Concepts for stereoselective acrylate insertion

Article abstract of DOI:10.1021/ja3101787

Various phosphinesulfonato ligands and the corresponding palladium complexes [{((P^aO)PdMeCl)-μ-M}_n] ([{( ^X1-Cl)-μ-M}_n], (P^aO) = κ²- P,O-Ar₂PC₆H₄SO₂O) with symmetric (Ar = 2-MeOC₆H₄, 2-CF₃C₆H₄, 2,6-(MeO)₂C₆H₃, 2,6-(iPrO)₂C ₆H₃, 2-(2′,6′-(MeO)₂C ₆H₃)C₆H₄) and asymmetric substituted phosphorus atoms (Ar¹ = 2,6-(MeO)₂C₆H ₃, Ar² = 2′-(2,6-(MeO)₂C ₆H₃)C₆H₄; Ar¹ = 2,6-(MeO)₂C₆H₃, Ar² = 2-cHexOC ₆H₄) were synthesized. Analyses of molecular motions and dynamics by variable temperature NMR studies and line shape analysis were performed for the free ligands and the complexes. The highest barriers of ΔG^a = 44-64 kJ/mol were assigned to an aryl rotation process, and the flexibility of the ligand framework was found to be a key obstacle to a more effective stereocontrol. An increase of steric bulk at the aryl substituents raises the motional barriers but diminishes insertion rates and regioselectivity. The stereoselectivity of the first and the second methyl acrylate (MA) insertion into the Pd-Me bond of in situ generated complexes ^X1 was investigated by NMR and DFT methods. The substitution pattern of the ligand clearly affects the first MA insertion, resulting in a stereoselectivity of up to 6:1 for complexes with an asymmetric substituted phosphorus. In the consecutive insertion, the stereoselectivity is diminished in all cases. DFT analysis of the corresponding insertion transition states revealed that a selectivity for the first insertion with asymmetric (P ^aO) complexes is diminished in the consecutive insertions due to uncooperatively working enantiomorphic and chain end stereocontrol. From these observations, further concepts are developed.

Full text of DOI:10.1021/ja3101787

Article
pubs.acs.org/JACS
Concepts for Stereoselective Acrylate Insertion
Boris Neuwald,† Lucia Caporaso,*,‡ Luigi Cavallo,§ and Stefan Mecking*,†
†Chair of Chemical Materials Science, Department of Chemistry, University of Konstanz, 78464 Konstanz, Germany
‡Department of Chemistry and Biology, University of Salerno, Via Ponte Don Melillo, 84084-Fisciano (SA), Italy
§Physical Sciences and Engineering, Kaust Catalysis Center, King Abdullah University of Science and Technology (KAUST), Thuwal
23955-6900, Saudi Arabia
S
* Supporting Information
ABSTRACT: Various phosphinesulfonato ligands and the
corresponding palladium complexes [{((P∧O)PdMeCl)-μ-
M}n] ([{(X1-Cl)-μ-M}n], (P∧O) = κ2-P,O-Ar2PC6H4SO2O)
with symmetric (Ar = 2-MeOC6H4, 2-CF3C6H4, 2,6-
(MeO)2C6H3, 2,6-(iPrO)2C6H3, 2-(2′,6′-(MeO)2C6H3)C6H4)
and asymmetric substituted phosphorus atoms (Ar1 = 2,6-
(MeO)2C6H3, Ar2 = 2′-(2,6-(MeO)2C6H3)C6H4; Ar1 = 2,6-
(MeO)2C6H3, Ar2 = 2-cHexOC6H4) were synthesized.
Analyses of molecular motions and dynamics by variable
temperature NMR studies and line shape analysis were
performed for the free ligands and the complexes. The highest barriers of ΔG⧧ = 44−64 kJ/mol were assigned to an aryl
rotation process, and the flexibility of the ligand framework was found to be a key obstacle to a more effective stereocontrol. An
increase of steric bulk at the aryl substituents raises the motional barriers but diminishes insertion rates and regioselectivity. The
stereoselectivity of the first and the second methyl acrylate (MA) insertion into the Pd−Me bond of in situ generated complexes
X1 was investigated by NMR and DFT methods. The substitution pattern of the ligand clearly affects the first MA insertion,
resulting in a stereoselectivity of up to 6:1 for complexes with an asymmetric substituted phosphorus. In the consecutive
insertion, the stereoselectivity is diminished in all cases. DFT analysis of the corresponding insertion transition states revealed
that a selectivity for the first insertion with asymmetric (P∧O) complexes is diminished in the consecutive insertions due to
uncooperatively working enantiomorphic and chain end stereocontrol. From these observations, further concepts are developed.
complexes,6 or by free radical polymerization at low temper-
INTRODUCTION
■
atures.7
The regularity of stereocenters in a polymeric chain can
Conclusively so far not only a broad and universal way of
profoundly affect the macroscopic material properties. This is
influencing the stereocontrol of alkyl acrylate polymerization is
illustrated most impressively by isotactic polypropylene,
missing, but there is also a conceptual gap to an effective
produced on the 107 ton/year scale. These materials are
tunable control mechanism. Insertion polymerization is a very
prepared by catalytic insertion polymerization, in principle the
effective and versatile approach to control the tacticity in
most powerful and generic concept for stereoregular polymer-
polypropylene homo- and copolymers as underlined by the
successful industrial application of metallocene catalysts.8,9 The
ization. However, stereoselective insertion polymerization is
restricted to date to nonpolar olefins, and excludes polar vinyl
monomers. Isotactic and syndiotactic PMMA can be prepared
clear advantage of insertion polymerization for stereocontrol is
that chain growth proceeds directly at the metal center to which
by a coordination−addition mechanism akin to anionic
polymerization with early transition metal catalysts. However,
acrylates are not amenable to these polymerizations, due to the
sterically less bulky and chemically more reactive proton in α-
position.1
Isotactic polymers can be obtained from alkyl acrylates by
anionic polymerization at low temperatures,2,3 with chiral
zirconocenes by a proposed site-controlled addition mecha-
nism,4 or with chiral auxiliary controlled free radical polymer-
ization.5 Preparation of syndiotactic material proved to be even
more difficult. Slightly syndioenriched (up to ∼63% r-dyads)
poly(tert-butyl acrylate) can be obtained if the polymerization is
initiated by tBuLi in the presence of organoaluminium
the growing chiral polymer chain and an (un)symmetric ligand
are attached in close proximity, enabling control of the
insertion mode of the next prochiral monomer via a chain
end or enantiomorphic site stereocontrol mechanism.
Before this general background, stereoselective alkyl acrylate
polymerization by insertion in principle appears a promising
approach. However, insertion polymerization of polar-sub-
stituted vinyl monomers had remained elusive for a long time.
Copolymerization of polar vinyl monomers with ethylene was
achieved first with cationic Pd(II) diimine complexes. In this
Received: August 16, 2012
Published: December 20, 2012
© 2012 American Chemical Society
1026
dx.doi.org/10.1021/ja3101787 | J. Am. Chem. Soc. 2013, 135, 1026−1036

Journal of the American Chemical Society
Article
case, highly branched copolymers, which consist of ethylene as
major component (≥75 mol %) and contain acrylate units at
the end of the branches preferentially, were obtained.10−12 The
extensive chain walking appears problematic in advancing
toward any stereoselective polymerizations. To this end, the
discovery that neutral arylphosphinesulfonato Pd(II) complexes
are able to copolymerize methyl acrylate (MA) and ethylene to
linear random copolymers represented a major improvement.13
The development of weakly coordinated catalyst precursors
([(P∧O)PdMe(L)] with L = DMSO, also cf., Scheme 1) or
center is significantly less crowded than for the aforementioned
early transition metal catalysts. However, the stereoselective Pd-
catalyzed copolymerization of CO with vinyl monomers like
propylene and styrene has been described. Achiral, symmetric
diimine, or bipyridine (N∧N) ligands lead to syndiotactic
styrene copolymers by a chain-end stereocontrol, whereas the
enantiomorphic site-controlled formation of isotactic polymer
was described for C2 and C1 symmetric (N∧N) and (P∧N)
ligands.1,21−23 In addition, also the formation of highly isotactic
propylene/CO copolymers has been described for bisphos-
phine and phosphine/phosphite ligands of various symme-
tries.23−26
Scheme 1. Numbering of Complexes
For (P∧O)Pd systems, the bidentate phosphine-sulfonato
ligand exhibits two completely different coordination sites at
the metal, and the exact combination of the hard sulfonate with
the soft phosphine donor seems to be a prerequisite for the
unique properties of this catalytic system.27,28 Whereas the
steric bulk of the phosphine substituents can be adjusted and
will have a distinct influence on the active center, the sulfonic
acid moiety is rather small and does not offer a possibility for
steric variation. This results in two completely open sides
around the SO3-group, and in general a relatively high flexibility
of the ligand (Figure 1). Crystallographic structures and DFT
Figure 1. Analysis of square planar (P∧O)Pd catalyst in regard to
possible stereoselectivity. The grayer top-left quadrant, where the
pseudo-axial Ar ligand is located, is sterically more crowded than the
other quadrants.
rather in situ generated complexes entirely free of an additional
monodentate ligand L allowed for isolation of ethylene
copolymers with polar monomer incorporation of up to 50
mol %. In addition, these weakly coordinated catalyst
precursors enabled insertion homopolymerization of MA to
products with degrees of oligomerization up to ca. DPn = 5−
7.14,15 A mechanistic study of the underlying steps clearly
revealed that this acrylate homopolymerization occurs by an
insertion mechanism. For example, a six-membered chelate
complex [(P∧O)Pd{κ2-C,O-CH(C(O)OMe)CH2CH(C(O)-
OMe)CH2CH3}] results from two consecutive MA insertions
into the Pd−Me bond.16 Interestingly, this six-membered
chelate was obtained as a mixture of two diastereomers in a ca.
2:1 ratio arising from the configuration of the two stereocenters
at the methyl acrylate derived methine groups.
For comparison to proven concepts of stereocontrol in olefin
polymerization, early transition metal metallocene and
phenoxy-imine/amine post metallocene catalysts are instruc-
tive. These catalysts all show a tetrahedral or octahedral
coordination sphere around the central atom, which allows for a
space filling surrounding of the active center. Typically, the
multidentate ligand and the growing chain arrange and
determine the preferred monomer enantioface. An illustrative
exception is the C2 symmetric unbridged phenoxy-imine
catalyst, which is regarded as rather fluxional resulting in the
production of a syndiotactic polymer by a chain end control
mechanism.1,8,17−20
studies indicate that one of the nonbackbone aryl substituents
at phosphorus adopts a pseudoaxial position and is twisted
toward the palladium center, while the other substituent adopts
a pseudoequatorial position with a larger distance to the central
atom.16 Overall, this can be visualized in a quadrant scheme as
depicted in Figure 1 where only one quadrant is accentuated as
distinctly sterically crowded.
In contrast to the migratory insertion mechanism for early
transition metals, monomer insertion at (P∧O)Pd system only
occurs after cis/trans isomerization of the growing chain and
the coordinated monomer, so that insertion always occurs from
the isomer in which the olefin is cis to the P-donor.16,27 The
structural features enabling this acrylate insertion mechanism
are a highly unsymmetric bidentate ligand not only in terms of
soft/hard donor characteristics but also in terms of the steric
demand of the two donors.29 We now report an analysis of
stereocontrol mechanisms in such unsymmetrical environ-
ments.
RESULTS AND DISCUSSION
■
Acrylate insertions into phosphinesulfonato palladium methyl
complexes (X1) to afford the single (X2) and double (X3)
insertion products were investigated (Scheme 1).
Stereoselective insertion polymerization with late transition
metal catalysts has been less studied. In this case, the metal
1027
dx.doi.org/10.1021/ja3101787 | J. Am. Chem. Soc. 2013, 135, 1026−1036

Journal of the American Chemical Society
Article
cHexO/(MeO)21-lut, the double MeO-substituted aryl moiety is
placed in the pseudoequatorial and the cyclohexyloxy-
substituted aryl moiety in the pseudoaxial position, with the
cyclohexyl-group aligned away from the Pd-center. In contrast,
in Ar/(MeO)21-py the double MeO-substituted aryl moiety at
phosphorus is arranged in the pseudoaxial position. The axial/
equatorial alignment of the substituents does not depend on
the configuration at the phosphorus atom (R/S), and the
substituents determine the configuration of the (P∧O) chelates
in the solid state. Here, the more constrained substituent is
always located in the less crowded pseudoequatorial ring-
position. Consequently, the steric bulk of the substituents
increased from 2-cHexOC6H4- to 2,6-(MeO)2C6H3 to 2-(2′,6′-
(MeO)2C6H3)C6H4, and the Ar/(MeO)21 fragment is substantially
more crowded than the cHexO/(MeO)21 fragment. This is also
indicated by shorter distances of the aryl substituents to the
Pd−Me in Ar/(MeO)21-py (Ar/(MeO)21-py, C(1)−C(8) =
3.138(6), C(1)−O(7) = 3.017(5) versus cHexO/(MeO)21-lut,
C(1)−C(20) = 3.242(5), C(1)−C(9) = 3.527(5); cf.,
Supporting Information).
A range of substitution patterns at phosphorus differing in
steric bulk, flexibility, and symmetry were studied (Figure 2),
Analysis of Complex Flexibility. An essential point for
this study is the analysis of the structure and possible motions
of the (P∧O)Pd system, as well as the time scale of such
motions.32−34 The (P∧O)PdMe complex is capable of at least
two independent motions, which can have a significant
influence on the surrounding of the active center: A ring flip
of the six-membered chelate formed by coordination of the
bidentate ligand to the palladium center, and a rotation of the
two non−chelating P-bound aryl moieties around the P−C
bond (Figure 4). A comparison of all available X-ray structures
Figure 2. Phosphorus substitution patterns studied.
including the previously reported compounds MeO1-L, (MeO)21-
L, and Ar1-L.13,30,31 Commonly, the [(P∧O)PdMe] fragment
was generated in situ by halide abstraction from the cation-
bridged complexes [{(X1-Cl)-μ-M}n] for insertion studies. For
synthetic procedures and full characterization of all compounds,
see the Supporting Information.
The crystal structures of Ar/(MeO)21-lut and cHexO/(MeO)21-py
both show a square-planar geometry around the palladium
center, with the methyl group trans to the sulfonate group, as
observed generally for (P∧O)PdMe complexes (Figure 3).
For the asymmetric complexes, a racemic 1:1 mixture of the
two configurations at the phosphorus (R/S) is found. In
Figure 4. Observed molecular motions of (P∧O)Pd complexes
(fragment shown extracted from the X-ray structure of MeO1-
OPBu3).43
exhibiting the (P∧O)Pd motif revealed that in nearly all cases a
boatlike ring configuration is adopted by the six-membered
(P∧O)Pd chelates (for a detailed discussion, see the Supporting
Information).15,16,27,31,32,34−43 These chelate possess chirality
and exist in two enantiomeric forms, which are also commonly
found in the X-ray structures (cf., Supporting Information
Figure S3). The phosphorus and the sulfur atom are situated at
the ends of a boatlike conformation, which can flip from above
to below the plane created by the other ring atoms (Figure 4).
If this motion is fast as compared to monomer insertion, it
scrambles the steric information, thus preventing a possible
stereocontrol (vide infra).
Arylphosphinesulfonato ligands resemble an Ar3ZX structure
(here, X = lone pair, or H in the free ligand phosphonium
tautomer), where all four ligands are arranged approximately
tetrahedrally around the central atom Z. Because of a uniform
Figure 3. Solid-state structure of Ar/(MeO)21-py (top) and of
cHexO/(MeO)21-lut. Ellipsoids represent 50% probability. Hydrogen
atoms are omitted for clarity.
1028
dx.doi.org/10.1021/ja3101787 | J. Am. Chem. Soc. 2013, 135, 1026−1036

Journal of the American Chemical Society
Article
aryl twist, these molecules show a helical configuration implying
axial chirality (P or M).44 As a result of the ortho substitution of
the aryl moieties, various stereoisomeric configurations can be
adopted. Experimentally only two different arrangements, exo3
and exo2, have been observed for (P∧O)PdMe complexes: In
the exo3-configuration, all ortho substituents are situated on the
same side as the central phosphorus above a plane defined by
the three aryl carbons bound to phosphorus. In the exo2-
configuration, one of the ortho-substituents is situated
underneath this plane (Figure 5).
leads to an inequivalence of the aryl substituents, and the
resonances in the 1H NMR split. These new resonances
provide further insights into the active process. In general, it is
found here that the pair of corresponding ortho-aryl protons
experiences the largest shift difference, whereas the para proton
resonances remain nearly unchanged (cf., Supporting Informa-
tion Figure S9). This is in line with an aryl rotation process,
where the para-protons are within the axis of rotation. In
contrast, a ring flip should have more pronounced influence on
these para-protons.47 For the asymmetric compound, Ar/(MeO)2
1
formation of two diastereomers is observed at low temperature.
The formation of diastereomers can be related to the
combination of the permanent asymmetric center at
phosphorus with the stereocenter derived at low temperature
due to a fixed conformation (Figure S10, cf., Supporting
Information for a detailed discussion, Figures S11−S30).
To gain deeper insights into the energetic barriers of the
observed processes, dynamic NMR studies of the ligands
X(P∧O)H and the corresponding complexes were performed.
Rate constants at variable temperatures were determined by a
full line shape analysis, and an Eyring plot yielded ΔH⧧ and
ΔS⧧, respectively (Table 1, cf., Supporting Information Figures
Figure 5. Helical chirality (P,M) as found for Ar3ZX motifs and
additional diastereomers for the case of ortho substitution of the aryl
rings for [(P∧O)PdMe] complexes.
S37−S48). For a comparison ΔG⧧ (at the coalescence
Tc
temperature Tc) was calculated from ΔH⧧ and ΔS⧧, and also
derived independently from the distance of the resonances in
the slow exchange region (Δν) according to kc = π × Δν/√2
The exo2 configuration is observed for all [MeO(P∧O)PdMe-
(L)] structures.16,27,30,32,38−40,43 Sterically more demanding aryl
moieties such as −C6H4Et, −C6H4[2,6-(MeO)2C6H3], and 1-
methoxynaphatalene generally favor the sterically less encum-
bered exo3 configuration.14,30,32,33 Consequently, a rigid
configuration of the P-bound aryl moieties results in a
stereocenter and may allow for a stereocontrol. Interconversion
of the different isomers is only possible by correlated rotations
of the aryl substituents.45
and ΔG⧧ = 19.14 × Tc(10.32 + log(Tc/kc) J mol−1.48 The
Tc
latter relationship was also used for determination of ΔG⧧
Tc
when line shape analysis was not applicable, as was the case for
X
most of the (P∧O)H compounds.
The ΔG⧧ values obtained by the two different approaches
Tc
agree well, and the comparison of ΔG⧧Tc for MeO1 (ΔG⧧Tc = 44
kJ/mol, Tc = −50 °C, Table 1) determined for the MeO-groups
nicely fits with the value determined by Jordan et al. for
Dynamic NMR Studies. With regard to a stereoselection in
catalysis, the dynamics in solution are relevant. Temperature-
dependent NMR studies for the free ligands as well as for the
MeO1-py (ΔG⧧ = 44 kJ/mol, Tc = −50 °C).32 A comparison
Tc
X
of the different ligands reveals that for the protonated ligands
X(P∧O)H the rotational barrier increases with the steric bulk:
complexes 1 (generated in situ by chloride abstraction from
[{(X1-Cl)-μ-M}n], cf., Supporting Information) revealed that
only one dynamic process is observable in the temperature
range from −90 to 130 °C. This is in line with related studies
for MeO1-py, Et1-py, and Ar1-py.32,34 It is assumed that an aryl
rotation and not a ring flip process is observed in the NMR
studies. This is based on the facts that, first, the non chelated
free ligands (P∧O)H and their respective salts (P∧O)Na show a
similar process.46 Further, the freezing of molecular motions
MeO < CF3 ≈ Ar.49 The rotational barriers for complexes 1
X
are always at least 10 kJ/mol higher than for the free ligands
X(P∧O)H. The highest barriers were found for CF31 for which
even at 130 °C no fast interconversion was observed (ΔG⧧
>
Tc
76 kJ/mol). Because for Et1 also a higher barrier was observed
(ΔG⧧ = 64 kJ/mol)32 in comparison to MeO1 (ΔG⧧ = 44
Tc
Tc
kJ/mol), it appears that small changes in the β-position of the
Table 1. Results of Line Shape Analysis and Derived Energy Barriers for Various (P∧O)Ligands and (P∧O)PdMe Complexes
a
a
b
c
compound
resonance
Tc [°C]
ΔH⧧ [kJ/mol]
ΔS⧧ [kJ/mol]
ΔG⧧ [kJ/mol] / [kJ/mol]
Tc
d
MeO(P∧O)H
MeO1
CF3(P∧O)H
CF31
Ar(P∧O)H
Ar1
Ar/(MeO)2(P∧O)H
Ar/(MeO)2(P∧O)H
Ar/(MeO)2(P∧O)Na
Ar/(MeO)21
−OMe
−OMe
−CF3
−CF3
−OMe
−OMe
−OMe
P
≪−90
−50
25
<35
39(1)
46(2)
−17(4)
43(2)/44
−/59
e
>130
20
>76
−/59
45(3)/46
−43
−10
≪−90
−58
25
5(9)
f
−/52
d
<35
g
P
34(1)
48(1)
−15(7)
−27(1)
38(3)/39
g
PdMe
56(1)/59
a
b
c
Determined from Eyring plot. Calculated from ΔG⧧ = ΔH⧧ − T × ΔS⧧. Calculated from kc = πΔν/√2 and ΔG⧧Tc = 19.14 × Tc(10.32 + log(Tc/
d
e
f
kc) J mol−1. No hindered rotation observed down to −90 °C. No coalescence observed up to 130 °C. No formation of diastereomers observed,
g
only rotation of 2,6-(OMe)2C6H3− is hindered, and a process different from that for the corresponding sodium salt and complex occurs. Formation
of diastereomers observed.
1029
dx.doi.org/10.1021/ja3101787 | J. Am. Chem. Soc. 2013, 135, 1026−1036

Journal of the American Chemical Society
Article
ortho substituents (OMe vs CH2Me vs CF3) can have a
significant impact on the rotational barriers. The increase of the
rotational barriers of around 10−20 kJ/mol upon coordination
of the free ligands to a Pd-center agrees with previous
observations on arylphosphine complexes (cf., Supporting
Information for a detailed discussion). However, for the very
bulky −C6H4(2,6-(OMe)2C6H3) substituent, barriers for the
complex Ar1 are significantly lower than for the free ligand
low temperature. The coalescence temperature for the MeO-
resonances is in the same range as for MeO1, indicating that
rotational barriers are roughly similar and consequently that the
steric flexibility of the ligand is probably unchanged during an
oligomerization process upon chain growth (cf., Supporting
Information Figures S31−S33).
For a systematic investigation of stereoselectivity with
different catalysts, a rapid and broad analysis protocol is
essential. The formation of the two diastereomeric complexes
X3MA‑meso and X3MA‑rac allows for a screening of stereoselectivity
for the first two insertions, as these compounds are rather stable
also in solution (t1/2 > 3h) and the ratio of diastereomers can
be analyzed by NMR spectroscopy (Figure 6). Analyzing the
(ΔG⧧
= 46 kJ/mol vs ΔG⧧
= 59 kJ/mol). For
Tc‑Ar1
Tc‑Ar(P∧O)H
the asymmetric compounds, the formation of diastereomers is
observed for Ar/(MeO)2(P∧O)Na and Ar/(MeO)21 with a higher
barrier for the complex (ΔG⧧
= 39 kJ/mol vs
∧
Tc‑Ar/(MeO)2(P O)Na
ΔG⧧
= 59 kJ/mol).50
Tc‑Ar/(MeO)21
In conclusion, of the two motional processes affecting
stereoselectivity, the aryl rotation possesses the significantly
higher barriers in the range of 40−60 kJ/mol, whereas the ring
flip process could not be observed in the accessible temperature
range and must exhibit barriers always below ca. 35 kJ/mol.
The observed barriers and corresponding coalescence temper-
atures are low in comparison to the temperatures required for
an effective insertion reaction and polymerization, 60−90 °C.
Only CF31 is an exception. Thus, in general the ligand
framework must be regarded as conformationally fluxional.47
Stereoselectivity of MA Insertion. The reaction of the
dimeric precursor [{(MeO1-Cl)-μ-Na}2] with AgBF4 in the
presence of MA in dichloromethane at 60 °C in a sealed tube
affords the two diastereomeric products of the consecutive MA
chelate formed by the MA insertion, the diastereomeric
MeO
methine groups at Cγ and Cγ′ in
3
give rise to
MA‑rac/meso
well separated proton resonances around 3 ppm (Figure 6). In
addition, also the Cγ 13C resonances show a pronounced shift
difference (Cγ, 51 ppm; Cγ′, 45 ppm), which allows for a
transferable assignment of the meso/rac stereochemistry (vide
infra). For a determination of stereoselectivity, also the low
field shifted (in comparison to the starting material) resonances
of the anisyl ortho protons 12-H of the (P∧O) ligand at around
8.4 ppm can be drawn upon. In addition, the 31P NMR
spectrum and integrals can be utilized as well. Note that, due to
underlying impurities, the exact ratios determined by
integration of the different resonances pairs can deviate from
each other.51 For a rapid assignment of the crucial MA derived
MeO
MeO
1
1
insertion into the Pd−Me bond
3
and
3
up
P
spin systems of new complexes, a combined H, H TOCSY/
COSY-NMR analysis starting from the unique low field shifted
γ-H resonances proved to be most effective.
MA‑rac
MA‑meso
to a 2:1 ratio (Figure 6).51 This ratio was confirmed by 1H, 31
Dependence of Stereoselectivity on Temperature. For
an effective homopolymerization of polar monomers, temper-
atures between 60 and 90 °C are generally required.14,16 Hence,
the influence of temperature on the stereoselectivity of the
consecutive MA insertion into MeO1 was studied at temper-
atures up to 95 °C. To a mixture of [{(MeO1-Cl)-μ-Na}2] and
AgBF4 in C2D2Cl4 was added 20 equiv of MA, and the reaction
mixture was kept at the corresponding temperature in an NMR
tube. Reaction times are strongly dependent on the reaction
temperature. Whereas consecutive insertion was completed
within 20 min at 95 °C, at 60 °C 90 min and at 25 °C 24 h
were required. Insertion at 4 °C was still not completed after a
week, and solvents and monomer were removed at 0 °C and
MeO
MeO
the remaining mixture of
2
and
3
was analyzed.
MA
MA
The slow progress of reaction at 4 °C indicates that a further
temperature decrease is not reasonable for the catalyst system
MeO1. The insertion experiments clearly reveal that an influence
of temperature on stereocontrol is negligible, and that the ratio
of the stereoisomers is temperature-independent within
experimental error (cf., Supporting Information Figures S49
and S50).
1
MeO
Figure 6. H NMR spectrum of
3
(600 MHz, CD2Cl2, 25 °C;
MA
inset 31P NMR, 168 MHz). Insets: NMR resonances relevant for
determination of stereoselectivity.16
Dependence of Stereoselectivity on Steric Bulk of the
Monomer. Because MeO1 exhibits no permanent chiral center,
and the above studies revealed a fast interconversion of
conformers, methyl acrylate insertion is likely dominated by a
chain-end stereocontrol mechanism as proposed previously.16
To further analyze this issue, the influence of the monomers
bulk on the stereoselectivity was studied. For this purpose, the
insertion of MA, isopropyl acrylate (iPrA), and tert-butyl
acrylate (tBuA) at 25 °C was compared under the same
experimental conditions as described above. The comparison of
the crude insertion products revealed that the steric bulk of the
monomer primarily influences the regioselectivity.
NMR, as well as X-ray crystallography.16,51 In the solid state,
the aryl moieties adopt an exo2 conformation. Because of the
chirality of the chelate formed by the inserted MA, the anisyl
moieties of the ligand are diastereotopic in MeO3 and can be
distinguished by NMR. An additional hindered rotation of
these groups is unlikely, because it would lead to an additional
stereocenter and hence the observance of four diastereomers
for MeO3. A qualitative temperature-dependent NMR study
reveals that up to 120 °C no intramolecular interconversion is
detectable, but that additional diastereomers are observed at
1030
dx.doi.org/10.1021/ja3101787 | J. Am. Chem. Soc. 2013, 135, 1026−1036

Journal of the American Chemical Society
Article
a
Table 2. Rate Constants of Insertion and β-H Elimination, and Stereoselectivity for Various Phosphinesulfonato Ligands
kfirst
regioselectivity
stereoselectivity
first insertion
ksecond
stereoselectivity
second insertion (rac:meso)
X
entry
compound
[10−3 s−1
]
first insertion 22,1
:
X21,2
[10−5 s−1
]
2-1
2-2
2-3
2-4
2-5
2-6
2-7
MeO1
CF31
(MeO)21
(iPrO)21
Ar1
Ar/(MeO)21
cHexO/(MeO)21
3.2
4.8
1.3
3.4
>15:1
4:1
9
43
4
2:1
2:1
1:1
1:1
d
>30:1
7:1
b
∼1
b
∼1
3
1:1
c
d
6
5
2:1
6:1
3:1
e
3
6:1
∼1:1
1:1
a
b
c
Conditions: [Pd] = 0.02 mol L−1, 1.1 equiv AgBF4, Pd:MA ∼ 1:15, solvent: CD2Cl2, T = 25 °C. Estimate, exact determination not possible. A 2:1
ratio of two
isomers is possible. No consecutive insertion observed. Additional 2,1-insertion products observed, clear assignment not possible.
Ar
2
products is found. However, diastereomer formation is probably due to hindered aryl rotation in Ar2, and interconversion of the
MA‑2,1
d e
For MA only a trace amount of the 1,2 insertion product
selectivity. This corresponds with the increased steric bulk at
the Pd−Me for Ar/(MeO)21 in the solid state and is reflected in
the different alignment (pseudoaxial/equatorial) of the second
2,6-(MeO)2C6H3 substituent in the X-ray structures for theses
two complexes (Figure 3).
MeO
MeO
2
was observed, whereas for tBuA a ratio
MA‑1,2
MeO
2
:
2
of ∼1:5 is found. For iPrA, the portion
tBuA‑1,2
tBuA‑2,1
of 1,2 insertion product cannot be determined due to
overlapping resonances, but it appears to be similar to that
for tBuA. Concerning the stereoselectivity of the consecutive
insertion, an unexpected trend was found: With increasing
steric bulk of the monomer, stereoselection is reduced, as
concluded from the integration of the ortho aryl protons (cf.,
Supporting Information Figures S51 and S52). The rac:meso
The insertion rates for the consecutive MA insertion are 1−2
orders of magnitude lower than for the first insertion. This is
due to a slower insertion into an α-substituted carbyl versus
insertion into a methyl group, and due to chelating
coordination of the carbonyl group of the incorporated acrylate
unit, which competes with the incoming monomer.16,43 Other
than for the first insertion, a more pronounced influence of the
steric bulk of the ligand on the rate constant ksecond is found. As
expected, ksecond decreases with steric bulk as illustrated by
comparison of MeO1 (ksecond = 9 × 10−5 s−1) and Ar/(MeO)21
(ksecond = 6 × 10−5 s−1). Introduction of a second 2,6-
(MeO)2C6H3 substituent in Ar1 even leads to a complete
inhibition of the consecutive insertion (entry 2-1 vs 2-5 vs 2-6).
That electronics can still dominate in the case of less
pronounced differences in steric bulk becomes clear by
considering that for CF31, exhibiting the highest rotational
barrier (vide supra), the consecutive insertion seems to be
accelerated due to electronic effects leading to the highest rate
constant ksecond = 40 × 10−5 s−1 of all investigated complexes,
whereas for the most electron-rich complex (MeO)21 the
insertion rate is significantly lower with ksecond = 4 × 10−5 s−1
(compare entry 2-1 to 2-3).
ratio decreases from 2.2:1 for MeO3MA over 1.5:1 for MeO3iPrA to
MeO
1.1:1 for
3
tBuA
. Hence, it appears here that chain-end
stereocontrol and any enantiomorphic site stereocontrol work
uncooperatively in different directions.
Influence of Ligand Structure and Complex Con-
formation on Stereoselectivity. To shed light on the
influence of the complex structure on the insertion behavior,
the reaction of MeO1, CF31, (MeO)21, (iPrO)21, Ar1, Ar/(MeO)21, and
cHexO/(MeO)21 (generated in situ from [{(X1-Cl)-μ-M}n]) with
X
MA to the mono insertion products 2MA and the products of
consecutive insertion X3MA were monitored over time by NMR
spectroscopy (Scheme 1).
The regioselectivity in the first insertion is determined from
X
the distribution of 1,2- to 2,1-insertion products 2MA. The
formation of the 1,2 insertion product is generally disadvanta-
geous because the resulting five-membered chelates inhibit
further insertions.16 Regarding the first insertion, the influence
of the phosphinesulfonato substitution pattern on the rate
constant kfirst is rather low for the complexes investigated, and
cannot always be distinguished from electronic effects. By
contrast, the steric bulk has a significant influence on the
regioselectivity. Here, an increase of the constraints around the
active center leads to a shift toward the 1,2 insertion product as
illustrated, for example, by the shift of the 2,1:1,2 ratio from
>15:1 for MeO2, over 4:1 for CF32, to ∼1:1 for Ar2 (Table 2). A
complete inversion of regiochemistry to 1,2 insertion versus the
2,1 insertion of MeO1 has been observed previously for the very
bulky (κ2-P,O)-diazaphospholidine-sulfonato ligands, which is
due to a destabilization of the 2,1-insertion transition states.40
For the asymmetric substituted compounds Ar/(MeO)21 and
cHexO/(MeO)21, several 2,1 insertion products are formed due to
the permanent stereo center at phosphorus. A selectivity of 6:1
was found for Ar/(MeO)21, indicating an effective enantiomorphic
site stereocontrol. For cHexO/(MeO)22, analysis is complicated due
to formation of additional species, probably due to hindered
rotation, but selectivity for the first insertion is roughly 1:1.
Consequently, the exchange of a cHexOC6H4 substituent by a
2-{2′,6′-(MeO)2C6H3}C6H4 moiety significantly increases
The stereoselectivity for the symmetric complexes was
determined from the product distribution after consecutive
X
MA insertion to form 3MA (vide supra). A comparison of the
MA insertion into MeO1 with CF31 at room temperature reveals
that the enhanced rotational barrier of the CF3-substituted
complex does not lead to a significant increase of selectivity
(∼2:1, Figure S53). In the complexes with C2v-symmetric aryl
(MeO)2
(iPrO)2
moieties
3
and
3 , the selectivity is even
MA
MA
diminished in comparison to the compounds with Cs-
symmetric aryl moieties, again a substitution of the MeO-
groups by the sterically more demanding iPrO-groups does not
influence the selectivity significantly (Table 2, Figure S53).
From the carbon shifts of the Cγ methine groups, an absolute
assignment of stereochemistry is possible (vide supra). An
increase of steric bulk leads to an equilibration of the rac:meso
product ratio, which was also observed for an increase of
monomer steric bulk (vide supra).
cHexO/(MeO)2
Ar/(MeO)2
For the compounds
3
and
3
,
MA
MA
analysis of the insertion products is more complex, because
the presence of the permanent stereocenter at phosphorus
X
leads to the formation of at least four diastereomers 3MA. For
1031
dx.doi.org/10.1021/ja3101787 | J. Am. Chem. Soc. 2013, 135, 1026−1036

Journal of the American Chemical Society
Article
a
cHexO/(MeO)2
1
Scheme 2. MA Coordination and Insertion with MeO1
3
MA
, all four diastereomers can be detected by H,
1H TOCSY-NMR analysis (cf., Supporting Information Figure
S73). Integration of the γ-H methine proton yields a
distribution of roughly 1:4:5:8. Assignment of the diaster-
eomers to meso and rac species is possible by analysis of the
13C NMR spectra, assuming that a similar chain configuration
reflects in similar carbon shifts. Consequently, the Cγ-methine
cHexO/(MeO)2
1H resonances of
3
at 3.27 (13C: 44 ppm) and
MA
2.82 (13C: 45 ppm) can be assigned to the meso isomers and
the resonances at 3.21 (13C: 51 ppm) and 3.00 ppm (13C: 52
ppm) to the rac isomers, corresponding to a rac:meso
Ar/(MeO)2
selectivity of ∼1:1. For
3 , analysis is complicated
MA
by overlapping resonances, but the insertion products evidently
are formed in rather similar proportions. In this case, the α-H
methine resonances integrals were employed (Figure 7).
a
Relative energies of the species are given in parentheses (kJ/mol;
black, intermediates; red, transition states).16
transition state TSins leads to the γ-agostic stabilized kinetic
X
X
product 2MA‑KP. The final insertion product is 2MA‑Ch, in
which the carbonyl group of the inserted MA is coordinated to
the metal center. The mechanism of further insertions is
similar, the major difference being the nature of the starting and
final complexes and their respective energy, due to the
formation of 4- or 6-membered chelates after the first and
the second insertion, respectively.
Site stereocontrol requires a permanent rigid chiral
configuration on the time scale of insertion. Hence, for a
symmetrically substituted phosphorus atom as in MeO1, (MeO)21,
and Ar1, at least a stable helical configuration at the phosphorus
atom or a stable ring configuration is necessary. If ring flip and
aryl rotation exhibit significantly lower barriers in comparison
to insertion, the first MA insertion cannot be enantioselective.
The corresponding transition states for ring flip (TSRF) and aryl
rotation (TSAR) in the MA complexes X1-MA are compared to
the transition state for the first 2,1-insertion (TSIns) in Table 3.
Figure 7. Distribution (1H NMR, 400 MHz, CD2Cl2) of the three
Ar/(MeO)2
diastereomers of
3
(for a complete assignment of the spin
MA
systems, see Figures S70 and S71).
1
1
1
Assignment of the spin systems by H, H TOCSY and H,
13C gHSQC experiments clearly shows the formation of only
three main diastereomers and yields a rac:meso distribution of
3:1. Thus, the high selectivity for the first insertion (6:1) is
diminished for the consecutive insertion, indicating that the
possible chain-end control, which becomes active from the
second insertion on, is opposite to the enantiomorphic site
stereocontrol, which controls the first insertion alone.
Table 3. Transition State Energies (kJ/mol) for MA
Insertion into X1 and for Ring Flip and Aryl Rotation for X1-
MA
compound
TSIns
ring flip-TS
aryl rotation-TS
MeO1
(MeO)21
Ar1
0
0
0
0
−49
−51
−49
−18
b
−34
−
−14
−5
b
Computational Studies. To gain a deeper understanding
of the dynamics, regio- and stereocontrol mechanisms in
(P∧O)PdMe complexes, alkyl acrylate insertion into the Pd−
Me bond of MeO1, (MeO)21, Ar1, and Ar/(MeO)21 was studied by
DFT (see the Supporting Information for computational
details). Considering the limited selectivity achieved exper-
imentally, the energy differences are expected to be rather low,
almost within the error expected for this kind of calculation,
and thus trends from the comparison of different systems are
more indicative than the energy values specific to a single
system.52
a
Ar/(MeO)21
a
Configuration at phosphorus: (R). Aryl rotation is not possible,
because the MeO-groups inevitably collide in the corresponding TS.
The reported values indicate that TSins is higher in energy than
TSAR and TSRF for MeO1 and Ar1, which indicates that here the
first MA insertion is not stereoselective. Further, in agreement
with the NMR studies, the energy barriers for aryl rotations are
always at least 13 kJ/mol higher than the barriers for the ring
flip and increase with the steric bulk of the nonchelating aryl
moieties.54 Considering that the ring flip is consistently
calculated to be a low energy process, it is assumed to be
generally rapid as compared to aryl rotation and acrylate
insertion and is not further discussed in the following.
The established theoretical understanding of the insertion
mechanism with (P∧O)Pd catalysts is briefly summarized using
insertion of MA into the Pd−Me bond of MeO1 as an example
(Scheme 2).16,27,28,40,53 The growing chain is always oriented
trans to the sulfonate group in the initial species, and
coordination of the monomer occurs cis to phosphorus as in
X1-MAcis. The preferred pathway for insertion into the Pd−C
bond proceeds through a Berry-pseudo rotation step that
For the bulky (MeO)21-MA system, rotation of the aryl groups
is hampered because the MeO-groups inevitably collide in the
corresponding TSAR and the fixed chiral conformation (P or M)
of the complex could in principle select between re- or si-MA
insertion. However, analysis of the TSins geometry for a P
X
X
isomerizes 1-MAcis to the less stable 1-MAtrans species. The
2,1-insertion of MA into the Pd−C bond of X1-MAtrans through
1032
dx.doi.org/10.1021/ja3101787 | J. Am. Chem. Soc. 2013, 135, 1026−1036

Journal of the American Chemical Society
Article
configuration at the active site reveals no significant differences
in the interaction between the two MA enantiofaces and the
ligand, due to the high symmetry of the system (Figure 8).
Consequently, the energy difference between the re- and si-MA
TSIns is negligible (<1 kJ/mol).
Table 4. Stereo- and Regioselectivity for the First and the
Second Acrylate Insertion into 1 (Energies in kJ/mol)
X
ΔEstereo
ΔEregio
a
b
d
compound
acrylate first ins second ins
first ins
second ins
c
c
MeO1
MA
MA
MA
MA
tBuA
−1(1)
9(10)
27
29
14
12
11
(MeO)21
Ar1
Ar/(MeO)21
MeO1
0
0
11
0
−1
3
11
7
2
−2
a
Positive numbers correspond to preference of the meso product
b
X3MA‑meso. Positive numbers correspond to preference of the 2,1 MA
c
insertion product. The values in parentheses are obtained with the
ADF program and were reported previously.16 ΔEregio = E2,1−1,2
−
d
X
Figure 8. Transition states for si- and re-MA insertion into (MeO)21
with a P helical configuration of the aromatic rings.
E2,1−2,1: positive numbers correspond to preference of 3MA. Helical
chirality for all compounds: (P).
The asymmetric system Ar/(MeO)21 exhibits a permanent
stereocenter at the phosphorus and can in principle distinguish
between the different MA enantiofaces even if molecular
motions are fast. In fact, a significant energy difference ΔEstereo
= 11 kJ/mol between the two respective lowest energy
transition states for acrylate insertion into the Pd−Me bond
was found. In case of an R configuration at the phosphorus
atom (which favors a P helical chirality), insertion of a re-MA is
favored (Figure 9).
places the functional group of the monomer away from the
growing chain (Figure S75).
For the asymmetric system Ar/(MeO)21, four different
consecutive insertion products Ar/(MeO)23 can be formed, due
to the chirality at the phosphorus. The energy differences
ΔEstereo between the four transition states are reported in Table
5. In the first insertion, a R chain is preferentially generated
from the most stable re-MA TSIns (Figure 9). For the second
MA insertion, stereoselectivity is diminished. In detail, other
than the NMR studies, the calculations indicate a slight
preference for the meso isomer. The loss of enantiomorphic
site control in favor of the re-MA enantioface in the second
insertion step can be explained by steric interaction between
the ethyl group of the chain and the ligand (Figure 10). The
2,6-(MeO)2(C6H3) moiety at the phosphorus is rotated in
comparison to the TSins of the first re-MA insertion, and one of
the methoxy groups is now arranged at a critical distance of
only 3.5 Å from the monomer (dihedral angle θ (Pd−P−C1−
C2): −33°(re-MAfirst) vs −9°(re-MAsecond); compare −9°(si-
MAfirst) vs −14°(si-MAsecond); Figure 9 vs Figure 10). As a
consequence, enantiomorphic site and chain end stereocontrols
act in opposite directions for Ar/(MeO)22 so that stereoselectivity
in the second MA insertion is definitively lost. Calculations
indicate that due to the flexibility of the system, the ligand
evades the growing chain and any kind of stereoselectivity in
the MA polymerization is prevented.
Figure 9. Transition states for si- and re-MA insertion into Ar/(MeO)21
with a P helical configuration of the aromatic rings.
The calculated ΔEstereo of 11 kJ/mol is in reasonable
agreement with the experimental stereoselectivity ratio of 6:1.
As depicted in Figure 9, the interaction between the 2,6-
(MeO)2(C6H3)-moiety and the monomer is responsible for the
higher energy of the si-MA insertion into the Pd−Me bond of
the Ar/(MeO)21 complex with a R configuration at the
phosphorus (shortest distances of 3.7 Å for si-MA as compared
to 3.9 and 3.8 Å in the favored re-MA insertion).
The first MA insertion generates a chiral carbon atom
directly bound to the metal center, with re- and si-MA insertion
resulting in R and S configuration of the chiral carbon atom,
respectively. Considering this, the second MA insertion differs
from the first in that either an interaction of the chiral chain
end with the ligand may force the catalytic site to assume a
specific chiral conformation or the chiral chain itself may select
between the monomer enantiofaces directly. Table 4 reports
the calculated ΔEstereo for the first two acrylate insertion steps.
In agreement with experiments, the calculations indicate that
the chiral chain end is unable to discriminate between the two
MA enantiofaces, even in case of bulky ligands. Further, these
results indicate that stereoselectivity due to direct interaction
between the chiral growing chain and the monomer is
negligible, which is reasonable considering that 2,1-insertion
Because the production of defined tactic polymers does not
only depend on stereoselectivity but also on regioregularity,
and experiments showed that both insertion modes (1,2 and
2,1) are observable, at least for the first insertion, the influence
of the ligand structure on regioselectivity was investigated.
Here, the energy difference between the transition states for 2,1
and 1,2 monomer insertion was calculated for the first insertion
X
with 1MA (ΔEregio,first, Table 4), as well as for MA insertion
X
with 22,1‑MA (ΔEregio,second, Table 4). In contrast, consecutive
1,2-insertion into X21,2‑MA is energetically not favorable because
of steric interaction between the monomer and the 1,2 chain
end, and was also not observed experimentally so far.55
As expected and in agreement with previous work,
calculations show that the shift from the electronically favored
2,1-transition state to the less sterically constrained 1,2-
transition state depends on the structure of the ligand and
the size of the monomer.40 The 1,2-transition state becomes
more favored by increasing either the size of the P-aryl
substituents of the chelating ligands from MeO1 (ΔEregio,first = 9
kJ/mol) to Ar/(MeO)21 (ΔEregio,first = 7 kJ/mol) to Ar1 (ΔEregio,first
1033
dx.doi.org/10.1021/ja3101787 | J. Am. Chem. Soc. 2013, 135, 1026−1036

Journal of the American Chemical Society
Article
Table 5. ΔEstereo (kJ/mol) for All Transition States to the Insertion Products of the First and Second MA Insertions into
Ar/(MeO)21 with a (R) Configuration at the Phosphorus and P Helical Configuration of the Aromatic Rings
Ar/(MeO)21
MA first insertion (2,1)
re si
MA second insertion (2,1; 2,1)
si/R re/S
re/R
si/S
ΔEstereo
product
0
11
Ar/(MeO)2
0
3
10
Ar/(MeO)2
4
Ar/(MeO)2
Ar/(MeO)2
Ar/(MeO)2
Ar/(MeO)2
2
2
′
3
3
3
′
3
MA‑meso
′
MA‑2,1
MA‑2,1
MA‑meso
MA‑rac
MA‑rac
phosphinesulfonato ligand or the monomer decreases the
regioselectivity of the insertion.
CONCLUSIONS AND OUTLOOK
■
As has been understood in the past few years, palladium
complexes with unsymmetrical, hard/soft chelating phosphine-
sulfonato ligands are unique in allowing for ethylene
copolymerization with a large range of polar vinyl monomers
and even enable insertion homopolymerization of acrylates.
An analysis of intramolecular interconversion processes
showed that the ligand framework is rather flexible and
undergoes several transformations including an aryl rotation in
a propeller-like conformation environment and a ring flip
process of the six-membered (P∧O)Pd chelate. The highest
barriers ΔG⧧ of these processes, of the aryl rotation, were found
to add up to 40−60 kJ/mol for the methyl complexes (X1).
Theoretical calculations revealed that also for the monomer-
coordinated complexes X1-MA the transition states for the
intramolecular motions are significantly lower than the
insertion transition states. An exception was found for systems
that bear double ortho substituted aryl moieties at phosphorus.
Here, an aryl rotation is inhibited, but still the transition states
for both enantiofaces of the monomer are rather similar due to
the symmetry of the ligand.
Figure 10. Transition states for the MA insertion into Ar/(MeO)22MA to
Ar/(MeO)2
Ar/(MeO)2
3
(left) and
3
(right); R configuration at
MA‑meso
MA‑rac
the phosphorus atom and P helical configuration of the aromatic rings.
= 0 kJ/mol) or the size of the monomer from MA to tBuA
(
MeO1 ΔEregio,first = −2 kJ/mol). The 2,1- and 1,2-transition
states for the first insertion of MA into Ar1 and of tBuA into
MeO1 are compared in Figure 11. In both cases, regioselectivity
Conclusively, a permanent chiral center at phosphorus by
asymmetric substitution is a prerequisite for a stereocontrol of
insertion. This concept is underlined by the experimental
observation of a 6:1 stereoselectivity in the first insertion with
the asymmetric complex Ar/(MeO)21. For the consecutive
insertions, stereocontrol is reduced. The theoretical calculations
are in line with these results and indicate that the high flexibility
of the ligand is responsible for the loss of stereocontrol. The
ligand evades the growing chain and transition states are
equilibrated for the consecutive insertion, because site stereo-
control and chain end stereocontrol work uncooperatively.
NMR-insertion studies with MeO1 showed that within the
accessible temperature range, stereoselectivity cannot be
influenced by temperature. Further, insertion studies of
monomers with variable steric bulk (MA, iPrA, tBuA) revealed
that stereocontrol decreases with increasing steric bulk, which
indicates that also in this case chain end stereocontrol and
enantiomorphic site stereocontrol work together uncooper-
atively.
Beyond the effects on stereocontrol, a clear relation between
steric bulk and the regioselectivity of insertion was found, such
that increased steric bulk in the ligand or in the monomer leads
to a shift from the desired 2,1- to the 1,2-insertion mode. The
resulting 1,2-insertion products are not amenable to further
insertions due to the stability of the five-membered chelate
formed by coordination of the carbonyl group of the
incorporated acrylate. Theoretical considerations clearly show
Figure 11. Regioselectivity of MA insertion into Ar1 (a,b) and of tBuA
insertion into MeO1 (c,d).
is completely lost (ΔEregio,first = 0−2 kJ/mol). As reported
previously, the decrease of regioselectivity can be rationalized
by an analysis of the deviation from planarity of the four center
́
Cossee−Arlman-like transition state that increases by increas-
ing the sterical pressure of the system.40 The extent of this
deviation can be estimated by comparing the Me−Pd−C1−C2
dihedral angles for the 2,1-transition states in Figure 11 with
the value for the 2,1-transition state for MeO1 reported
previously.40 The deviation from 0° for insertion into MeO1
increases from MA insertion (−4°)40 to tBuA insertion (8°)
(Figure 11). Conclusively, in agreement with the experimental
observations, an increase of steric constrains via the
́
that deviation from the planarity of the four center Cossee−
Arlman-like transition state increases the sterical pressure of the
1034
dx.doi.org/10.1021/ja3101787 | J. Am. Chem. Soc. 2013, 135, 1026−1036

Journal of the American Chemical Society
Article
Notes
system and results in energetically similar 1,2- and 2,1-insertion
transition states. That is, site differentiation by extreme steric
bulk is not a feasible approach to tactic polymers because the
regioregularity is lost. In addition, the insertion studies also
disclosed an intrinsic problem of these catalysts: Increase of
steric bulk always leads to a reduction of the acrylate insertion
rates, and thus hinders chain growth. This culminates in a
complete suppression of consecutive insertions for Ar1.
Further Concepts. These limitations understood it is also
evident that there is a “window” of intermediate steric bulk,
which allows stereoselection without compromising regioregu-
larity. In more general terms, it is remarkable that any
stereoselection at all is possible in the open environment of
the hard/soft unsymmetric phosphinesulfonato ligand (Figure
1). One decisive feature appears to be the insertion from the
alkyl-olefin complex with the π-bound acrylate in cis-position to
the P-donor. This allows for influencing the stereoselectivity of
the insertion step via the P-substituents. With the concept of an
unsymmetric substitution, stereocontrol was realized. A
reduction of flexibility in such asymmetric complexes appears
a possible approach toward stereoselective polymerization.
Reduction of flexibility has already been shown to be possible
by a double ortho substitution of the aryl moieties at
phosphorus. However, a more effective concept may be the
introduction of a rigid spacer between the two different aryl
moieties (Figure 12). This could further allow one to place the
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was financially supported by the DFG (Me 1388/
10). B.N. acknowledges support by the state of Baden-
Wurttemberg by a Landesgraduiertenforderung-Stipend. We
̈
̈
thank Anna-Lena Steck for high resolution ESI−MS measure-
ments, Thomas Wiedemann for participation in this research as
a part of his undergraduate studies, Inigo Gottker-Schnetmann
̈
for fruitful discussions, and the HPC team of Enea for use of
the ENEA-GRID and the HPC facilities CRESCO in Portici,
Italy.
REFERENCES
■
(1) Baugh, L. S.; Canich, J. M. Stereoselective Polymerization with
Single-Site Catalysts; CRC Press: London, 2008.
(2) Matsuzaki, K.; Uryu, T.; Ishida, A.; Ohki, T.; Takeuchi, M. J.
Polym. Sci., Part A-1: Polym. Chem. 1967, 5, 2167−2177.
(3) Liu, W.; Nakano, T.; Okamoto, Y. Polym. J. 1999, 31, 479−481.
(4) Deng, H.; Soga, K. Macromolecules 1996, 29, 1847−1848.
(5) Porter, N. A.; Allen, T. R.; Breyer, R. A. J. Am. Chem. Soc. 1992,
114, 7676−7683.
(6) Liu, W.; Nakano, T.; Okamoto, Y. Polymer 2000, 41, 4467−4472.
(7) Matsuzaki, K.; Uryu, T.; Kanai, T.; Hosonuma, K.; Matsubara, T.;
Tachikawa, H.; Yamada, M.; Okuzono, S. Makromol. Chem. 1977, 178,
11−17.
(8) Brintzinger, H.-H.; Fischer, D.; Mulhaupt, R.; Rieger, B.;
̈
Waymouth, R. Angew. Chem., Int. Ed. Engl. 1995, 34, 1143−1170.
(9) Resconi, L.; Cavallo, L.; Fait, A.; Piemontesi, F. Chem. Rev. 2000,
100, 1253−1346.
(10) Johnson, L. K.; Mecking, S.; Brookhart, M. J. Am. Chem. Soc.
1996, 118, 267−268.
(11) Mecking, S.; Johnson, L. K.; Wang, L.; Brookhart, M. J. Am.
Chem. Soc. 1998, 120, 888−899.
Figure 12. Concepts toward stereoselective catalysts.
(12) Ittel, S. D.; Johnson, L. K.; Brookhart, M. Chem. Rev. 2000, 100,
1169−1204.
chiral information in the spacer, which might reduce steric bulk
at the active center. Preliminary analysis of the transition state
configuration reveals that for a rigid aryl configuration
presumably no large difference in size between the substituents
at the phosphorus is required, which again could help to keep
the steric bulk at the active center moderate.
Note that in parallel to this work, Nozaki et al.
communicated that (P∧O)PdMe complexes with asymmetric
substitution at phosphorus like Ar/H1 are capable of asymmetric
copolymerization of CO with polar monomers. In this case, the
copolymerization with CO introduces a nonchiral CO-spacer
between the chiral side chains. This apparently advantageously
reduces a chain end stereocontrol, which in our work
counteracts the enantiomorphic site stereocontrol of the
homoinsertions.56
(13) Drent, E.; Dijk, R. v.; Ginkel, R. v.; Oort, B. v.; Pugh, R. I. Chem.
Commun. 2002, 744−745.
(14) Guironnet, D.; Roesle, P.; Runzi, T.; Gottker-Schnetmann, I.;
̈
̈
Mecking, S. J. Am. Chem. Soc. 2009, 131, 422−423.
(15) Piche, L.; Daigle, J.-C.; Rehse, G.; Claverie, J. P. Chem.-Eur. J.
2012, 18, 3277−3285.
(16) Guironnet, D.; Caporaso, L.; Neuwald, B.; Gottker-Schnetmann,
̈
I.; Cavallo, L.; Mecking, S. J. Am. Chem. Soc. 2010, 132, 4418−4426.
(17) Corradini, P.; Guerra, G.; Cavallo, L. Acc. Chem. Res. 2004, 37,
231−241.
(18) Tshuva, E. Y.; Goldberg, I.; Kol, M. J. Am. Chem. Soc. 2000, 122,
10706−10707.
(19) Mitani, M.; Saito, J.; Ishii, S.-I.; Nakayama, Y.; Makio, H.;
Matsukawa, N.; Matsui, S.; Mohri, J.-i.; Furuyama, R.; Terao, H.;
Bando, H.; Tanaka, H.; Fujita, T. Chem. Rec. 2004, 4, 137−158.
(20) Makio, H.; Terao, H.; Iwashita, A.; Fujita, T. Chem. Rev. 2011,
111, 2363−2449.
(21) Brookhart, M.; Wagner, M. I.; Balavoine, G. G. A.; Haddou, H.
A. J. Am. Chem. Soc. 1994, 116, 3641−3642.
(22) Aeby, A.; Consiglio, G. Inorg. Chim. Acta 1999, 296, 45−51.
(23) Bianchini, C.; Meli, A. Coord. Chem. Rev. 2002, 225, 35−66.
(24) Nozaki, K.; Sato, N.; Tonomura, Y.; Yasutomi, M.; Takaya, H.;
Hiyama, T.; Matsubara, T.; Koga, N. J. Am. Chem. Soc. 1997, 119,
12779−12795.
(25) Sesto, B.; Consiglio, G. Chem. Commun. 2000, 1011−1012.
(26) Sesto, B.; Consiglio, G. J. Am. Chem. Soc. 2001, 123, 4097−
4098.
ASSOCIATED CONTENT
* Supporting Information
■
S
Text, tables, figures, and CIF files giving detailed experimental
procedures and analytical data, structural diagrams, selected
bond lengths and angles, and crystallographic data/processing
parameters for all structures. This material is available free of
charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
stefan.mecking@uni-konstanz.de; lcaporaso@unisa.it
■
(27) Noda, S.; Nakamura, A.; Kochi, T.; Chung, L. W.; Morokuma,
K.; Nozaki, K. J. Am. Chem. Soc. 2009, 131, 14088−14100.
1035
dx.doi.org/10.1021/ja3101787 | J. Am. Chem. Soc. 2013, 135, 1026−1036

Journal of the American Chemical Society
Article
(28) Nozaki, K.; Kusumoto, S.; Noda, S.; Kochi, T.; Chung, L. W.;
Morokuma, K. J. Am. Chem. Soc. 2010, 132, 16030−16042.
(29) This consideration also applies to recently reported phosphine-
phosphineoxide coordinated catalysts: Carrow, B. P.; Nozaki, K. J. Am.
Chem. Soc. 2012, 134, 8802−8805.
mol is found for an isolated process only affecting the rotation of the
(MeO)2C6H3 aryl substituent at the phosphorus atom (cf., Supporting
Information).
(51) The rac/meso designation is based on consideration of the
stereocenters as part of a growing chain. The two stereocenters can
MeO
possess identical (
3
) or opposite (MeO3MA‑rac) stereo-
(30) 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.
MA‑meso
configuration. The absolute assignment is based on a reinvestigation
of the corresponding X-ray structure (ref 16) and is in line with
NOESY/ROESY experiments in solution. Note that within exper-
imental error the ratios of the integrals of different resonances in
Figure 7 are all identical, and agree with a slight preference for the rac
isomer (ca. 2:1).
(31) Liu, S.; Borkar, S.; Newsham, D.; Yennawar, H.; Sen, A.
Organometallics 2007, 26, 210−216.
(32) Vela, J.; Lief, G. R.; Shen, Z.; Jordan, R. F. Organometallics 2007,
26, 6624−6635.
(33) Anselment, T. M. J.; Wichmann, C.; Anderson, C. E.;
Herdtweck, E.; Rieger, B. Organometallics 2011, 30, 6602−6611.
(34) Piche, L.; Daigle, J.-C.; Poli, R.; Claverie, J. P. Eur. J. Inorg. Chem.
2010, 4595−4601.
(35) Kochi, T.; Yoshimura, K.; Nozaki, K. Dalton Trans. 2006, 25−
27.
(52) Ziegler, T.; Autschbach, J. Chem. Rev. 2005, 105, 2695−2722.
(53) Haras, A.; Anderson, G. D. W.; Michalak, A.; Rieger, B.; Ziegler,
T. Organometallics 2006, 25, 4491−4497.
(54) Differences between the ring flip and aryl rotation barriers for
the olefin complexes are nearly identical to the corresponding
X
monomer-free complexes 1 (cf., Supporting Information).
(55) For all systems, the TS for the consecutive 1,2−1,2 MA
insertion is at least 25 k/J mol higher than for the corresponding
consecutive 2,1−2,1 MA insertion TS.
(56) Nakamura, A.; Kageyama, T.; Goto, H.; Carrow, B. P.; Ito, S.;
Nozaki, K. J. Am. Chem. Soc. 2012, 134, 12366−12369.
(36) Kochi, T.; Nakamura, A.; Ida, H.; Nozaki, K. J. Am. Chem. Soc.
2007, 129, 7770−7771.
(37) Anselment, T. M. J.; Anderson, C. E.; Rieger, B.; Boeddinghaus,
M. B.; Fassler, T. F. Dalton Trans. 2011, 40, 8304−8313.
(38) Kochi, T.; Noda, S.; Yoshimura, K.; Nozaki, K. J. Am. Chem. Soc.
2007, 129, 8948−8949.
(39) (a) Nakamura, A.; Munakata, K.; Kochi, T.; Nozaki, K. J. Am.
Chem. Soc. 2008, 130, 8128−8129. (b) Conley, M. P.; Jordan, R. F.
Angew. Chem., Int. Ed. 2011, 50, 3744−3746. (c) Newsham, D. K.;
Borkar, S.; Sen, A.; Conner, D. M.; Goodall, B. L. Organometallics
2007, 26, 3636−3638.
(40) 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.
(41) (a) Hearley, A. K.; Nowack, R. J.; Rieger, B. Organometallics
2005, 24, 2755−2763. (b) Runzi, T.; Guironnet, D.; Gottker-
̈
̈
Schnetmann, I.; Mecking, S. J. Am. Chem. Soc. 2010, 132, 16623−
16630. (c) Friedberger, T.; Wucher, P.; Mecking, S. J. Am. Chem. Soc.
2012, 134, 1010−1018. (d) Ito, S.; Munakata, K.; Nakamura, A.;
Nozaki, K. J. Am. Chem. Soc. 2009, 131, 14606−14607.
(42) Note that for bulky nonchelating aryl moieties, for example,
C6H4Et, C6H4(2,6-(OMe)2C6H3), and 1-methoxynaphatalene, a
distortion of the chelate ring toward a half-boat conformation with
the palladium, or the oxygen bound to the palladium, placed at the top
position is observed.30,32,33
̈
(43) Neuwald, B.; Olscher, F.; Gottker-Schnetmann, I.; Mecking, S.
̈
Organometallics 2012, 31, 3128−3137.
(44) Mislow, K. Acc. Chem. Res. 1976, 9, 26−33.
(45) The mechanism of transformation is discussed in the Supporting
Information.
(46) Differences are found for coalescence temperature and the
adopted exo configuration. The free ligands MeO(P∧O)H, CF3(P∧O)H,
Ar(P∧O)H and their salts adopt an exo3 configuration in solution,
whereas the complexes MeO1, CF31 adopt an exo2 configuration in line
with the solid-state structures.
(47) In addition, DFT calculations for MeO1, (MeO)21, Ar1, and
Ar/(MeO)21 are in line with the experimental results. Energy barriers for
the ring flip transition state were found to be always below 30 kJ/mol,
whereas barriers for the aryl rotation process are at least 10 kJ/mol
higher in energy and range from 36 to 63 kJ/mol (cf., Supporting
Information Table S24).
(48) Friebolin, H. Basic One- and Two-Dimensional NMR Spectros-
copy, 4th ed.; Wiley-VCH: Weinheim, Germany, 2005.
(49) Note that as a whole group CF3 is much smaller than 2,6-
(MeO)2C6H3, but that for the steric interaction the steric bulk directly
attached at the P-aryls appears decisive, which leads to a similar
influence on the rotational barrier for the tetrahedral CF3- versus the
flat aryl-substituent.
(50) However, for the protonated ligand Ar/(MeO)2(P∧O)H, no
formation of diastereomers can be observed down to −90 °C
(ΔG⧧Tc‑Ar/(MeO)2(P O)H < 35 kJ/mol), but a rather high barrier of 52 kJ/
∧
1036
dx.doi.org/10.1021/ja3101787 | J. Am. Chem. Soc. 2013, 135, 1026−1036