Synthesis of allyl-terminated syndiotactic polypropylene: Macromonomers for the synthesis of branched polyolefins

Article abstract of DOI:10.1021/ma050735u

Low molecular weight syndiotactic polypropylene (syndio-PP) with an olefin end group was synthesized using methylaluminoxane-activated bis(phenoxyimine)titanium dichloride ((PHI)₂TiCl₂) catalysts. Propylene enchainment occurs by a 2,

Full text of DOI:10.1021/ma050735u

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Macromolecules 2005, 38, 6259-6268
6259
Articles
Synthesis of Allyl-Terminated Syndiotactic Polypropylene:
Macromonomers for the Synthesis of Branched Polyolefins
Anna E. Cherian, Emil B. Lobkovsky, and Geoffrey W. Coates*
Department of Chemistry and Chemical Biology, Baker Laboratory, Cornell University,
Ithaca, New York 14853-1301
Received April 8, 2005; Revised Manuscript Received May 3, 2005
ABSTRACT: Low molecular weight syndiotactic polypropylene (syndio-PP) with an olefin end group was
synthesized using methylaluminoxane-activated bis(phenoxyimine)titanium dichloride ((PHI)2TiCl2)
catalysts. Propylene enchainment occurs by a 2,1-insertion mechanism, and termination by â-hydride
elimination produces low molecular weight syndio-PP with allyl end groups. Several new (PHI)2TiCl2
complexes with various ligand modifications were found to display a wide range of activities (turnover
frequency (TOF) ) 3-1200 h-1) and syndiospecificities ([rrrr] ) 0.46-0.93) for propylene polymerization.
While the TOF increases with increasing reaction temperature and propylene concentration, the molecular
weight of the resulting macromonomer shows little variation. This provides strong support for a chain-
transfer mechanism involving one molecule of propylene. The allyl-terminated PPs reported herein can
be used to synthesize branched polyolefins or other new polyolefin architectures.
Scheme 1. Branched Polyolefins via Alkene/
Macromonomer Copolymerization
Introduction
Characteristics of polymer architecture such as mo-
lecular weight, regiochemistry, stereochemistry, and
branching can dramatically influence polymer proper-
ties.1 The pursuit of single-site polymerization catalysts
capable of controlling such polymer characteristics has
been a major focus of research over the past two
decades.2 Recent advances in the development of single-
site metallocene catalysts have led to the production of
polymers with high stereoregularity and regioregularity,
which can give rise to improved mechanical properties.3
In addition, numerous non-metallocene catalysts have
also been developed that are highly stereoselective for
olefin polymerizations.4-6 The discovery of catalysts
capable of living olefin polymerization has made possible
the synthesis of block copolymers, which can exhibit
greatly improved mechanical properties compared with
their homopolymer counterparts.7-9
Although drastic improvements in the mechanical
properties of polyolefins synthesized with single-site
catalysts have been noted (such as impact resistance,
toughness at low temperatures, and resistance to stress
cracking), the narrow molecular weight distribution
(Mw/Mn) of polymers obtained in these single-site poly-
merizations (Mw/Mn e 2) often renders them difficult
to process in extensional flow processes.10,11 The addition
of even small amounts of long chain branched (LCB)
polymers can improve processability in polyolefins.10,12,13
Branches are typically considered to be long chain
branches when the branch length is at least 2.5 times
the entanglement molecular weight (Me; “the molecular
weight between adjacent temporary entanglement
points”, as defined by Eckstein and co-workers14) for the
given polymer composition,15-17 resulting in greatly
increased melt strength and improved processabil-
ity.10,12,18-22
LCB polymers are typically manufactured using
electron-beam irradiation,23 peroxide decomposi-
tion,20 grafting branches from
a polymer back-
bone,24-29 addition of a branching agent to a polymeri-
zation,25,30-33 addition of a chain transfer agent to a
polymerization,34-36 or copolymerization of a monomer
with a macromonomer.37-65 Single-site catalysts can be
used for the synthesis of LCB polyolefins through the
copolymerization of an alkene-terminated macromono-
mer with a comonomer (Scheme 1).37-40,42,43,46,48-58,60-74
This method allows for the control of branch length and
distribution. In addition, the ability to synthesize LCB
polymers with different compositions on the side chains
and backbone allows for the precise design of polymer
architecture for applications in thermoplastic elasto-
mers75-79 and compatibilizers.80 Accessibility to mac-
romonomers with substantial molecular weight and
polymerizable end groups is advantageous for the
development of LCB polymers.
Isotactic polypropylene (iso-PP) is an important com-
mercial polymer, with nearly 18 billion pounds produced
in the United States in 2003.81 iso-PP has a high melting
point (Tm), tensile strength, and chemical resistance but
a low melt strength.82 Many of the applications for
* Corresponding author. E-mail gc39@cornell.edu.
10.1021/ma050735u CCC: $30.25 © 2005 American Chemical Society
Published on Web 06/21/2005

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
6260 Cherian et al.
Macromolecules, Vol. 38, No. 15, 2005
Scheme 2. Chain Termination in Propylene Polymerization with 1,2-Insertions
Scheme 3. Chain Termination in Propylene
Polymerization with 2,1-Insertions
Scheme 4. Syndiospecific Propylene Polymerization
which iso-PP is suitable, such as appliance and auto-
mobile parts, require injection-molding and melt-
processing methods. The inherent low melt strength of
iso-PP makes this type of processing difficult; however,
the addition of long chain branches to PP has been
shown to improve the melt strength greatly.22 Although
LCB-PP has been made using macromonomer incor-
poration methods,46,49-51,55-58,62,63,65 it is far less straight-
forward than the synthesis of LCB polyethylene. The
challenge lies in obtaining PP macromonomers with a
polymerizable end group and with an Mn that is
>2.5Me.17 Stereoregularity in PP has a significant
impact on Me and the plateau modulus. Isotactic and
atactic PP exhibit similar rheological behavior, with Me
determined to be ∼7000 g/mol,14 while for syndiotactic
PP Me has been reported to be below 3500 g/mol.14,83
The polymerization of propylene usually occurs through
1,2-insertions; thus, termination by â-hydride (â-H)
transfer gives the sterically congested vinylidene end
group84,85 (Scheme 2), which resists subsequent inser-
tion polymerization. Only â-CH3 elimination yields an
allylic end group in this primary insertion mech-
anism.86-89 Although most metallocenes exhibit low
selectivity for allylic end groups in polypropylene syn-
thesis,87 Cp*2MCl2 (M ) Zr, Hf) and bridged bis-
(fluorenyl)zirconocenes exhibit between 70 and 90%
â-CH3 elimination to give allyl-terminated atactic
polypropylene,87 while select chiral metallocenes exhibit
up to 80% â-CH3 elimination to give allyl-terminated
isotactic polypropylene.87,90,91 Subsequent polymeriza-
tion steps incorporate propylene as well as allyl-
terminated polypropylene chains, leading to the forma-
tion of LCB-PP.46,49-51,55-58,62,63,65
atactic to partially isotactic ([mmmm] e 0.67) PP
macromonomers by this method.62,92-94 These PP mac-
romonomers have been successfully copolymerized with
propylene using an isospecific metallocene complex.46
However, the Mn of these iso-PP macromonomers (e6500
g/mol) is substantially below the 2.5Me that is required
for LCB entanglements.14,17
Group IV complexes bearing chelating phenoxyimine
(PHI) ligands have been studied extensively for alkene
polymerization.96,97 Titanium(IV) complexes 1-3/
methylaluminoxane (MAO) give highly syndiotactic PP
(syndio-PP; Scheme 4) through chain-end control in a
2,1-insertion mechanism.95,98-102 Furthermore, com-
plexes with ortho fluorination on the N-aryl ring (e.g.,
2) produce PP with narrow molecular weight distribu-
tions.103-108 1H and 13C NMR spectroscopic end-group
analyses of PP synthesized with 3/MAO showed an
equal number of saturated and unsaturated end groups.95
The only unsaturated end groups observed were allylic,
consistent with â-H transfer following a secondary
insertion.95,107,109 Saturated end groups were composed
of a 1:1 mixture of n-propyl and n-butyl chain ends. This
composition was proposed to result from a 1,2-insertion
of propylene into the Ti-H bond, followed by a 50:50
distribution of 1,2- and 2,1-insertions. A chain-termina-
tion mechanism involving direct â-H transfer to mono-
mer has recently been proposed.107 Given this high
selectivity for allyl end groups, we reasoned that
catalyst 3/MAO could be used to make highly syndio-
tactic macromonomers with substantial molecular
weights. These macromonomers could subsequently be
used in the synthesis of LCB polymers with crystalline
side chains. Herein we describe the synthesis of a
variety of new catalysts and the study of different
reaction conditions to optimize the polymerization for
macromonomer production.
Another synthetic approach to allyl-terminated PP is
the use of catalysts in which propylene 2,1-insertions
predominate. As shown in Scheme 3, â-H transfer from
a secondary metal alkyl can produce either a 1-propenyl
end group or an allyl end group.84,87 Interestingly, allyl
end groups are observed exclusively in propylene poly-
merizations in which 2,1-insertions predominate.62,92-95
The resulting metal-hydride can then undergo propy-
lene insertion to generate a new polymer chain. Py-
ridinediimine-Fe(II) complexes, which polymerize pro-
pylene through a 2,1-insertion mechanism, generate
Results and Discussion
Complex Synthesis and Characterization. While
numerous isomeric forms of (PHI)2TiCl2 complexes could

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Macromolecules, Vol. 38, No. 15, 2005
Syndiotactic Polypropylene 6261
of this second isomeric form could be indicative of a
nonliving polymerization catalyst, we synthesized sev-
eral other (PHI)2TiCl2 complexes to determine the
significance of the isomeric forms.
In an effort to find a catalyst that was even more
active and stereoselective than is 3/MAO, systematic
steric and electronic modifications to the ligand struc-
ture were made. The phenoxyimine ligand synthesis
allows for a wide variety of ligands to be prepared from
commercially available anilines and readily synthesized
salicylaldehydes. Standard synthesis of (PHI)2TiCl2
complexes involves the deprotonation of the phenoxy-
imine ligand with n-butyllithium, followed by addition
of 0.5 equiv of TiCl4 (Scheme 5). Complexes 1-8 were
synthesized by varying the aniline to study the effect
of electron-withdrawing substituents on the N-aryl ring
(Scheme 5). When R1 ) R2 ) R3 ) F (2), the complex
exists only as the C2 isomer with trans oxygen atoms
and nitrogen and chlorine atoms oriented in a cis
configuration (no C1 isomer is observed using 1H NMR
spectroscopy).103 When R1 ) H (1, 3-8), the complex
exists as a mixture of isomers in solution, with 18-29%
C1 isomer. In complexes 9-12, the salicylaldehyde was
varied to study the effect of steric bulk in the R4
i
substituent. When R4 is Pr or a smaller substituent,
Figure 1. Molecular structure of 3 with thermal ellipsoids
at the 40% probability level (hydrogen atoms have been
omitted for clarity). Selected bond lengths [Å] and angles [deg]:
Ti(1)-O(1) 1.842(1), Ti(1)-O(2) 1.850(1), Ti(1)-N(1) 2.214(2),
Ti(1)-N(2) 2.239(2), Ti(1)-Cl(1) 2.285(1), Ti(1)-Cl(2) 2.307-
(1); O(1)-Ti(1)-O(2) 163.3(1), N(1)-Ti(1)-N(2) 84.2(1), Cl(1)-
Ti(1)-Cl(2) 95.8(1).
only the C2 isomer is observed using 1H NMR spectros-
copy, whereas if R4 is Bu or a larger substituent, the
t
complex is a mixture of C2 and C1 isomers. In fact, 11
exists primarily as the C1 isomer in solution, as deter-
mined by 1H NMR spectroscopy, and analysis of a single
crystal of 11 using X-ray crystallography (Figure 2)
revealed C1 symmetry, with nitrogen, oxygen, and
chlorine atoms cis.
exist, those complexes exhibiting living polymerization
behavior exist as a single C2-symmetric isomer, both in
the solid state and in solution.103,104 Analysis of a single
crystal of 3 using X-ray crystallography shows C2-
symmetric geometry similar to that of our previously
reported complexes,103,104 with oxygen atoms trans and
chlorine and nitrogen atoms cis (Figure 1). In solution,
however, 3 is determined to be a mixture of C2-
symmetric (73%) and C1-symmetric (27%) isomers using
1H NMR spetroscopy.110 Suspecting that the presence
To determine whether the formation of the C1 isomer
was a result of the preparation method, an alternate
synthetic route for 3 was investigated. Reaction of 2
equiv of (PHI)H with Ti(OiPr)4 and subsequent treat-
ment with excess Me3SiCl gave 3 with C2- and C1-
symmetric isomers in a 73:27 ratio (Scheme 6). A
NOESY NMR experiment indicates rapid exchange
between the C1-symmetric enantiomers of 3, but no
Scheme 5. Synthesis of (PHI)2TiCl2 Complexes 1-12

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
6262 Cherian et al.
Macromolecules, Vol. 38, No. 15, 2005
Figure 3. Representative 1H NMR spectrum (500 MHz,
1,1,2,2-tetrachloroethane-d2, 135 °C) of allyl-terminated syn-
diotactic polypropylene made using 3/MAO (S ) 1,1,2,2-
tetrachloroethane-d1).
Scheme 7. Synthesis of Allyl-Terminated
Syndiotactic Polypropylene Using (PHI)2TiCl2/MAO
Figure 2. Molecular structure of 11 with thermal ellipsoids
at the 40% probability level (hydrogen atoms and a toluene
molecule of solvation have been omitted for clarity). Selected
bond lengths [Å] and angles [deg]: Ti(1)-O(1) 1.833(2), Ti-
(1)-O(2) 1.845(2), Ti(1)-N(1) 2.228(3), Ti(1)-N(2) 2.262(3), Ti-
(1)-Cl(1) 2.299(1), Ti(1)-Cl(2) 2.341(1); O(1)-Ti(1)-O(2) 98.1-
(1), N(1)-Ti(1)-N(2) 88.2(1), Cl(1)-Ti(1)-Cl(2) 94.8(1).
Scheme 6. Alternate Synthesis of Complex 3
ficities. Electron-withdrawing groups in the R3 position
(5 and 8) increase the polymerization rate to a small
degree, and the resulting polymers are more stereoreg-
ular and have higher Mn than those produced by 1.
When these effects are combined, with R2 ) R3 ) F (4),
the polymerization rate is even higher while stereocon-
trol appears unchanged and the Mn of the polymer is
slightly lower than those from complexes 3 or 5.
Fluorination in the R1 positions (2), as already de-
scribed, decreases activity and leads to a polymer with
a very narrow molecular weight distribution and high
syndiotacticity.
evidence of exchange between the C2- and C1-symmetric
isomers was observed (see Supporting Information).
However, analysis of a single crystal of complex 6 by
1H NMR spectroscopy shows the C2-symmetric isomer
upon initial dissolution, with peaks corresponding to the
C1-symmetric isomer growing in over ≈30 min. Whereas
C1-symmetric (and presumably C2-symmetric) isomers
racemize quickly, interconversion between C2- and C1-
symmetric isomers occurs much more slowly.
Polymerization Results. Complexes 1-12 were
studied for the polymerization of propylene to determine
which factors influence the polymerization rate, termi-
nation rate, tacticity, and Mn of the desired syndio-PP
macromonomer (Scheme 7).111 â-H transfer during po-
lymerization leaves an allyl end group that is easily
identifiable using 1H NMR spectroscopy, with the vinyl
peaks appearing at 5.0 and 5.8 ppm (Figure 3).3,86,112
Ligand substitution has a dramatic effect on catalyst
activity as well as on the tacticity and Mn of the
resulting polymer (Table 1). The presence of electron-
withdrawing substituents in the R2 position (3, 6, and
7) increases the polymerization rate by almost an order
of magnitude (compared to 1) while leaving Mn es-
sentially unchanged. These effects are most pronounced
with R2 ) F (3), whereas larger groups, such as Cl (6)
and CF3 (7), are less active and have lower syndiospeci-
Polymerization behavior is dramatically influenced by
the steric bulk of the R4 substituent on the phenoxy-
imine ligand. Decreasing the steric bulk of R4 from tBu
i
(3) to Me (9) or Pr (10) greatly increases the catalyst
activity. These complexes produce polymer with much
lower tacticity than that of 3 (Table 1). Unexpectedly,
complexes 9 and 10 produce high molecular weight
polymer, show no evidence of â-H transfer, and have
narrow polymer molecular weight distributions. To the
best of our knowledge, these are the first bis(phenoxy-
imine)titanium complexes without ortho fluorination
that exhibit living behavior for propylene polymeriza-
tion, and they will be the subject of future investigation
in our lab.
t
Increasing the steric bulk from Bu (3) to SiMe3 (12)
has little effect on the catalyst activity, but the Mn of
the polymer increased (7100 g/mol) as did the tacticity
of the polymer, with [rrrr] of 0.94 and Tm of 146 °C. We
expected 11/MAO to perform similarly, but instead very
little polymer was formed, and the molecular weight
distribution was multimodal. As 11 is 91% C1 isomer
and complexes 1-10 and 12 give polymers with a Mw/
Mn e 2sconsistent with a single active speciesswe infer