Ethylene polymerization reactions with multicenter Ziegler-Natta catalysts-manipulation of active center distribution

Article abstract of DOI:10.1002/pola.24208

This article describes ethylene/1-hexene copolymerization reactions with a supported titanium-based, multicenter Ziegler-Natta catalyst. The catalyst was modified by pretreating its solid precursor with AlEt₂Cl and with similar organoaluminum chlorides, Al₂Et₃Cl₃, AlEtCl₂, and AlMe₂Cl. Testing of the untreated and the pretreated catalysts in copolymerization reactions under standard reaction conditions demonstrated that the modifying agents produce two changes in the catalyst. First, the pretreatment significantly reduces the reactivity of active centers that produce high molecular weight, highly crystalline copolymer components with a low 1-hexene content. Second, the pretreatment noticeably increases the reactivity of active centers that produce low molecular weight copolymer components with a high 1-hexene content. The first effect is caused by Lewis acidbase interactions of the modifiers with the active centers, whereas the second (activating) effect is due to the removal of catalyst poisons (organosilicon compounds generated in the process of the catalyst synthesis) by AlEt₂Cl.

Full text of DOI:10.1002/pola.24208

Ethylene Polymerization Reactions with Multicenter Ziegler-Natta
Catalysts—Manipulation of Active Center Distribution
1
2
YURY V. KISSIN, ROBERT I. MINK
1
Department of Chemistry and Chemical Biology, Rutgers, The State University of New Jersey, 610 Taylor Road, Piscataway,
New Jersey 08854
2
Edison Research Laboratory, Mobil Chemical Company, P.O. Box 3029, Edison, New Jersey 08818
Received 20 May 2010; accepted 29 June 2010
DOI: 10.1002/pola.24208
Published online in Wiley Online Library (wileyonlinelibrary.com).
ABSTRACT: This article describes ethylene/1-hexene copolymer-
ization reactions with a supported titanium-based, multicenter
Ziegler-Natta catalyst. The catalyst was modified by pretreating
ment noticeably increases the reactivity of active centers that
produce low molecular weight copolymer components with a
high 1-hexene content. The first effect is caused by Lewis acid-
base interactions of the modifiers with the active centers,
whereas the second (activating) effect is due to the removal of
catalyst poisons (organosilicon compounds generated in the pro-
its solid precursor with AlEt
chlorides, Al Et Cl , AlEtCl
2
Cl and with similar organoaluminum
, and AlMe Cl. Testing of the un-
2
3
3
2
2
treated and the pretreated catalysts in copolymerization reactions
under standard reaction conditions demonstrated that the modi-
fying agents produce two changes in the catalyst. First, the pre-
treatment significantly reduces the reactivity of active centers
that produce high molecular weight, highly crystalline copolymer
components with a low 1-hexene content. Second, the pretreat-
cess of the catalyst synthesis) by AlEt
2
Cl. VC 2010 Wiley Periodi-
cals, Inc. J Polym Sci Part A: Polym Chem 48: 4219–4229, 2010
KEYWORDS: copolymerization reactions; gel permeation chro-
matography; molecular weight distribution; Ziegler-Natta catalysts
INTRODUCTION Supported Ti-based Ziegler-Natta catalysts
used for ethylene polymerization reactions consist of two com-
ponents, solid components called catalyst precursors and coca-
talysts, trialkylaluminum compounds. The active catalyst species
are formed when the two catalyst components are combined,
usually immediately before the polymerization reactions. A large
number of competing synthetic procedures for the preparation
of the Ti-containing catalyst precursors are described in numer-
ous patents. Several of the techniques are reviewed in ref. 1.
ular weights of the polymer components produced by differ-
ent centers are quite large, as several examples below dem-
onstrate; for example, the ratio between the M (Center V)
w
and the M (Center I) values can be as high as 80 to 90.
w
2. Centers of different types vary greatly in their ability to
copolymerize a-olefins with ethylene. This ability steadily
decreases from Center I to Centers IV and V. Centers IV
and V have poor copolymerization ability. Even when a-
olefins are present in the copolymerization reactions in
significant concentrations, the copolymer molecules pro-
duced by Centers IV and V have low a-olefin contents,
and these materials are highly crystalline.
. Centers of different types significantly vary in their kinetic
behavior; they are formed and decay at different rates.
Centers I and II are formed rapidly but they decay rapidly
and become practically inactive after 20 to 30 min. On the
other hand, Centers IV and V are formed much slower,
during a period of 15 to 30 min, but they are much more
stable and can polymerize ethylene for several hours
without any significant loss of activity.
Our earlier work on the kinetics of ethylene homopolymeri-
zation and copolymerization reactions with different Ti-
based catalysts has demonstrated that they all exhibit several
common features. The catalyst performance can be described
3
1
–4
in terms of a multicenter reaction scheme as follows
:
1
. All the catalysts contain several (usually four or five) types
of active centers. One of the principal differences between
the centers is the average molecular weight of the polymer
molecules they produce under given reaction conditions.
Each type of center produces polymer molecules with a
narrow molecular weight distribution described by the Flory
4. Ziegler-Natta catalysts are sensitive to the presence of the
majority of inorganic and organic chemicals. Most chemical
compounds poison the active centers. Some chemical com-
pounds do it indiscriminately, whereas others selectively
poison only certain types of the centers. Although the poi-
soning selectivity is never complete, it is still significant
5
theory, with the M
w
/M
n
ratio of 2.0. Similar to the earlier
1–4
publications, different centers in this article are arbitrarily
numbered from I to V in the order of increasing mole-
cular weights of the polymer components they produce:
M (Center I) < M (Center II) < M (Center III) < M
w
w
w
w
1
(
Center IV) < M (Center V). The differences between molec-
enough to be useful as a means of modifying the catalysts.
w
Correspondence to: Y. V. Kissin (E-mail: ykissin@rci.rutgers.edu)
Journal of Polymer Science: Part A: Polymer Chemistry, Vol. 48, 4219–4229 (2010) V
C
2010 Wiley Periodicals, Inc.
ETHYLENE POLYMERIZATION REACTIONS, KISSIN AND MINK
4219

JOURNAL OF POLYMER SCIENCE: PART A: POLYMER CHEMISTRY DOI 10.1002/POLA
ꢀ
ꢀ
This article describes the modification of one type of a
precursor of a supported Ti-based Ziegler-Natta catalyst by
pretreating it with organoaluminum chlorides. The principal
experimental tool for the study of the modification effects
is the comparison of the untreated and the pretreated
catalysts in ethylene/1-hexene copolymerization reactions
under standard reaction conditions. Analysis of the produced
copolymers by three techniques, the gel permeation chro-
matographic method (GPC), the crystallization fractionation
method (Crystaf), and the differential scanning calorimetry
columns were heated from 40 to 300 C at a rate of 5 C/
min. GC analysis of model reactions between Si(OEt)4 and
MgBu was carried out at different [Si]:[Mg] ratios, with and
2
without silica. It required the estimation of response factors
for various organosilicon compounds. Some of them were
determined experimentally [for Si(OEt) , Si(n-Bu) , Si(OE-
4
4
t) Et, Si(OEt) Et ], others were calculated based on the addi-
3
2
2
tivity principle.
Copolymerization Reactions
(
DSC), provided detailed information about different types of
The ethylene/1-hexene copolymerization reactions were car-
ried out in a 3.8-L autoclave under standard conditions. The
detailed description of the copolymerization experiments
active centers in the catalyst and about the extent and the
nature of the catalyst modification.
2
–4,7
was presented earlier.
3
AlEt (Akzo Nobel, 25-wt % solu-
EXPERIMENTAL
tion in hexane) was used as a cocatalyst in an amount of 3
to 4 mmol. The amount of the catalyst precursors varied
from 0.02 to 0.05 g. Hydrogen was used in the copolymeriza-
tion reactions as a chain transfer agent to reduce the molec-
ular weight of the copolymers. Typical reactions conditions
Catalyst Precursor
The majority of the copolymerization experiments were car-
ried out with a single supported catalyst precursor. Its prep-
aration procedure was reported previously.
was amorphous microspherical porous silica, which was pre-
1
,2,6,7
The support
ꢀ
were temperature 85 C, 1-hexene concentration in solution
16 and 45 vol %, the total reaction pressure 0.90 to 0.93
ꢀ
calcined at 700 C to remove most of the silanol groups
from its surface. Heptane slurry of the calcined silica was
contacted at 50 C for 40 min with a solution of MgBu2 in
MPa, the hydrogen partial pressure P
H
from 0.12 to 0.18
ꢀ
MPa, the ethylene partial pressure P
E
ꢁ 0.7 to 0.8 MPa, and
hexane (purchased from FMC) at a [MgBu ]:[silica] ratio of
reaction time 60 min.
2
1
.0 mmol/g. Commercial MgBu2 is a mixture of MgR2 com-
pounds with R ¼ n-Bu (45 mol %), sec-Bu (45 mol %), and
n-C 17 (10 mol %). Next, Si(OEt) was added to the slurry
Polymer Analysis
8
H
4
Infrared spectra of the copolymers (thin films hot-pressed at
at a [Si(OEt) ]:[silica] ratio of 0.44 mmol/g, and the mixture
ꢀ
4
ꢁ
140 C) were recorded with a Perkin–Elmer Paragon 1000
ꢀ
was stirred at 50 C for 40 min. Then the mixture was
FTIR spectrophotometer. Composition of the copolymers was
reacted with TiCl at a [TiCl ]:[silica] ratio of 1.0 mmol/g for
8
4
4
calculated from IR data as described previously ; it is reported
as the average molar content of 1-hexene, (C ) , mol %.
ꢀ
4
0 min at 50 to 60 C. Finally, the liquid phase of the slurry
cop av
Hex
was removed by evaporation, and the remaining solid mate-
rial, the catalyst precursor, was dried under a nitrogen flow.
It contains ꢁ3.0 to 3.5 wt % of Ti and can be used in ethyl-
Molecular weights and molecular weight distributions of the
polymers were determined by two techniques. The first one
was GPC. It was performed at 145 C with a Waters 150C
ꢀ
ꢀ
ene polymerization reactions at 80 to 90 C in combination
6
4
3
6
,7
Liquid Chromatograph (two columns 10 , 10 , and 10 Å)
using 1,2,4-trichlorobenzene as a solvent. Resolution of GPC
curves into elemental components (Flory components, see
the definition in refs. 1 and 9) was carried out with the Sci-
entist program (MicroMath Scientific Software), as described
with any standard organoaluminum cocatalyst,
such as
AlEt in this study.
3
This precursor was pretreated with ethylaluminum chlorides,
AlEt Cl (x ¼ 2, 1.5, and 1), with AlMe Cl, and with two
x
3-x
2
other organoaluminum compounds AlEt F and Al(n-C H ) .
9
2
10 21 3
earlier. The second technique used rheological properties of
ꢀ
The pretreatment was carried out in hexane slurry at 55 C
for 30 min at different molar ratios between the organoalu-
polymer melts for measuring weight average molecular
weights of polyethylenes, M , and the widths of their molec-
ular weight distributions, M /Mn ratios. It was based on a
correlation between Mw values of polyethylenes and their
w
minum compounds and TiCl . After removing the liquid
4
w
ꢀ
phase from the slurries, the precursors were dried at 55 C
for 2 h under a nitrogen flow.
melt indices I
2 w n
and a correlation between M /M values and
melt index ratios, I /I . Both melt indices were measured at
2
1
2
Catalyst Analysis
ꢀ
190 C with a Kayeness extrusion plastometer according to
Chemistry of the preparative steps of the catalyst precursor
was studied by gas chromatography (GC). Two Hewlett–
Packard gas chromatographs were used. HP 5880A was
equipped with a wide-bore, 60-m, 0.75-mm i.d. SUPELCO
SPB-1 column and was used in splitless mode with He as a
ASTM method D-1238, Conditions E and F. In the case of
polymers with a medium-broad molecular weight distribu-
tion prepared with typical Ti-based Ziegler-Natta catalysts
(
M /M ꢁ 4–5, I /I ꢁ 25–35), the following empirical cor-
w
n
21
2
2
,8
ꢄ5
relations were found : log(I
M /M ꢂ 0.238ꢅ(I /I ) to 2.4.
2
) ꢂ 1.440 to 1.286 ꢃ 10 ꢅM
w
;
ꢀ
carrier gas (5 cc/min). The column was held at 30 C for 15
min and then heated at a rate of 2 C/min. HP 5890A was
ꢀ
w
n
21
2
equipped with two columns, 60-m and 30-m, 0.28-mm i.d.
capillary RESTEK MTX-1 columns; both were used in split-
flow mode (1:70) with N2 as a carrier gas (1 cc/min). The
Crystaf analysis of ethylene/1-hexene copolymers was
carried out with a Model 110 Crystaf instrument by Polymer
Char. A small amount of a copolymer material, ꢁ0.05 g,
1
0
4220
WILEYONLINELIBRARY.COM/JOURNAL/JPOLA

ARTICLE
ꢀ
was dissolved in 15 mL of 1,2,4-trichlorobenzene at 150 C;
reaction significantly exceeds what would be expected if only
ꢀ
the solution was rapidly cooled to 100 C and then slowly
2
MgBu participated in Reactions 3 to 5, which signifies that
Mg(OEt)Bu also reacts with Si(OEt)4:
ꢀ
ꢀ
cooled at a rate of 0.2 C/min to 30 C, while the concen-
tration of the copolymer remaining in solution was continu-
ously recorded. Resolution of Crystaf curves into peaks
of individual components was carried out similarly to the
SiðOEtÞ4 þ MgðOEtÞBu ! SiðOEtÞ3Bu þ MgðOEtÞ2 (6)
At a [Si]:[Mg] ratio of 0.5, similar to that in the real precur-
1
0
previously described procedure.
ꢀ
sor, the conversion of Si(OEt)4 at 55 C is ꢁ90%, and the
DSC melting curves of the copolymers were recorded using a
DuPont 9900 DSC System. First, the samples were melted
at 140 C, slowly crystallized, and annealed by cooling to
30 C at a rate of 0.5 C/min. Then, the melting curves
were recorded at a heating rate of 2 C/min.
main products are Si(OEt) Bu (80%) and Si(OEt)Bu (15%).
2
2
3
Apparently, the driving force in Reactions 3 to 6 is the high
thermodynamic stability of the Si(C) bond.
ꢀ
ꢀ
ꢀ
ꢁ
ꢀ
When the real catalyst precursor is synthesized in the pres-
2
ence of silica, a significant fraction of all the MgBu is al-
ready spent in Reaction 1 and, therefore, the number of the
Chemistry of Catalyst Synthesis Steps
Every chemical step of the precursor synthesis was studied
by GC. In most cases, model mixtures without silica were
also analyzed by GC.
remaining Mg(Bu) bonds is decreased from ꢁ2.0 to ꢁ1.4
mmol/g silica. In this case, the conversion of Si(OEt) at
4
ꢀ
[
Si(OEt) ]:[MgBu ] of ꢁ0.5 at 60 C after 30 min is only 45
4
2
to 55%, which is much lower than in the experiments with-
Reactions of Calcined Silica with MgR
Calcination of silica at 700 C leaves a small quantity of
2
ꢀ
out silica. The main Si-containing products are Si(OEt) Bu
3
1
1,12
(ꢁ45%) and Si(OEt) Bu (ꢁ5%). They are generated in
silanol groups on its surface, 0.5 to 0.6 mmol/g.
Accord-
2
2
Reactions 3, 4, and 6, as well as the reaction:
ing to refs. 1 and 13, MgBu2 reacts with such silica at two
different sites, one containing a silanol group and another a
siloxane group:
ꢆSiꢄꢄOꢄꢄMgꢄꢄBu þ SiðOEtÞ ! þ ꢆSiꢄꢄOꢄꢄMgꢄꢄOEt
4
þ SiðOEtÞ Bu (7)
3
ꢆSiꢄꢄOH þ MgBu ! ꢆSiꢄꢄOꢄꢄMgꢄꢄBu þ C H10 (1)
2
4
…
Only ꢁ55% of all the silicon species remain in solution, the
rest are absorbed on the silica surface; the fact supported in
separate model experiments.
ꢆSiꢄꢄOꢄꢄSi þ MgBu ! ꢆSiꢄꢄO fMgBu gꢄꢄSiꢆ
(2)
2
2
These reactions can be quantitatively examined by GC,
if MgBu is replaced with Mg(C H ) to avoid the loss of n-
2
6 13 2
Reactions with TiCl4
butane produced in Reaction 1. Several model experiments
TiCl
4
vigorously reacts with all the silicon compounds
ꢀ
were carried out at 55 C in which the calcined silica was
involved in the precursor synthesis, both the starting
Si(OEt)4 and the products of Reactions 3 to 7. Reactions
between solutions of Si(OEt)4 and TiCl4 produced viscous
heptane-immiscible liquids. GC-MS analysis identified two sil-
icon compounds in them, Si(OEt) Cl and Si(OEt) Cl :
reacted with different amounts of Mg(C H ) , then all liquid
6
13 2
reaction products were removed, and the solid, after thorough
washing with heptane, was reacted with an excess of ethanol
to solvolize the remaining MgAC bonds. Analysis of the
evolved n-hexane showed that the conversion of the :SiAOH
groups in Reaction 1 is high and can approach ꢁ90%.
3
2
2
SiðOEtÞ þ TiCl4 ! SiðOEtÞ Cl þ TiðOEtÞCl3
(8)
(9)
4
3
Reactions with Si(OEt)
4
SiðOEtÞ Cl þ TiCl4 ! SiðOEtÞ Cl2 þ TiðOEtÞCl3
3
2
It is well known in the literature that the interaction of a
MgAR moiety with an alkoxysilane leads to an exchange
Similar model reactions between TiCl and two silicon com-
4
0
14
reaction between the MgAR and the SiAOR bonds. Our
experiments with mixtures of solutions of MgBu2 and
Si(OEt)4 in the absence of silica showed that these com-
pounds imitating the products of Reactions 3 and 4, Si(OEt) Et
3
and Si(OEt) Et , produced single exchange products, appa-
2
2
rently, Si(OEt) (Et)Cl and Si(OEt)Et Cl, respectively. The
2
2
ꢀ
pounds rapidly and irreversibly interact even at 25 C with
products of all these reactions can be converted back to the
products of Reactions 3 to 5 if the reaction with TiCl is fol-
lowed by extensive solvolysis with ethanol-water mixtures.
the formation of Si(OEt) Bu, Si(OEt) Bu , Si(OEt)Bu , and a
3
2
2
3
4
small quantity of SiBu4:
TiCl also readily reacts with all the products of Reactions 3
4
SiðOEtÞ þMgBu ! SiðOEtÞ Bu þ MgðOEtÞBu
(3)
(4)
(5)
4
2
3
to 5 formed in the Si(OEt) –MgBu2 system. These reactions
4
SiðOEtÞ Bu þ MgBu ! SiðOEtÞ Bu2 þ MgðOEtÞBu
3
2
2
are rapid and produce voluminous dark-brown solids indica-
tive of the reduction of Ti(IV). Reactions 3 to 5 produce two
families of products, the silicon compounds and the magne-
sium compounds. The reactions of TiCl4 with the silicon
SiðOEtÞ Bu2 þ MgBu ! SiðOEtÞBu þ MgðOEtÞBu
2
2
3
Several isomers of each product with n-Bu and sec-Bu
groups are formed. Conversion of Si(OEt) in Reactions 3 to
depends on the [Si]:[Mg] ratio, is only 25 to 30% at
Si]:[Mg] ꢁ 4, when the dominant product is Si(OEt) Bu
4
compounds give Si(OEt)
9) and SiBu (OEt) Cl compounds. The reaction of a
4-x-y
x
Cl4-x compounds (Reactions 8 and
5
[
(
x
y
MgAOEt moiety (formed in Reactions 3–7) with TiCl should
3
4
ꢁ90%). The number of the formed Si(Bu) bonds in this
produce MgACl bonds and Ti(OEt)Cl . In addition, any
3
ETHYLENE POLYMERIZATION REACTIONS, KISSIN AND MINK
4221

JOURNAL OF POLYMER SCIENCE: PART A: POLYMER CHEMISTRY DOI 10.1002/POLA
a
TABLE 1 AlR
x
Cl3-x-Pretreated Catalysts; Composition and Performance in Ethylene/1-Hexene Copolymerization Reactions
cop av
Hex
[
Al]:[Ti]
Productivity
(C
)
av
w
x
AlEt Cl3-x
(mol/mol)
(g/g catꢅh)
(mol %)
M
M /M
w n
mon
b
Part 1. Reaction conditions: (CHex/C
E
)
)
¼ 5.7-5.9, P
H
¼ 0.186 MPa
b
b
b
b
c
2
AlEt Cl
0
3
1
5,580
3.5
5.1
ꢁ8
108,500
106,900
102,900
4.02 (4.0 )
b
c
4,750
4,150
4.47 (4.8 )
b
0
5.20
mon
c
Part 2. Reaction conditions: (CHex/C
E
¼ 2.1, P
H
¼ 0.134 MPa
2
AlEt Cl
0
1
1
1
2
5,680
1.4
3.7
3.4
3.6
3.5
108,800
115,500
112,000
115,500
119,500
3.7
5.1
5.5
0
3,400
3,550
3,100
2,680
2.5
5
5
mon
c
Part 3. Reaction conditions: (CHex/C
E
)
¼ 2.8, P
H
¼ 0.186 MPa
Al
Et
2 3
Cl
3
3
3
5,450
4.8
100,600
96,100
5.8
7.3
AlEtCl
2
4,360
ꢁ5
a
ꢀ
b
av
Reaction conditions: cocatalyst AlEt
from 0.69 to 0.74 MPa.
3
, 85 C, run time 60 minutes, P
E
M
and M
w
/M
n
values from GPC data.
w
c
av
w n
Estimations of Mw and M /M based on rheological measurements.
MgABu moiety remaining in the system at the end of Reac-
in the solvolysis/hydrolysis products of the final precursor
indicating the absence of any remaining MgAC or TiAC
bonds.
tions 3 to 7 will reduce TiCl , resulting in the formation of
4
MgACl bonds and the reduced Ti-chloride species.
Ti(OEt)Cl , the product of Reaction 8, may also undergo sim-
ilar reactions. All these solid products absorb most of the sil-
icon compounds formed in Reactions 3 to 9.
3
Polymerization Reactions: Pretreatment of Catalyst
Precursor with AlEt Cl
2
Two aspects of the ethylene/1-hexene copolymerization
reactions give the most significant information about the
performance of the multicenter catalysts. The first one is the
structure of the produced copolymers, the presence in
the components with different molecular weight and differ-
ent composition. The second source is the copolymerization
kinetics, which reflects differences in stabilities of different
types of the centers.
The reaction between TiCl
4
and the reaction products in the
systems was also studied with
ternary MgBu /silica/Si(OEt)
2
4
GC. Even before the addition of TiCl , these systems contain
only a fraction of the silicon derivatives remaining in the
supernatant solution, mostly Si(OEt)4 itself and Si(OEt) Bu.
4
3
Addition of TiCl results in a nearly complete disappearance
4
from solution of all silicon compounds. Apparently, these
reactions produce principally the same species as those
formed in the absence of silica, but they all either precipitate
in the silica pores or are strongly adsorbed on the silica sur-
face. The :SiAOAMgAOEt species, the product of Reaction
General Effects
Parts 1 and 2 in Table 1 give the pretreatment conditions
with AlEt
2
Cl and the basic testing results in the copolymer-
ꢀ
mon
ization reactions at 85 C at two different monomer concen-
7, can also undergo exchange reactions with TiCl4 or with
tration ratios (CHex/CE) , one high, ꢁ5.8, and another low,
Ti(OEt)Cl3:
2.1. In general, the pretreatment with AlEt Cl has a signifi-
2
cant effect on the performance of the catalyst system. Its
TiCl4þ ꢆSiꢄꢄOꢄꢄMgꢄꢄOEt ! TiðOEtÞCl3þ
activity deteriorates slightly, especially at higher [AlEt Cl]:
2
ꢆSiꢄꢄOꢄꢄMgꢄꢄCl (10)
[
Ti] ratios, whereas the catalyst becomes much more respon-
sive to the presence of 1-hexene in the copolymerization
To summarize, the solid catalyst precursor used in this study
contains several families of compounds, which affects the
formation and functioning of active centers. They include
such species (all either bonded to the silica surface or
cop av
reaction: the (CHex
in Part 1 of the table and from 1.4 to 3.5 mol % in Part
. Concurrently, the molecular weight distributions of the
)
value increases from ꢁ3.5 to > 8 mol
%
2
copolymers become broader.
strongly adsorbed on it) as MgACl moieties, Si(OEt) Cl4-x,
x
SiBu (OEt) Cl4-x-y, as well as titanium chlorides and titanium
To evaluate the effects of the copolymerization conditions,
x
y
ethoxychlorides in which the Ti atoms are in various oxida-
tion states. All the Mg species and all the Si species are
absorbed within the silica pores in the precursor. Virtually,
no organometallic compounds were found in the supernatant
the precursor pretreated at [AlEt Cl]:[Ti] ¼ 10 was tested in
2
mon
a series of reactions at three different (CHex/C
E
)
ratios,
mon
5.85, 2.60, and 2.17. As expected, as the (CHex/CE)
ratio
value also decreases from
cop av
decreases, so does the (CHex
)
after the TiCl addition and practically no butane was found
ꢁ8 to 3.4 mol %. This dependence was analyzed with the
4
4222
WILEYONLINELIBRARY.COM/JOURNAL/JPOLA

ARTICLE
copolymer composition equation in the case of low a-olefin
1
cop
mon
contents, [CHex/(1 ꢄ CHex)]
ꢂ (1/r
1
)ꢅ(CHex/C
E
)
. The
reactivity ratio r1 for the untreated catalyst is ꢁ80 to 100,
that is, ethylene is much more reactive in the copolymeriza-
1
,3
tion reaction compared with 1-hexene.
However, the r1
value for the pretreated catalyst decreased to ꢁ50 reflecting
the increased sensitivity to 1-hexene.
Part 3 in Table 1 gives the copolymerization data for the
same catalyst precursor pretreated with two other ethyl-
aluminum chlorides, Al Et Cl and AlEtCl . The results are
2 3 3 2
similar: the pretreated catalysts become more responsive to
the presence of 1-hexene in the copolymerization reactions
and their molecular weight distribution becomes broader.
Nature of Catalyst Modification
Up to this point, the general results on the pretreatment
with AlEt Cl do not differ much from several similar studies
2
of other ethylene polymerization catalysts. The most perti-
nent results of this type were published in 1992 to 2000 in
a series of articles, which examined the AlEt Cl pretreatment
2
of a well-defined crystalline Ti-Mg complex [Mg (l-Cl) ꢅ
2
3
þ
ꢄ
(
THF) ] [TiCl ꢅTHF] and silica-supported catalysts using
6
5
15–19
this complex.
The pretreatment of the complex at
[
AlEt Cl]:[Ti] ratios from 1.4 to 8.6 was also accompanied by
2
significant changes in the catalyst behavior. In particular, this
pretreatment also significantly increased the sensitivity of
the catalysts to a-olefins and broadened the molecular
weight distribution of the copolymers. These are the same
effects that were found in this study, although the catalyst
precursor we used was synthesized from different reagents
and in a completely different procedure.
Catalyst modification is usually viewed in the literature as
chemical reactions between precursors of active centers and
the modifying chemical compounds, that is, as reactions which
change the nature of the active centers. For example, the
FIGURE 1 GPC curves of ethylene/1-hexene copolymer pro-
duced with the untreated catalyst (top) and with the precursor
pretreated at [AlEt
2
Cl]:[Ti] ¼ 10 (bottom) (see Table 1).
changes in the behavior of the AlEt Cl-pretreated TiAMg com-
2
plex described in refs. 15–19 were ascribed to a change in the
1
,9
3þ
molecular weight distribution (M /Mn ꢁ2.0) ; their GPC
ratio between isolated and clustered Ti
species after the
w
curves are narrow and have a characteristic asymmetric
reaction of the pretreated complex with the cocatalyst AlEt3.
9
appearance. The M
w
/M
n
ratios of the polymers prepared
However, the use of modern polymer characterization techni-
ques provides more nuanced information about the nature of
the catalyst modification. The modifiers interact with the pre-
cursors of active centers and their action can be partially
selective; that is, they may reduce the number or reactivity of
some types of active centers, while leaving other centers
mostly intact or even increasing their number or reactivity.
The data discussed in the following sections demonstrate that,
indeed, viewing catalyst modifiers as the agents of redistribu-
tion in the relative populations of different active centers can
explain the experimental findings presented in Table 1.
with supported Ti-based Ziegler-Natta catalysts usually vary
between 3.5 and 5.5 (Table 1). GPC curves of such polymers
are quite broad and can be resolved into GPC curves of
their elemental components, each component representing a
polymer material produced by a single type of active center
or by several types of centers producing polymer compo-
nents with similar average molecular weights.
1,9
Figure 1 shows two examples of such resolution, one for an
ethylene/1-hexene copolymer prepared with the untreated
catalyst and another a copolymer prepared with the catalyst
pretreated at [AlEt Cl]:[Ti] ¼ 10. Table 2 gives the results of
2
Changes in Molecular Weight Distribution
the GPC resolution. Each catalyst contains five types of active
centers, which differ greatly in the Mw values of the polymer
components they produce, from ꢁ5000 for the component
produced by Center I to ꢁ420,000 for the component pro-
duced by Center V. Taking into account, the unavoidable
uncertainty in resolution of curves of a complex shape into
As described in Introduction, all heterogeneous Ziegler-Natta
polymerization catalysts contain several populations of active
centers which differ in the molecular weights of polymer
components they produce. If only one type of center exists
in a catalyst, polymers prepared with it have a narrow
ETHYLENE POLYMERIZATION REACTIONS, KISSIN AND MINK
4223

JOURNAL OF POLYMER SCIENCE: PART A: POLYMER CHEMISTRY DOI 10.1002/POLA
2
TABLE 2 GPC Curve Resolution Results for Ethylene/1-Hexene Copolymers Produced with Untreated and AlEt Cl-Pretreated
Catalysts at a High 1-Hexene Concentration (Part 1 in Table 1)
[
AlEt
2
Cl]:[Ti] ¼ 03.5 mol %
[AlEt
2
Cl]:[Ti] ¼ 35.1 mol %
[AlEt
2
Cl]:[Ti] ¼ 10 ꢁ 8 mol %
cop av
Hex
(
C
)
Center
M
w
Fraction (%)
M
w
Fraction (%)
M
w
Fraction (%)
I
ꢁ5,600
15,700
ꢁ1
4,700
2.2
13.4
42.7
34.0
7.7
4,800
19,200
4.3
19.6
41.8
27.2
7.1
II
10.6
39.4
36.6
12.2
17,000
53,300
III
IV
V
47,900
57,700
124,400
342,000
144,700
430,800
154,300
452,700
9
several closely spaced elemental peaks (Fig. 1), it can be
is mostly produced by Center II. In addition, a significant
fraction of the copolymers does not crystallize even at the
lowest temperature in the Crystaf procedure, 30 C. This
dissolved material is nearly completely amorphous, it is
mostly produced by Center I.
stated that modification with AlEt Cl does not affect the
2
ꢀ
10
performance of individual centers: the average molecular
weight of each polymer component remains approximately
the same. What changes with the pretreatment are the con-
tents of the polymer components: as the [AlEt Cl]:[Ti] ratio
2
Table 3 lists parameters of elemental peaks in the Crystaf
curves of ethylene/1-hexene copolymers prepared under
conditions given in Part 1 in Table 1 with both the untreated
increases, the fractions of the components produced by
Centers IV and V gradually decrease and the fractions of the
components produced by Centers I and II increase.
catalyst and the one pretreated at [AlEt
2
Cl]:[Ti] ¼ 10. Both
This redistribution of the copolymer components is univer-
sal. For example, the same effect was observed in the copoly-
curves contain peaks with maximums at approximately the
same crystallization temperatures; therefore, the composi-
tions of the elemental components (estimated as described
in ref. 10) are also close for the two copolymers. As shown
in Table 3, the most obvious difference between the copoly-
mers is the relative areas of different Crystaf peaks. The pre-
treated catalyst produced a copolymer with a lower fraction
of components A–C and with a much higher fraction of the
soluble amorphous material. Of course, these changes are
mon
merization reactions at a significantly lower (CHex/CE)
ratio of 2.1: the catalyst pretreated at [AlEt Cl]:[Ti] ¼ 15
2
produced a copolymer with 6% less of components IV and V.
This change was sufficient to increase the 1-hexene content
in the copolymer from 1.4 to 3.6 mol %.
Changes in Compositional Distribution
The Crystaf method is an effective technique for the analy-
sis of the compositional distribution in ethylene/a-olefin
1
0,20,21
copolymers.
Crystaf curves of compositionally uniform
copolymers prepared with catalysts containing only one type
1
,10
of active center have single narrow crystallization peaks.
In contrast, Crystaf curves of ethylene/a-olefin copolymers
produced with multicenter Ziegler-Natta catalysts are broad
and consist of several overlapping peaks at crystallization
ꢀ
temperatures from 83 to ꢁ35 C (Fig. 2).
These Crystaf data were quantified using the earlier devel-
1
0
oped computerized method, which is based on resolution
of the Crystaf curves into their elemental components. Each
such component consists of copolymer molecules of approxi-
mately the same composition; the lower is the a-olefin con-
tent in a copolymer, the higher is its crystallization tempera-
ture. The components in the Crystaf curves are arbitrarily
marked as A, B, C, etc., in the order of decreasing crystal-
lization temperature. Complex cocrystallization phenomena
1
0
between copolymer components of different compositions
result in a strong dependence of half-widths of the elemental
peaks on their mutual proximity obvious from Figure 2. The
range of compositions for different components is quite
broad, from 0.3 to 0.5 mol % of 1-hexene for the highly crys-
talline material of peaks A and B (the material produced by
Centers IV and V) to ꢁ5.5 to 6 mol % for the peak F, which
FIGURE 2 The Crystaf curve of an ethylene/1-hexene copoly-
mer produced with the untreated catalyst and its resolution
into elemental components.
4224
WILEYONLINELIBRARY.COM/JOURNAL/JPOLA

ARTICLE
TABLE 3 Parameters of Individual Peaks in Crystaf Curves of Two
a
Ethylene/1-Hexene Copolymers
Crystaf
Peak
Fraction (%)
Fraction (%)
ꢀ
cop
Hex
T ( C)
C
(mol %)
at [Al]:[Ti] ¼ 0
at [Al]:[Ti] ¼ 10
A
83.1–83.2
81.5–81.7
76.6–77.2
68.2–69.4
58.3–60.3
35.2–40.2
0.3
0.5
9.0
27.3
8.8
5.2
14.9
5.6
B
C
1.0–1.1
1.9–2.0
3.0–3.2
5.3–5.9
D
10.5
12.8
15.1
16.6
6.0
E
10.2
15.6
42.6
F
Soluble fraction
a
Reaction conditions are in Part 1 of Table 1.
reflected in the change in the average copolymer composi-
ferent meanings depending on the nature of interactions
between the active centers in the catalysts and the modifiers.
The active centers in Ziegler-Natta catalysts are coordina-
tively unsaturated, and the most common type of such an
interaction is poisoning of the active centers in Lewis acid-
cop av
tion: the (CHex
)
value increases from 3.5 to ꢁ8 mol %.
Cl effect on the compo-
The conclusions concerning the AlEt
2
sitional distribution of the copolymers were corroborated by
the data on the melting points and the crystallinity degree of
the copolymers, which were measured with the DSC method
2
2
base reactions. Such reactions can be either completely
reversible (with such poisons as CO or allene) or practically
(
Table 4). The position of the main melting peak on the DSC
1
,3,22,23
irreversible (with amines, conjugated dienes, etc.).
All the data in Tables 2 to 4 demonstrate that increasing the
AlEt Cl]:[Ti] ratio results in a decrease in the fraction of
curve of any ethylene/a-olefin copolymer prepared with a
Ti-based catalyst reflects the properties of copolymer mate-
rial with the lowest a-olefin content and the highest crystal-
[
1
2
linity degree. These copolymer fractions are produced by
cop
Hex
highly crystalline polymer components with low C
and
Centers IV and V; they appear in GPC curves as polymer
components IV and V (Fig. 1) and in Crystaf curves as com-
ponents A and B (Fig. 2). The analysis of the GPC and the
Crystaf data from Tables 2 and 3 suggests that the compo-
nents produced by Centers IV and V under the same copoly-
merization conditions should have the same characteristics,
although the fractions of these materials decrease as the
high molecular weights and an increase in the fraction of the
cop
components of low crystallinity with high C
and low
Hex
molecular weights. In principle, all these data can be inter-
preted either as gradual poisoning of the precursors of Cen-
ters IV and V, or as gradual activation of the precursors of
Centers I and II, or as both the changes occurring in parallel.
[
2
AlEt Cl]:[Ti] ratio in the pretreated catalyst increases. The
The following approach was used to discriminate between
these three possibilities. Table 5 compares absolute yields of
different copolymer components produced in reactions under
identical conditions (listed in Part 1 of Table 1). Two sets of
data are compared; one derived from the Crystaf data and
another from the GPC data. To avoid uncertainties related to
separation of closely spaced peaks (see Figs. 1 and 2), both
the GPC and the Crystaf results were rearranged to produce
the data for several distinct large fractions of the copoly-
mers. The results clearly show that the pretreatment of the
DSC data confirm this prediction: the peak melting points of
the copolymers remain practically the same irrespective of
the amount of AlEt Cl used for the catalyst pretreatment.
However, the crystallinity degree of these copolymers (also
estimated from DSC data) decreases in parallel with the
increase of the [AlEt Cl]:[Ti] ratio; it reflects the decrease of
2
2
the fraction of the highly crystalline components produced
by Centers IV and V.
Polymerization Reactions: Different Types of
Catalyst Modification
catalyst precursor with AlEt Cl produces two different
2
The term ‘‘catalyst modification’’, when applied to hetero-
geneous Ziegler-Natta catalysts, is general and can have dif-
effects. First, it indeed significantly reduces the productivity
of active Centers IV and V, those centers which generate high
TABLE 4 Crystaf and DSC Data for Ethylene/1-Hexene Copolymers (Part 1 in Table 1)
cop av
Hex
(
C
)
Amorph.
Highly Crystalline
Crystallinity
Degree (%)
ꢀ
m
T ( C)
[
AlEt
2
Cl]:[Ti]
(mol %)
Material (%)
Fractions AþB (%)
0
3
1
3.5
5.1
ꢁ8
16.6
26.5
42.6
36.3
23.9
20.2
126.5
127.1
126.5
40-42
ꢁ39
ꢁ13
0
ETHYLENE POLYMERIZATION REACTIONS, KISSIN AND MINK
4225

JOURNAL OF POLYMER SCIENCE: PART A: POLYMER CHEMISTRY DOI 10.1002/POLA
1
–4
TABLE 5 Yields of Various Components of Ethylene/1-Hexene
fin copolymerization reactions.
Figure 3 gives one example
a
2
Copolymers Produced with Untreated and AlEt Cl-Pretreated
of kinetic curves of ethylene consumption by different types
of active centers in ethylene/1-hexene copolymerization
reactions with the untreated catalyst at 80 C. The method
Catalysts (Part 1 in Table 1); Data from Crystaf and
GPC Analysis
ꢀ
used for the generation of these kinetic curves was thor-
oughly described in refs. 3 and 4. When such copolymeriza-
tion reactions are carried out at 80 to 90 C, Centers I and II
[
AlEt
2
Cl]:
2
[AlEt Cl]:
Catalyst:
[Ti] ¼ 0
[Ti] ¼ 10
ꢀ
are formed rapidly within ꢁ10 min, but they decompose
rapidly and become inactive after 20 to 30 min. In contrast,
Centers IV and V are formed much slower, during a period of
Total yield (g/g catꢅh)
Crystaf data, yields (g/g catꢅh)
Highly cryst. fraction (A þ B)
Medium-cryst. Fraction (C–E)
Low-cryst. fraction (F)
Amorphous material
GPC data, yields (g/g catꢅh)
IV þ V
5,580
4,150
2,025
1,790
840
840
900
ꢁ
25 to 30 min, but they are significantly more stable,
whereas Centers III exhibit an intermediate behavior.
640
These kinetic differences provide additional information
930
1,770
about the nature of the catalyst modification with AlEt Cl.
2
Figure 4(A) shows kinetic curves of ethylene consumption in
the copolymerization reactions with the untreated and the
2,720
2,200
650
1,420
1,740
990
III
AlEt Cl-pretreated catalysts. When the untreated catalyst
2
I þ II
precursor is used, the kinetic curve (which is the sum of the
molecular weight copolymer components with a low 1-hex-
ene content. The amounts of Crystaf components A and B
and the amounts of GPC components IV and V are both
reduced by a factor of 2 to 2.5. The data in Table 5 also
reveal the second pretreatment effect: the productivity of
active centers I and II, the centers producing low molecular
weight copolymer components with a high 1-hexene content
(
the amorphous material in the Crystaf analysis, GPC compo-
nents I and II), increases by a factor of 1.5 to 2.
Polymerization Reactions: Kinetic Manifestations of
Catalyst Modification
Different types of active centers in supported Ziegler-Natta
catalysts exhibit different kinetic behavior in ethylene/a-ole-
FIGURE 4 Kinetic curves of ethylene consumption in ethylene/
FIGURE 3 Kinetic curves of ethylene consumption by different
1-hexene copolymerization reactions. Reaction conditions are
types of active centers in ethylene/1-hexene copolymerization
in Part 1 of Table 1. A: Standard catalyst; numbers give [AlEt
ratios. B: Catalyst precursor prepared at [Si(OEt) ]:[MgBu
] ꢁ0.68,
pretreatment at [AlEt
Cl]:[Ti] ¼ 10.
2
Cl]:[Ti]
ꢀ
mon
reactions with the untreated catalyst at 80 C at (CHex/C
.0.
E
)
¼
4
2
2
2
4226
WILEYONLINELIBRARY.COM/JOURNAL/JPOLA

ARTICLE
kinetic curves of all five types of active centers, as shown in
with the data for the untreated precursor. Crystaf analysis
confirmed these results: the fraction of components A and B
decreased from >67% to ꢁ62%, the fraction of components
of low crystallinity increased from ꢁ16 to ꢁ23%, and the
fraction of the soluble amorphous material increased from
Fig. 3) reaches a maximum at ꢁ8 to 10 min and then gradu-
ally declines. When this precursor is pretreated with AlEt Cl,
2
two changes in the catalyst behavior become apparent:
1. At the end of the copolymerization reactions (60 min in
ꢁ
5 to 14.5%.
these experiments), the reaction rates decrease as the
[
AlEt Cl]:[Ti] ratio increases:
In contrast, the pretreatment of the catalyst precursor with
2
Al(n-C H ) and AlEt F resulted in severe deterioration of
1
0
21 3
2
½
AlEt2Clꢇ:½Tiꢇ in the precursor:
0
1
1
3
10
the catalyst activity. Both compounds indiscriminately
destroyed precursors of the active centers of all types and
produced no changes in the contents of the polymer
components.
Relative rate after 60 minutes:
0:78 0:72 0:58
As Figure 3 shows, Centers I and II and a large fraction of
Centers III have already decomposed after 60 min; there-
fore, this decline in activity reflects the reduction in the
number of Centers IV and V, in agreement with the data
in Table 5.
DISCUSSION
The catalyst described in this article was designed to pro-
duce linear low-density polyethylene (LLDPE) resins, that is,
semicrystalline ethylene/a-olefin copolymers with the a-ole-
fin content of 3.0 to 3.5 mol %. The main practical require-
ments for such catalysts are high productivity, the ability to
copolymerize well a-olefins with ethylene, and the ability to
produce copolymers with a moderately broad molecular
weight distribution characterized by a M /M ratio of 4.0 to
2
. In the beginning of the copolymerization reactions, the
reaction rates increase as the [AlEt Cl]:[Ti] ratio increases.
2
This change supports the second conclusion from the data
in Table 5: the pretreated catalysts contain a higher num-
ber of Centers I and II, which become active soon after
the start of the polymerization reaction.
Both effects were even more apparent when another similar
catalyst precursor was tested. It was prepared following the
same procedure (see Experimental section) but using a
w
n
5.0.
The application of various distributive analytical techniques,
such as GPC and Crystaf, allows reformulation of these gen-
eral requirements into specific, active center-oriented goals:
a deliberate manipulation of the active center distribution.
From the practical point of view, it is desirable to reduce the
productivity of Centers IV and V (the centers producing high
molecular weight and highly crystalline copolymer compo-
nents) and to enhance the productivity of Centers II and III.
lower [MgBu ]:[silica] ratio, 0.85 mmol/g, and a higher
2
[
Si(OEt) ]:[silica] ratio, 0.68. This precursor was prepared at
4
the same [TiCl ]:[silica] ratio of 1.0 mmol/g, and therefore,
4
the principal chemical difference between it and the stand-
ard precursor is the higher molar [Si(OEt) ]:[MgBu ] ratio,
4
2
ꢁ
0.68 versus 0.44. This second precursor was also modified
with AlEt Cl. The untreated precursor with this formulation
2
contains a low number of Centers I and II and its activity is
dominated by Centers IV and V, and the respective kinetic
curve in Figure 4(B) is nearly flat (see Fig. 3 for compari-
The aforementioned results show that the pretreatment
with AlEt Cl or with AlMe Cl does not substantially change
2 2
son). The pretreatment at [AlEt Cl]:[Ti] ¼ 10 brought about
the chemical structure of individual active centers in the cat-
alyst, but it changes the proportion among the different
types of centers. Two major changes occur in parallel: a sig-
nificant reduction in the activity of Centers IV and V and a
noticeable increase in the activity of Centers I and II. The
existence of these two unrelated effects calls for two differ-
ent explanations.
2
the same two changes as those in the standard catalyst pre-
cursor. First, the reaction rate at the end of the copolymer-
ization reaction decreased by ꢁ55% due to poisoning of
Centers IV and V, and second, the initial rate increased by
ꢁ
80% due to activation of Centers I and II. Both these
changes transformed a relatively stable catalyst of moderate
activity into a rapidly decaying one.
According to ref. 22 active Centers IV and V contain reduced
Ti atoms. These centers readily polymerize ethylene, but
they copolymerize a-olefins with ethylene poorly, and they
do not homopolymerize a-olefins. Our approximate estima-
tions of the productivity of these two centers in several
untreated catalysts prepared according to the same recipe
(described in Experimental section) with different amounts
of MgBu2 and Si(OEt)4 showed that the productivity of
Centers IV and V in the copolymerization reactions does not
depend on the [Si(OEt) ]:[MgBu ] ratio and remains nearly
Other Organoaluminum Modifiers
Three other organoaluminum compounds were tested as
modifiers for the standard catalyst precursor. Its treatment
with AlMe
2
Cl was carried out at [AlMe
2
Cl]:[Ti] ¼ 10, and the
pretreated precursor was tested under the copolymerization
conditions in Part 2 of Table 1. As expected, the pretreat-
ment with AlMe Cl produced the results similar to those
2
caused by the pretreatment with AlEt Cl: the 1-h productiv-
ity decreased from 5680 to 4950 g/g cat, and the (CHex
value increased from 1.4 to 3.3 mol %. GPC resolution data
also showed a similar pattern: the combined content of com-
ponents IV and V decreased by ꢁ7% and the combined con-
tent of components I and II increased by ꢁ30% compared
2
4
2
cop av
)
constant, 2500 to 2700 g/g cat h. Thus, the reactivities of
Centers IV and V are independent of the presence of Si(OEt)
4
and the products of its reactions during the synthesis of the
precursor (Reactions 3–9). Judging by the data in Table 5,
selective poisoning of these centers with AlEt Cl, which is
2
ETHYLENE POLYMERIZATION REACTIONS, KISSIN AND MINK
4227

JOURNAL OF POLYMER SCIENCE: PART A: POLYMER CHEMISTRY DOI 10.1002/POLA
also obvious from the kinetic data in Figure 4, is most prob-
ably caused by Lewis-acidic properties of the modifier.
ꢆSiꢄꢄOEt þ AlEt2Cl ! SiꢄꢄEt þ AlðEtÞðOEtÞCl
(11)
Active Centers I and II behave differently compared with
Centers IV and V. These centers easily copolymerize a-olefins
with ethylene, and they can homopolymerize a-olefins,
mostly to atactic polymers and polymers of a low isotacticity
level. When the untreated catalyst precursors are prepared
However, the final products could be species containing both
Al and Si atoms, which are linked via bridging OEt groups.
The modifying effect of AlEt Cl on supported Ziegler-Natta
2
catalysts in ethylene/a-olefin copolymerization reactions has
important practical implications. These catalysts are often
used to copolymerize ethylene and 1-hexene in gas-phase,
fluidized-bed reactors. An improvement in the overall ability
of such catalysts to effectively copolymerize 1-hexene with
ethylene allows the operation of the reactors at lower [1-
hexene]/[ethylene] ratios and, therefore, at higher ethylene
partial pressures and with higher productivity. Another
advantage of the catalysts that incorporate 1-hexene well is
their ability to produce low density LLDPE resins in gas-
phase fluidized-bed reactors, a target that is difficult to reach
when using unmodified catalysts.
with different amounts of Si(OEt) , the higher the amount of
4
the used Si(OEt) the lower the productivity of Centers I and
4
II. Centers I and II are especially active in the beginning of
the copolymerization reactions, the time periods when Cen-
ters IV and V are only starting to form (Fig. 3). The initial
polymerization rate for the untreated precursor prepared at
[
Si(OEt) ]:[MgBu ] ¼ 0.44 is ꢁ110 g/g cat min [Fig. 4(A)],
4
2
but it decreases to 45 to 50 g/g cat min at
Si(OEt) ]:[MgBu
] ꢁ0.68 [Fig. 4(B)].
[
4
2
As discussed in Section 3.1, reactions between silica, MgBu2,
and Si(OEt) in the course of the precursor synthesis gener-
4
ate several organosilicon compounds of the general formula
Si(OEt) Bu in Reactions 3 to 7. Then these compounds are
The authors are indebted to J. L. Garlick (Edison Research Labo-
ratory, Mobil Chemical Company) for experimental assistance.
W. Harrison and J. Orvos (Mobil Chemical Company) carried
out GPC measurements and Dr. H. A. Fruitwala (Mobil Chemical
Company) carried out Crystaf analysis of ethylene/1-hexene
copolymers. Dr. D. D. Hills-Haney (Mobil Chemical Company)
performed GC/MS analysis of catalyst synthesis products.
x
4-x
converted by TiCl4 into organosilicon chlorides Si(OEt)x-
Bu Cl in Reactions 8 and 9 Apparently, some of these
y
4-x-y
organosilicon products coordinate with the predecessors of
Centers I and II and either prevent the formation of the
active centers from them or partially poison the centers after
they are formed. The higher is the amount of Si(OEt)4 with
respect to the number of the MgAC bonds present during
the precursor synthesis the higher the poisoning degree.
REFERENCES AND NOTES
This type of selective poisoning by alkylalkoxysilanes is well
researched. When Ti-based Ziegler-Natta catalysts are used
1
Kissin, Y. V. Alkene Polymerization Reactions with Transition
for homopolymerization of propylene and other a-olefins,
Metal Catalysts. Elsevier: Amsterdam, 2008; Chapter 4.
0
alkylalkoxysilanes of the general formula Si(OR) R
(where
4-x
x
2
Kissin, Y. V.; Brandolini, A. J.; Mink, R. I.; Nowlin, T. E.
0
x ¼ 2 and 3; R ¼ Me, Et; R ¼ isoBu, isoPr, cyclohexyl,
and cyclopentyl) are widely used for selective poisoning of
active Centers I and II, the centers which form atactic
polymers. This poisoning improves the fractional isotacticity
J Polym Sci Part A: Polym Chem 2009, 47, 3271–3285.
3
Kissin, Y. V.; Mink, R. I.; Nowlin, T. E. J Polym Sci Part A:
Chem 1999, 37, 4255–4272.
1
,24–26
0
of a-olefin polymers.
The effects of Si(OR)xR
com-
4 Kissin, Y. V.; Mink, R. I.; Nowlin, T. E.; Brandolini, A. J.
4-x
pounds in ethylene/a-olefin copolymerization reactions are
Topics Catal 1999, 7, 69–88.
2
2
also known. When these compounds are admitted into the
copolymerization reactions, the average catalyst activity
decreases (all these silicon compounds are poisons), the ini-
tial polymerization rates decrease, a-olefin contents in the
copolymers decrease, and the average molecular weights of
the copolymers increase. All these data are in agreement
5
Flory, P. J. Principles of Polymer Chemistry. Cornell Univer-
sity Press: Ithaca, New York, 1953.
6
1
5
5
(a) Nowlin, T. E.; Mink, R. I. (Mobil Oil) US Patent 5,336,652,
994; (b) Nowlin, T. E.; Mink, R. I. (Mobil Oil) US Patent
,470,812, 1995; (c) Nowlin, T E; Mink, R. I. (Mobil Oil) US Patent
,939,348, 1999; (d) Nowlin, T. E.; Mink, R. I. (Mobil Oil) US
with the conclusion that the compounds Si(OEt) Bu
x
4-x
Patent 6, 2001, 291, 384.
Hagerty, R. O.; Husby, P. K.; Kissin, Y. V.; Mink, R. I.; Nowlin,
T. E. (Mobil Oil) US Patent 5, 1997, 693, 583.
Kissin, Y. V.; Nowlin, T. E.; Wagner, K. P. J Polym Sci Part A:
Chem 1988, 26, 755–764.
formed in Reactions 3 to 7 and their chlorinated derivatives
Si(OEt) Bu Cl4-x-y may indeed be selective poisons of active
7
x
y
Centers I and II.
8
Apparently, AlEt
2
Cl and other organoaluminum chlorides
react with these compounds, convert them into nonpoison-
ous products, and thus restore the predecessors of Centers I
9
Kissin, Y. V. J Polym Sci Part A: Polym Chem 1995, 33,
27–237.
2
and II. AlEt Cl is usually used in a significant excess with
2
1
0 Kissin, Y. V.; Fruitwala, H. A. J Appl Polym Sci 2007, 106,
respect to the SiAOEt bonds remaining in the catalyst pre-
cursor after Reactions 3 to 7 (Table 1), and one can assume
that the primary reaction between the SiAOEt bond and
3
872–3883.
11 Soga, K.; Hyakkoku, K.; Izumi, K.; Ikeda, S. J. Polym Sci,
2
7
AlEt Cl is similar to its reaction with the Grignard reagent.
Polym Chem Ed 1978, 16, 2383–2394.
2
4228
WILEYONLINELIBRARY.COM/JOURNAL/JPOLA

ARTICLE
12 Soga, K.; Izumi, K.; Terano, M.; Ikeda, S. Makromol Chem
21 Soares, J. B. P.; Monrabal, B.; Nieto, J.; Blanco, J. Macro-
1980, 181, 657–666.
mol Chem Phys 1998, 199, 1917–1927.
1
3 Nowlin, T. E.; Mink, R. I.; Lo, F. Y.; Kumar, T. J Polym Sci
22 Kissin, Y. V. Alkene Polymerization Reactions with Transi-
tion Metal Catalysts. Elsevier: Amsterdam, 2008; Chapters 5
and 6.
Part A: Polym Chem 1991, 29, 1167–1173.
14 (a) George, P. D.; Strong, W. A.; Sommer, L. H.; Whitmore, F. C.
J. Am Chem Soc 1946, 68, 1380–1387; (b) George, P. D.; Sommer,
L. H.; Whitmore, F. C. J Am Chem Soc 1955, 77, 6647–6653.
2
3 Kissin, Y. V. Makromol Chem, Macromol Symp 1995, 89,
1
13–123.
15 Chang, H.-S.; Song, W.-D.; Chu, K.-J.; Ihm, S.-K. Macromole-
2
4 Sacchi, M. C.; Forlini, F.; Tritto, I.; Mendici, R.; Zannoni, G.;
cules 1992, 25, 2086–2097.
Noristi, L. Macromolecules 1992, 25, 5914–5932.
5 Di Noto, V.; Cecchini, G.; Zannetti, R.; Viviani, M. Macromol
Chem Phys 1994, 195, 3395–3402.
6 Chadwick, J. C.; van der Burgt, F. P. T. J.; Rastogi, S.;
16 Kim, J. H.; Han, T.-K.; Choi, H. K.; Kim, I.; Woo, I. Macromol
2
Rapid Commun 1995, 16, 113–117.
1
1
7 Karol, F. J. J Catal Rev—Sci Eng 1984, 26, 557–578.
2
8 Chu, K.-J.; Soares, J. B. P.; Penlidis, A.; Ihm, S.-K. Macromol
Busico, V.; Cipullo, R.; Talarico, G.; Heere, J. J. R Macromole-
cules 2004, 37, 9722–9732.
Chem Phys 1999, 200, 1298–1314.
9 Chu, K.-J.; Soares, J. B. P.; Penlidis, A.; Ihm, S.-K. Eur
Polym Mater 2000, 36, 3–12.
0 Monrabal, B. J Appl Polym Sci 1994, 52, 491–513.
1
2
7 Lindsell, W. E. In Comprehensive Organometallic Chemistry;
Wilkinson, G.; Stone, F. G. A.; Abel, E. W., Eds.; Pergamon: Oxford,
1986; Vol. 1, Chapter 4.
2
ETHYLENE POLYMERIZATION REACTIONS, KISSIN AND MINK
4229