
ing pyridine or 2,6-lutidine without additional silver salts results
in the direct formation of the corresponding mononuclear
pyridine or lutidine complexes 1-pyr and 1-lut in 56% to 95%
yield. All ligands and complexes were fully characterized by 1D-
and 2D-NMR spectroscopy and elemental analysis (see
Supporting Information).
RESULTS AND DISCUSSION
■
Preparation of Diazaphospholidine Ligands and
Complexes. Diazaphospholidine-sulfonates with various N-
aryl substitution patterns and their Pd(II) complexes were
prepared similarly to the procedures reported previously for 1d-L
and 1e-L:7 The diamines 5 were prepared in a two-step reaction
in 32% to 80% yield, starting with the condensation of glyoxal
and aniline derivative 4, followed by reduction of the
intermediate diimines with NaBH4. The isolated diamines 5
were reacted with PCl3 in the presence of an excess of
triethylamine to yield the 2-chloro-diazaphospholidines 6,8
which were further reacted with o-dilithiobenzenesulfonate to
form the anionic ligands 7. These lithium salts form the LiCl-
bridged binuclear Pd(II) complexes 1a-ClLi(acetone), 1b-
ClLi(acetone), 1c-ClLi(acetone), and 1f-ClLi(MeOD) in 63%
to 88% yield by a stoichiometric reaction with [(1,5-
cyclooctadiene)Pd(Me)Cl]9 (Scheme 1). The corresponding
In addition, X-ray diffraction analyses of complexes 1a-lut, 1b-
lut, 1c-lut, 1d-lut, 1d-pyr, 1e-lut, 1e-dmso, 1e-pyr, and 1b-
ClLi(acetone) (Figure 2) confirm their identity. Complexes 1a-
lut and 1c-lut crystallize in the triclinic P1 space group, whereas
̅
1b-lut, 1d-lut, 1d-pyr, 1e-lut, 1b-ClLi(acetone), and 1e-dmso
crystallize in the monoclinic P21/c or C2/c space group. All
complexes exhibit a square-planar environment around the Pd
C
dx.doi.org/10.1021/om300755j | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Table 1. Selected Bond Distances for Diazaphospholidine Complexes 1-L
Pd−L [Å]
2.134(6)
Pd−P [Å]
2.191(7)
Pd−O [Å]
2.147(5)
Pd−C [Å]
2.030(0)
2.048(7)
1a-lut
1b-lut
2.134(0)
2.197(3)
2.141(5)
1c-lut
2.120(7)/2.119(7)
2.139(2)
2.191(3)/2.190(6)
2.228(6)
2.147(7)/2.151(2)
2.153(4)
2.034(7)/2.028(2)
2.027(8)
1d-lut
1d-pyr
1d-dmso7
1e-lut
2.107(7)
2.205(7)
2.165(4)
2.032(2)
2.149(6)
2.187(9)
2.159(6)
2.187(9)
2.136(9)
2.212(8)
2.174(9)
2.018(9)
1e-pyr
1e-dmso
2.112(2)
2.202(8)
2.178(2)
2.021(7)
2.129(8)
2.179(1)
2.181(2)
2.019(3)
Figure 3. Synthesis of 9e-lut and 1H NMR spectrum (400 MHz, CD2Cl2, 298 K) of 9e-lut.
Figure 4. ORTEP plots of two rotamers (rotational disorder of the sulfonate group and the hydride atom) of 9e-lut drawn with 50% probability
ellipsoids. Hydrogen atoms (except Pd-H) are omitted for clarity. Hydrides H40 and H41 were located in the electron density map after the other
hydrogen atoms were treated as a riding model. Selected bond length [Å]: Pd1−H40 1.559(0), Pd1−H41 1.514(7), Pd1−N3 2.125(3), Pd1−P1,
2.176(8), Pd1−O2 2.190(5), Pd1−O5 2.230(6).
D
dx.doi.org/10.1021/om300755j | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Scheme 2. Regioselectivity of Insertion of Methyl Acrylate into a Pd−Me Bond of Complex 1-L
Scheme 3. Observed Regioselectivity of Diazaphospholidine Sulfonato Pd(II) Complexes 1a−f with MA
center, with the methyl group and the phosphorus atom located
mutually cis to each other. The crystal structure of complex 1c-lut
contains two rotamers per unit cell, with both methoxy-anisyl
groups occupying the same and the opposite site, respectively, of
the diazaphospholidine ring.
C6H4]2(o-SO2O-C6H4)PdMe]-dmso, where dmso was found to
bind κ-S.4b
In addition to these palladium methyl complexes 1-L, the Pd-
hydride and Pd-deuteride complexes 9e-lut and 10e-lut (Ar =
iPr2Ph) were synthesized by oxidative addition of P-protonated
(7e-H) or deuterated ligand (7e-D), respectively, to [Pd0(dba)2]
in the presence of excess 2,6-lutidine (Figure 3). The deuteride
complex 10e-lut contained about 30% of the corresponding
hydride complex 9e-lut, indicated by the characteristic high-field
1H NMR resonance of the hydride at −19.42 ppm (see
Supporting Information).
Crystals of 9e-lut suitable for single-crystal X-ray diffrac-
tometry were grown by layering a solution of the complex in
acetone with pentane in an NMR tube (Figure 4). The crystal
structure shows a rotational disorder of the sulfonate group,
which was modeled by two sets of oxygen positions, O1, O2, O3
and O4, O5, O6, refining to an occupancy of 61:39. After
modeling this disorder the split hydride positions H40 and H41
The bond lengths around the Pd atom (Table 1) are all in the
expected range, with palladium phosphorus distances between
2.18 and 2.22 Å.10 For the 2,6-lutidine-coordinated complexes
1a−e-lut, the Pd−N and the Pd−P distances increase slightly in
the order o-MeO < Ph < o-Tol < iPr < Mes from 2.120 Å to 2.139
Å and from 2.191 Å to 2.228 Å, respectively. Concerning the
effect of the labile ligand L, the Pd−P distance (trans to the labile
ligand) elongates in the order dmso < pyridine < 2,6-lutidine
since dmso exerts less of a trans influence on the phosphorus
atom than an N-donor. Furthermore, dmso exhibits κ-O
coordination to the Pd(II) center in the solid-state structure of
1d-dmso7 and 1e-dmso, which is in contrast to the phosphine-
sulfonato Pd(II) complex [(κ2-P,O)-P[o-(2′-6′-(OMe)2C6H3)-
E
dx.doi.org/10.1021/om300755j | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Figure 5. Transition-state geometries of 1,2-insertion [left: a and c] and 2,1-insertion [middle: b and d] of MA into fragments 1a (top) and 1e (bottom).
Right: Steric maps of fragments 1a and 1e. The colored scale indicates the isocontour levels, in angstroms. Orientation of the ligands in the steric maps as
in (a) and (b), (c) and (d). Data for fragment 1e were taken from ref 7. Energies in kJ mol−1.
were located in the electron density map, while all other
hydrogen atoms were refined by use of a riding model. As in the
other crystallized diazaphospholidine-sulfonato complexes (vide
supra), the labile neutral monodentate ligand 2,6-lutidine is
coordinated trans to the phosphorus atom.
within minutes at room temperature, and the reaction mixture
was monitored by proton NMR spectroscopy at 50 °C. In all
cases, the Pd-CH3 signal decreases and a simultaneous increase of
characteristic signals for 2,1- and 1,2-insertion was observed.15 β-
Hydride elimination after 2,1-insertion to methyl crotonate (13)
took place immediately after some 2,1-insertion product 11 had
formed, whereas the 1,2-insertion product 12 was stable under
these conditions. Due to this complex reaction scheme, only the
decrease of the Pd-CH3 signal could be analyzed by a linear fit for
a pseudo-first-order reaction. The obtained rate constants at 50
°C are kobs = (5.0 0.1) × 10−5 s−1, kobs = (3.7 0.1) × 10−5 s−1,
and kobs = (1.2 0.1) × 10−4 s−1 for the decay of the Pd-CH3
signal of 1a-ClLi(acetone), 1b-ClLi(acetone), and 1c-ClLi-
(acetone), respectively. These numbers compare to the reported
values of 1d-ClLi(actone) at 25 °C and 1e-ClLi(thf) at 45 °C of
kobs = (6.0 0.1) × 10−4 s−1 and kobs = (4.8 0.1) × 10−4 s−1,
respectively.7
A quantitative analysis of the formed insertion products 12
(after 1,2-insertion) and methyl crotonate (13) (after 2,1-
insertion and β-hydride elimination) by 1H NMR spectroscopy
reveals that a decrease of the steric bulk of the N-aryl moieties in
diazaphospholidine-sulfonato palladium methyl complexes 1-L
indeed inverts the insertion regiochemistry of methyl acrylate
from 1,2-insertion (1d-L and 1e-L)7 to 2,1-insertion: The
observed ratios of 12 to methyl crotonate (13) are 20:1 (for 1d,
1e), ca. 1:1 (for 1b, 1c), and 1:23 (for 1a) (Scheme 3).
Regiochemistry of Insertion. For polar vinyl monomers,
the regioselectivity of insertion into a transition metal carbon
bond, and in particular a palladium−carbon bond, is usually
electronically controlled. This applies to polymerization as well
as transition metal catalyzed cross-coupling reactions.2i,11 Thus,
electron-deficient olefins such as methyl acrylate selectively
insert in a 2,1-fashion,3,4,12 whereas electron-rich olefins such as
vinyl ethers have a strong preference for 1,2-insertion.5a,13,14
As we have shown previously in a combined experimental and
computational study,7 the regiochemistry of insertion of
electron-deficient polar monomers such as methyl acrylate can
be inverted from an electronically preferred 2,1- to a 1,2-insertion
by an appropriately arranged steric bulk of the ligand. Complexes
1e-ClLi(acetone) and 1d-ClLi(acetone) yielded the “regioirre-
gular” 1,2-insertion product 12 in over 93% NMR yield, due to a
destabilization of the transition state of 2,1-insertion. This
proposed concept of sterical destabilization of the transition state
of 2,1-insertion also implies that by decreasing the steric bulk of
the ligand, the insertion should be electronically controlled to
favor the 2,1-insertion product (11) of MA into the Pd−CH3
bond of diazaphospholidine-sulfonato complexes 1-L, which
upon β-hydride elimination, yield methyl crotonate (13) and the
corresponding hydride complex 9-L (Scheme 2).
Complex 1f-ClLi(CD3OD) was not stable after silver-
mediated chloride abstraction and decomposed during the
insertion experiment at 50 °C, evident by the rapid formation of
palladium black. Therefore, the more stable pyridine complex 1f-
pyr was investigated in its reactivity toward MA. Insertion of MA
into the Pd−methyl bond was monitored in C2D2Cl4 via proton
NMR spectroscopy at elevated temperature (80 °C). The Pd-
CH3 signal decays under these conditions with a pseudo-first-
To this end, the reaction of the electron-deficient polar olefin
methyl acrylate toward the sterically less demanding complexes
1a-ClLi(acetone), 1b-ClLi(acetone), and 1c-ClLi(acetone) was
monitored by NMR spectroscopy. To a methylene chloride-d2
solution of the complex (30 μmol in 0.6 mL), one equivalent of
silver tetrafluoroborate and 20 to 25 equivalents of MA were
added. The silver-mediated chloride abstraction took place
F
dx.doi.org/10.1021/om300755j | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Scheme 4. Possible Insertion Pathways of MA into the Pd(II)-Hydride Complex 9e-lut
a
Scheme 5. Reaction of 10e-lut with 12 Equivalents of MA in C6D6 at Room Temperature
a
Exclusive 2,1-insertion of methyl acrylate into 10e-lut is proven by H/D exchange only at the terminal position of methyl acrylate.
order rate constant of kobs = (1.2 0.1) × 10−4 s−1, while only the
formation of the 1,2-insertion product was observed.
formation of palladium black was observed during reaction
overnight. This implies that the investigated diazaphospholidine-
sulfonato complexes are somewhat unstable under these
conditions. Also, in all cases, no insertion of MA into a Pd−
hydride bond, which should be present in the reaction mixture
after β-hydride elimination, was observed. To further elucidate
this issue, the reactivity of the isolated hydride complex 9e-lut
toward MA was studied. NMR spectroscopic monitoring of a
solution of 9e-lut and MA in CD2Cl2 at 40 °C did not provide any
evidence of an insertion, even at prolonged reaction times
(overnight). Only slow decomposition by reductive elimination,
evidenced by the formation of palladium black and the lutidinium
salt of the diazaphospholidine ligand 7e ([{N-(2,6-
iPr2C6H3)2C2H4N2P}C6H4SO3]−[C7H10NH]+), was observed.
Even with a large excess of MA (ca. 200 equiv), no insertion
products were observed over a wide temperature range studied
(−80 to 25 °C).
When, however, the analogous deuteride complex 10e-lut is
reacted with 12 equivalents of MA in benzene-d6, an immediate
increase of the Pd-hydride signal from 30% (from 9e-lut as an
impurity in the deuteride complex, vide supra) to over 90% is
observed within a few minutes at room temperature. Also, the
phosphorus resonance at 88.39 ppm, a 1:1:1 triplet with a
coupling constant of 3.6 Hz (2JP−D), decreases in favor of a
distinct doublet at 88.18 ppm with a coupling constant of 26.4 Hz
(2JP−H). This clearly shows that insertion into the Pd−D/H bond
takes place very rapidly, but since no insertion product was
detected, immediate β-hydride elimination appears to be even
faster such that the insertion/elimination equilibrium is on the
side of the hydride complex 9e-lut and free MA (Scheme 4).
2H NMR spectroscopy of the reaction mixture in C6D6
revealed the nature of the previous insertion mode. Only two
resonances, at 5.26 and 6.23 ppm, for the terminal olefin, but no
resonance for the internal olefin (expected at 5.91 ppm) were
observed. This clearly shows that only 2,1-insertion into the Pd−
D bond of 10e-lut occurs (Scheme 5). The intermediate
formation of the electronically favored 2,1-insertion product is
further backed up by the observed ingrowth of two triplets at
The observed ratios of 1,2- and 2,1-insertion are in qualitative
agreement with density functional theory calculations at the
BP86 generalized gradient approximation level16 performed on
some of these complexes. MA insertion proceeds by η2-
coordination of MA to the palladium(II) center trans to the
phosphorus donor atom, followed by cis−trans isomerization,
and insertion in the less stable cis isomer.
Previous calculations revealed that the mesityl and 2,6-
diisopropylphenyl moieties of fragments 1d and 1e are in close
proximity to the methoxycarbonyl group of the incoming MA
substrate, which forces the latter to rotate out of plane. This
́
destabilizes the Cossee−Arlmann-like transition state of the
electronically preferred 2,1-insertion. As a result, the transition
state of 1,2-insertion is kinetically favored over the transition
state of 2,1-insertion by an energy difference of 3 and 9 kJ mol−1
for the fragments 1d (N-aryl = mesityl) and 1e (N-aryl = 2,6-
diisopropylphenyl), respectively.7
DFT calculations for fragment 1a (N-aryl = phenyl) now
confirm that for this sterically less constrained metal center
indeed the transition state for 2,1-insertion of methyl acrylate is
favored by 12 kJ mol−1 over 1,2-insertion (Figure 5). This agrees
with the experimentally observed regiochemistry of methyl
acrylate insertion into 1a (vide supra). Further calculations for an
N-methyl-substituted diazaphospholidine show that for this even
smaller substituent 2,1-insertion is even more pronouncedly
favored (18 kJ mol−1 vs 1,2-insertion), which further illustrates
the role of these substituents for regioselectivity (cf. Supporting
Information, Figure S7.3).
The influence of the appropriately arranged steric bulk of the
ligands of complexes 1a (Ar = phenyl) and 1e (Ar = 2,6-
diisopropylphenyl) on the transition state of MA insertion is
illustrated in the calculated steric maps (Figure 5). Fragment 1a
features a very flat topology, which shows no strong influence of
steric factors.
Besides the two aforementioned opposite insertion products
and the β-hydride elimination product methyl crotonate, the
G
dx.doi.org/10.1021/om300755j | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Figure 6. Transition-state geometries for MA insertion in phosphine-sulfonato Pd(II) hydride complex (top) in 1,2-mode (a) and 2,1-mode (b). MA
insertion into diazaphospholidine-sulfonato Pd(II) hydride complex 9e-MA (bottom) in 1,2-mode (c) and 2,1-mode (d). Right: Steric maps of the
ligand spheres of complex 2 (top) and 1e/9e (bottom). The colored scheme for the isocontour profiles, in Å, is reported on the right. Orientation of the
ligands in the steric maps as in (a) and (b), (c) and (d). Energies in kJ mol−1.
130.3 ppm with a characteristic coupling constant of 24.7 Hz
barrier is higher compared to the related phosphine-sulfonato
system 2 (15 kJ mol−1).
(1JC−D) in the 13C NMR spectrum of a stoichiometric reaction of
2
10e-lut and MA in CH2Cl2. The H NMR spectrum of this
Ethylene Homopolymerization Studies. Beyond these
stoichiometric studies of insertion, the catalytic properties of
complexes 1-L were studied. All compounds are precursors to
active catalysts for ethylene homopolymerization (Table 2).
Productivities are limited by comparison to the phosphine-
sulfonato complex 2-L (L = dmso, pyridine, 2,6-lutidine) at
comparable conditions.4 The low activity also arises from a
reduced stability at elevated temperatures for the diazaphospho-
lidine complexes, particularly 1d-L and 1e-L, as evidenced by a
decreasing ethylene uptake with time during polymerization
experiments (Figure 7).
stoichiometric reaction mixture features the terminal olefinic
resonances (6.38 and 5.83 ppm in CH2Cl2) and the Pd-deuteride
complex 10e-lut (−19.32 ppm) in a statistical ratio of
approximately 1:1:1 (see Supporting Information).
This 2,1-insertion of methyl acrylate into the palladium
hydride bond of 10e-lut is unexpected since methyl acrylate
insertion into the corresponding palladium methyl fragment 1e
proceeds with 1,2-insertion regioselectivity.7 The origin of this
opposite reactivity was revealed by DFT studies. For the methyl
complexes 1e-MA and 2-MA, in agreement with calculations
reported for 2-MA17 and for 1e-MA,7 methyl acrylate initially
coordinates trans to the phosphorus atom, then undergoes a trans
to cis isomerization. Insertion proceeds from this cis-coordinated
intermediate since the corresponding cis transition state is lower
in energy than the trans transition state (Figure S7.6, top).7,17
The regioselectivity is controlled by the steric bulk of the ligand,
and the 2,1-insertion mode is disfavored by 9 kJ mol−1 vs the 1,2-
insertion for 1e (vide supra).7 In the case of the hydride or
deuteride complex (9e-MA or 10e-MA), consistently with
Nozaki’s results on the olefin insertion into the hydride complex
analogous to 2,17b calculations reveal that the trans transition
state for 2,1-insertion of MA is favored over the cis transition state
(Figure S7.5 and Figure 6). As a consequence, reaction occurs by
MA coordination trans to the phosphorus atom18 followed by
2,1-insertion from this trans geometry (Figure S7.6, bottom). In
other words, no trans to cis isomerization is necessary. In the trans
complex, 1,2-insertion is disfavored by 24 kJ mol−1 vs 2,1-
insertion (Figure 6). Due to the cooperative electronic and steric
effects (see Figure 6, steric maps), this difference in insertion
Complexes 1d-lut (Ar = mesityl) and 1e-lut (Ar = 2,6-
iPr2C6H3) for example exhibit their highest productivity at 50 °C
(Table 2, entries 2−18 and 2−24), whereas complexes 1a-lut (Ar
= phenyl) and 1b-lut (Ar = o-tolyl) reach their highest
productivity of 2.0 × 103 molC2H4 molPd and 3.6 × 103
−1
molC2H4 molPd−1, respectively, at 90 °C (Table 2, entries 2-3
and 2-14). The terphenyl-substituted complex 1f-pyr did not
yield significant amounts of polymer.
As expected, complexes of the weaker coordinating ligand
DMSO display higher activities (compare for example entries 2-3
and 2-8 or entries 2-18 and 2-21), as it competes less with
monomer binding by comparison to lutidine or pyridine. Also,
higher ethylene pressure leads to an increase in catalyst activity
(compare entries 2-4 and 2-6).
Molecular weights of the isolated polymers depend strongly
on the diazaphospholidine aryl substitution pattern. The least
bulky complex 1a-lut (Ar = phenyl) produced polyethylene with
a molecular weight of up to 35 000 g mol−1, which is on the same
order of magnitude as found for the phosphine-sulfonato
H
dx.doi.org/10.1021/om300755j | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
a
Table 2. Ethylene Polymerization by Neutral Diazaphospholidine-Sulfonato Pd(II) Complexes 1-L
c
entry
cat. precursor
pressure [bar]
temperature [°C]
yield [mg]
TON [(mol C2H4) (mol Pd)−1
]
Mn [g mol−1
]
2-1
1a-lut
1a-lut
1a-lut
1a-lut
1a-lut
1a-lut
10
10
10
5
50
70
90
90
90
90
70
90
50
70
90
50
70
90
90
90
90
50
70
90
50
70
40
50
70
50
50
90
2-2
40
275
167
354
534
40
146
984
n.d.
34 600
2-3
2-4
599
30 900
38 300
35 900
2-5
20
40
10
10
10
10
10
10
10
10
5
1268
1909
650
2-6
b
2-7
1a-dmso
n.d.
b
d
2-8
1a-dmso
120
46
1948
167
24 300
2-9
1b-lut
1b-lut
1b-lut
1c-lut
1c-lut
1c-lut
1c-lut
1c-lut
1c-lut
1d-lut
1d-lut
1d-lut
2580
1800
1800
2-10
2-11
2-12
2-13
2-14
2-15
2-16
2-17
2-18
2-19
2-20
2-21
2-22
2-23
2-24
2-25
2-26
2-27
2-28
131
103
468
369
91
501
277
748
944
900
784
428
278
64
326
1793
991
27 500
19800
20 200
20 000
1000
20
40
10
10
10
10
10
10
10
10
20
10
10
2673
3372
3217
2800
1531
4513
1039
3514
4992
3964
5107
600
600
b
1d-dmso
1500
b
d
1d-dmso
1200
1e-lut
1e-lut
1e-lut
1e-lut
1f-lut
1f-lut
983
1397
1109
1430
n.d.
900
700
600
700
63
228
4400
a
b
c
1
Reaction conditions: 100 mL of toluene, 30 min reaction time, 10 μmol of Pd(II). 2.2 μmol of Pd(II). Determined by H NMR in C2D2Cl4 at
d
130 °C. Determined by GPC in 1,2,4-trichlorobenzene at 160 °C vs linear polyethylene.
enhances productivity and molecular weight compared to the
methyl-substituted complex 1b-lut (compare entries 2-11 and 2-
14).
All catalyst precursors produce highly linear polyethylenes
according to quantitative high-temperature 13C NMR spectros-
copy. Only for polyethylenes formed by 1a-lut and 1b-lut a small
amount of methyl branches was detectable, two methyl branches
per 1000 carbon atoms, but no higher alkyl branches were found.
Copolymerization Studies. The most active catalyst, 1e-lut,
was also studied in copolymerization experiments of ethylene
and various comonomers, namely, methyl acrylate, styrene,
norbornene, 5-norbornene-2,3-dicarboxylic anhydride, vinyl
acetate, and vinyl ether (Table 3). Only the strained norbornene
was incorporated (Table 3, entries 3-6 and 3-7), while in the
presence of all other comonomers polyethylene homopolymer
was formed. Molecular weights of these polyethylenes are
between 600 and 900 g mol−1, which is in the same range as in
ethylene homopolymerization in the absence of comonomers.
Taking into account these molecular weights, the detection limit
of the aforementioned composition analysis is less than one
comonomer repeat unit per chain. That is, even comonomer-
derived end groups can be excluded. The polymer yield
decreased significantly in the presence of polar-substituted
comonomers by comparison to polymerization without
comonomer (entry 3-1). This decrease in productivity is likely
due to competing σ-coordination of the polar moiety, as
observed in detailed studies of the phosphine-sulfonato
complexes 2.21 This interpretation is also supported by the
finding that the presence of styrene did not reduce the yield as
Figure 7. Typical mass flow plot of ethylene homopolymerization with
1e-lut (20 μmol of Pd, 100 mL of toluene, 50 °C, 10 bar of ethylene, 30
min).
complexes 2. The more bulky complexes 1d-lut (Ar = mesityl)
and 1e-lut (Ar = 2,6-iPr2C6H3) are more active (vide supra), but
produce oligomeric material with molecular weights of about
1000 g mol−1 or less. This is somewhat unexpected since it is
thought that more bulky ligands increase catalyst activity and
molecular weight of polyethylene produced by phosphine-
sulfonato Pd(II) complexes,19 although this is not a clear trend.20
The electron-donating methoxy group of complex 1c-lut
I
dx.doi.org/10.1021/om300755j | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
a
Table 3. Results of Polymerization in the Presence of Ethylene and Various Comonomers by 1e-lut
b
CComonomer
yield
[mg]
TON ethylene [(mol C2H4)
TON comon. [(mol X)
incorp. comon.
[mol %]
Mn
entry
comonomer
[mol L−1
]
(mol Pd)−1
]
(mol Pd)−1
]
[g mol−1
]
3-1
3-2
3-3
3-4
3-5
3-6
3-7
3-8
2950
544
4213
777
900
1000
1000
800
methyl acrylate
0.1 M
0.3 M
0.1 M
0.3 M
0.1 M
0.3 M
0.1 M
<0.1%
<0.1%
<0.1%
<0.1%
3.5%
methyl acrylate
styrene
303
433
1869
1275
1744
2325
744
2670
1821
2404
3119
1063
styrene
800
norbornene
norbornene
26
60
700
6.1%
700
5-norbornene-2,3-dicarboxylic
anhydride
<0.1%
700
3-9
5-norbornene-2,3-dicarboxylic
anhydride
0.3 M
293
418
<0.1%
600
3-10 vinyl acetate
3-11 vinyl acetate
3-12 ethyl vinyl ether
3-13 ethyl vinyl ether
0.05 M
0.1 M
535
312
413
194
764
446
590
277
<0.1%
<0.1%
<0.1%
<0.1%
800
900
900
800
0.05 M
0.1 M
a
Reaction conditions: 100 mL total volume (comonomer + toluene), 30 min reaction time, 50 °C, 25 μmol of 1e-lut, 10 bar of ethylene, 200 mg of
b
1
BHT. Determined by H NMR in C2D2Cl4 at 130 °C.
pronouncedly (entries 3-4 and 3-5) and that productivity
decreases with increasing concentration of the polar como-
nomer.
polymerization conditions, incorporation of ethylene is much
preferred. This chemoselectivity can be related to a preference
for the less bulky ethylene monomer.
Copolymerizations of ethylene with 0.1 and 0.3 mol L−1
norbornene (entries 3-6 and entry 3-7) yielded oligomeric
ASSOCIATED CONTENT
* Supporting Information
■
1
materials as highly viscous oils with a molecular weight (by H
S
NMR) of about 700 g mol−1 and a norbornene content of 3.5 and
6.1 mol %, respectively. Isolated norbornene units are
incorporated into the linear ethylene backbone.22 This
incorporation of norbornene is about four times lower compared
to phosphine-sulfonato complexes 2, which produced copoly-
mers with 48 000 g mol−1 and 12 mol % norbornene
incorporation under similar conditions (24 μmol of Pd, 0.1
mol L−1, 20 bar).23 In general, the steric bulk of the
diazaphospholidine moiety appears to hamper incorporation of
comonomer vs ethylene incorporation.
Supplemental tables and figures, CIF files, general experimental
procedures, synthesis, additional NMR spectra, crystal struc-
tures, Cartesian coordinates, and energies of DFT calculations.
This material is available free of charge via the Internet at http://
pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
■
*Fax: +49 7531 88-5152. Tel: +49 7531 88-5151. E-mail: stefan.
mecking@uni-konstanz.de; lcaporaso@unisa.it.
Notes
SUMMARY AND CONCLUSIONS
■
The authors declare no competing financial interest.
This comprehensive study of new diazaphospholidine-sulfonato
Pd(II) complexes reveals that within this class of compounds the
regioselectivity of acrylate insertion can be inverted via the
bidentate ligand. Depending on the steric bulk of the N-aryl
substituents, insertion can occur very selectively (>95%) in either
a 2,1- or 1,2-fashion. This is due to a destabilization of the 2,1-
insertion transition state by interference of bulky substituents
with the coordinated acrylate substrate that overrides the
electronic preference for this insertion mode,7 while for less
sterically demanding substituents the 2,1-insertion transition
state was found to be energetically favorable vs the transition
state of 1,2-insertion by DFT studies. In stark contrast, acrylate
insertion into diazaphospholidine hydride (or deuteride)
complexes was observed to follow the common 2,1-insertion
pathway even for very bulky substituted complexes (R = 2,6-
iPr2C6H3). Theoretical studies reveal the origin of this different
behavior. Other than insertion into a metal−alkyl bond, insertion
into the hydride occurs from the species in which the π-
coordinated acrylate is trans to the P-donor, with the hydride in
cis position. Thus, the olefin is more remote from the P-donor
and its substituents, and the insertion step is less sensitive to their
steric bulk. All representatives of diazaphospholidine-sulfonato
complexes studied were found to be active for ethylene
polymerization. In the presence of various comonomers, under
ACKNOWLEDGMENTS
■
Financial support by the DFG (Me1388/10-1) is gratefully
acknowledged. The authors thank Lars Bolk for GPC, Anke
Friemel and Ulrich Haunz for support with NMR measurements,
and the HPC team of Enea (www.enea.it) for use of the ENEA-
GRID and the HPC facilities CRESCO (www.cresco.enea.it) in
Portici, Italy.
REFERENCES
■
(1) Mulhaupt, R. Macromol. Chem. Phys. 2003, 204, 289−327.
̈
(2) (a) Ittel, S. D.; Johnson, L. K.; Brookhart, M. Chem. Rev. 2000, 100,
1169−1203. (b) Gibson, V. C.; Spitzmesser, S. K. Chem. Rev. 2003, 103,
283−316. (c) Mecking, S. Coord. Chem. Rev. 2000, 203, 325−351.
(d) Mecking, S. Angew. Chem., Int. Ed. 2001, 40, 534−540. (e) Guan, Z.
Chem. Eur. J. 2002, 8, 3086−3092. (f) Domski, G. J.; Rose, J. M.; Coates,
G. M.; Bolig, A. D.; Brookhart, M. Prog. Polym. Sci. 2007, 32, 30−92.
(g) Berkefeld, A.; Mecking, S. Angew. Chem., Int. Ed. 2008, 47, 2538−
2542. (h) Chen, E. Y.-X. Chem. Rev. 2009, 109, 5157−5214.
(i) Nakamura, A.; Ito, S.; Nozaki, K. Chem. Rev. 2009, 109, 5215−5244.
(3) (a) Johnson, L. K.; Mecking, S.; Brookhart, M. J. Am. Chem. Soc.
1996, 118, 267−268. (b) Mecking, S.; Johnson, L. K.; Wang, L.;
Brookhart, M. J. Am. Chem. Soc. 1998, 120, 888−899. (c) Johnson, L. K.;
Bennett, A.; Dobbs, K.; Hauptamnn, E.; Ionkin, A.; McCord, E.;
J
dx.doi.org/10.1021/om300755j | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
McLain, S.; Radzewich, C.; Yin, Z.; Wang, L.; Wang, Y.; Brookhart, M.
Polym. Mater. Sci. Eng. 2002, 86, 319.
(4) (a) Drent, E.; van Dijk, R.; van Ginkel, R.; van Oort, B.; Pugh, R. I.
Chem. Commun. 2002, 744−745. (b) Guironnet, D.; Roesle, P.; Runzi,
̈
T.; Gottker-Schnetmann, I.; Mecking, S. J. Am. Chem. Soc. 2009, 131,
422−423.
̈
(5) (a) Luo, S.; Vela, J.; Lief, G. R.; Jordan, R. F. J. Am. Chem. Soc. 2007,
129, 8946−8947. (b) Weng, W.; Shen, Z.; Jordan, R. F. J. Am. Chem. Soc.
2007, 129, 15450−15451. (c) Skupov, K. M.; Piche, L.; Claverie, J. P.
Macromolecules 2008, 41, 2309−2310. (d) Ito, S.; Munakata, K.;
Nakamura, A.; Nozaki, K. J. Am. Chem. Soc. 2009, 131, 14606−14607.
(e) Bouilhac, C.; Runzi, T.; Mecking, S. Macromolecules 2010, 43, 3589−
̈
3590. (f) Runzi, T.; Guironnet, D.; Gottker-Schnetmann, I.; Mecking, S.
̈
̈
J. Am. Chem. Soc. 2010, 132, 16623−16630. (g) Runzi, T.; Frohlich, D.;
̈
̈
Mecking, S. J. Am. Chem. Soc. 2010, 132, 17690−17691. (h) Daigle, J.-
C.; Piche, L.; Claverie, J. P. Macromolecules 2011, 44, 1760−1762.
(i) Daigle, J.-C.; Piche, L.; Arnold, A.; Claverie, J. P. ACS Macro Lett.
2012, 1, 343−346.
(6) Nozaki, K.; Kusumoto, S.; Noda, S.; Kochi, T.; Chung, L. W.;
Morokuma, K. J. Am. Chem. Soc. 2010, 129, 16030−16042.
(7) Wucher, P.; Caporaso, L.; Roesle, P.; Ragone, F.; Cavallo, L.;
Mecking, S.; Gottker-Schnetmann, I. Proc. Natl. Acad. Sci. U. S. A. 2011,
̈
108, 8955−8959.
(8) Abrams, M. B.; Scott, B. L.; Baker, R. T. Organometallics 2000, 19,
4944−4956.
(9) Rulke, R. E.; Ernsting, J. M.; Spek, A. L.; Elsevier, C. J.; van
̈
Leeuwen, P. W. N. M.; Vrieze, K. Inorg. Chem. 1993, 32, 5769−5778.
(10) Pd−N and Pd−P distances of phosphine-sulfonato Pd(II)
lutidine and pyridine complexes: Pd−P: 2.2 Å, Pd−N: 2.1 Å; ref 20. For
further examples cf. refs 5a, d, 6, 11, and 19.
(11) Beletskaya, I. P.; Cheprakov, A. V. Chem. Rev. 2000, 100, 3009−
3066.
(12) Kochi, T.; Noda, S.; Yoshimura, K.; Nozaki, K. J. Am. Chem. Soc.
2007, 129, 8948−8949.
(13) Luo, S.; Jordan, R. F. J. Am. Chem. Soc. 2006, 128, 12072−12073.
(14) Cabri, W.; Candiani, I. Acc. Chem. Res. 1995, 28, 2−7.
(15) The intermediately formed 2,1-insertion product was not fully
characterized due to its low concentration. However, the formation of
the β-hydride elimination product methyl crotonate together with a
simultaneous decrease of a characteristic high-field triplet (0.3−0.5
ppm) unambiguously confirm the identity of this intermediate.
(16) (a) Becke, A. D. Phys. Rev. A 1988, 38, 3098−3100. (b) Perdew, J.
P. Phys. Rev. B 1986, 33, 8822−8824. (c) Perdew, J. P. Phys. Rev. B 1986,
34, 7406−7406.
(17) (a) Guironnet, D.; Caporaso, L.; Neuwald, B.; Gottker-
̈
Schnetmann, I.; Cavallo, L.; Mecking, S. J. Am. Chem. Soc. 2010, 132,
4418−4426. (b) Noda, S.; Nakamura, A.; Kochi, T.; Chung, L. W.;
Morokuma, K.; Nozaki, K. J. Am. Chem. Soc. 2009, 131, 14088−14100.
(c) Haras, A.; Anderson, G. D. W.; Michalak, A.; Rieger, B.; Ziegler, T.
Organometallics 2006, 25, 4491−4497.
(18) In line with Nozaki’s results (ref 17b), the trans-coordinated
intermediate is lower in energy than the cis intermediate for both
complexes 2 and 9e-MA. In detail, the trans intermediate is lower than
the cis intermediate by about 42 and 69 kJ mol−1 for Pd-H complex 2 and
complex 9e-MA, respectively (Figure S7.7).
(19) Skupov, K. M.; Marella, P. R.; Simard, M.; Yap, G. P. A.; Allen, N.;
Conner, D.; Goodall, B. L.; Claverie, J. P. Macromol. Rapid Commun.
2007, 28, 2033−2038.
(20) Piche, L.; Daigle, J.-C.; Poli, R.; Claverie, J. P. Eur. J. Inorg. Chem.
2010, 29, 4595−4601.
(21) Friedberger, T.; Wucher, P.; Mecking, S. J. Am. Chem. Soc. 2012,
134, 1010−1018.
(22) For characterization cf.: (a) Wehrmann, P.; Zuideveld, M.;
Thomann, R.; Mecking, S. Macromolecules 2006, 39, 5995−6002.
(b) Tritto, I.; Marestin, C.; Boggioni, L.; Zetta, L.; Provasoli, A.; Ferro,
D. R. Macromolecules 2000, 33, 8931−8944.
(23) Skupov, K. M.; Marella, P. R.; Hobbs, J. L.; McIntosh, L. H.;
Goodall, B. L.; Claverie, J. P. Macromolecules 2006, 39, 4279−4281.
K
dx.doi.org/10.1021/om300755j | Organometallics XXXX, XXX, XXX−XXX