Ligand exchange and abstraction reactions involving titanium isopropoxide with tris(pentafluorophenyl)borane and -alane: Ramifications for ring-opening polymerization of propylene oxide

Article abstract of DOI:10.1016/j.ica.2004.05.015

In situ mixing of Ti(OⁱPr)₄ and B(C₆F ₅)₃ generates a very efficient catalyst system for ring-opening polymerization (ROP) of propylene oxide (PO) with a turnover frequency (TOF) >1350/h, whereas the mixture of Ti(OⁱPr) ₄ and Al(C₆F₅)₃ is inactive for the same polymerization. The inactivity of the Ti(OⁱPr) ₄/Al(C₆F₅)₃ mixture is due to the formation of the stable isopropoxy-bridged bimetallic species Ti(O ⁱPr)₃(μ-OⁱPr)Al(C₆F ₅)₃ (1), the structure of which has been confirmed by X-ray diffraction. The products of the Ti(OⁱPr)₄+B(C ₆F₅)₃ reaction, however, depend on the Ti(OⁱPr)₄:B(C₆F₅)₃ ratio. The 1:1 ratio reaction in toluene at ambient temperature is rapid and produces the ligand exchange products: Ti(OⁱPr)₃C₆F ₅ and ⁱPrOB(C₆F₅)₂ (2), along with a small amount of (ⁱPrO)₂BC₆F ₅. The two resulting boranes are inseparable by recrystallization or vacuum distillation, and the formation of the undesired (ⁱPrO) ₂BC₆F₅ is either significantly enhanced upon heating the reaction in toluene to 80°C or nearly exclusive by carrying out the reaction in THF. By employing 1.2 equiv. of B(C₆F ₅)₃ in the reaction with Ti(OⁱPr)₄, however, the formation of (ⁱPrO)₂BC₆F ₅ is suppressed, enabling the isolation of the new borane 2 in its pure state. The excess of B(C₆F₅)₃ added to the reaction apparently slows down the exchange reaction by stabilizing the intermediate Ti(OⁱPr)₃(μ-OⁱPr)B(C ₆F₅)₃ (4), as shown by the 1:2 Ti(O ⁱPr)₄:B(C₆F₅)₃ reaction which initially forms the ligand abstraction product 4 followed by subsequent slow ligand exchange to give the final products Ti(OⁱPr) ₃C₆F₅ and 2. The studies of these individual reactions, in combination with control polymerization runs, reveal that the active species responsible for the catalytic activity of the Ti(O ⁱPr)₄/B(C₆F₅)₃ mixture is the isopropoxy borane 2. Thus, the isolated 2, in the absence or presence of a hydroxylic initiator, serves as a very effective catalyst for the ROP of PO, producing PPOs with M_n=2000-3000, M_w/M_n=1.30-1. 43, and TOF >1400/h. The MALDI-TOF MS analyses of the PPOs formed show the linear PPO structures having the initiator and water molecules as end groups, demonstrating the control over the PPO structure.

Full text of DOI:10.1016/j.ica.2004.05.015

Inorganica Chimica Acta 357 (2004) 3911–3919
www.elsevier.com/locate/ica
Ligand exchange and abstraction reactions involving titanium
isopropoxide with tris(pentafluorophenyl)borane and -alane:
ramifications for ring-opening polymerization of propylene oxide q
*
Antonio Rodriguez-Delgado, Eugene Y.-X. Chen
Department of Chemistry, Colorado State University, Fort Collins, CO 80523-1872, USA
Received 30 March 2004; accepted 1 May 2004
Available online 20 July 2004
Dedicated to Professor T.J. Marks on the occasion of his 60th birthday
Abstract
In situ mixing of Ti(OiPr)4 and B(C6F5)3 generates a very efficient catalyst system for ring-opening polymerization (ROP) of
propylene oxide (PO) with a turnover frequency (TOF) >1350/h, whereas the mixture of Ti(OiPr)4 and Al(C6F5)3 is inactive for the
same polymerization. The inactivity of the Ti(OiPr)4/Al(C6F5)3 mixture is due to the formation of the stable isopropoxy-bridged
bimetallic species Ti(OiPr)3(l-OiPr)Al(C6F5)3 (1), the structure of which has been confirmed by X-ray diffraction. The products of
the Ti(OiPr)4 + B(C6F5)3 reaction, however, depend on the Ti(OiPr)4:B(C6F5)3 ratio. The 1:1 ratio reaction in toluene at ambient
temperature is rapid and produces the ligand exchange products: Ti(OiPr)3C6F5 and iPrOB(C6F5)2 (2), along with a small amount of
(iPrO)2BC6F5. The two resulting boranes are inseparable by recrystallization or vacuum distillation, and the formation of the un-
desired (iPrO)2BC6F5 is either significantly enhanced upon heating the reaction in toluene to 80 °C or nearly exclusive by carrying out
the reaction in THF. By employing 1.2 equiv. of B(C6F5)3 in the reaction with Ti(OiPr)4, however, the formation of (iPrO)2BC6F5 is
suppressed, enabling the isolation of the new borane 2 in its pure state. The excess of B(C6F5)3 added to the reaction apparently slows
down the exchange reaction by stabilizing the intermediate Ti(OiPr)3(l-OiPr)B(C6F5)3 (4), as shown by the 1:2 Ti(OiPr)4:B(C6F5)3
reaction which initially forms the ligand abstraction product 4 followed by subsequent slow ligand exchange to give the final products
Ti(OiPr)3C6F5 and 2. The studies of these individual reactions, in combination with control polymerization runs, reveal that the
active species responsible for the catalytic activity of the Ti(OiPr)4/B(C6F5)3 mixture is the isopropoxy borane 2. Thus, the isolated 2,
in the absence or presence of a hydroxylic initiator, serves as a very effective catalyst for the ROP of PO, producing PPOs with
Mn ¼ 2000–3000, Mw=Mn ¼ 1:30–1.43, and TOF >1400/h. The MALDI-TOF MS analyses of the PPOs formed show the linear PPO
structures having the initiator and water molecules as end groups, demonstrating the control over the PPO structure.
Ó 2004 Elsevier B.V. All rights reserved.
Keywords: Polymerization of propylene oxide; Titanium isopropoxide; Tris(pentafluorophenyl)borane; Tris(pentafluorophenyl)alane; Ligand
exchange; Ligand abstraction; Poly(propylene oxide)
1. Introduction
molecular weight PPOs of a few thousand Da are key
intermediates in polyurethane production [2] in which
primary-hydroxyl-terminated polyols undergo polycon-
densation reactions with diisocyanates in certain rapid
processes such as polyurethane reaction injection
molding (Eq. (1)),
Poly(propylene oxide) (PPO) is a useful material in a
number of applications [1]. In particular, low to medium
q
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.ica.2004.05.015.
* Corresponding author. Tel.: +1-970-491-5609; fax: +1-970-491-
1801.
O
C
O
C
+
HO
R
OH
OCN R' NCO
O
R
O
NH R' NH
E-mail address: eychen@lamar.colostate.edu (E.Y.-X. Chen).
ð1Þ
0020-1693/$ - see front matter Ó 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.ica.2004.05.015

3912
A. Rodriguez-Delgado, E.-Y.X. Chen / Inorganica Chimica Acta 357 (2004) 3911–3919
Ti(OiPr)4/B(C6F5)3
Base, acid, and coordination catalyses are three
mechanistically distinct chemical processes commonly
used for the production of PPOs through ring-opening
polymerization (ROP) of propylene oxide (PO). In the
base-catalyzed, anionic ROP process, PO is combined
with a hydroxylic initiator and a strongly basic catalyst
(e.g., potassium hydroxide) [1], whereby the initiator
compound determines the functionality of PPO. Ethyl-
ene oxide is often added in a subsequent step at the end
of the polymerization of PO to afford the reactive, pri-
mary-hydroxyl-terminated PPOs. A side reaction of
concern in this process is the formation of allyl alcohol
that leads to a terminally unsaturated, monofunctional
polyol [3].
O
iPrOB(C6F5)2
iPrOB(C6F5)2
iPrOH, HO(CH2)4OH
HOCH(CH2OH)2
O
(R)H
OR(H)
n
5000 equiv
80-91 % conversion
(toluene, 23 oC, 3 h)
Scheme 1.
tems, as well as the implications of these reactions for
PO polymerization (Scheme 1).
PPOs produced by acidic catalysts such as boron
trifluoride diethyl etherate and HBF4 [1,4] typically
contain considerable amounts of byproducts, including
dimethyldioxane and various concurrent cyclic oligo-
mers [5] due to ‘‘backbiting’’ reactions. To substantially
reduce the formation of such byproducts, Penczek and
co-workers [6] have developed cationic ‘‘activated
monomer’’ ring-opening polymerization where a het-
erocyclic monomer is activated via protonation by
Brønsted acids such as HBF4 and the growing chain end
is the neutrally charged hydroxyl group, instead of the
highly reactive, electrophilic oxonium ion (i.e., active-
chain end). To suppress the reactions proceeding with
the active-chain end mechanism, the ‘‘activated mono-
mer’’ ROP is typically carried out under monomer
starving conditions.
2. Experimental
2.1. Materials and methods
All syntheses and manipulations of air- and moisture-
sensitive materials were carried out in flamed Schlenk-
type glassware on a dual-manifold Schlenk line, on a
high vacuum line (10ꢀ5–10ꢀ7 Torr) or in an argon-filled
glovebox (<1.0 ppm oxygen and moisture). NMR scale
reactions were conducted in Teflon-valve-sealed sample
J. Young-type NMR tubes. Anhydrous or HPLC-grade
organic solvents were first saturated with nitrogen dur-
ing filling of the solvent reservoir and then dried by
passage through activated alumina and Q-5 supported
copper catalyst in stainless steel columns prior to use.
Benzene-d6 and toluene-d8 were dried over Na/K alloy
and distilled and/or filtered prior to use, whereas CDCl3
Coordination catalysis typically produces PPOs hav-
ing high molecular weight or with controlled structures
[7]. For example, discrete (porphyrin)Al–X (X ¼ Cl, OR,
SR) complexes promote living or immortal coordination
polymerization of PO [8]. Other examples of the discrete
coordination catalysts investigated for PO polymeriza-
tion include (Salen)Al–X [9] bulky bisphenoxide alumi-
num chloride dimers [10], isobutylalumoxane/PO
complex [11]. The proposed bimetallic nature [12] of the
PO polymerization by Lewis acidic metal complexes has
been recently confirmed with the microstructure analysis
of PPOs [10] and with the synthetic and polymerization
study that PO polymerization by neutral aluminum Le-
wis acid/anionic aluminate pair is much faster and more
controlled than that by the neutral catalyst alone [13].
We previously reported the PO polymerization cata-
lyzed by strong organo-Lewis acids such as M(C6F5)3
(M ¼ B, Al) and initiated by hydroxylic initiators such
as 1,4-butanediol [14]. Because of the instability of the
alane catalyst in the presence of a hydroxylic initiator,
we sought to explore the possibility of combining the
Lewis acid catalysts M(C6F5)3 with an organometallic
complex carrying an initiating ligand such as Ti(OiPr)4.
We report here the PO polymerization behavior of the
Ti(OiPr)4/M(C6F5)3 catalyst systems, the ligand ex-
change and abstraction reactions involved in these sys-
ꢀ
was dried over activated Davison 4 A molecular sieves.
NMR spectra were recorded on a Varian Inova 300 (FT
300 MHz, 1H; 75 MHz, 13C; 282 MHz, 19F; and 96
MHz, 11B) spectrometer. Chemical shifts for 1H and 13
C
spectra were referenced to internal solvent resonances
and are reported as parts per million relative to tetra-
methylsilane. 19F NMR and and 11B spectra were ref-
erenced to CFCl3 and BF3 ꢁ OEt2 external standards,
respectively. Elemental analyses were performed by
Desert Analytics, Tucson, AZ.
Propylene oxide (PO), titanium isopropoxide, anhy-
drous isopropanol, 1,4-butanediol, and glycerin were
purchased from Aldrich Chemical Co. and used as re-
ceived unless otherwise indicated. PO and 1,4-butane-
diol were first degassed and dried over CaH2 overnight
and then vacuum-distilled before use.
Tris(perfluorophenyl)borane, B(C6F5)3, was obtained
as a research gift from Boulder Scientific Company and
further purified by recrystallization from hexanes at
)35 °C. Tris(pentafluorophenyl)alane, Al(C6F5)3, as a
0.5 ꢁ toluene adduct, was prepared according to the lit-
erature procedure [15], which is the modified synthesis of
the alane first disclosed by Biagini et al. [16] Extra
caution should be exercised when handling this material
because of its thermal and shock sensitivity.

A. Rodriguez-Delgado, E.-Y.X. Chen / Inorganica Chimica Acta 357 (2004) 3911–3919
3913
2.2. Synthesis of Ti(OiPr)3(l-OiPr)Al(C6F5)3 (1)
2.4. Reaction of Ti(OiPr)4 and 1.2 equiv. of B(C6F5)3:
i
Isolation of the pure PrOB(C6F5)2 (2)
In an argon-filled glovebox, a 20-mL glass reactor
was equipped with a magnetic stir bar and charged
with Al(C6F5)3 ꢁ 0.5 toluene (0.24 g, 0.42 mmol) and
10 mL of toluene. To this reactor was added Ti(OiPr)4
(0.12 mL, 0.42 mmol) at ambient temperature. The
resulting yellow clear solution was stirred for 45 min,
the volatiles of which were removed in vacuo to ob-
tain a pale yellow solid. The crude material was re-
dissolved in a minimum amount of hexanes and the
solution was cooled to )30 °C inside a freezer of the
glovebox. Colorless needles formed were filtered and
dried to give 0.22 g (75%) of the pure product 1.
Single crystals suitable for X-ray diffraction were ob-
tained from a solution of 1 in a 3:1 toluene/hexanes
solvent mixture at )30 °C in 3 h.
The reaction was carried out in the similar manner as
the 1:1 ratio reaction shown above, except for the
change of the amount of the borane to 1.2 equiv. and the
reaction time to 24 h. The crude product of 2 (89.2 %)
was vacuum-distilled at 100–110 °C/0.1 Torr. The first
two minor fractions contained predominant 2 plus a
trace amount of 3, whereas the third fraction gave the
spectroscopically pure 2; yield 62.9 %. The 1H, 19F, 13C,
and 11B NMR data of 2 are reported above.
2.5. Reaction of Ti(OiPr)4 with 2 equiv. of B(C6F5)3: in
situ generation of Ti(OiPr)3(l-OiPr)B(C6F5)3 (4)
In an argon-filled glovebox, a 6-mL glass vial was
charged with B(C6F5)3 (20.4 mg, 0.04 mmol) and 0.7 mL
of benzene-d6. To this vial was added Ti(OiPr)4 (5.8 lL,
0.02 mmol), and the resulting mixture was transferred to
a Teflon-valve-sealed sample J. Young NMR tube via
1H NMR (C6D6, 23 °C) for Ti(OiPr)3(l-
OiPr)Al(C6F5)3 (1): d 4.71 (sept, J ¼ 6:3 Hz, 1H,
CHMe2), 4.03 (sept, J ¼ 6:0 Hz, 3H, CHMe2), 1.26 (d,
J ¼ 6:0 Hz, 6H, CHMe2), 1.26 (d, J ¼ 6:0 Hz, 18H,
CHMe2). 19F NMR (C6D6, 23 °C): d )121.41 (d,
3JF–F ¼ 16:9 Hz, 6F, o-F), )153.23 (t, 3JF–F ¼ 19:52 Hz,
2F, p-F), )161.93 (m, 6F, m-F). 13C NMR (C6D6, 23
°C): d 152.17, 148.70, 138.92, 135.52 (C6F5), 82.57
(CHMe2), 76.94 (CHMe2), 25.45 (CHMe2), 23.80
(CHMe2). Anal. Calc. for C30H28AlF15O4Ti: C, 44.35;
H, 3.47. Found: C, 43.95; H, 3.47%.
pipette. The reaction was then monitored by 1H and 19
F
NMR spectroscopy at 23 °C. The formation of 4 was
observed within 10 min, while the other minor species
(Ti(OiPr)3C6F5 and 2) were also detected. After sitting
in the NMR tube for 24 h, the reaction gave
Ti(OiPr)3C6F5 and 2, plus a trace amount of 4.
1H NMR (C6D6, 23 °C) Ti(OiPr)3(l-OiPr)B(C6F5)3
(4): d 4.48 (sept, J ¼ 6 Hz, 1H, CHMe2), 4.11 (sept,
J ¼ 6:0 Hz, 3H, CHMe2), 1.29 (br, 6H, CHMe2), 0.91
(d, J ¼ 6:0 Hz, 18H, CHMe2). 19F NMR (C6D6, 23 °C):
d )130.41 (br, 6F, o-F), )156.76 (br, 2F, p-F), )164.27
(b, 6F, m-F).
2.3. Reaction of Ti(OiPr)4 with 1 equiv. of B(C6F5)3
In a glovebox, a 60-mL glass reactor was equipped
with a magnetic stir bar and charged with a 25 mL
solution B(C6F5)3 (2.00 g, 3.34 mmol) in toluene. To
this reactor was added Ti(OiPr)4 (0.95 g, 3.34 mmol)
at ambient temperature. The resulting mixture was
stirred for 45 min, upon which time the mixture
turned to yellow. The volatiles were removed in vacuo
2.6. Polymerization procedures and polymer character-
izations
Polymerization of propylene oxide was carried out in
50-mL Schlenk flasks on a Schlenk line. A weighed,
flame-dried flask equipped with a magnet stirrer and
capped with a septum was charged with PO (5.00 mL,
4.15 g, 71.5 mmol). For polymerizations employing a
hydroxylic initiator, to this flask was added via micro-
syringe the degassed hydroxylic initiator in the desired
PO/initiator ratio (see Table 1). After a desired tem-
perature (23 °C) was reached using the external tem-
perature-controlled bath, 2.0 mL of the freshly prepared
catalyst solution in toluene (14.3 lmol, [PO]o:[Cat]o ¼
5000:1) was added into the rapidly stirred flask using a
gastight syringe. The reaction was stirred for the mea-
sured time interval and unreacted monomer was quickly
removed under reduced pressure. The residue was dried
under 0.1–0.2 Torr at ambient temperatures for 2 h,
yielding a viscous oil. After the polymer yield was
measured, the residue was dissolved in 25 mL of methyl-
ene chloride and washed with 15 mL of 0.1 M HCl in a
1
to obtain a pale yellow oil, the H and 19F NMR of
which showed the formation of three species:
Ti(OiPr)3C6F5
[17],
iPrOB(C6F5)2
(2),
and
(iPrO)2BC6F5 (3) [18]. The crude material was ex-
tracted with hexanes and filtered, and removal of the
solvent of the filtrate produced 2 in ꢂ94% purity, with
a small amount of 3 still present. Vacuum distillation
of the oily product twice at 120–130 °C/0.2 Torr gave
the slightly purer 2, but it was still contaminated with
a trace amount 3.
i
1H NMR (C6D6, 23 °C) for PrOB(C6F5)2 (2): d 4.14
(sept, J ¼ 6:0 Hz, 1H, CHMe2), 1.07 (d, J ¼ 6:0 Hz, 6H,
CHMe2). 19F NMR (C6D6, 23 °C): d )133.14 (br, 6F, o-
3
F), )149.70 (t, JF–F ¼ 19:5 Hz, 3F, p-F), )161.26 (br,
6F, m-F). 13C NMR (C6D6, 23 °C): d 149.02, 144.50,
139.32 (C6F5, Cipso obscured), 74.42 (CHMe2), 23.93
(CHMe2). 11B NMR (C6D6, 23 °C): d 39.58.

3914
A. Rodriguez-Delgado, E.-Y.X. Chen / Inorganica Chimica Acta 357 (2004) 3911–3919
dried and degassed at 120 °C/10ꢀ6 Torr for 24 h). The
crystals were then mounted on thin glass fibers and
transferred into the cold-steam of a Siemens SMART
CCD diffractometer. The structures were solved by di-
rect methods and refined using the Siemens SHELXTL
program library [19]. The structure was refined by full-
matrix weighted least-squares on F 2 for all reflections.
All non-hydrogen atoms were refined with anisotropic
displacement parameters. Hydrogen atoms were in-
cluded in the structure factor calculations at idealized
positions unless otherwise indicated. Selected crystal
data and structural refinement parameters are collected
in Table 1.
Table 1
Crystal data and structure refinements for 1
Empirical formula
Formula weight
Temperature (K)
C30H28AlF15O4Ti
812.40
167(2)
ꢀ
Wavelength (A)
0.71073
triclinic
Crystal system
Space group
ꢁ
P1
ꢀ
a (A)
10.0821(16)
10.9600(17)
16.715(3)
ꢀ
b (A)
ꢀ
c (A)
a (°)
b (°)
c (°)
77.904(3)
89.719(3)
72.961(3)
1723.6(5)
3
ꢀ
Volume (A )
Z
2
1.565
Dcalc (mg/m3)
Absolute coefficient (mmꢀ1
F ð000Þ
)
0.389
820
3. Results and discussion
Crystal size (mmꢀ3
)
0.30 ꢃ 0.30 ꢃ 0.25
1.99 to 28.38
)12 6 h 6 13, )14 6 k 6 14,
)21 6 l 6 22
15 999
3.1. PO polymerization catalyzed by Ti(OiPr)4/
h Range for data collection (°)
Index ranges
M(C6F5)3 (M ¼ B, Al)
Reflections collected
In situ mixing of Ti(OiPr)4 and B(C6F5)3 generates an
efficient catalyst system for ROP of PO. In a ratio of
[PO]o:[Cat]o ¼ 5000:1, a PO conversion >80% in 3 h of
polymerization time was achieved, reflecting a turnover
frequency (TOF) >1350/h (entries 1–3, Table 2). The PPO
obtained was analyzed by MALDI-TOF MS, giving a Mn
of 2170 Da and a molecular weight distribution of
Mw=Mn ¼ 1:41. As can be seen from the table, the pre-
mixing time, which ranged from 2 min to 3 h, had virtually
no effect on polymerization activity in terms of TOF
numbers (entries 1–3), whereas increasing the Ti(OiPr)4:
B(C6F5)3 ratio from 1 to 10 significantly reduced the
polymerization activity (entries 4–6 versus 1–3).
In sharp contrast, the system derived from in situ
mixing of Ti(OiPr)4/Al(C6F5)3 was completely inactive
for PO polymerization (entry 7). Control runs using
Ti(OiPr)4 or Ti(OiPr)3C6F5 showed no activity for either
reagent (entries 8 and 9), whereas the polymerization
using B(C6F5)3 or Al(C6F5)3 alone produced only dimer,
trimer, and other low molecular weight oligomers [14].
Thus, the substantially increased molecular weight of the
resulting PPO with Ti(OiPr)4/B(C6F5)3 and the inactivity
for Ti(OiPr)4/Al(C6F5)3 pointed to an important role of
Ti(OiPr)4 in these two catalyst systems. To understand
their catalytic behavior, individual reactions involved in
these systems were investigated in detail, the results of
which are discussed in the following two sections.
Independent reflections
Absorption correction
Data/restraints/parameters
Goodness-of-fit on F 2
Final R indices [I > 2rðIÞ]
R indices (all data)
8220 [Rint ¼ 0:0403]
SADABS
8220/0/460
1.047
R1 ¼ 0:0661, wR2 ¼ 0:1717
R1 ¼ 0:0831, wR2 ¼ 0:1866
1.439 and )0.601
Largest difference peak and
3
ꢀ
hole (e/A )
separatory funnel. The organic layer was washed with
2 ꢃ 10 mL distilled water and dried over anhydrous
MgSO4. The mixture was filtered; the solvent of the
filtrate was removed in vacuo to yield PPOs, typically as
colorless, viscous oils.
1
All PPOs produced from this study have similar H
1
NMR spectra. H NMR (CDCl3, 23 °C): d 3.8–3.2 (m,
3H, CHMe and CH2), 1.2–1.0 (m, 3H, Me). The PPO
samples were analyzed by Matrix-assisted laser desorp-
tion/ionization time-of-flight mass spectroscopy (MAL-
DI-TOF MS). MALDI-TOF MS was performed on a
Voyager DE Pro (Perseptive Biosystems) mass spec-
trometer operated in linear, delayed extraction, positive
ion mode using a N2 laser at 337 nm and 20 kV accel-
erating voltage. Matrix dithranol in THF was mixed
with NaI (aqueous solution), followed by mixing with
PPO (THF solution) before this mixture was added to
the target.
3.2. Reaction of Ti(OiPr)4 with Al(C6F5)3 and X-ray
structure of Ti(OiPr)3(l-OiPr)Al(C6F5)3 (1)
2.7. X-ray crystallographic analysis of 1
Single crystals suitable for X-ray diffraction studies
were obtained by recrystallization from a saturated so-
lution of 1 in a 3:1 toluene/hexanes solvent mixture at
)30 °C inside of a freezer of the glovebox. The solvents
were decanted in the glovebox and the crystals were
quickly covered with a layer of Paratone-N oil (Exxon,
The reaction of Ti(OiPr)4 with Al(C6F5)3 in toluene
proceeds cleanly to produce the microcrystalline, iso-
propoxy-bridged bimetallic species: Ti(OiPr)3(l-OiPr)-
Al(C6F5)3 (1). As already shown in Table 2, the
Ti(OiPr)4/Al(C6F5)3 mixture is inactive for PO poly-

A. Rodriguez-Delgado, E.-Y.X. Chen / Inorganica Chimica Acta 357 (2004) 3911–3919
3915
Table 2
Summary of PO polymerization catalyzed by Ti(OiPr)4/M(C6F5)3
a
Entry Catalyst
Premixing time (min)
Polymerization time (h)
Conversion (%)
TOFb (/h)
1.
2.
3.
Ti(OiPr)4/B(C6F5)3
Ti(OiPr)4/B(C6F5)3
Ti(OiPr)4/B(C6F5)3
2
3
3
3
82.4
81.9
81.4
1370
1365
1356
15
180
4.
5.
6.
2 Ti(OiPr)4/B(C6F5)3
3 Ti(OiPr)4/B(C6F5)3
10 Ti(OiPr)4/B(C6F5)3
180
180
180
3
3
3
38.6
31.8
12.8
640
530
210
7.
8.
9.
Ti(OiPr)4/Al(C6F5)3
Ti(OiPr)4
Ti(OiPr)3C6F5
120 or 180
N.A.
N.A.
3
3
3
0
0
0
0
0
0
a Carried out at 23 °C, 5.00 mL (4.15 g) of PO, [PO]o:[Cat]o ¼ 5000.
b Turnover frequency (TOF) ¼ moles of PO consumed per mole of catalyst per hour; calculations were based on polymer yield.
merization; likewise, the PO polymerization using the
isolated complex 1 showed no activity, indicative of the
high stability of this adduct.
3.3. Reaction of Ti(OiPr)4 with B(C6F5)3: isolation of
iPrOB(C6F5)2 (2) and observation of unstable interme-
diate Ti(OiPr)3(l-OiPr)B(C6F5)3 (4)
The bimetallic complex 1 crystallizes in the triclinic
ꢁ
space group P1. The crystal structure of 1 (Fig. 1) fea-
The reaction of Ti(OiPr)4 with B(C6F5)3 is sensitive
to the ratio of the two reactants and solvent. The 1:1
ratio reaction in toluene at ambient temperature un-
dergoes fast ligand exchange producing a product mix-
tures the tight ion pair in which one isopropoxy ligand is
unsymmetrically bonded to the distorted tetrahedral ti-
tanium and aluminum centers with a Ti–O–Al vector
angle of 116.58(9)°. This bridging isopropoxy ligand to
i
ture consisting of Ti(OiPr)3C6F5 and PrOB(C6F5)2 (2),
ꢀ
Ti distance is ꢂ0.238 A longer than the distance of Ti to
along with
a
small amount of the undesired
the rest of three terminal isopropoxy ligands. The av-
ꢀ
erage Al–C(aryl) distance (1.998 A) compares well with
(iPrO)2BC6F5 (3) (Scheme 2). The boranes 2 and 3 can
be readily separated from the titanium species by hexane
extraction; however, all attempts to separate themselves
using either recrystallization or vacuum distillation af-
those in other Al(C6F5)3 complexes with metallocene
alkyl [20], imidazole [21], toluene [22], THF [23], water
ꢀ
[24], and methanol [24]; the Al–O distance (1.834(2) A)
is noticeably shorter than those observed in the adducts
1
forded 2 in ꢂ94% purity based on H NMR. The 1:1
ratio reaction at 80 °C substantially enhanced the for-
mation of 3, reaching a 2:3 ratio of 3:1 in the product
mixture, whereas the reaction in THF selectively pro-
duced 3, along with the co-product Ti(OiPr)3C6F5 and a
trace amount of 2 (Scheme 2).
ꢀ
of Al(C6F5)3 with THF (1.860(2) A), water (1.857(3) A),
ꢀ
ꢀ
and methanol (1.858(3) A).
Interestingly, employing 1.2 equiv. of B(C6F5)3 in the
reaction with Ti(OiPr)4 in toluene considerably slows
down the exchange reaction, and more significantly, it
suppresses the formation of the undesired 3. Using this
strategy, the pure borane 2 was finally isolated in 63%
yield through the 1:1.2 ratio reaction (Scheme 3), fol-
lowed by recrystallization and vacuum distillation of the
crude reaction products. Addition of B(C6F5)3 to a 2:3
mixture of a known ratio did not change the original
ratio after more than 4 h of reaction time; this experi-
ment suggests that the cleaner formation of 2 in the 1:1.2
reaction than in the 1:1 reaction is not due to the pos-
sible conversion of 3 to 2 by B(C6F5)3 via ligand redis-
tribution (Scheme 3).
To further probe the role of an excess of B(C6F5)3 in
suppressing the formation of 3 and in slowing down the
ligand exchange reaction, we investigated the reaction
of Ti(OiPr)4 and B(C6F5)3 in a 1:2 ratio. This reac-
tion is fast, but it produces the ligand abstraction
product Ti(OiPr)3(l-OiPr)B(C6F5)3 (4), which eventually
ꢀ
Fig. 1. Molecular structure of 1. Selected bond lengths (A) and angles
(°): Ti–O(4) ¼ 1.979(2), Al–O(4) ¼ 1.834 (2), Ti–O(1) ¼ 1.741(2), Ti–
O(2) ¼ 1.739(2), Ti–O(3) ¼ 1.741(2), Al–C(13) ¼ 2.007(2), Al–C(19)
¼ 1.997(3), Al–C(25) ¼ 1.992(2); Ti– O(4)–Al ¼ 116.58(9).

3916
A. Rodriguez-Delgado, E.-Y.X. Chen / Inorganica Chimica Acta 357 (2004) 3911–3919
toluene
(C6F5)Ti(OiPr)3 + iPrOB(C6F5)2 (2) + (iPrO)2BC6F5 (3)
RT, 45min
10
:
1
toluene
80 oC, 14 h
Ti(OiPr)4 + B(C6F5)3
(C6F5)Ti(OiPr)3 + iPrOB(C6F5)2 (2) + (iPrO)2BC6F5 (3)
3
:
1
THF
(C6F5)Ti(OiPr)3 + iPrOB(C6F5)2 (2) + (iPrO)2BC6F5 (3)
RT, 90min
trace
Scheme 2.
toluene
Ti(OiPr)4 + 1.2 B(C6F5)3
(C6F5)Ti(OiPr)3 + iPrOB(C6F5)2 (2) + (iPrO)2BC6F5 (3)
RT
trace
toluene
B(C6F5)3 + (iPrO)2BC6F5 (3)
2 iPrOB(C6F5)2 (2)
RT
Scheme 3.
3. In the absence of any hydroxylic initiators, 2 effec-
toluene
RT, 6 h
Ti(OiPr)4 + 2 B(C6F5)3
+ B(C6F5)3
O
tively polymerizes PO to moderate molecular weight
PPO (Mn ¼ 2550, Mw=Mn ¼ 1:43) with a TOF of 1438/h
(entry 1, Table 3). This result is in sharp contrast to the
PO polymerization result by B(C6F5)3, which produced
only dimer, trimer, and other low molecular weight
oligomers [14]. This marked difference is originated from
their substantially dissimilar reactivity patterns towards
the PO monomer and the alcohol initiator: while
B(C6F5)3 forms stable adducts with the initiator and
reacts vigorously with PO in the absence of the initiator
to produce low molecular weight oligomers and the
isomerization product – propionaldehyde [14], 2 does
not form adducts with the initiator and reacts smoothly
with PO in the absence of the initiator to produce only
PPO. Nevertheless, the polymerization results by 2
compare well with – despite slightly better than – those
obtained directly by the Ti(OiPr)4/B(C6F5)3 mixture
(entry 1, Table 2). The decreasing of polymerization
activity with increasing the Ti(OiPr)4:B(C6F5)3 ratio
from 1 to 10 (entries 4–6, Table 2) correlates with the
amount of 2 produced from the ligand exchange reac-
tion; with an excess of Ti(OiPr)4 employed, significant
amounts of other species such as Ti(OiPr)3C6F5,
(iPrO)2BC6F5, and B(iPrO)3 are formed, along with the
unreacted Ti(OiPr)4, all of which exhibit negligible PO
polymerization activity by control experiments. These
(iPrO)3Ti
B(C6F5)3
4
RT
28 h
(C6F5)Ti(OiPr)3 + iPrOB(C6F5)2 (2)
Scheme 4.
undergoes slow ligand exchange to give Ti(OiPr)3C6F5
and 2 (Scheme 4). This experiment suggests that the li-
gand exchange reaction observed for the 1:1 reaction
occurs via the unstable intermediate 4, and that in the 1:2
reaction the excess of B(C6F5)3 stabilizes the bimetallic
intermediate, allowing its direct detection in solution.
This stabilization also renders the ligand exchange in the
1:1.2 Ti(OiPr)4:B(C6F5)3 reaction to occur in more
controlled the fashion and suppress the formation of any
further exchanged, undesired products such as (iPrO)2-
BC6F5 (3).
i
3.4. PO polymerization catalyzed by PrOB(C6F5)2 (2)
and the resulting PPO structures
We examined the PO polymerization behavior using
the isolated 2 in the absence or presence of hydroxylic
initiators, the results of which were summarized in Table
Table 3
Summary of PO polymerization catalyzed by 2a
TOFb(/h)
Mn (Mw=Mn)
c
Entry
Catalyst/initiator
Polymerization time (h)
Conversion (%)
1.
2.
3.
4.
2/none
i
3
3
3
3
86.3
91.3
84.0
85.4
1438
1522
1400
1423
2550 (1.43)
1830 (1.30)
2210 (1.32)
3060 (1.43)
2/120 PrOH
2/60 HO(CH2)4OH
2/40 HOCH(CH2OH)2
a Carried out at 23 °C, 5.00 mL (4.15 g) of PO, [PO]o:[B]o ¼ 5000.
b Turnover frequency (TOF) ¼ moles of PO consumed per mole of catalyst per hour; calculations were based on polymer yield.
c Determined by MALDI-TOF MS in dithranol matrix.

A. Rodriguez-Delgado, E.-Y.X. Chen / Inorganica Chimica Acta 357 (2004) 3911–3919
3917
Fig. 2. MALDI-TOF MS spectrum of the PPO produced by with A and B representing the mass distributions for two primary PPO structures shown
in Scheme 5.
observations, along with the individual reaction results
discussed above, strongly suggest the real active species
for the Ti(OiPr)4/B(C6F5)3 mixture is the borane 2.
Next, the PO polymerization behavior of the catalyst
2 in combination with hydroxylic initiators was also
examined. Thus, addition of 120 equiv. mono-functional
higher intensities (A series) in Fig. 2 represents the linear
PO structure A shown in Scheme 5, where Mend group is
the molar mass of isopropanol initiator (M ¼ 60:10) and
n is the degree of polymerization. The calculated mass
values (e.g., M27 ¼ 1651:26 Da) correlates well with the
measured mass values (e.g., M27 ¼ 1651:93 Da) listed in
Fig. 2. The second PO structure is identified as linear
structure B, where Mend group is the molar mass of water
molecule (M ¼ 18:02), the mass distribution of which is
labeled as B series in Fig. 2. The calculated (e.g.,
M22 ¼ 1318:77 Da) and the found (M22 ¼ 1319:09 Da)
mass values match well. The formation of this secondary
structure is presumably due to the initiation by water
contained in the isopropanol initiator. The MALDI-
TOF MS spectra of the PPO sample prepared with 2 in
the presence of di- and tri-functional initiators features
similarly two primary mass distributions for PPO
structures A and B having the corresponding initiator
end groups. These results are significant, because they
show the borane 2-catalyzed PO polymerization can
control over the PPO structure.
i
initiator PrOH increased the polymerization activity
and, as expected, significantly decreased the PPO mo-
lecular weight (entry 2 versus 1, Table 3). The use of 60
equiv. of bifunctional initiator HO(CH2)4OH did not
affect the polymerization characteristics noticeably,
whereas employing 40 equiv. of trifunctional initiator
HOCH(CH2OH)2 increased Mn of the PPO to 3060 Da
(entries 3 and 4, Table 3). These results show the ad-
vantage of using multifunctional initiators such as
HOCH(CH2OH)2 for enhancing the PPO molecular
weight in the acid-catalyzed ROP of PO.
Fig. 2 depicts an expanded range of the MALDI-TOF
MS spectrum of the PPO prepared with 2/iPrOH. There
are two primary mass distributions, both of which have
mass differences between the peaks representing the
molar mass of PO (58.08 g/mol). The distribution with
4. Conclusions
The mixture of Ti(OiPr)4/B(C6F5)3 is an effective
catalyst system for ROP of PO, while the Ti(OiPr)4/
Al(C6F5)3 mixture is inactive; the inactivity of the
latter mixture is due to the formation of a stable,
catalytically inactive isopropoxy-bridged bimetallic
species Ti(OiPr)3(l-OiPr)Al(C6F5)3 (1). On the other
Scheme 5.

3918
A. Rodriguez-Delgado, E.-Y.X. Chen / Inorganica Chimica Acta 357 (2004) 3911–3919
hand, the reaction of Ti(OiPr)4 and B(C6F5)3 produces
the ligand exchange products via the unstable inter-
mediate Ti(OiPr)3(l-OiPr)B(C6F5)3 (4). The excess of
B(C6F5)3 stabilizes this intermediate, thereby slowing
down the ligand exchange reaction and rendering the
reaction more selective. This strategy was used to
(b) S.D. Gagnon, in: J.I. Kroschwitz (Ed.), Encyclopedia of
Polymer Science and Engineering, vol. 6, second ed., Wiley, New
York, 1986, pp. 273–307.
[2] (a) M. Szycher, Szycher’s Handbook of Polyurethanes, CRC
Press, New York, 1994;
(b) G. Oertel, Polyurethane Handbook: Chemistry, Raw Materi-
als, Processing, Application, Properties, second ed., Hanser-
Gardner Publications, Cincinnati, 1994.
i
successfully isolate the borane PrOB(C6F5)2 (2) in its
pure state.
[3] D.J. Sparrow, D. Thorpe, in: E.J. Goethals (Ed.), Telechelic
Polymers: Synthesis and Applications, CRC Press, Boca Raton,
1989, pp. 181–228.
Control polymerization runs, in combination of the
above studies on individual reactions, show that the
active species responsible for the catalytic activity of
the Ti(OiPr)4/B(C6F5)3 mixture is the neutral borane
2. When used in its pure form, isopropoxy borane 2,
either by itself or when combined with hydroxylic
initiators, is highly effective for PO polymerization,
producing PPOs with Mn ¼ 2000–3000 and TOF
>1400/h. The MALDI-TOF MS analyses of the PPOs
obtained by 2/initiator show the well-defined linear
PPO structures having the initiator end groups. The
effectiveness of this isopropoxy borane, when used
alone for PO polymerization, is in sharp contrast to
B(C6F5)3; while B(C6F5)3 forms stable adducts with
alcohol initiators and reacts vigorously with PO in the
absence of the initiator to produce only low molecular
weight oligomers (dimer, trimer, etc.) and the isom-
erization product (propionaldehyde), 2 does not form
adducts with the initiator and reacts with PO in the
absence of any initiator in a controlled fashion to
produce only PPO. These results show better poly-
merization control over the PPO structure demon-
strated by the borane 2-catalyzed PO polymerization
than that by B(C6F5)3-catalyzed polymerization.
[4] P. Kubisa, in: K. Matyjaszewski (Ed.), Cationic Polymerizations:
Mechanisms, Synthesis, and Applications, Marcel Dekker, New
York, 1996, pp. 437–553.
[5] (a) E.J. Goethals, Adv. Polym. Sci. 23 (1977) 103;
(b) R.J. Katnik, J. Schaefer, J. Org. Chem. 33 (1968) 384;
(c) R.J. Kern, J. Org. Chem. 33 (1968) 388.
[6] (a) S. Penczek, J. Polym. Sci., Part A: Polym. Chem. 38 (2000)
1919;
(b) P. Kubisa, S. Penczek, Prog. Polym. Sci. 24 (1999) 1409;
(c) M. Bednarek, P. Kubisa, S. Penczek, Makromol. Chem. Suppl.
15 (1989) 49;
(d) M. Wojtania, P. Kubisa, S. Penczek, Makromol. Chem.
Macromol. Symp. 6 (1986) 201.
[7] (a) For representative reviews, see H. Sugimoto, S. Inoue, Adv.
Polym. Sci. 146 (1999) 39;
(b) W. Kuran, Prog. Polym. Sci. 23 (1998) 919.
[8] (a) T. Aida, S. Inoue, Acc. Chem. Res. 29 (1996) 39;
(b) H. Sugimoto, C. Kawamura, M. Kuroki, T. Aida, S. Inoue,
Macromolecules 27 (1994) 2013;
(c) T. Aida, Y. Maekawa, S. Asano, S. Inoue, Macromolecules 21
(1988) 1195;
(d) T. Aida, S. Inoue, Macromolecules 14 (1981) 1166.
[9] (a) A. Le Borgne, V. Vincens, M. Jouglard, N. Spassky,
Macromol. Symp. 73 (1993) 37;
(b) V. Vincens, A. Le Borgne, N. Spassky, Macromol. Chem.
Rapid Commun. 10 (1989) 623;
(c) A. Le Borgne, N. Spassky, C.L. Jun, A. Momtaz, Macromol.
Chem. 189 (1988) 637.
[10] (a) M.H. Chisholm, D. Navarro-Llobet, W.J. Simonsick, Macro-
molecules 34 (2001) 8851;
5. Supplementary materials
(b) B. Antelmann, M.H. Chisholm, S. Iyer, J.C. Huffman, D.
Navarro-Llobet, M. Pagel, W.J. Simonsick, W. Zhong, Macro-
molecules 34 (2001) 3159.
Crystallographic data for the structural analysis of
complex 1 have been deposited at the Cambridge
Crystallographic Data Center (CCDC). CCDC refer-
ence number is 234877.
[11] B. Wu, C.J. Harlan, R.W. Lenz, A.R. Barron, Macromolecules 30
(1997) 316.
[12] (a) E.J. Vandenberg, J. Polym. Sci. Part A-1 7 (1969) 525;
(b) C.C. Price, R. Spector, J. Am. Chem. Soc. 87 (1965)
2069;
(c) E.J. Vandenberg, J. Polym. Sci. 47 (1960) 489.
[13] W. Braune, J. Okuda, Angew. Chem., Int. Ed. 42 (2003) 64.
[14] D. Chakraborty, A. Rodriguez, E.Y.-X. Chen, Macromolecules
36 (2003) 5470.
Acknowledgements
This work was supported by The Dow Chemical Co.
and Colorado State University. We thank W. Jack
Kruper, Pete N. Nickias, and John Weston of Dow
Chemical for helpful discussions and Boulder Scientific
Co. for the gift of B(C6F5)3. E.Y.C. gratefully ac-
knowledges an Alfred P. Sloan Research Fellowship.
[15] S. Feng, G.R. Roof, E.Y.-X. Chen, Organometallics 21 (2002)
832.
[16] (a) P. Biagini, G. Lugli, L. Abis, P. Andreussi, U.S. Patent
5,602269, 1997;
(b) C.H. Lee, S.J. Lee, J.W. Park, K.H. Kim, B.Y. Lee, J.S. Oh, J.
Mol. Cat., Sect. A: Chem. 132 (1998) 231.
[17] M.D. Rausch, H.B. Gordon, J. Organomet. Chem. 74 (1974)
85.
[18] H.J. Frohn, H. Franke, P. Fritzen, V.V. Bardin, J. Organomet.
Chem. 598 (2000) 127.
References
[19] G.M. Sheldrick, SHELXTL, Version 5, Siemens, Madison, WI,
1996.
[1] (a) S.D. Gagnon, in: J.I. Kroschwitz, M. Howe-Grant (Eds.),
Kirk-Othmer Encyclopedia of Chemical Technology, vol. 19,
fourth ed., Wiley, New York, 1996, pp. 722–743;
[20] E.Y.-X. Chen, W.J. Kruper, G. Roof, D.R. Wilson, J. Am. Chem.
Soc. 123 (2001) 745.

A. Rodriguez-Delgado, E.-Y.X. Chen / Inorganica Chimica Acta 357 (2004) 3911–3919
3919
[21] R.E. LaPointe, G.R. Roof, K.A. Abboud, J. Klosin, J. Am.
Chem. Soc. 122 (2000) 9560.
[23] T. Belgardt, J. Storre, H.W. Roesky, M. Noltemeyer, H.-G.
Schmidt, Inorg. Chem. 34 (1995) 3821.
[22] G.S. Hair, A.H. Cowley, R.A. Jones, B.G. McBurnett, A. Voigt,
J. Am. Chem. Soc. 121 (1999) 4922.
[24] D. Chakraborty, E.Y.-X. Chen, Organometallics 22 (2003)
207.