C1-Bridged fluorenylidene cyclopentadienylidene complexes of the type (C13H8-CR1R2-C5H 3R)ZrCl2 (R1, R2=alkyl, phenyl, alkenyl; R=H, alkyl, alkenyl, substituted silyl) as catalyst precursors for the polymerization of ethylene and propylene

Article abstract of DOI:10.1016/S0022-328X(98)00740-2

The synthesis and characterization of 14 new C₁-bridged fluorenylidene cyclopentadienylidene complexes of zirconium is described. After activation with methylaluminoxane (MAO), these metallocene complexes can be used for the polymerization of ethylene and propylene. Various substituents in the bridge and in position 3 of the cyclopentadienylidene ligand have a strong influence on the activity of the corresponding catalysts and the molecular weights of the formed polyolefins. The bulkiness of the substituents in position 3 of the cyclopentadienylidene ring determines the extent of the stereospecific polymerization of propylene. Substituents can be selected that produce hemitactic, syndiotactic or isotactic polypropylenes and/or the corresponding block copolymers.

Full text of DOI:10.1016/S0022-328X(98)00740-2

Journal of Organometallic Chemistry 568 (1998) 87–112
C1-Bridged fluorenylidene cyclopentadienylidene complexes of the type
(C13H8–CR1R2–C5H3R)ZrCl2 (R1, R2=alkyl, phenyl, alkenyl; R=H,
alkyl, alkenyl, substituted silyl) as catalyst precursors for the
polymerization of ethylene and propylene
Helmut G. Alt *, Michael Jung
Laboratorium fu¨r Anorganische Chemie der Uni6ersita¨t Bayreuth, Postfach 10 12 51, D-95440 Bayreuth, Germany
Received 13 March 1998
Abstract
The synthesis and characterization of 14 new C1-bridged fluorenylidene cyclopentadienylidene complexes of zirconium is
described. After activation with methylaluminoxane (MAO), these metallocene complexes can be used for the polymerization of
ethylene and propylene. Various substituents in the bridge and in position 3 of the cyclopentadienylidene ligand have a strong
influence on the activity of the corresponding catalysts and the molecular weights of the formed polyolefins. The bulkiness of the
substituents in position 3 of the cyclopentadienylidene ring determines the extent of the stereospecific polymerization of propylene.
Substituents can be selected that produce hemitactic, syndiotactic or isotactic polypropylenes and/or the corresponding block
copolymers. © 1998 Elsevier Science S.A. All rights reserved.
Keywords: Catalysis; Polymerization; Metallocene complexes; Zirconium; Polyethylene; Polypropylene
1. Introduction
substituents in the bridge and substituents in position 3
of the cyclopentadienylidene moiety on the polymeriza-
tion of ethylene and propylene.
The C1-bridged fluorenylidene cyclopentadienylidene
complex (C13H8–CMe2–C5H4)ZrCl2 is regarded as the
classical metallocene catalyst precursor for the synthesis
of syndiotactic polypropylene [1–4]. Recently we
showed that substitutions on the fluorenylidene ligand
[5–7] or on the bridging carbon atom [6,8] strongly
influence the catalyst activity as well as the molecular
weight of the produced syndiotactic polypropylene.
Substituents in position 3 of the cyclopentadienylidene
ligand of these complexes play key roles in determining
the stereoselectivity of polypropylenes produced with
the respective catalysts [9–12]. In this paper we report
the synthesis and the characterization of new C1-
bridged fluorenylidene cyclopentadienylidene com-
plexes. We have investigated the influence of various
2. Results and discussion
2.1. Synthesis of C1-bridged ligand precursors of the
type C13H9–CR1R2–C5H4R (R1, R2=alkyl, phenyl,
alkenyl; R=H, alkyl, alkenyl, substituted silyl)
The synthesis of C1-bridged ligand precursors was
performed with the ‘fulvene method’ [13,14]. According
to this method, the lithium salts of fluorene, substituted
fluorenes, indene or substituted indenes are reacted with
a fulvene or a substituted fulvene. The anion adds
nucleophilic to the exocyclic double bond of the fulvene
to yield anion A. After hydrolysis, the corresponding
neutral ligand precursor B is obtained. This method has
* Corresponding author. Fax: +49 921 552157.
0022-328X/98/$19.00 © 1998 Elsevier Science S.A. All rights reserved.
PII S0022-328X(98)00740-2

88
H.G. Alt, M. Jung / Journal of Organometallic Chemistry 568 (1998) 87–112
Scheme 1. General synthesis of the ligand precursors of the type C13H9–CR1R2–C5H4R (R1, R2=alkyl, phenyl; R=H, alkyl, alkenyl, substituted
silyl).
unlimited application, if the acidity of the cyclopentadi-
ene derivative and the fulvene are higher than that of
the ansa-metallocene dichloride complex (Scheme 2)
[17,19,20].
the anion used (pKa
15).
ca. 22, pKa
ca. 19, pKa ca.
Starting with the appropriate ligand precursors, the
ansa-metallocene dichloride complexes listed in Fig. 2
and Table 2 were synthesized. The complexes indicated
with a star (*) are diastereomers.
FluH
IndH
CpH
Reactive halocarbons and trialkylchlorosilanes can
be reacted with the cyclopentadienyl anion A. This high
yield one-pot synthesis route allows the substitution of
position 3 of the cyclopentadienyl ring to yield the
neutral ligand precursor C (Scheme 1).
The substituted C1-bridged ligand precursors 1/1*–
13/13* were obtained by reacting fluorenyllithium with
6,6-dimethylfulvene, 3-(3-butenyl)-6,6-dimethylfulvene,
6-(3-butenyl)-6-methylfulvene, 6,6-tetramethylene ful-
vene, 6,6-pentamethylene fulvene, 6,6-hexamethylene
fulvene, or 6,6-heptamethylene fulvene and reacting this
product with water, methyliodide, trimethylchlorosilane
or allyldimethylchlorosilane (Fig. 1).
The large rings of 6,6-undecamethylene fulvene and
6,6-tetradecamethylene fulvene prevent the reaction of
these compounds with fluorenyllithium. Even the ad-
dition of hexamethylphosphoric triamide (HMPA) to
the reaction mixtures did not give the desired products.
The sterical demand of these large rings completely
inhibits the nucleophilic addition of the fluorenyl anion
(Table 1).
We have recently demonstrated [22] that the
fluorenylidene ligand in C1-bridged metallocene com-
plexes is p3 bonded in solution. The 13C-NMR chemical
shift of the quaternary carbon 9 of the fluorenylidene
ligand is indicative of the bonding mode. The 13C-
NMR shifts for these carbon atoms in the C1-bridged
metallocene complexes reported here are around l=78
ppm while the corresponding shifts of C1-bridged p5-
bonded fluorenylidene complexes are found between
1
l=105–115 ppm [18]. In the H-NMR spectrum of 23
(Fig. 3a), the individual signal groups are separated.
The trimethylsilyl group shows a signal at l=0.10
ppm. The signals for the 12 protons of the six
methylene groups of the cycloheptylidene ring are
found at l=3.41 (2H), 2.36 (2H), 2.04 (4H) and 1.77
ppm (4H).
The three protons of the cyclopentadienylidene li-
gand (l=6.34, 5.98 and 5.67 ppm) appear as virtual
triplets. The 1H-NMR signals for the fluorenylidene
hydrogen atoms are found in the region of l=7.20–
8.10 ppm. The distance of the two non-equivalent
benzo rings of the fluorenylidene ligand from the center
of asymmetry of the molecule results in similar chemi-
cal shifts. Therefore, four signal groups are found, and
two of these signal groups overlap (Fig. 3a, Fig. 3b).
The J-modulated 13C-NMR spectrum of 23 (Fig. 3b)
displays the expected signals for the nine quaternary, 11
tertiary and the six methylene carbon atoms, as well as
the signal for the trimethylsilyl group (l= −1.0 ppm).
Due to the asymmetry of the complex, a separate signal
2.2. Synthesis of the C1-bridged ansa-metallocene
dichloride complexes (C13H8–CR1R2–C5H3R)ZrCl2
(R1, R2=alkyl, phenyl, alkenyl; R=H, alkyl, alkenyl,
substituted silyl) (14–26/26*)
Bridged metallocene dichloride complexes were syn-
thesized according to known procedures [15,16]. The
selected ligand precursor is reacted with two equivalents
n-butyllithium in diethyl ether to form the dianion. In a
second step, the dianion is reacted with ZrCl4 to form

H.G. Alt, M. Jung / Journal of Organometallic Chemistry 568 (1998) 87–112
89
Table 1
NMR data of the C1-bridged ligand precursors 1/1*–13/13*
.

90
H.G. Alt, M. Jung / Journal of Organometallic Chemistry 568 (1998) 87–112
Table 1 (Continued)
.

H.G. Alt, M. Jung / Journal of Organometallic Chemistry 568 (1998) 87–112
91
Table 1 (Continued)
a)
CDCl3 (saturated solution) at 25°C; b) indicated as a shift range due to signal overlap; c) further signals could not be recorded. d) l
(
29Si)=2.7ppm; e) l (29Si)=5.5, 2.7 ppm; f) l (29Si)=3.2 ppm; g) l (29Si)=3.6 ppm; h) l (29Si)=4.2, 1.1 ppm; i) l (29Si)=3.8, 3.7 ppm.
appears for every carbon atom. The C and CH2 carbon
123.7 ppm correspond to the tertiary carbon atoms,
and the resonances at l=123.5, 123.5, 122.7, 121.7 and
79.5 ppm to the quaternary carbon atoms of the
fluorenylidene ligand. The signal at l=79.5 ppm for
the quaternary carbon atom 9 is typical and indicative
for all C1-bridged trihapto bonded fluorenylidene
metallocene complexes. The signals at l=123.4, 107.0
and 106.8 ppm and at l=128.9 and 117.7 ppm derive
from the tertiary and quaternary carbon atoms of the
cyclopentadienylidene ligand. The signals for the
methylene groups of the cycloheptylidene bridge appear
at l=39.1, 38.8, 27.5, 27.4, 23.9 and 23.5 ppm.
atoms (negative signals) are in anti phase to the CH
and CH3 carbon atoms (positive signals). The signals at
l=129.0, 128.7, 126.6, 125.0, 124.8, 124.7, 124.5 and
2.3. Polymerization of ethylene
All synthesized metallocene complexes reported here
polymerize ethylene after activation with methylalumi-
noxane (MAO). The metallocene catalyst precursors 15,
16 and 26/26*, which contain an ꢀ-alkenyl substituent,
are able to copolymerize themselves during the poly-
merization of h-olefins. In this way, the homogeneous
metallocene catalysts are self-immobilizing and catalyze
the further polymerization heterogeneously [18]; the
formed polymer serves as organic support.
In Fig. 4 and Fig. 5, the intrinsic viscosity molecular
weights of the polyethylenes that were produced with
the C1-bridged fluorenylidene cyclopentadienylidene
complexes 14–27 are compared. Until now, these cata-
lysts have primarily been tested for propylene polymer-
ization [5–8,11,21,23]. In Fig. 4, the molecular weights
of the polyethylenes that were produced with complexes
17–20 and 24 are given (unsubstituted cyclopentadi-
enylidene ligands), and in Fig. 5, those that were pro-
duced with complexes 14–16 and 21–27 (substituent in
position 3 of the cyclopentadienylidene ligand). Com-
plexes 17–20 each contain a ring system of different
size in the bridge. The size of the ring has an influence
on the molecular weight. The polyethylene prepared
with the cyclohexylidene bridged complex 18 has the
lowest molecular weight (Mp=140×103 g mol−1). The
complexes with smaller or larger bridging rings produce
Fig. 1. Overview of the synthesized ligand precursors 1/1*–13/13*.

92
H.G. Alt, M. Jung / Journal of Organometallic Chemistry 568 (1998) 87–112
higher molecular weights. The methyl phenyl methylene
bridged complex 24 performs the highest molecular
weight (Mp=280 103 g mol−1 [21]) after its activation
with MAO (Fig. 4).
The melting points of polyethylenes made with com-
plexes 17–20 and 24 are in the range of 13691.5°C
(Table 3). The corresponding enthalpy values of 80–
115 J g−1 indicate crystallinity degrees of 28–40%,
which are unusually low values for polyethylenes. The
polymer obtained with 18 has the lowest molecular
weight and the highest enthalpy value.
Fig.
5 displays the molecular weights of the
polyethylenes synthesized with the fluorenylidene cy-
clopentadienylidene complexes 14–16 and 21–27, all
substituted in position 3 of the cyclopentadienylidene
moiety (continuous line). The molecular weights range
from 125×103–175×103 g mol−1 except for complex
27 which produces polyethylene with an inherent
molecular weight of Mp=330×103 g mol−1. Complex
27 is a metallacycle derived from complex 15 and it
produces a molecular weight that has doubled com-
pared with that of 15. The cyclohexylidene bridged
complex 21 produces the lowest molecular weight
(Mp=125×103 g mol−1). For comparison, the molec-
ular weights of the polyethylenes that were obtained
with the corresponding unsubstituted complexes 17–20
and 24, are also shown (dotted line) in Fig. 5. The
unsubstituted complexes produce higher molecular
weights than their substituted homologues. Obviously a
substituent in position 3 of the cyclopentadienylidene
ligand favors termination over propagation.
The melting points of the polyethylenes made with
complexes 14–16 and 21–27 are in the range of 134.0–
138.0°C (Table 3). Most of these polymers have melting
points around 13791°C. The corresponding fusion
enthalpies range from 140 to 160 J g−1 (ca. crystallinity
degrees of 48–55%) and are significantly higher than
those for the unsubstituted complexes. Substitution of
the cyclopentadienylidene ligand increases the crys-
tallinity of the polyethylene by an average of 18%.
Whether the substituent is a trimethylsilyl group, a
methyl group or a butenyl group is insignificant.
Fig. 2. Overview of the synthesized metallocene complexes 14–27.
Scheme 2. Synthesis of the metallocene complexes.

H.G. Alt, M. Jung / Journal of Organometallic Chemistry 568 (1998) 87–112
93
Fig. 3. (a) 250.13 MHz 1H-NMR spectrum of 23 (CDCl3, 25°C). (b) 62.9 MHz J-modulated 13C{1H}-NMR spectrum of 23 (CDCl3, 25°C).
Quarternary and CH2 carbon atoms (negative signals) as well as CH and CH3 carbon atoms (positive signals) are in phase; S=CDCl3.

94
H.G. Alt, M. Jung / Journal of Organometallic Chemistry 568 (1998) 87–112
Table 2
NMR data of the C1-bridged metallocene complexes 14–27
.

H.G. Alt, M. Jung / Journal of Organometallic Chemistry 568 (1998) 87–112
95
Table 2 (Continued)
.

96
H.G. Alt, M. Jung / Journal of Organometallic Chemistry 568 (1998) 87–112
Table 2 (Continued)
a)
In CDCl3 (saturated solution) at 25°C; b) indicated as a shift range because of resonance overlap; c) 13C-NMR signals not completely separated
because of overlap; d) l (29Si) (16)=2.8; l (29Si) (21)=−5.5, l (29Si) (22)=−5.5, l (29Si) (23)=−5.5; l (29Si) (25/25*)=−5.3; l (29Si)
(26/26*)=−5.4 ppm; e) complex less soluble in chloroform. not all resonances observed; f) decomposition during measurement; g) the complete
spectroscopic characterization of the isomers was not carried out; h) no further Cq found.
2.4. Polymerization of propylene
The molecular weights of the polypropylene samples
were also measured using gel permeation chromatogra-
phy (GPC). The GPC equipment available could only
be operated with chloroform as a solvent at room
temperature (r.t.). Therefore, large portions of the re-
spective samples remained insoluble and were not avail-
able for measurements. The polymer samples made
with complexes 14, 15, 23, 25/25* and 27 were suffi-
ciently soluble to perform the GPC experiment (Table
3). The remaining samples were too insoluble for
measurements.
2.4.1. Molecular weights of the produced polypropylenes
All C1-bridged fluorenylidene cyclopentadienylidene
complexes were activated with MAO and successfully
tested for propylene polymerization. The molecular
weights of the polypropylenes are dependent on the size
of the bridging cycloalkylidene ring in the catalyst
molecule (Fig. 6). The molecular weights of the
polypropylenes synthesized with complexes 17–20 vary
from 54000 (18) to 110000 g mol−1 (19), and the resin
made with complex 24 reaches 160000 g mol−1
.
Substituents in position 3 of the cyclopentadienyli-
dene moiety have little effect on the molecular weights
(Fig. 7) of the produced polypropylenes (48000 (25/25*,
26/26*, 14, 16)–63000 g mol−1 (21, 27)). For compari-
son, the molecular weights of polypropylenes made
with the homologous unsubstituted complexes are in-
cluded in Fig. 7 (dotted line).
2.4.2. Configuration of the polypropylene chains
The polypropylene samples were also studied by
NMR spectroscopy to determine their stereospecifity.
The pentad distribution of the methyl groups in the
13C-NMR spectrum gives information about the struc-
ture of the polymers [24–27]. The methyl groups can be
arranged in two ways in the polymer: two neighboring
Fig. 4. Mean intrinsic viscosity Mp of the polyethylenes synthesized with the complex type (C13H8–CR1R2–C5H4)ZrCl2/MAO.

H.G. Alt, M. Jung / Journal of Organometallic Chemistry 568 (1998) 87–112
97

98
H.G. Alt, M. Jung / Journal of Organometallic Chemistry 568 (1998) 87–112
Fig. 6. Mp of the polypropylenes synthesized with the complex type (C13H8–CR1R2–C5H4)ZrCl2/MAO.
methyl groups of the same configuration (meso (m) or
isotactic), two neighboring methyl groups of the oppo-
site configuration (racemic (r) or syndiotactic). One
pentad consists of four diades. There are ten possible
combinations of diades in one pentad. The total r and
m portion of a polypropylene sample was calculated
according to: ꢀ r [%]=12 mr+rr; mr=mmrm+
mmrr+rmrm+rmrr; rr=mrrm+mrrr+rrrr. ꢀ m
[%]=12 mr+mm; mr=mmrm+mmrr+rmrm+rmrr;
mm=mmmm+mmmr+rmmr. The 13C-NMR spec-
trum of polypropylene produced with 19 is illustrated
in Fig. 8 as an example. The unsubstituted cyclopenta-
dienylidene complexes produce polypropylenes which
are almost exclusively racemic, e.g. syndiotactic
polypropylenes [1,13,15]. The remaining signals derive
from mistakes in the polymer structure.
(17.1%) pentads and rrrr pentads (19.5%); this is the
same behavior as seen for the corresponding
polypropylene produced with complex 27. Small iso-
and syndiotactic alternating blocks are found in this
polypropylene (stereoblock polymer). This polymer is
almost identical in structure to that produced with
complex 14. The 13C-NMR spectrum of the polypropy-
lene synthesized with complex 15 is shown in Fig. 10.
A tert-butyl group in position 3 of the cyclopentadi-
enylidene ligand leads to highly isotactic polypropylene
[11,12]. The tert-butyl group sterically blocks one side
of the complex. The growing polymer chain cannot
move to the side of the metal atom where the tert-butyl
group is located. This fixes the chain and the propylene
insertion occurs from one side (chain stationary inser-
tion). As a consequence polypropylene with an isotactic
configuration is formed.
If a hydrogen in position 3 of the cyclopentadienyli-
dene ligand is exchanged for a methyl group (14), this
catalyst produces a block copolymer with small iso-
and syndiotactic polypropylene units [10,11]. In this
hemiisotactic polypropylene [11,23,28–30] every other
methyl substituted carbon atom has the same configu-
ration and all other methyl substituted carbon atoms
have a statistic configuration. The methyl group of the
cyclopentadienylidene ligand can influence the coordi-
nation sphere of the metal and, therefore, it has a
significant impact on the structure of the polymer [11].
The polypropylene produced with 14 has an mmmm
pentad content of 9.7% and an rrrr pentad content of
41.0%. The total r diade content is 65.6%, and the total
m diade content is 34.4%. Thus the highly syndiotactic
nature of polymers made with the unsubstituted species
is lost by a methyl substituent at the cyclopentadi-
enylidine ring. Fig. 9 contains the 13C-NMR spectrum
of polypropylene made with 14 (compare to [10,11]).
The analysis of the polypropylene obtained with
complex 15 shows that it is similarly high in mmmm
The isotactic content decreases to approximately 80–
85% [11] if a trimethylsilyl or an allyldimethylsilyl
group is located in position 3 instead of a tert-butyl
group. In this case, 15–20% of the polymer consists of
syndiotactic blocks. Fig. 11 shows the 13C-NMR spec-
trum of the polypropylene synthesized with complex 16.
Fig. 12 summarizes the results of the pentad analyses
for all the polypropylenes synthesized. The exact pen-
tad distribution is listed in Table 6.
Fig. 13 shows the amount of racemic and isotactic
diades found in polymers made from complexes with
selected structures. The tacticities of the various
polypropylenes are compared. It is obvious that the size
of the h group of the substituent in position 3 of the
cyclopentadienylidene ligand is decisive in determining
the resulting tacticity of the polypropylene chain.
An allyldimethylsilyl group on the cyclopentadienyli-
dene ligand (16) has the same effect on tacticity as a
trimethylsilyl group [11], and a butenyl group (15) has
the same effect as a methyl group (14). Therefore, one

H.G. Alt, M. Jung / Journal of Organometallic Chemistry 568 (1998) 87–112
99

100
H.G. Alt, M. Jung / Journal of Organometallic Chemistry 568 (1998) 87–112
Table 3
Overview of the polymerization experiments with ethylene and polymer analysis
.

H.G. Alt, M. Jung / Journal of Organometallic Chemistry 568 (1998) 87–112
101
Table 3 (Continued)
a)
[Zr]:[Al]=1:17000; b) Ti,max=maximum inside temperature of the reactor, during polymerisation; c) the values of the second heating course of
the DSC were indicated as fusion enthalpies DHm; d) see [31]; e) the maximum of the melting peak of the second heating course of the DSC was
selected as melting point; f) ability of the catalyst to copolymerize: ++=very good, +=good, 0=sufficient,– =none; g) Polymerization time 15
min; h) Polymerization time 80 min; i) Polymerization time 30 min; j) Polymerization time 25 min; k) Polymerization time 10 min.

102
H.G. Alt, M. Jung / Journal of Organometallic Chemistry 568 (1998) 87–112
Fig. 8. 67.8 MHz 13C{1H}-NMR spectrum of polypropylene synthesized with complex 19/MAO; in C2D2Cl4/C6H3Cl3 (1/2; v/v) at 120°C (55306
scans; total measuring time 12.5 h.
Fig. 9. 67.8 MHz 13C{1H}-NMR spectrum of polypropylene synthesized with complex 14/MAO; in C2D2Cl4/C6H3Cl3 (1/2; v/v) at 120°C (43172
scans; total measuring time 9 h 50 min.

H.G. Alt, M. Jung / Journal of Organometallic Chemistry 568 (1998) 87–112
103
Fig. 10. 67.8 MHz 13C{1H}-NMR spectrum of polypropylene synthesized with complex 15/MAO; in C2D2Cl4/C6H3Cl3 (1/2; v/v) at 120°C (100000
scans; total measuring time 22 h 45 min.
can control the tacticity of the polypropylene chains
made with C1-bridged fluorenylidene cyclopentadienyli-
dene complexes by the choice of the substituent in
position 3 of the cyclopentadienylidene ring. At the same
time the metallocene catalyst becomes self-immobilizing
because of the m-alkenyl substituent (Tables 3–6).
enes with syndiotactic blocks. The fusion enthalpies of
these polymers vary over a wide range (14.6–37.5 J g−1
and approach the high enthalpies found for syndiotactic
polypropylenes (Fig. 14).
)
2.4.4. Comparison of the polymerization acti6ities
Unsubstituted C1-bridged fluorenylidene cyclopenta-
dienylidene complexes 17–20 and 24 have higher poly-
merization activities for propylene than for ethylene (14).
This is unusual and cannot be explained satisfactorily
except when electronic parameters determine the kinetics
of the polymerization.
The trimethylsilyl substituted complexes 21–23, 25/
25*, 26/26* and 16 exhibit considerably higher polymer-
ization activities for ethylene than for propylene. When
the substituent is a sterically smaller group like alkenyl
(15), alkyl (27) or methyl (14), the activities for ethylene
and propylene are both low. Smaller groups have no
influence on the polymerization activity of ethylene,
larger groups do. Unfortunately, the most active propy-
lene catalysts tended to begin polymerizing in the burette
as soon as they were charged to the reactor filled with
propylene. This sometimes blocked some of the catalyst
from reaching the reactor and resulted in low activity
2.4.3. Fusion enthalpies of the polypropylenes
The fusion enthalpies, DHm, of the polypropylenes
vary from 6.4 to 37.5 J g−1 (see Table 6). The polypropy-
lene produced with complex 14 has a fusion enthalpy of
only 6.4 J g−1. This indicates a highly disordered and
amorphous structure. Accordingly, the polymer has a
high glass transition enthalpy of 0.40 J g−1. The DSC
fusion enthalpies were not obtained for polymer samples
produced with complexes 15 and 27. The glass transition
enthalpies of 0.42 (15) and 0.46 J g−1 (27) indicate a
fusion enthalpy as low as that for the polypropylene
made with complex 14. The syndiotactic polypropylenes
made with complexes 17–20 and 24 have fusion en-
thalpies of 20.8–29.0 J g−1 and low glass transition
enthalpies of 0.06–0.21 J g−1. The trimethylsilyl substi-
tuted complexes 17–19, 25/25*, 26/26* and 16, with an
allyldimethylsilyl substituent, yield isotactic polypropyl-

104
H.G. Alt, M. Jung / Journal of Organometallic Chemistry 568 (1998) 87–112
Fig. 11. 67.8 MHz 13C{1H}-NMR spectrum of polypropylene synthesized with complex 16/MAO; in C2D2Cl4/C6H3Cl3 (1/2; v/v) at 120°C (40176
scans; total measuring time 9 h 10 min.
numbers. Therefore, it is possible that some of the
measured activities are too low.
combination with a Varian MAT 312 mass spectro-
meter.
3.3. Gas chromatography
3. Experimental
A Carlo Erba HRGC gas chromatograph with flame
ionization detector was used to analyze organic com-
pounds. The gas chromatograph was equipped with a
30 m long J&W fused silica column (DB1, film thick-
ness 0.25 mm). Helium served as carrier gas; the flow
through the column was 3.8 ml min−1, split 1:30,
septum flushing 1.3 ml min−1. The following tempera-
ture program was routinely used: 3 min at 50°C (start-
ing phase), 5 K min−1 (heating phase), 15 min at 310°C
(plateau phase). The retention time was indicated in
seconds.
3.1. NMR spectroscopy
The instruments Jeol JNM-EX 270 E, Bruker ARX
250 and Bruker DRX 500 were available to record the
NMR spectra. The samples were filled under argon and
routinely measured in CDCl3 at 25°C. The chemical
1
shifts in the H-NMR spectra are referred to the resid-
ual proton signal of the solvent (l=7.24 for CHCl3), in
the 13C-NMR spectra to the solvent signal (l=77.0 for
CDCl3) and in the 29Si-NMR spectra to the resonance
of external TMS (l=0.0) [32].
3.4. General synthesis procedure for the C1-bridged
ligand precursors of the type C13H9–CR1R2–C5H5
(R1, R2=Me, cycloalkylidene, phenyl) (1/1*–5/15*)
3.2. Mass spectroscopy
Routine measurements were conducted with
a
VARIAN MAT CH7 instrument (direct inlet system,
electron impact ionization 70 eV). GC/MS spectra were
recorded using a Varian 3700 gas chromatograph in
A total of 6.0 g (36 mmol) fluorene was dissolved in
100 ml diethyl ether and slowly mixed with 22.5 ml (36
mmol) n-butyllithium (1.6 M in hexane) at r.t. The

H.G. Alt, M. Jung / Journal of Organometallic Chemistry 568 (1998) 87–112
105

106
H.G. Alt, M. Jung / Journal of Organometallic Chemistry 568 (1998) 87–112
Fig. 13. Comparison of the tacticities of the polypropylenes synthesized with substituted fluorenylidene cyclopentadienylidene complexes/MAO.
reaction mixture was stirred for at least 6 h. An
equimolar amount of the corresponding fulvene deriva-
tive was added and stirred overnight. Then a 0.1 equiv-
alent butyllithium was added to convert the excess
fulvene into a more soluble compound. The mix-
ture was stirred for another 30 min and hydro-
lyzed with 50 ml water. The organic phase was dried
over sodium sulfate, and the solvent evaporated in
vacuo. For purification, the residue was dissolved in
pentane, the solution was filtered over silica and the
compound was crystallized at −18°C. The yields were
70–90%.
1-(9-Fluorenyl)-1-[1-(1,3-cyclopentadienyl)]cyclopen-
tane and isomer (1/1*): GC: 2436 s. MS: m/e 298 (M+).
1-(9-Fluorenyl)-1-[1-(1,3-cyclopentadienyl)]cyclohexane
and isomer (2/2*): GC: 2540 s, 2555 s. MS: m/e 312
(M+). 1-(9-Fluorenyl)-1-[1-(1,3-cyclopentadienyl)]cyclo-
heptane and isomer (3/3*): GC: 2804 s, 2811 s. MS:
Table 4
Pentad analysis of the polypropylenes produced with the catalysts 14–27/MAO
Isotactic portion
mr
Syndiotactic portion
Sra,b
Sma,b
mmmm
(%)
mmmr
(%)
rmmr
(%)
mmrr
(%)
mmrm+
rmrr (%)
rmrm
(%)
rrrr
(%)
mrrr
(%)
mrrm
(%)
l
13C
21.83
9.7
21.59
9.5
21.37
4.2
7.8
21.05
20.8
26.0
7.2
1.5
2.2
2.4
2.3
8.5
4.7
20.85
1.1
20.60
20.31
41.0
19.5
10.1
92.4d
97.8
93.2
97.7
8.9
20.10
9.8
12.8
19.95
3.8
5.3
14
15
16
17
18
19
20
21
22
23
24
25/25*
26/26*
27
65.6
50.6
16.3
98.2
98.9
95.1
98.8
15.6
19.5
9.0
34.4
49.4
83.7
1.8
1.1
4.3
17.1
11.4
80.1c
80.1c
2.6
0.8
1.0
0.5
1.4
4.9
1.4
1.2
72.1
75.9
78.6
8.0
7.0
2.4
3.5
84.4
80.6
91.0
4.7
2.4
7.4
7.4
10.9
99.0
9.3
5.8
99.0
22.9
12.7
51.9
52.4
74.4
17.8
13.4
8.3
7.8
2.0
3.0
16.3
9.3
39.4
2.1
9.0
2.2
2.2
3.0
77.1
87.3
48.3
20.3
a Sr (%)=0.5mr+rr; mr=mmrm+mmrr+rmrm+rmrr; rr=mrrm+mrrr+rrrr; Sm (%)=0.5mr+mm; mm=mmmm+mmmr+rmmr.
b 13C-NMR data of the polypropylenes: spectra recorded in 1,2,4-trichlorobenzene/1,1,2,2-tetrachlorodeuteroethane (3:1 v/v) at 120°C.
c No signal resolution.
d Error91.5%.

H.G. Alt, M. Jung / Journal of Organometallic Chemistry 568 (1998) 87–112
107
Table 5
Molecular weight determinations for the polypropylene samples produced with 14, 15, 23, 25/25*, 27/MAO using gel permeation chromatography
a
b
c
d
Mn (g mol−1
)
Mw (g mol−1
)
Mz (g mol−1
)
Dispersity (Mw/Mh)
Mh (g mol−1
)
(
(
(
(
(
(
Complex
14
15
23
25/25*
27
20 382
27 654
25 595
13 716
31 333
44 403
68 507
64 353
41 479
73 106
71 643
124 984
119 057
74 187
2.18
2.47
2.51
3.02
2.33
48 000
58 000
54 000
48 000
63 000
132 120
a
b
c
Mn, number average molecular weight. Mw, weight average molecular weight. Mz, z-average molecular weight. d For comparison, the mean
(
(
(
(
intrinsic viscosity Mh is included.
m/e 327 (M+). 1-(9-Fluorenyl)-1-[1-(1,3-cyclopentadi-
enyl)]cyclooctane and isomer(4/4*): GC: 2878 s, 2891 s.
1-(9-Fluorenyl)-1-[1-(1,3-cyclopentadienyl)]-1-phenyleth-
ane and isomer (5/5*): GC: 2832 s, 2850 s. MS: m/e 334
(M+).
100 ml diethyl ether and slowly mixed with 22.5 ml (36
mmol) n-butyllithium (1.6 M in hexane) at r.t. The
reaction mixture was stirred for at least 6 h. An
equimolar amount of the corresponding fulvene deriva-
tive was added, and the mixture was stirred overnight.
At −78°C, an equimolar amount of methyl iodide,
allyldimethylsilyl chloride or trimethylsilyl chloride was
added to the reaction mixture, and the mixture was
stirred overnight. For processing, the mixture was hy-
drolyzed with 50 ml water, the organic phase was dried
over sodium sulfate, and the solvent evaporated in
vacuo. For purification, the residue was dissolved in
3.5. General synthesis procedure for C1-bridged ligand
precursors of the type C13H9–CR1R2–C5H4R; (R1,
R2, R=alkyl, alkenyl, phenyl, substituted silyl)
(6/6*–13/13*)
A total of 6.0 g (36 mmol) fluorene was dissolved in
Fig. 14. Polymerization activities for substituted metallocene complexes of the type (C13H8–CR1R2–C5H4)ZrCl2/MAO: ethylene compared with
propylene.

108
H.G. Alt, M. Jung / Journal of Organometallic Chemistry 568 (1998) 87–112
Table 6
Overview of the polymerization experiments with propylene and polymer analysis

H.G. Alt, M. Jung / Journal of Organometallic Chemistry 568 (1998) 87–112
109
Table 6 (Continued)
a)
[Zr]:[AI]=1:17000; b) Ti,max=maximum inside temperature of the reactor during polymerisation; c) DHm, describes the cristallization enthalpy
for the second cooling phase of the DSC: d) the minimum of the crystallization peak of the second cooling phase was selected as melting point;
e) the changing point of the glass transition of the DSC was selected as glass temperature; f) glass transition enthalpies; g)ꢀ r [%]=12 rr,
mr=mmrm+mmrr+rmrm+rmrr, rr=mrrm+mrrr+rrr, ꢀ m [%]=12 mr+mm, mm=mmmm+mmmr+rmmr; h) Melting temperature of the
second heating course due to missing crystallization peak.

110
H.G. Alt, M. Jung / Journal of Organometallic Chemistry 568 (1998) 87–112
(l-cyclopentadienylidene)cycloheptane zirconium dich-
loride (19): red crystals. MS: m/e 486 (M+). 1-p3-(9-
Fluorenylidene)-l-p5-(l-cyclopentadienylidene)cyclooct-
ane zirconium dichloride (20): orange crystals. MS: m/e
500 (M+). 1-p3-(9-Fluorenylidene)-1-p5-[1-(3-trimethyl-
silyl)cyclopentadienylidene]cyclopentane zirconium di-
chloride (21): red crystals. MS: m/e 530 (M+). 1-p3-(9-
Fluorenylidene)-1-p5-[1-(3-trimethylsilyl)cyclopentadi-
enylidene]cyclohexane zirconium dichloride (22): red
crystals. MS: m/e 544 (M+). 1-p3-(9-Fluorenylidene)-l-
p5 -[1-(3-trimethylsilyl)cyclopentadienylidene]cyclohep-
tane zirconium dichloride (23): red crystals. MS: m/e
559 (M+). 1-p3-(9-Fluorenylidene)-l-p5-(l-cyclopentadi-
enylidene)-l-phenylethane zirconium dichloride (24): red
crystals. MS: m/e 494 (M+). 1-p3-(9-Fluorenylidene)-l-
p5-[1-(3-trimethylsilyl)cyclopentadienylidene]-l-pheny-
lethane zirconium dichloride and isomer (25/25*): red
crystals. MS: m/e 566 (M+). 5-p3-(9-Fluoorenylidene)-
5-p5-[1-(3-trimethylsilyl)cyclopentadienylidene]-1-hex-
ene zirconium dichloride and isomer (26/26*): red crys-
tals. MS: m/e 544 (M+).
pentane, the solution was filtered over silica, and the
compound crystallized at −18°C. The yields were 70–
90%.
2-(9-Fluorenyl)-2-[1-(3-methyl)-1,3-cyclopentadienyl]-
propane and isomer (6/6*): GC: 2273 s. MS: m/e 286
(M+). 2-(9-Fluorenyl)-2-{1-[3-(3-butenyl)]-1,3-cyclopen-
tadienyl}propane and isomer (7/7*): GC: 2665 s, 2680
s. 2-(9-Fluorenyl)-2-{1-[3-(allyl-dimethylsilyl)]-1,3-cyclo-
pentadienyl}propane and isomer (8/8*): GC: 2666 s.
MS: m/e 370 (M+). 1-(9-Fluorenyl)-l-[l-(3-trimethylsi-
lyl)-1,3-cyclopentadienyl]cyclopentane and isomer (9/
9*): GC: 2650 s. 1-(9-Fluorenyl)-1-[1-(3trimethy-
lsilyl)-1,3-cyclopentadienyl]cyclohexane and isomer (10/
10*): GC: 2766 s, 1-(9-Fluorenyl)-l-[l-(3-trimethylsilyl)-
l,3-cyclopentadienyl]cycloheptane and isomer (11/11*):
GC: 3119 s. MS: m/e 398(M+). 1-(9-Fluorenyl)-1-[1-
(3-trimethylsilyl)-1,3-cyclopentadienyl]-1-phenylethane
and isomer (12/12*): GC: 3360 s. MS: m/e 406 (M+).
5-(9-Fluorenyl)-5-[1-(3-trimethylsilyl)-1,3-cyclopenta-
dienyl]-1-hexene and isomer (13/13*): GC: 2824 s, MS:
m/e 384 (M+).
3.6. General synthesis procedure for the bridged
metallocene complexes 14–26/26*
3.7. Synthesis procedure for metallocene complex 27
A total of 3.0 mmol of the metallocene dichloride
complex 15 and 0.79 g (3.11 mmol) lithium aluminum-
tri-tert-butoxy hydride was dissolved in 50 ml THF,
and the mixture was stirred at least overnight at r.t.
The solvent was evaporated in vacuo, and the resi-
due was extracted with toluene. Complex 27 forms
orange crystals at −25°C when hexane is added. Yield:
40%.
A total of 1.0 g of the corresponding ligand precursor
was dissolved in 40 ml diethyl ether and stirred with
exactly two equivalents of n-butyllithium (1.6 M
in hexane) for at least 8 h at r.t. Then one equiva-
lent zirconium tetrachloride or hafnium tetrachlo-
ride was added, and the mixture stirred over-
night. Further processing was conducted based on the
solubility of the product: for ether soluble com-
plexes, the mixture was filtered directly from the lit-
hium chloride by-product; for the less soluble com-
plexes, either the solvent was evaporated and the
residue extracted with methylene chloride, or the com-
plex solution was filtered over sodium sulfate, and the
product extracted with methylene chloride or toluene
from the frit. The solvent was then evaporated in
vacuo.
3.8. Self-immobilization
Investigations on self-immobilization were conducted
in Schlenk tubes. Approximately 10 mg of the corres-
ponding metallocene complex were activated with 10 ml
MAO (30% solution in toluene), diluted with 40 ml
toluene and then exposed to an ethylene pressure of
0.4–0.6 bar. The incorporation of the ꢀ-alkenyl substi-
tuted metallocene complexes into the polymer chains
was indicated by the characteristic color of the formed
polymer precipitate.
2-p3-(9-Fluorenylidene)-2-p5-[1-(3-methyl)cyclopenta-
dienylidene]propane zirconium dichloride (14): red crys-
tals. MS: m/e 446 (M+). 2-p3-(9-Fluorenylidene)-2-p5-
{1 - [3 - (3 - butenyl)lcyclopentadienylidene}propane zir-
corium dichloride (15): red crystals. 2-p3-(9-Fluorenyli-
dene)-2-p5-{1-[3-(allyldimethylsilyl)]cyclopentadienyli-
dene}propane zirconium dichloride (16): red crystals,
elemental analysis, found: C 50.97, H 4.97. C26
H28SiZrCl2 ·1.3 CH2Cl2 calc.: C 51.14, H 4.81. MS: m/e
530 (M+). 1-p3-(9-Fluorenylidene)-l-p5-(l-cyclopentadi-
enylidene)cyclopentane zirconium dichloride (17): red
crystals. MS: m/e 458 (M+). 1-p3-(9-Fluorenylidene)-1-
p5-(l-cyclopentadienylidene)cyclohexane zirconium di-
chloride (18): red crystals. 1-p3-(9-Fluorenylidene)-1-p5-
3.9. Polymerization experiments
3.9.1. Acti6ation of the catalyst precursors
The respective metallocene complex was weighed un-
der inert gas (ca. 8–1290.1 mg) and activated with
MAO (1 ml MAO (30% in toluene) per mg metallocene
dichloride complex. The solution was diluted with
toluene in a way that ca. 0.2–0.5 mg of the metallocene
complex were dissolved in 1 ml of toluene. From this

H.G. Alt, M. Jung / Journal of Organometallic Chemistry 568 (1998) 87–112
111
To determine the crystallinity degree, h, the relation-
solution, an aliquot containing ca. 1 mg catalyst was
removed and used for polymerization. These solutions
were used for polymerization within 60 min after
preparation.
ship h=DHm/DH° was selected. DHm is derived from
m
the data of the second heating course of the DSC. The
fusion enthalpy for 100% crystalline polyethylene was
assumed as 290 J g−1 [31].
3.9.2. Polymerization of ethylene
3.10.2. Viscosimetry
A 1 l Bu¨chi laboratory autoclave BEP 280 was filled
with 500 ml pentane, 7 ml MAO (30% in toluene) and
the corresponding amount of catalyst solution ([Zr]:[Al]
1:17000). The reactor was thermostated at 60°C, and a
constant ethylene pressure of 10 bar was applied. The
polymerization was terminated after 1 h by venting the
ethylene.
The intrinsic viscosity was determined using an Ub-
belohde precision capillary viscometer in cis/trans de-
calin at 13590.1°C. Prior to the measurements, the
samples were weighed into sealable small flasks and
dissolved in an exactly measured amount of decalin at
140–150°C over a period of 3–4 h. Calibration curves
were available for the determination of Mp. Every
polymer sample was weighed and measured twice to
reduce the error.
3.9.3. Polymerization of propylene
A total of 500 ml propylene (‘polymerization grade’)
was condensed into the reactor, stirred for 20 min with
5 ml MAO (30% in toluene) at 20°C and then cooled to
−5°C. The prepared catalyst solution (toluene/
MAO=10:1) was pressed into the cooled stirring vessel
with 6.5 bar argon pressure from a burette. The reactor
temperature was increased to 60°C within 15 min. Then
the polymerization was conducted at ca. 25 bar propy-
lene pressure and terminated after 1 h by releasing the
unreacted propylene.
Acknowledgements
We thank Phillips Petroleum Company (Bartlesville,
OK, USA) and the Deutsche Forschungsgemeinschaft
for their financial support and Witco company for a
donation of methylaluminoxane.
References
3.10. Characterization of the polymer samples
[1] J.A. Ewen, R.L. Jones, A. Razavi, J. Am. Chem. Soc. 110 (1988)
6255.
[2] (a) H.-H. Brintzinger, D. Fischer, R. Mu¨lhaupt, B. Rieger, R.
Waymouth, Angew. Chem. 107 (1995) 1255. (b) H.-H.
Brintzinger, D. Fischer, R. Mu¨lhaupt, B. Rieger, R. Waymouth,
Angew. Chem. Int. Ed. Engl. 34 (1995) 1143.
[3] M. Aulbach, F. Ku¨ber, Chemie in Unserer Zeit 4 (1994) 197.
[4] W. Kaminsky, M. Arnd, Adv. Polym. Sci. 127 (1997) 143.
[5] H.G. Alt, R. Zenk, J. Organomet. Chem. 514 (1996) 257.
[6] H.G. Alt, R. Zenk, J. Organomet. Chem. 518 (1996) 7.
[7] H.G. Alt, R. Zenk, J. Organomet. Chem. 522 (1996) 33.
[8] K. Patsidis, H.G. Alt, W. Milius, S.J. Palackal, J. Organomet.
Chem. 509 (1996) 63.
[9] J.A. Ewen, M.J. Elder, R.L. Jones, et al., Macromol. Chem.
Macromol. Symp. 48/49 (1991) 253.
[10] A. Razavi, J.L. Atwood, J. Organomet. Chem. 497 (1995) 105.
[11] A. Razavi, L. Peters, L. Nafpliotis, et al., Macromol. Symp. 89
(1995) 345.
[12] A. Razavi, J.L. Atwood, J. Organomet. Chem. 520 (1996) 115.
[13] A. Razavi, J. Ferrara, J. Organomet. Chem. 435 (1992) 299.
[14] (a) A. Winter, J. Rohrmann, M. Antberg, V. Dolle, W. Spaleck,
Ger. Offen. (1991) GE 3, 907, 965. (b) A. Winter, J. Rohrmann,
M. Antberg, V. Dolle, W. Spaleck, Chem. Abstr. 114 (1991)
165103w.
3.10.1. Differential scanning calorimetry
A Perkin Elmer calorimeter DSC-7 was available to
measure the thermal properties of the polymer samples.
Prior to the measurements, the polymer samples were
dried in vacuo. To determine the fusion enthalpies of
the polyethylenes and polypropylenes, 3–5 mg of each
of the polymers were fused into standard aluminum
pans and measured using the following temperature
program: first heating phase (20 K min−1) from 50 to
200°C, cooling phase (−20 K min−1) to 50°C, second
heating phase (20 K min−1) from 50 to 200°C, secondly
the cooling phase (−20 K min−1) to 50°C. The tem-
perature was linearly corrected relative to indium (m.p.
156.6°C); the fusion enthalpy of indium (DHm=28.45 J
g
−1) was used for calibration. The maximum of the
melting temperatures during the second heating run
was selected for the polyethylenes; for the polypropyle-
nes, the minimum of the crystallization peak during the
second cooling phase.
To determine the glass transition enthalpies of the
polypropylene samples, 30–50 mg of the polymer were
fused into standard aluminum pans and measured using
the following temperature program: first heating phase
(20 K min−1) from −40 to 40°C, cooling phase (−20
K min−1) to −40°C, second heating phase (20 K
min−1) from −40 to 40°C, 2nd cooling phase to
−40°C.
[15] H.G. Alt, R. Zenk, J. Organomet. Chem. 514 (1996) 113.
[16] H.G. Alt, S.J. Palackal, J. Organomet. Chem. 472 (1994) 113.
[17] M. Antberg, V. Dolle, R. Klein, J. Rohrmann, W. Spaleck, A.
Winter, in: T. Keii, K. Soga (Eds.), Catalytic Olefin Polymeriza-
tion, Proceedings of the International Symposium on Recent
Developments in Olefin Polymerization Catalysts, Tokyo, 1989,
Kodansha (1990) 501.
[18] B. Peifer, W. Milius, H.G. Alt, J. Organomet. Chem. 553 (1998)
205.

112
H.G. Alt, M. Jung / Journal of Organometallic Chemistry 568 (1998) 87–112
[19] N. Inoue, T. Shiomura, T. Asanuma, M. Jinno, Y. Sonobe, K.
Mizutani, Jpn. Kokai Tokkyo Koho JP 03, 193, 797 [91, 193,
797]; N. Inoue, T. Shiomura, T. Asanuma, M. Jinno, Y. Sonobe,
K. Mizutani, Chem. Abstr. 116 (1992) P119765g.
(1980) 81.
[25] T. Hayashi, Annu. Rep. NMR Spectrosc. 29 (1992) 326.
[26] A. Zambelli, P. Locatelli, G. Bajo, F.A. Bovey, Macromolecules
8 (1975) 689.
[20] N. Inoue, T. Shiomura, T. Asanuma, M. Jinno, Y. Sonobe, K.
Mizatani, Jpn. Kokai Tokkyo Koho JP 03, 195, 706 [91, 195,
706]; N. Inoue, T. Shiomura, T. Asanuma, M. Jinno, Y. Sonobe,
K. Mizatani, Chem. Abstr. 117 (1992) P8703m.
[27] A. Zambelli, C. Pellecchia, L. Oliva, Macromol. Chem. Macro-
mol. Symp. 48/49 (1991) 297.
[28] M. Farina, G. Di Silvestro, P. Sozzani, B. Savare´, Macro-
molecules 18 (1985) 923.
[21] A. Winter, V. Dolle, W. Spaleck (Hoechst A.-G.), Eur. Pat.
Appl. (1992), 516019; A. Winter, V. Dolle, W. Spaleck (Hoechst
A.-G.) Chem. Abstr. 119 (1993) 28794m.
[22] H.G. Alt, M. Jung, J. Organomet. Chem. (in press).
[23] N. Herfert, Dissertation, University of Du¨sseldorf, 1992.
[24] I. Ando, T. Asakura, Annu. Rep. NMR Spectrosc. 10a
[29] A. De Marco, P. Sozzani, G. Di Silvestro, M. Farina, Macro-
molecules 22 (1989) 2154.
[30] M. Farina, G. Di Silvestro, P. Sozzani, Macromolecules 26
(1993) 946.
[31] G. Luft, M. Dorn, Angew. Macromol. Chem. 188 (1991) 177.
[32] J. Mason, Multinuclear NMR, Plenum Press, N.Y., 1987.
.
.