Molecular Behaviour of Soluble Catalysts for Olefin Polymerization Part I: Ethylene insertion with soluble Ziegler Catalysts

MOLECULAR BEHAVIOUR OF SOLUBLE CATALYSTS FOR OLEFIN POLYMERIZATION Part I: Ethylene insertion with soluble Ziegler Catalysts G. FINK, W. FENZL and R. MYNOTT

Max-Planck-Institut fiir Kohlenforschung, D-4330 Miilheim a. d. Ruhr, Federal Republik of Germany ABSTRACT

The development of the oligomer distribution during the polymerization of I3C enriched ethylene by soluble Ziegler catalyst systems of the type Cp2TiMeC1/A1MenCln was followed by I3C NNR spectroscopy. It is shown that the rate of formation of new chains can be monitored directly from the spectra. The concentrations of Ti-propyl and Tipentyl species during the polymerization were followed: both attain a steady state concentration. These results give a greater insight into the way that the oligomer distribution develops and into the dependence on the chain length of the first insertion steps. INTRODUCTION One suitable method for obtaining information on catalytically active systems without disturbing the reaction is I3C NMR spectroscopy. In an earlier publication’) we reported our studies on the polyrnerization of 3C-enriched ethylene using the soluble Cp2TiEtC1/ A1EtC12 catalyst in an NMR sample tube. Experiments using a batch reactor have shown that this Ti-Et system is a much more active polymerization catalyst than Ti-Me systems (see Fig. 1). The lowest curve (solid line) is for the Cp2TiWeC1/A1MeC12 catalyst system, with which the polymerization proceeds much mor slowly. This catalyst is ideally suitable for the I3C NMR investigations reported here and with the right choise of experimental parameters it has proved possible to obtain considerable detail of the reaction path of the insertion.

2 15

216

G. F i n k , W. Fenzi a n d R. M y n o t t

APPLICATION OF

3C-ENRICHED ETHYLENE I3

In these experiments we used ethylene to over 90 atom % C. Besides providing a considerable gain in sensitivity over ethylene with 1 3 C at natural abundance this allows the carbons in the polymer chain derived from l3C-enriched ethylene to be distinguished from those from the Ti-Me or A 1 - M e carbons on the basis of their signal intensities. Whereas 3C-13C spin-spin couplings are observed in natural abundance I3C NMR spectra only as very weak satellites, in enriched samples they may cause the signals to appear as multiplets. The proton-decoupled spectrum of enriched ethylene free in solution is a singlet because the two 13C nuclei are magnetically equivalent. Once incorporated into a chain, these carbons are no longer chemically equivalent and the coupling between neighbouring I3C nuclei is observed. Table 1 surnmarizes the multiplet structures expected for a Tialkyl chain formed by the repeated insertion of ethylene enriched to 91 atom % 13C at both carbons start.ing with a Ti-methyl at the natural isotopic abundance. For those molecules containing 3C in the O(-position there is a probability of 91 % that the P-carbon will also be I3C and the O(-carbon signal will therefore be split into a doublet. The remaining 9 % of the molecules have 12C/3-neighbours and show no coupling. The %-carbon signal is thus observed to be a 6ouSlet superimposed on a small signlet slightly shifted to the left of centre as a result of isotope shifts. The /jl-carbon has a probability of 82.8 % that both the ol- and &'-carbons are 13C anda triplet will be observed, while those isotopomers with only the O ( - or only the carbon 13C (combined probability 1 6 . 4 % ) will give a doublet.

$-

Inspection of the I3C chemical shifts of the n-alkanes shows that - and & -carbons have characteristic shift ranges, but that the q - , for acrbon atoms from the /-position to the chain centre the environments become too similar to produce any significant differences in their chemical shifts. These carbons in the middle of the chain have resonances at approximately 30 ppm. S?.milarly, for Ti-alkyls, the presence of the metal suhstituent on the Q-carbon will produce no significant effect beyond the g-carbon. Thus a signal at 30 ppm indicates that chains at least eight carbon atoms long are present: the larger its relative intensity the greater the fraction of even longer chains which are present. This peak is labelled in the spectra

/3

217

m

u

-

g

o! Q

5

3 )

N

3 .

I .(=

Q

0

h

0.c

w 4 J 0

aa a

c , Q

v m a,

P,u Xrn

.-

I-

9

0,

acn 3

3

>\

ul C

0, c

C

H d

0,

CK

0

ucI

v

c

do

ItlE

a

218

G. Fink, W. Fenzl and R. Mynott

with the symbol .**.”-

CH2-*.

*.

POLIMERIZATION I ”

13

Figure 2 illustrates the first of a series of C NYR spectra of the system reacting with ethylene. The upper spectrum was recorded at 213 K before the start of the reaction. The lower spectrum was measured after the sample had been kept at 258 K for 30 minutes and shows that the first insertion steps have occured as indicated by the small peaks of the p-carbons of Ti-propyl ar,d Ti-pentyl. The ethylene signal at 123 ppm remains a sharp singlet and is unshifted, demonstrating that even at the chosen ratio Ti : A1 : C2H4 = 1.0 : 0.95 : 0.7 no interaction with the catalyst is detectable.

/$,XI

In Figure 3, after a reaction ttme of 70 minutes, the o(, W - 2 a n d W - I signals of the different chains up to Ti-heptyl are to be discerned. In addition, the presence of a small peak at 30.7 ppm due to central CH2-groups indicates that chains with a least 9 carbon atoms are present. Figure 4 shows the situation after 1 1 0 minutes. The signals reveal the developing oligomer distribution and the size of the peak of the central CH2-groups has increased considerably. By the time that the reaction has been proceeding for approximately 3 hours (Fig. 5) this peak has become the major feature of the oligomer spectrum, showing that most chains are longer than Ti-nonyl. Let us now consider how this result could arise when the initial concentration ratio Ti : C2H4 was 1.0 : 0.7. If all the initially added Ti had been active then at the end of the reaction on average less than one ethylene per Ti-CH would have undergone insertion. In 3 this case we would find mostly Ti-propyl chains and possibly a small quantity of Ti-pentyl chains. This assumes that all the ethylene has been consumed, but as can be seen in Fig. 5 , there are much longer oligomer chains present even though there is a considerable amount of unreacted ethylene left. The observation that longer oligomer chains have been formed is very important since it proves that not all the Ti has been able to undergo insertion. A large amount of Ti-CH3 groups must therefore still be present. This is confirmed by the corresponding signal at 64 ppm, which represents a considerable concentration of Ti-methyl groups because the methyl group has natural I3C abundance. This leads again to the conclusion, that the primary complex formed between Ti-

Molecular Behaviour of Soluble Catalysts Part I

219

and Al-compound cannot be already the active species. Figure 6 gives a brief synopsis of the true reaction scheme which has been deduced from our previous experiments2 ) 4 ) 5 ) and which is confirmed by the experiments in this paper. The most important component of this reaction scheme for soluble Ziegler catalyst systems is the formation of the active species in two successive equilibrium steps. The first equilibrium lies well to the right, the second well to the left. Consequently at low ratios Al/Ti there is only a very small concentration of active species Cx . A further consequence is that the propagation process itself is then an intermittent process which causes the molecular weight distribution to undergo a particular type of development4)5). The I 3 C NMR spectrum recorded towards the end of the reaction (after 14.5 hours at 258 K ) is illustrated in Fig. 7 . In addition to the oligomer distribution and the strong "polyethylene" peak, the following features should be noted: (I) Comparing the intensities of all Ti-0(-peaks (i.e., Ti-%propyl, Ti-O(-(& pentyl)) at 90 pprn with the intensity of the Ti-methyl peak at 64 ppm and taking into account the I3C abundances of 91 % and 1.1 %, respectively, one obtains the true concentration ratio of about 0.1 : 1.0. This indicates that only 10 % of the initial Ti-methyl compound has been involved in insertion reactions. This is to be expected from the location of the successive equilibria at the chosen ratio of Ti/A1 = 1

:

1.

(11) The signal of the W-carbon at approximately 14 ppm is very weak since it is at the natural abundance of I3C. (111) The ( W - I ) peak for Ti-heptyl and longer chains is clearly a doublet, produced by coupling with only one adjacent I3C nucleus (see Table 1). Both (11) and (111) prove that the insertion has occurred into the Ti-carbon bond. (IV) After the long reaction time of 15.5 hours, small quantities of q-olefines are formed as indicated by the weak signals at 114 and 140 ppm. This means that a transfer reaction to the -elimination has occurred. monomer via H-

P

220

G. Fink, W. Feiizl and R. Mynott

The experiment described above has provided general information on the course of the polymerization reaction. We have carried out further investigations aimed at obtaining more details on the chain propagation itself. Figure 8 illustrates the series of spectra recorded for a sample of Cp2TiMeC1/A1Me2C1/13C2H4 in the ratio 1 : 2 : 2 , to increase the concentration of active species. The potential of this approach is apparent, These spectra, each measured over a period of an hourF were recorded succcssively in order to follow the development in chain growth. The regions depicted are for the Ti- M r e s o nances and for the signals for the

P

to&-1

positions.

First of all it is to notice in Figure 8 the rapid increase of the peak of the central CH2-groups. This is again in full agreement with our reaction scheme because now we have more active species. Furthermore it is to see in this presentation, that we were successful1 in better separating the peaks of the different chain carbon atoms. Hence, we discover that the propyl chains (as to see by the Ti-w - or Ti-p -propyl positions in the spectrum) and the pentyl chains (as to see by the Ti-( +& ) - or (&-l)-pentyl positions in the spectrum) remain constant during the reaction; i-e., here is visible a steady state of the concentration of these chains.

P

The integration+’ of the Ti-propyl, the Ti-pentyl peaks and the peaks!fo -heptyl and longer chains leads to the important result in Figure 9: after the starting phase of the chain development the concentrations of propyl chains and pentyl chains attain a steady state. The reason for i s the dependence on the chain length of the first insertion steps. For this steady state now the equations hold written at the top of the diagram. The evaluation shows, that the ratio of the propais about equal 2. That means, gation constants k ProPYl and kpentyl the insertion into a Ti-propyl chain is twice faster than the insertion into a Ti-pentyl chain.

/3

+)This is a preliminary evaluation because the - and &‘-resonances of the Ti-pentyl species lie very close together. A detailled separation by spectra simulation is in working.

Molecular Behaviour of Soluble Catalysts Part I

221

This confirms exactly our former results of oligomer kinetics by means of a p!.ug flow reactor and GPC analysis3)5 , , where we did find that k has a ~ 7 a l ~of e 9 6 l/mol-sec and kpentyl has a value of FroPY1 48 l/mol.sec. EXPERIMENTAL The I3C NMR spectra were recorded on a Bruker WM-3C0 spectrorr.eter at a frequency of 7 5 . 5 MHz. Samples were measured in 1 0 mm tubes dissolued in E8-toluene which had been dried over LiA1H4, degassed several times and distilled in vacuum. Chemical shifts were recorded relative to the CD -signal of the D -toluene solvent as internal 3 8 = 2 0 . 4 7 at 2 5 8 K). reference and converted to the TMS scale (&cD 3

For preparation of the NFR-samples see ref. 1. REFEREN CCS

Fink and R. Rottler, Angew. Makromol. Chemie 2 , 2 5 ( 1 9 8 1 ) . Fink and W. Zoller, Makromol. Chem. 182, 3 2 6 5 ( 1 9 8 1 ) . 3. G. Fink and G. Dollinger, J. Appl. Polymer Sci.: Appl. Polymer Symp. 36, 6 7 ( 1 9 8 1 ) . 4. G. Fink and D. Schnell, Angew. Makromol. Chemie 105, 1 5 ( 1 9 8 2 ) . 5. G. Fink and D. Schnell, Angew. Makromol. Chemie 105, 31 ( 1 9 8 2 ) . 1.

G.

2.

G.

222

G. Fink, W. Fenzl and R. Mynott

9

N

\

I I

I

I

I I

I I I

\ \

\

I

I

Cp2Ti Et CI I A l Et CI,

\ I

\

\

\ \

I

I I

I

9

L

I

I

I

I

1/

Figure 1.

Overall polymerization rate of ethylene versus time for

various Ti/Al catalysts determined in a batch reactor by following the rate of consumption of ethylene.

(For experimental method see reference2)). Al/Ti = 4; [Ti] = 3.10-3 mol/l; [CzH4] = 0.089 mol/l;

solvent toluene; T = 283 K.

M o l e c u l a r B e h a v i o u r Of S o l u b l e C a t a l y s t s P a r t I

1

100

120

110

Figure 2.

120

100

7. 5_. 5 MHz

80

10

b0

LO

LO

LO

Upper

= 0.05

b

10

10

0

3C NMR spectra o f t h e s y s t e m Cp2TiMeC1/AlFeC12

r e a c t i n g w i t h 13C2H4 i n t o l u e n e - d g r e c o r d e d a t 213 K ; [Cp,TiMeCl]

223

nol/l; [ T i ] : [ A l l : [ I 3 C 2 H 4 ] = 1 : 0 . 9 5 : 0.7. s p e c t r u m : a t s t a r t of r e a c t i o n ( 0 m i n u t e s ) . Lower s p e c t r u m : a f t e r t h e s a m p l e h a d r e a c t e d for 30 m i n u t e s a t 258 K a n d t h e n b e e n cooled r a p i d l y t o 213 K .

224

G.

F i n k , W.

F e n z l and R. Mynott

CpZTifleCl

Ti - P - h o p

/ t4eAlC12

Ti:Al:13C2H4= Toluene-dg

1

:

/

13C2k

0.95:0.7

[Ti]= 0.05moll[

r.t.: 2YlK after 70rnin. rn.t.: a3 K

1 140

Figure 3.

M

35

40

1s

20

25

CP I

-

T i a -(>Pent1 AI-Me

1

120

100

A s for Fig.

m i n u t e s a t 258 K .

80

60

40

20

2 , b u t a f t e r a t o t a l r e a c t i o n t i m e of 70

0

M o l e c u l a r B e h a v i o u r of Soluble C a t a l y s t s P a r t I

io

35

1:

lir YO

Figure 4.

n,

1s

20

i5

225

CPl

1

120

I 100

A s f o r Fig.

m i n u t e s a t 258 K .

80

60

40

20

0

2 , b u t a f t e r a t o t a l r e a c t i o n t i m e of 110

226

G.

Fink, W.

1

i0

35

F e n z l a n d R.

"

.

I

30

Mynott

--.

I

'

25

I

20

'

'

"

I

15

'

~

CP I

120

140

F i g u r e 5.

WO

80

60

b0

20

0

13C NMR s p e c t r u m of t h e s y s t e m Cp2TiMeC1/A1MeC12 r e a c -

t i n g w i t h I 3 C 2 H 4 i n t o l u e n e - d 8 a t 258 K . [Cp7TiMeC1] = 0 . 0 5 m o l / l ; [Ti]

: [All

: [l3C2H4I = 1 : 0 . 0 5

: 0.7.

S p e c t r u m r e c o r d e d a t 2 5 8 K a t b e t w e e n 2 . 5 a n d 3 . 5 h o u r s of r e a c t i o n

time

.

'

~

M o l e c u l a r B e h a v i o u r of S o l u b l e C a t a l y s t s Part I

Formation of active species

227

For 283 K is valid

kl

SJ

10".

. . 10"

k; = l o 6 . . . lo7

llrnol

-s

5-l

k2 =3k l

Chain propagation 'C

+

c'

nM

P,

in detail respectively C'P,

+M

kP

CP' ,1

kp

-

140 (Ethyl). ..50(Hexyll l l m o l - s

+M

kP

C'P,+*

+M

kP

and

50 0"

Figure 6. Reaction scheme of the successive equilibria using soluble Ziegler catalysts €or ethene polymerization (Cp2TiRC1/A1RxC1 / Y toluene).

228

Fink, W.

G.

F e n z l a n d R. M y n o t t

15

I

J. 140

Figure 7.

im

do

80

60

[CpZTiMeC1] = 0.05 m o l / l ; : [All

20

6

1 3 C NPR s p e c t r u m of t h e s y s t e m Cp2TiMeC1/A1MeCl2

r e a c t i n g w i t h 13C2H4 i n t o l u e n e - d 8 a t 2 5 8 K . [Ti]

u)

: [13C2H4] = 1 : 0 . 9 5

R e a c t i o n t i m e from 1 4 . 5 t o 1 5 . 5 h .

: 0.7.

Molecular Behaviour of Soluble Catalysts part I

CpZTiMeCl / Me2AlCl I I3C2Hk TI .Ai:13C2Hh: Toluene

chain propagation in dependence on time

1: 2 : 2

- d8 ITi1:O 0 5 m o i I I

mt:258K

229

Ti -(w-ZI-lrHeptyl)

Ti - y - b H e p t y l i

I

75.5 VHZ 13C NMR spectra of the system cp TiKeCl/ Figure 8. 2 A1Me2C1 reacting with I3C2H4 in toluene-d8 at 258 K. Initial [Cp2TiMeC1] = 0.05 mol/l; [Ti] : [All : [I3C2H4] = 1 : 2 : 2. These spectra, each measured over a period of an hour, were recorded successively in order to follow the development in chain growth. The regions depicted are €or the T i - q resonances and for the signals to (3-1 positions. for the

/3

230

G.

Fink, W.

F e n z l and R. Mynott

Relative unirs

of peak areas

*

p-Heptyl and

*

longer chains

+

+

+ 0

c

0

Figure 9 .

I

100

tI

200

0

0

x

X

II

300

0 x

0

Q

X

X

0

I

I

I

I

400

500

600

I

I

0

0 x

x 1 I

700

I

800

0

n-Pentyi

x

n -Propyl I 900 min.

Integration of t h e spectra (preliminary evaluation) of

Fig. 8; steady state concentrations for Ti-propyl and Ti-pentyl chains.