Borane-functionalized oxide supports: Development of active supported metallocene catalysts at low aluminoxane loading

Article abstract of DOI:10.1016/S1381-1169(98)00341-0

Treatment of hydroxylated silica or alumina with tris(perfluorophenyl)borane (1), bis(perfluorophenyl)borane (2) or bis(perfluorophenyl)boron chloride (3), provides borane-functionalized supports of variable composition, as revealed by in situ monitoring

Full text of DOI:10.1016/S1381-1169(98)00341-0

Ž
.
Journal of Molecular Catalysis A: Chemical 144 1999 137–150
www.elsevier.comrlocatermolcata
Borane-functionalized oxide supports: development of active
supported metallocene catalysts at low aluminoxane loading
)
Jun Tian, Shaotian Wang, Yuding Feng, Jieming Li, Scott Collins
Department of Chemistry, UniÕersity of Waterloo, Waterloo, Ontario, Canada N2L 3G1
Received 17 July 1998; accepted 30 September 1998
Abstract
Ž
.
Ž . Ž .
Ž
.
Treatment of hydroxylated silica or alumina with tris perfluorophenyl borane 1 , bis perfluorophenyl borane 2 or
Ž
.
Ž .
bis perfluorophenyl boron chloride 3 , provides borane-functionalized supports of variable composition, as revealed by in
situ monitoring by 19F NMR spectroscopy andror elemental analysis. These chemically treated supports can be impregnated
with Cp2ZrMe2 to provide supported catalysts for ethylene polymerization. Although some of these supported catalysts are
Ž
.
active for ethylene polymerization in the presence of alkylaluminum compounds e.g., TMA, TIBAL , all of these catalysts
are more efficiently activated in the presence of small quantities of methyl aluminoxane, even at very low Al:Zr ratios of
Ž
.
10:1. The polymer properties are quite similar to those produced using the soluble catalyst Cp2ZrMe2rB C6F5 3, again in
the presence of MAO at low loading. A variety of experiments suggest that minimal leaching of the metallocene complex
from the support occurs under the conditions studied; in particular, production of polyethylene with high bulk densities
3
Ž
.
)0.2 grcm and little reactor fouling, even at elevated temperature in toluene slurry, is observed. q 1999 Elsevier
Science B.V. All rights reserved.
Keywords: Boron; Supported-metallocene; Olefin polymerization
1. Introduction
The development of effective, supported metallocene catalysts for olefin polymerization continues
w
x
to attract attention 1–3 . The most common approach involved impregnation of hydroxylated silica
Ž
.
w
x
with methyl aluminoxane MAO , followed by treatment with a metallocene complex 4–16 . These
materials are active for olefin polymerization, particularly in the context of gas-phase polymerization
w
x
w
x
processes 15,16 , or in slurry polymerizations in the presence of additional MAO 4–10,14,16 or
w
x
other alkylaluminum compounds 11–13 . More recently, alternative approaches have been described,
Ž
w
Ž
.
xw
x.
where silica is functionalized with a single component co-catalyst e.g., SiO2–OB C6F5 3 Ph3C
and this chemically modified support is used in combination with a metallocene dialkyl complex in
)
Corresponding author
1381-1169r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved.
Ž
.
PII: S1381-1169 98 00341-0

(
)
138
J. Tian et al.rJournal of Molecular Catalysis A: Chemical 144 1999 137–150
w
x
olefin polymerization 17–21 . Although such supported catalysts are active for ethylene polymeriza-
tion in the absence of MAO, they are less active than those based on MAO-treated silica.
It had occurred to us that a complementary approach might involve functionalization of hydroxyl-
w
x
ated silica or alumina with organoboranes. By analogy to the work of Siedle and coworkers 22,23 ,
Ž
.
Ž .
use of B C6F5 3 1 might lead to surface-bound, silanol-borane adducts which might activate
Ž
Ž ..
dialkylzirconocene complexes for ethylene polymerization Eq. 1 .
Ž .
1
Ž
.
Ž .
w
x
Ž
. Ž . w x
Alternatively, use of reactive organoboranes such as HB C6F5 2
2
24 or ClB C6F5 2 3 25
Ž
Ž ..
might provide surface-bound, borinic esters Eq. 2 , that might be sufficiently Lewis acidic to also
activate dialkylmetallocene complexes.
Ž .
2
w
x
w x
Based on the work of Feher and coworkers 26–29 and Hermann et al. 30 using model
Ž
.
sesquisiloxanes, the electron-withdrawing power of a tri-functional surface site on silica e.g., as in 4
is similar to that of a CF3 group and this should render the surface-bound organoborane much more
Lewis acidic.
Herein, we report the results of both synthetic studies and olefin polymerization experiments that
reveal that this approach can be effective, particularly in the presence of dramatically reduced
quantities of MAO.
2. Functionalization of silica and alumina with organoboranes
Partially dehydroxylated silica was prepared by calcination of fully hydroxylated material, while
w
x
fully hydroxylated alumina was prepared from dehydroxylated alumina as previously described 16 .
The surface hydroxyl group concentration was determined by titration with AlMe3 and the properties
of the supports employed are summarized in Table 1.

(
)
J. Tian et al.rJournal of Molecular Catalysis A: Chemical 144 1999 137–150
139
Table 1
Properties of supports used
Support
Code
Surface area
Pore volume
Pore radius
OH content
2
˚
Ž
.
Ž
.
Ž
.
Ž
.
m rg
mlrg
A
mmol OHrg
PQ SiO2
Ketjen Al2O3
PQ
KG
440
326
3.1
0.73
144
45
1.30
3.02
Ž
These hydroxylated supports were then treated with an equivalent of organoboranes 1–3 based on
.
surface OH content in toluene slurry to provide borane functionalized supports. Following filtration,
washing and drying, chemically modified silicas and aluminas were produced and based on fluorine
1
Ž
.
analyses of some of these materials, significant organoborane was incorporated Table 2 .
These reactions were also indirectly monitored by 19F NMR in toluene-d8 suspension in the
presence of an internal standard to determine the extent of incorporation of the borane as well as the
production of any soluble by-products arising from chemical reaction with the support. The amount of
Ž
.
borane incorporated on the support, as measured by this technique Table 2 , was found to be much
higher than that based on chemical analysis for reactions involving borane 1; we suspect this arises
Ž
Ž ..
from reversible binding of borane 1 to hydroxylated silica or alumina Eq. 1 , so that during
washing of the support, significant loss of borane occurs, compared to the ‘equilibrium’ value as
measured by NMR.
Ž
Quite high levels of compounds 2 and 3 could be deposited on silica ca. 70–80% yield with
.
Ž
.
respect to surface OH content, Table 2 , whereas only small amounts on a per mole basis of
organoborane 1 were adsorbed on this support ca. 7 wt.% . There was little, if any, by-product
formation observed in all three cases as judged by the absence of C6F5H, HOB C6F5 2 or the
Ž
.
Ž
.
2
w
Ž
.
x
Ž
.
corresponding anhydride O B C6F5 2 2 hydrolysis products of 1–3 .
Interestingly, quite different behavior was observed using alumina; here, 1 was adsorbed in the
Ž
.
largest quantities again without decomposition, as judged by the absence of any soluble by-products
whereas compound 3 appears to dehydrate the alumina as judged by the quantitative and clean
2,3
w
Ž
.
x
Ž
Ž
.
.
formation of O B C6F5 2 2 formed from reaction of 3 with HOB C6F5 2 in solution .
3. Adsorption of dimethylzirconocene on borane-functionalized supports
Ž
The chemically modified supports PQ-1 to PQ-3, and KG-1 were contacted with an excess based
.
on surface OH content of the untreated supports of Cp2 ZrMe2 in toluene suspension to provide
supported catalysts, following filtration, washing and drying. Again, in addition to Zr analysis on the
1
final materials, the reaction of excess Cp2 ZrMe2 with these supports was indirectly monitored by H
NMR in benzene-d6 suspension to determine the extent of reaction and the formation of any soluble
by-products. In all cases, less Cp2 ZrMe2 was adsorbed, based on final Zr-content, as determined by
Ž
.
analysis, than was predicted based on the NMR experiments Table 3 ; we suspect this discrepancy
1 Reproducible analyses for B in the presence of Al or Si proved problematic.
2
Ž
.
Independent synthesis of C6 F5 2 BOH can be performed through either hydrolysis of 2 and 3 at room temperature, or hydrolysis of 1 at
Ž
.
elevated temperatures; the corresponding anhydride is available by reaction of C6 F5 2 BOH with 3 at room temperature in toluene solution
Ž
.
Ž
.
Ž
.
see Section 8 . We thank Prof. Warren Piers U. Calgary for providing spectroscopic data for authentic samples of HOB C6 F5 .
2
3
Ž
Compound 2 appeared to show very complicated behavior with alumina i.e., appearance of multiple by-products not corresponding
.
solely to hydrolysis and the resulting support was not further investigated.

(
)
140
J. Tian et al.rJournal of Molecular Catalysis A: Chemical 144 1999 137–150
Table 2
Adsorption of boranes on supports
Support
Borane
Wt.%a
Wt.%b
Surface B:OHc
Code
PQ
PQ
PQ
KG
KG
1
2
3
1
3
6.80
26.3
24.2
20.6
trace
18.2
25.7
25.4
47.3
0.0
0.10
0.80
0.71
0.17
0.0
PQ-1
PQ-2
PQ-3
KG-1
KG-3
a Wt.% BR3 or BR2 incorporated as determined by neutron activation analysis.
b Wt.% BR3 or BR2 incorporated as determined indirectly by NMR.
cCalculated value based on neutron activation analysis.
Table 3
Adsorption of Cp2 ZrMe2 on borane-treated supports
Support
Wt.% Zra
Wt.% Zrb
Wt.% Bc
Zr:Bd
Code
PQ-1
PQ-2
PQ-3
KG-1
PQ
4.41
4.68
3.63
3.81
6.42e
6.74e
–
6.80
19.4
13.9
13.3
–
3.57
0.91
0.96
1.16
–
PQ-1-Zr
PQ-2-Zr
PQ-3-Zr
KG-1-Zr
PQ-Zr
4.98
4.66
4.04
8.25
8.86
KG
–
–
KG-Zr
a Wt.% Zr present as determined by neutron activation analysis.
b Wt.% Zr incorporated as determined indirectly by NMR.
c Wt.% BR3 or BR2 calculated based on F-analysis.
dCalculated values based on neutron activation analysis.
e Wt.% Zr present as determined by weight difference.
arises, in some cases, because of reversible binding andror decomposition during isolation and
washing.
Monitoring of these reactions by NMR proved informative as, in certain cases, soluble byproducts
Ž
.
were detected. For reactions involving either PQ-1 or KG-1 or PQ and KG , the only by-product
4
detected was methane. Qualitatively, less was evolved in the reactions involving the chemically
treated supports compared with their hydroxylated counterparts. As might be expected, larger
quantities of Cp2 ZrMe2 can be deposited on untreated PQ silica and alumina. In these cases, it is
Ž
.
Ž
.
known that the surface-bound species are principally SiO2–OZr Me Cp2 and Al2O3–OZr Me Cp2
w
x
31,32 .
In the case of PQ-1 it is clear that the dominant process occurring involves reaction of Cp2 ZrMe2
with surface-OH groups as considerably more Zr was deposited onto this support than would be
Ž . Ž
predicted if the reaction proceeded according to Eq. 1 i.e., the surface Zr:B ratio was much greater
.
than 1, Table 3 ; in contrast, KG-1 reacted with Cp2 ZrMe2 in a manner consistent with the
stoichiometry indicated in Eq. 1 as the final Zr:B ratio was close to 1.
Ž .
In the case of both PQ-2 and PQ-3 significant amounts of by-products were detected in solution.
Ž
.
Ž
.
These were identified as Me2 B ArF and Cp2 ZrMe ArF and were formed in the amounts indicated in
Ž .
Ž
.
Eq. 3 . These products also form from MeB ArF and Cp2 ZrMe2 in toluene solution at room
2
4
Ž
. Ž
.
3
Interestingly, the compound Cp2 ZrMe m-Me B C6 F5 was not detected in solution; this would be expected product from reaction of
Ž
.
Ž
.
Cp2 ZrMe2 with B C6 F5 using PQ-1 or KG-1 .
3

(
)
J. Tian et al.rJournal of Molecular Catalysis A: Chemical 144 1999 137–150
141
Ž .
a
Ž .
b zirconium distribution as measured by energy-dispersed, X-ray
Fig. 1.
SEM micrograph of PQ-3-Zr catalyst particle and
microanalysis.
w
xw
Ž
. x
2
5
temperatures, via the ion pair Cp2 ZrMe Me2 B ArF , a species, which was also detected at low
Ž
.
Ž
.
conversion in a separate experiment see Section 8 . Thus, the formation of Me2 B ArF and
Ž
.
Ž
.
Cp2 ZrMe ArF suggests that MeB ArF is produced on reaction of Cp2 ZrMe2 with these supports
and a plausible mechanism for this is depicted in Eq. 4 .
2
Ž .
Ž .
3
Ž .
4
Ž
.
In agreement with this hypothesis, the fluorine content i.e., BR2 content of both PQ-2-Zr and
Ž
.
PQ-3-Zr was considerably lower Table 3 than would be predicted if the only reaction occurring was
ion-pair formation on the surface Eq. 2 . In fact, when the formation of MeB ArF is taken into
Ž
Ž ..
Ž
.
2
Ž
.
account, the surface stoichiometry of Zr:B is close to 1:1 in the final supported catalysts Table 3 .
Of the various supports prepared, PQ-2-Zr, PQ-3-Zr and KG-1-Zr were obtained as yellow-orange
powders whereas PQ-1-Zr, PQ-Zr, and KG-Zr were formed as off-white solids. This suggests that
only in the former three cases, were cationic complexes formed on the support; clearly the origins of
the differences seen, both during reaction of these supports with boranes, and also with Cp2 ZrMe2 are
of some fundamental interest.
4. Zirconium distribution on the supported catalysts
The spatial distribution of zirconium on the porous silica particles is illustrated by SEM micro-
Ž
.
graphs of the supported catalyst PQ-3-Zr, where the particle size was ca. 50 mm Fig. 1 . Upon
Y
5 For related observations using CpY2 ZrMe2 Cp s1,3- Me3Si 2Cp and MeB ArF 2 , see Ref. 33 .
w
Ž
.
x
Ž
.
w x

(
)
142
J. Tian et al.rJournal of Molecular Catalysis A: Chemical 144 1999 137–150
focusing on the same observation region, the macroscopic distribution of zirconium was detected by
Ž
.
means of an energy-dispersed X-ray microanalyzer EDX ; as shown in Fig. 1b, the black patches,
which represent Zr-rich sites, are dispersed uniformly on the support.
5. Ethylene polymerization with supported catalysts
Ethylene polymerizations using the catalysts prepared were conducted in toluene slurry using
Ž
.
TIBAL as a scrubbing agent Al:Zrs10:1 under a variety of conditions, and the results are
summarized in Table 4. Of the various materials prepared, only PQ-2-Zr, PQ-3-Zr and KG-1-Zr
Ž
exhibited any activity for ethylene polymerization in the absence of MAO Table 4 entries 1, 2 and
.
14 ; having said that the productivities based on both Zr or the amount of support employed, leave
much to be desired. These experiments suggest that very little, ‘active’ cationic zirconocene complex
is present on these three supports. However, as indicated in Table 4, pre-contacting PQ-1-Zr,
Ž
PQ-2-Zr, PQ-3-Zr, and KG-1-Zr with small quantities of MAO in toluene typically at 50:1 Al:Zr
.
Ž
.
levels , led to dramatic improvements in catalytic activity Table 4, entries 3–7, 15 . Of these
supports, PQ-1-Zr is the most active based on Zr or on the amount of supported catalyst entry 3
while the productivities of PQ-3-Zr and KG-1-Zr are comparable entries 5 and 15 and PQ-2-Zr is
the least active entry 4 . The activity of all of these catalysts is superiorrcomparable to a
Ž
.
Ž
.
Ž
.
Ž
.
Table 4
Ethylene polymerization with supported catalystsa
Al:Zrb
A1c
Ad2
Ž .
T 8C
Entry
Catalyst
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
PQ-2-Zr
PQ-3-Zr
PQ-1-Zr
PQ-2-Zr
PQ-3-Zr
PQ-3-Zr
PQ-3-Zr
PQ-Zr
0
0
30
30
70
70
70
30
30
70
70
70
70
70
70
30
70
70
70
70
70
70
70
31e
0.016
0.014
5.07
2.45
3.72
0.50
0.13
5.88
2.85
2.82
1.35
0.033
0.023
0.006
3.57
1.67
2.10
0.267
0.010
–
37e
10.6
50
50
50
25
10
50
25
10
50
50
50
0
50
50
50
ca. 20g
50
50
50
4.75
9.33
1.26
0.329
7.48
3.63
3.58
4.88
–
PQ-Zr
PQ-Zr
PQ-MAO-Zr
PQ-3-Zrf
PQ-Zrf
KG-1-Zr
KG-1-Zr
KG-Zr
KG-MAO-Zr
KG-MAO-Zr
KG-1-Zrf
Cp2 ZrMe2
Cp2 ZrMeh2
–
42e
10.2
2.10
7.89
1.01
–
0.92
4.52
–
a
Ž
.
Ž
.
Ž
.
Conditions: toluene 500 ml , TIBAL 10:1 Al:Zr , C2 H4 75 psig , 1000 rpm.
b
Ž
.
Total Al:Zr ratio as MAO .
cActivity in 103 kg PErmol Zr=h.
dActivity in kg PErg support=h.
eActivity in kg PErmol Zr=h.
f Polymerization with toluene soluble portion, following filtration.
g
w
x
Surface Al:Zr ratio—see Ref. 16 .
hCp2 ZrMe2 treated with a slight excess 1.1 equiv. of B C6 F5
.
.
3
Ž
Ž

(
)
J. Tian et al.rJournal of Molecular Catalysis A: Chemical 144 1999 137–150
143
Table 5
Properties of PE prepared with supported catalysts
5
b
Conditionsa
Mw 10
Mw rMn
Tm 8C
Bulk densityc
Ž
.
Ž
.
Entry
Catalyst
3
Ž
.
grcm
1
2
3
4
5
6
7
PQ-1-Zr
PQ-2-Zr
PQ-3-Zr
PQ-Zr
KG-1-Zr
Cp2 ZrMe2
Cp2 ZrMe2
3
4
5
1.97
2.11
1.89
1.71
1.89
2.57
2.88
2.04
2.10
2.10
2.53
1.90
2.16
2.89
129.8
127.7
129.0
134.9
126.2
129.0
122.8
0.43
0.39
0.41
0.49
0.32
0.17
0.16
8
15
20
21
a The number refers to the corresponding entry in Table 4.
b Peak melting temperature as measured by DSC.
c
w
x
Bulk density measured according to Ref. 35 .
Ž
.
conventional supported catalyst based on MAO-impregnated silica or alumina entries 11 and 17 at
Ž
.
similar total loadings of MAO.
Quite surprisingly, PQ-Zr proved to be an active catalyst in the presence of reduced quantities of
MAO. This finding suggests that MAO is capable of cleaving Si–O–Zr bonds to provide active
6
Ž
.
catalyst perhaps via formation of Cp2 ZrMe2 and SiO2–OMAO, followed by ionization .
A
Ž
.
comparison of this catalyst with, e.g., PQ-3-Zr Table 4, entries 5–7 and 8–10 , reveals that
pre-treatment of silica with borane 3 appears to confer little, if any, advantage compared to direct
reaction of silica with Cp2 ZrMe2. In fact, at lower Al:Zr ratios, the performance of PQ-Zr is superior.
Somewhat different results were observed with the alumina-based catalysts. Of these KG-1-Zr is
Ž
the most active formulation, being marginally superior to a conventional MAO-treated alumina Table
.
4, entries 15 and 17 and considerably more active than alumina that is directly treated with
Ž
.
Cp2 ZrMe2 entry 16 .
While the productivities observed are much higher than that observed using Cp2 ZrMe2rMAO
Ž
.
under the same conditions Table 4, entry 20 , there is the obvious concern that significant leaching of
7
soluble catalyst occurs during the pre-contacting period andror during polymerization. Although not
exhaustively pursued for all combinations, control experiments that involved pre-contacting the
support with MAO, filtration and use of the filtrate revealed that minimal leaching of metallocene
8, 9
Ž
.
complex was observed at this stage Table 4, entries 12, 13, and 19 .
6. Polymer properties
As is evident from the results summarized in Table 5, the nascent polyethylene produced by these
3
Ž
.
supported catalysts has high bulk density )0.3 grcm , and good particle morphology for
6 We note that PQ-Zr appears to be unreactive towards a large excess of AlMe3 i.e., no Cp2 ZrMe2 produced in toluene solution ,
Ž
.
Ž
.
suggesting that a component of MAO other than AlMe3 is responsible for this alkylationrionization process.
7 A recent report suggests that supported catalysts analogous to PQ-Zr undergo extensive leaching of metallocene complex; these results
Ž
appear to be contrary to the findings reported here but may reflect the different treatments employed i.e., large excess of MAO at elevated
.
w x
temperatures . See Ref. 34 .
8 Also, pre-polymerization with ethylene at an initial temperature of 258C and 75 psig in a sample bomb, followed by filtration, washing
Ž
.
Ž
and use of the filtrate in a subsequent polymerization at 708Cr75 psig also indicated minimal leaching trace quantities of PE formed with
.
either KG-1-Zr or PQ-Zr . We cannot, however, eliminate the possibility that leaching does occur under the conditions reported in Table 5
Ž
.
w x
708Cr75 psig . See Ref. 34 .
9 1H NMR experiments also showed that no Cp2 ZrMe2 or Cp2 Zrq Me appeared in solution within 20 min at room temperature on
treatment of PQ-Zr with excess MAO in toluene-d8.

(
)
144
J. Tian et al.rJournal of Molecular Catalysis A: Chemical 144 1999 137–150
wŽ .
Ž .
x
Fig. 2. SEM and optical microscope images of PQ-3-Zr and its nascent polyethylene a and b , respectively and KG-1-Zr and its nascent
wŽ . Ž .
x
polyethylene c and d , respectively .
Ž
.
silica-based catalysts Fig. 2a and b , as shown by SEM and optical microscopy. However, the
Ž
.
polymer product from the alumina-based catalyst KG-1-Zr had irregular shape and a broad size
Ž
.
distribution, which more or less mirrored the properties of the alumina support itself Fig. 2c and d .
Also, minimal, if any, reactor fouling was observed, even at elevated temperatures in toluene slurry,
conditions where leaching should be most problematic. Finally, as is evident from the MW data
summarized in Table 5, comparable MWs and MWDs were observed with these supported catalysts
as compared to use of a soluble catalyst under similar conditions.
7. Conclusions
As can be appreciated from this work, active formulations for supported catalysts can be derived
Ž .
from either silica or alumina by treatment with a either an organoborane, followed by treatment with
Ž .
a dialkylzirconocene or b direct treatment with dimethylzirconocene. Both approaches require
pre-contacting with MAO for useful activity and paradoxically, this approach appears to be more
efficient than first depositing MAO on the support in many cases. Having said that, there appears to

(
)
J. Tian et al.rJournal of Molecular Catalysis A: Chemical 144 1999 137–150
145
Ž
be no practical advantage in chemically pre-treating silica with an organoborane except perhaps with
.
borane 1 , while for alumina, the borane-treated support gives rise to a catalyst with marginally
improved productivity. Quite clearly, further progress along these lines is hampered by the lack of
knowledge concerning the Zr and B species that are present on the surface of these materials. Future
work will focus on this important issue.
8. Experimental
All chemicals were reagent grade and purified as required prior to use. Diethyl ether,
dichloromethane, tetrahydrofuran, hexane and toluene were dried and deoxygenated using a solvent
w
x
purification system similar to that reported in the literature 36 . Benzene and pentane were dried and
deoxygenated by distillation from sodium and benzophenone under nitrogen. Deuterated toluene and
benzene for NMR spectroscopy were pre-dried and distilled in the same manner. All manipulations
with air and moisture sensitive compounds were performed using standard Schlenk techniques under
nitrogen. Air and moisture sensitive compounds were stored in either a Vacuum Atmosphere glove
Ž
.
box or a Braun MB 150 M glove box. Methyllithium Aldrich was titrated immediately prior to use
w
x
w x
with 4-biphenylmethanol as described in the literature 37 . The compounds Cp2 ZrMe2 38 ,
HB C6F5 2 2 24 , and ClB C6F5 2 3 25 were synthesized according to literature procedures.
Ž
. Ž .
w
x
Ž
. Ž . w x
1
Routine H NMR spectra were obtained on Bruker AM-250 or AC-200 spectrometer in the
appropriate deuterated solvents; chemical shifts are referenced to residual undeuterated solvent. 19F
NMR spectra were recorded on Bruker AC-200 spectrometer and chemical shifts are referenced to
2,3,5,6-tetrafluoro-p-xylene, which was added as an internal standard and had already been referenced
to CFCl3 at the same recording conditions. 11B NMR spectra were recorded on the same instrument
and chemicals shifts were referenced to external standard F3B OEt2. The 11B NMR spectra were
.
obtained using a background substraction routine to remove resonances duo to borosilicate glass in the
NMR probe and NMR tube. This was accomplished by recording the spectrum of an NMR tube
containing the same height of pure solvent and with identical collection parameters to that of the
sample and subtracting this background FID from that of the sample. Chemical shift data are
´
summarized in Table 6. Neutron activation analyses were performed at the Ecole Polytechnique,
Montreal, PQ while conventional combustion analyses were performed by M.H.W. Laboratories of
´
Phoenix, AZ.
Table 6
11
19 F and B NMR data for compounds prepareda
19
11
Ž
.
Ž
.
Compound
F ppm
B ppm
o-F
p-F
m-F
1
y130.0
y134.8
y133.1
y129.6
y133.1
y132.6
y139.2
y143.1b
y153.6
y146.7
y144.0
y148.3b
y144.3b
y154.3
y161.3
y162.3
y159.2
y160.7
y161.4
y160.2
y162.5
57.6
5.34
1PH2O
2
3
18.0c
58.8
40.2
40.1
–
Ž
.
.
C6 F5 2 BOH
wŽ
x
C6 F5 2 B 2O
C6 F5H
aAll spectra were recorded in toluene-d8 solution at 258C at 188.0 and 64.2 MHz, for 19 F and 11 B, respectively. For the acquisition of 11 B
spectra, a blank sample was recorded in toluene-d8 and then the same tube was used for sample preparation and the baseline was corrected
by spectral subtraction.
b The chemical shift andror line-width of this signal are sensitive to the presence of adventitious water in the sample andror NMR solvents.
c
Ž
w x.
Major component corresponding to dimer; minor component at 60.0 ppm corresponding to monomer see Ref. 24 .

(
)
146
J. Tian et al.rJournal of Molecular Catalysis A: Chemical 144 1999 137–150
8.1. Support preparation
Silica available from Quantum Chemicals under the designation of PQ 3010, has a surface area of
2
3
˚
440 m rg and a pore volume of 3.10 cm rg with an average pore radius 144 A. Unless indicated
otherwise, the silica used was washed with deionized water at room temperature, dried at 808C under
vacuum to constant weight, heated in a quartz tube inside a horizontal tube furnace to 500–6008C for
4 h under oxygen flow and then cooled down to room temperature under oxygen flow to produce
Ž
.
partially dehydroxylated silica PDS . The hydroxyl group content was determined by titration with
trimethylaluminum, and the content was 1.30 mmol OHrg support.
2
3
Ž
.
Ketjen G alumina Akzo Nobel had a surface area of 326 m rg and a pore volume of 0.73 cm rg
˚
with an average pore radius 45 A. It was washed with deionized water and then heated at 5008C for 3
Ž
.
h under oxygen flow, to form partially dehydroxylated alumina PDA with a hydroxyl content of
0.19 mmolrg support. PDA was then heated under water-saturated nitrogen for 48 h at 3008C to
Ž
.
provide fully hydroxylated alumina FHA . The hydroxyl group content was 3.02 mmol OHrg.
8.2. Preparation of PQ-1
Ž
.
Ž
. Ž
.
To a mixture of 1.000 g silica 1.305 mmol OH and 0.690 g B C6F5 3 1.305 mmol was added 20
ml toluene at room temperature. The slurry was stirred in the glove box for 18 h. A white solid was
Ž
.
collected on a medium fritted funnel, washed with toluene 4=5 ml and dried in vacuo at room
temperature. The weight was 1.056 g. Analytical data appear in Table 2.
8.3. Preparation of PQ-2 and PQ-3
The same procedure as reported for PQ-1 was followed, except that the reaction time was 3 h, with
Ž
.
Ž
.
the following reagents: 1.000 g silica 1.305 mmol in each case, and 0.451 g of 2 1.305 mmol and
Ž
.
Ž
.
Ž
.
0.496 g of 3 1.305 mmol , for the preparation of PQ-2 1.384 g and PQ-3 1.480 g , respectively.
Analytical data appear in Table 2.
8.4. Preparation of KG-1
Twenty milliliters of toluene was added with vigorous stirring to a mixture of 0.788 g alumina
Ž
.
Ž
.
2.630 mmol OH and 1.347 g of 1 2.630 mmol in a two-neck round-bottom flask attached to a
Schlenk filter at y788C. The mixture was warmed up under nitrogen to room temperature over a
period of 7 h. The slurry was then filtered under nitrogen, and the white solid was washed with
Ž
.
toluene 3=5 ml and then dried under vacuum at room temperature. The yield was 0.815 g.
Analytical data appear in Table 2.
8.5. Preparation of PQ-1-Zr
Twenty milliliters of toluene was added slowly with vigorous stirring to a mixture of 0.800 g PQ-1
and 0.280 g Cp2 ZrMe2 in a two-neck round-bottom flask attached to a Schlenk filter at y788C under
nitrogen. After 1 h at 258C, the off-white solid was filtered and washed with hexane and dried under
vacuum at room temperature. The weight was 0.805 g. For analytical data, see Table 3.

(
)
J. Tian et al.rJournal of Molecular Catalysis A: Chemical 144 1999 137–150
147
8.6. Preparation of PQ-2-Zr and PQ-3-Zr
The same procedure as reported for PQ-1-Zr was followed with the following reagents: 0.640 g
PQ-2 and 0.231 g Cp2 ZrMe2, resulting in 0.590 g of PQ-2-Zr as a yellow solid; 0.891 g PQ-3 and
0.293 g Cp2 ZrMe2, providing 0.864 g of yellow solid. For analytical data, see Table 3.
8.7. Preparation of KG-3-Zr
The same procedure as reported for PQ-1-Zr was followed with the following reagents: 0.512 g of
Ž
.
KG-3 and 0.175 g Cp2 ZrMe2. The isolated product was a pale yellow solid 0.480 g . Analytical data
appears in Table 3.
8.8. Preparation of PQ-Zr and KG-Zr
Ž
.
A solution of Cp2 ZrMe2 in toluene 0.252 g, 1.00 mmol, 15 ml was added to 0.500 g of PDS in a
round-bottom flask attached to a Schlenk filter and stirred under nitrogen at room temperature in the
Ž
dark for 3 h. The slurry was filtered under nitrogen, and the solid was washed with dry toluene 3=5
.
Ž
.
Ž
.
ml and hexane 5 ml . The solid was dried under vacuum 0.601 g . The same procedure was
followed for preparation of KG-Zr with the amount of alumina and Cp2 ZrMe2 being 0.309 g and
0.350 g, respectively. The weight of the final product was 0.374 g.
8.9. Preparation of PQ-MAO-Zr and KG-MAO-Zr
Adsorption of MAO on PDS and FHA, and impregnation of Cp2 ZrCl2 on these MAO-modified
w
x
supports involved a literature procedure 16 . The amount of grafted MAO and supported metallocene
w
x
complex was determined as described previously 16,39 .
(
)
8.10. Reaction of ClB C6 F5 2 with water
Ž
.
Ž
.
To a solution containing 0.02 mmol 7.6 mg ClB C6F5 2 in 0.5 ml of toluene-d8, 0.02 mmol of
water was added via syringe. The resulting solution had its 19F NMR spectra recorded at room
19
Ž
.
Ž
.
Ž
.
temperature, showing clean formation of HOB C6F5 2 4b . F NMR toluene-d8, 258C d y132.6
Ž
.
Ž
.
Ž
.
w
Ž
. x
dd, 2F , y149.0 t, F , y160.9 m, 2F ppm. The anhydride, O B C6F5 2 2, was formed at room
temperature by adding more ClB C6F5 2 or by separate reaction of isolated borinic acid with 3. 19F
Ž
.
11
Ž
.
Ž
.
Ž
.
Ž
.
Ž
NMR toluene-d8, 258C d y132.8 dd, 2F , y144.3 t, F , y160.2 m, 2F . B NMR toluene-d8,
q
.
Ž
.
Ž
.
258C d 40.1. MS EI : mre 706 M . Anal. calcd. for C24B2 F20O: C 40.79, F 53.82. Found: C
40.69, F 53.47.
Ž
.
The reaction of HB C6F5 2 with one equiv. of water under the same conditions also provides the
borinic acid.
(
)
8.11. Reaction of B C6 F5 3 with water
Ž
.
Ž
.
To a solution containing 0.02 mmol 10 mg B C6F5 in 0.5 ml of toluene-d8, 0.02 mmol of water
3
was added via syringe. The resulting solution had its 19F NMR spectra recorded at room temperature
19
Ž
.
Ž
.
Ž
.
Ž
.
showing clean formation of H2OPB C6F5 3 4b : F NMR C7D8, 258C d y134.8 dd, 2F ,
10
Ž
.
Ž
.
y153.6 t, F , y162.3 m, 2F ppm.
On heating the solution at 1008C, slow formation of
10
Ž
.
3
We note that commercial samples of B C6 F5 sometimes contain trace amounts of water that leads to line-broadening of the resonance
Ž
.
due to the p-F atom. The water may be removed by treatment with excess BCl3 in toluene, followed by filtration to remove boric acid and
recrystallization from hexane–toluene mixtures.

(
)
148
J. Tian et al.rJournal of Molecular Catalysis A: Chemical 144 1999 137–150
19
Ž
.
w
Ž . Ž . Ž
.x
HOB C6F5 2 and C6F5H
observed.
F NMR d y139.2 m, 2F , y154.3 t, F , y162.5 m, 2F was
(
)
8.12. Reaction of MeB C6 F5 2 with Cp2 ZrMe2
Ž
.
Ž
.
A cationic species was formed immediately on mixing MeB C6F5 2 0.1 mmol with Cp2 ZrMe2
1
Ž
.
w
Ž
.
Ž
.
0.1 mmol in benzene-d6 at room temperature H NMR benzene-d6, 258C d 5.38 s, Cp, 10H
0.37 exchange broadened s, 9H . After 60 min at room temperature, this species completely
decomposed into Me2 B C6F5 and Cp2 ZrMe C6F5 . Me2 B C6F5 : H NMR benzene-d6, 258C d
0.97 t, Js2 Hz, 6H . Cp2 ZrMe C6F5 : d 5.67 s, Cp, 10H and 0.31 t, Js4 Hz, 3H .
Ž
.x
1
Ž
.
Ž
.
Ž
.
Ž
.
Ž
.
Ž
.
Ž
.
Ž
.
8.13. 19F NMR spectroscopic monitoring of the reaction between boranes 1–3 and the supports
These monitoring experiments were conducted in a Teflon valve-sealed NMR tube with a mole
ratio of boronrOH content of the supports1:1 in toluene-d8 solution at 258C. 19F NMR spectra were
referenced to tetrafluoro-p-xylene, added as an internal standard in known amounts, to quantify the
reaction. This data is summarized in Table 2.
1
8.14. H NMR spectroscopic monitoring of the reaction between Cp2 ZrMe2 and Õarious supports
These experiments were conducted in a Teflon valve-sealed NMR tube with addition of excess
Cp2 ZrMe2, based on the amount of OH content of the support and the modified supports, in
1
benzene-d6 solution at 258C. H NMR spectra were referenced to undeuterated benzene, added as an
internal standard in known amounts, to quantitate the extent of reaction. The results are summarized
in Table 3.
11
(
)
8.15. Preparation of MeB ArF 2
Ž
.
MeB C6F5 2 was synthesized from reaction of 3 and Cp2 ZrMe2 at room temperature in toluene,
following removal of the by-product Cp2 ZrCl2. The final pure product was obtained by sublimation
1
y
3
Ž
.
Ž
.
Ž
.
at 50–608C under vacuum 10 mm Hg . H NMR C6D6, 258C d 1.39 quintet, JHF s2 Hz, 3H .
19
Ž
.
Ž
.
Ž
.
Ž
.
F NMR C6D6, 258C d y128.6 m, 2F , y145.5 tt, F , y159.9 m, 2F .
8.16. Polymerization procedures and polymer characterization
8.16.1. Ethylene polymerization
Ž
.
Toluene solvent Ominisol grade, VWRrCanlab used in all polymerizations was freshly purified
w
x
using a solvent purification system similar to that described in the literature 36 and was delivered to
the reactor under pressure from a calibrated sight-glass. Nitrogen and ethylene Praxair, C.P. grade
was purified by passed through a mixed bed column containing activated BASF R3-11 catalystr4 A
Ž
.
˚
11
w
x
We thank Prof. Warren Piers for providing details for this preparation. See also Ref. 40 .

(
)
J. Tian et al.rJournal of Molecular Catalysis A: Chemical 144 1999 137–150
149
U
.
Ž
.
molecular sieve, 1:2 wrw . Methylaluminoxane MAO, 9.6 wt.% Al in toluene; PMAOr Tol-781
y
3
Ž
.
was obtained from Akzo Nobel and concentrated in vacuo to dryness 10 mm Hg to provide solid
Ž
.
MAO for subsequent use. Triisobutylaluminum TIBAL from Aldrich was used as received.
Polymerizations were conducted in a 1-l autoclave equipped with an overhead D.C. motor stirring
assembly and a bottom drain valve. The stirring rate was maintained at 500 rpm for all the
polymerizations and was initially calibrated with a strobe light. The polymerization temperature was
Ž
controlled using a Neslab RTE-100 constant temperature bath by circulation of coolant ethylene
.
glycolrdeionized water, 1:1 vrv through the external, spiral jacket welded to the reactor body. The
temperature was monitoredrcontrolled by using a Neslab RTD-220 temperature controller with a Pt
RTD sensor immersed in a thermowell inside the reactor. The monomer flow rate was measured with
Ž
.
a calibrated Matheson mass flow transducer Model 8172-0413 connected to a Matheson Multiple
Ž
.
Ž
Flow Controller Model 8274 . The polymerization profiles monomer flow rate and solution
temperature were recorded by an IBM PC with a PCL-711s data acquisition card. The progress of the
.
polymerization was monitored in real time using Labtech Acquire data acquisition software, which
converted the digital input readings to the desired values using appropriate conversion factors.
In a typical polymerization experiment, the reactor was first conditioned under dynamic vacuum
Ž
.
and high temperature bath temperature around 1008C for ca. 2 h, during which time the reactor was
purged with high purity nitrogen three times. The reactor was then charged with 500 ml of toluene. A
desired amount of TIBAL in toluene was added to the stirred reactor via a N2-pressurized stainless
Ž
.
steel cylinder 50 ml prior to heating to the desired temperature. At this point the reactor was
saturated with ethylene at 75 psig. The required weight of the supported catalyst was slurred in
toluene and was then added to the reactor via the same pressurized cylinder to initiate the
polymerization. After a specific period of time, the ethylene flow was stopped and the reactor was
vented. The polymer was filtered off, washed with methanol, and then dried in vacuo to constant
weight.
8.16.2. Polymer characterization
Polymer molecular weights and distributions were determined by gel permeation chromatography
3
˚
Ž
.
Ž
.
GPC using a Water 150C chromatograph equipped with a Jordi mixed-bed column 10–10 A
employing a differential refractive index detector at 1458C in 1,2,4-TCB solution. Sample dissolution
Ž
.
0.1% wrv was accomplished by rotating the samples in an oven operating at 1608C for a few h with
0.1% BHT as antioxidant. Samples were eluted at a flow rate of 1.0 mlrmin, and columns were
calibrated using both a broad MWD PE and narrow MWD PS standards.
DSC analysis was conducted on a Perkin–Elmer DSC-2 Model calibrated using indium metal. A
scanning rate of 108Crmin from 25–1808C was employed, and the sample was then cooled to 508C at
108Crmin to remove all prior thermal history. The second trace was recorded using the same heating
program.
The morphologies of the catalyst surface and the nascent polymer particle were examined on a
Ž
.
Olympus SZH Optical Microscope and a JEOL JSM-840 scanning electron microscope SEM ,
Ž
.
equipped with an energy-dispersed X-ray microanalyzer EDX , to detect the element distribution on
the catalyst surface. SEM was obtained at 2 kV and varying levels of magnification, and the
˚
specimens were coated with a 400 A gold layer by using an electron impact evaporator to increase the
conductivity.
The bulk density of polyethylene was determined from the dry weight and the volume of the
w
x
sample in a volumetric tube by liquid displacement 35 .

(
)
150
J. Tian et al.rJournal of Molecular Catalysis A: Chemical 144 1999 137–150
Acknowledgements
The authors would like to thank the Natural Sciences and Engineering Research Council of Canada
and Nova Chemicals of Calgary, for financial support of this work. The technical assistance of
Novachem Research and Technology with some of the SEC analyses is gratefully acknowledged.
Finally, we thank Prof. Warren Piers of the University of Calgary for sharing some spectroscopic data
for compounds independently prepared and characterized there.
References
w x
Ž .
H.H. Brintzinger, D. Fisher, R. Mulhaupt, B. Rieger, R.M. Waymouth, Angew. Chem., Int. Ed. Engl. 34 1995 143.
¨
1
w x
2
Ž .
W. Kaminsky, Macromol. Chem. Phys. 197 1996 3907.
w x
Ž
.
3
K. Soga, Macromol. Symp. 101 1996 281.
w x
4
H.C. Welborn, Jr., US Patent 4,701,432, Exxon Chemical Patents, 1987.
H.C. Welborn, Jr., US Patent 4,808,561, Exxon Chemical Patents, 1989.
H.C. Welborn, Jr., US Patent 5,077,255, Exxon Chemical Patents, 1991.
H.C. Welborn, Jr., US Patent 5,183,867, Exxon Chemical Patents, 1993.
M. Kioka, M. Kashiwa, Eur. Pat. 0-285-443, Mitsui Petrochemical Industries, 1988.
T. Takahashi, Eur. Pat. 0-313-386, Mitsubishi Petrochemical Industries, 1988.
w x
5
w x
6
w x
7
w x
8
w x
9
w
w
w
w
w
w
w
w
w
w
w
w
w
w
w
w
w
w
w
w
w
w
w
w
w
w
w
w
w
w
w
x
Ž
.
10 J.C.W. Chien, D.W. He, J. Polym. Sci., Part A: Polym. Chem. 215 1991 47.
x
Ž
.
11 M. Kaminaka, K. Soga, Makromol. Chem., Rapid Commun. 12 1991 367.
x
Ž
Ž
.
.
12 K. Soga, M. Kaminaka, Makromol. Chem., Rapid Commun. 13 1992 221.
x
13 K. Soga, M. Kaminaka, Makromol. Chem., Rapid Commun. 15 1994 593.
x
Ž
.
14 C. Janiak, C.B. Rieger, Angew. Makromol. Chem. 215 1994 47.
x
Ž
.
15 P. Roos, G.B. Miere, J.J.C. Samson, G. Weichert, K.R. Westerterp, Macromol. Rapid Commun. 18 1997 319.
x
Ž
.
16 D. Harrison, I.M. Coulter, S. Wang, S. Nistala, B.A. Kuntz, M. Pigeon, J. Tian, S. Collins, J. Mol. Catal., Chem. 128 1998 65.
x
Ž
.
17 G.G. Hlatky, D.J. Upton, Macromolecules 29 1996 8019.
x
18 G.G. Hlatky, D.J. Upton, H.W. Turner, WO Patent 91r09882, Exxon Chemical Patents, 1991, and Eur. Pat. 507876B1, 1995.
x
19 H.W. Turner, US Patent 5,427,991, Exxon Chemical Patents, 1995.
x
20 J.F. Walzer, Jr. WO Patent 96r04319, Exxon Chemical Patents, 1996.
x
21 D.G. Ward, E.M. Carnahan, WO Patent 96r23005, W.R. Grace and Co.-CONN., 1996.
x
Ž
.
22 A.R. Siedle, R.A. Newmark, W.M. Lamanna, J.C. Huffman, Organometallics 12 1993 1491.
x
23 A.R. Siedle, W.M. Lamanna, US Patent 5,296,433, Minnesota Mining & Manufacturing Co., 1994.
x
Ž
.
24 D.J. Parks, R.E. von, H. Spence, W.E. Piers, Angew. Chem., Int. Ed. Engl. 24 1995 809.
x
Ž
.
25 R.D. Chambers, T.J. Chivers, J. Chem. Soc. 1965 3933.
x
Ž
.
26 F.J. Feher, T.A. Budzichowski, Polyhedron 14 1995 3239.
x
Ž
.
27 T.A. Budzichowski, S.T. Chacon, M. Chisholm, F.J. Feher, W.J. Streib, J. Am. Chem. Soc. 113 1991 689.
x
Ž
.
28 F.J. Feher, K.J. Weller, Inorg. Chem. 30 1991 882.
x
Ž
.
29 F.J. Feher, D.A. Newman, J. Am. Chem. Soc. 111 1998 1741.
x
Ž
.
30 W.A. Herrmann, R. Anwander, V. Dufaud, W. Scherer, Angew. Chem., Int. Ed. Engl. 33 1994 1285.
x
Ž
.
31 T.J. Marks, Acc. Chem. Res. 25 1992 57.
x
Ž
.
32 K.-H. Dahmen, D. Hedden, R.L. BurWell Jr., T.J. Marks, Langmuir 4 1988 1212.
x
Ž
.
33 X. Yang, C.L. Stern, T.J. Marks, J. Am. Chem. Soc. 116 1994 10015.
x
Ž
.
34 N.V. Semikolenova, V.A. Zakharov, Makromol. Chem. Phys. 198 1997 2889, and references therein.
x
35 G.E. Manger, Porosity and Bulk Density of Sedimentary Rocks, United States Government Printing Office, Washington.
x
Ž
.
36 A.B. Pangborn, M.A. Giardello, R.H. Grubbs, R.K. Rosen, F.J. Timmers, Organometallics 15 1996 1518.
x
Ž
.
37 E. Juaristi, A. Martinez-Richa, A. Garcia-Rivera, J.S. Cruz-Sanchez, J. Org. Chem. 48 1983 2603.
´
ˆ
x
Ž
.
38 E. Samuel, M.D. Rausch, J. Am. Chem. Soc. 95 1973 6263.
x
Ž
.
39 S. Collins, W.M. Kelly, D.A. Holden, Macromolecules 25 1992 1780.
x
40 P. Biagini, G. Lugli, F. Garbassi, P. Andreussi, Eur. Pat. Appl. EP 667,357, 1995.