Ethene Polymerization Activity and Coordination Gap Aperture in Non

Home , Zirconocene

FULL PAPER Ethene Polymerization Activity and Coordination Gap Aperture in non-ansa Alk yI-Su bs tituted Cyclopent adieny1- and Phospholyl-Zirconium/MAO Cata1 yst s* Christoph Janiak'", Katharina C. H. Langea, Uwe Versteega, Dieter Lentzb, and Peter H. M. Budzelaar'

lnstitut fur Anorganische Cheinie, Techiiische Universitat Berlin", StraDe des 17. Juni 135, D-10623 Berlin, Germany Telefax: (internat.) +49(0)30314-22168 Institut fur Anorganische und Analytische Chemie, Freie Universitat Berlinh, FabeckstraDe 34-36, D-14195 Berlin, Germany Department of Inorganic Chemistry, University of Nijmegenc, P. 0. Box 9010, 6500 GL Nijmegen, The Netherlands

Received June 28, 1996

Key Words: Zirconocenes / Metallocene-methylalumoxane catalysts / Ethene polymerization / Structure-reactivity correlation / Coordination gap aperture / Cone angles / 91Zr NMR

The ethene polymerization activities of a series of brium reaction at high Al/Zr ratios, as shown by 31PNMR. Ab Cp'(C5H5)ZrC12and CpiZrC12 pre-catalysts (Cp' = C5HMe4, initio calculations on model LzTiMe+ complexes (L = C5H5, C,Me4P, C5Me5, C5H4tBu, C5H3-1,3-tBu2, C5H2-1,2,4-tBu,) C5H4Nand C5H4P)for the insertion of ethene in the Ti-Me together with (C5H&ZrC12 have been correlated with the bond suggest a high electronic similarity for phospholyl and coordination gap aperture. In the case of the mixed substi- cyclopentadienyl. The glZr-NMR data (chemical shift and tuted C,R,-,R; ligands, the coordination gap aperture has line width) for the above zirconocene series are reported. The been determined with the help of the cyclopentadienyl cone glZr chemical shift values increase with a good linear correla- angle and the bending angle at zirconium. A hindered rotation to the number of methyl or tert-butyl groups, which is tability of the fert-butyl substituted systems has been eva- traced to the additive electron donating effect of the alkyl luated by molecular modeling calculations and was included groups. The hyperconjugative donor effect of a tert-butyl in the gap aperture estimation. Deviations from the activity- group is found here to be 1.25 times that of a methyl group. -gap aperture correlation in the case of the phospholyl The X-ray structure of (C5H3-l,3-tBu2)(C5H5)ZrC12has been (C4Me4P)systems could be accounted for in terms of adduct determined (monoclinic, P2,/n,(I = 6.631(3), b = 18.843(9), formation between aluminum species and the phosphorus c = 15.265(5) A, p = 91.3", Z = 4). donors. These Lewis acid/base adducts form in an equili-

Introduction stituted parent system zirconocene dichloride (1) was included as a reference point. Single-site metallocene-methylalumoxane (MAO) catalysts are currently being introduced in industry as a new Results generation of Ziegler-Natta catalysts for the polymerization Coordination Gap Apertures of olefins['-']. Relationships between the structures of the zirconocene complexes and their properties as polymeri- This geometric parameter is defined as "the largest pos- zation catalysts are of central interest in the majority of sible angle spanned by those two planes through the metal studies in this Determinations of the steric (and center which touch the inner van der Waals outline of each electronic) parameters of zirconocenes and their corre- of the C5-ring ligands" (cf. Figure l)[I3]. At the beginning lations with activity have been The most of our study, our aim was a determination of the coordi- successful approach to such a correlation appears to be the nation gap apertures for the zirconocenes 1-12. Only the use of the coordination gap aperture (Figure 1) put forward gap apertures for compounds 1, 6 and 7, which contain the by Brintzinger et al.[13].We have used this concept in a sys- fivefold symmetrical C5H5or C5Me5 ligands, could be ob- tematic study concerning the comparative performance of tained by simply drawing the van der Waals outlines based methyl- and tert-butyl-substituted zirconocenes in ethene on the experimental X-ray structural geometries and meas- polymerization. Metal complexes bearing bulky cyclopen- uring the opening angle. The gap apertures for complexes tadienyl ligands often exhibit properties different from with the unsymmetrical substituted C5HMe4, C4Me4P, and those of the corresponding unsubstituted cyclopentadienyl the various tert-butylcyclopentadienyl ligands were not ac- analogs[14].Here we report the coordination gap apertures, cessible in this way due to the different rotamers possible. 91Zr-NMR chemical shifts, and the ethene polymerization To average the gap aperture for the different ring rotamers activity (in combination with MAO) of 11 different system- we utilized the cyclopentadienyl ring cone angles with the atically alkyl-substituted zirconocenes (2- 12). The unsub- metal as apex of the cone (Figure 2). This cyclopentadienyl

Clzern. Bey. 1996,129,1517 - 1529 0 VCH VerlagsgesellschaftmbH, D-69451 Weinhem. 1996 0009-2940/96!1212-1517 $15.00+.25/0 1517 FULL PAPER C. Janiak et al.

Scheme 1. Methyl-substituted zirconocene dichlorides and parent Figure 1. Schematic representation of the coordination gap aper- zirconocene dichloride used in the catalytic studies ture (cga) of a bent metallocene based on the van der Waals surface of the ring ligands. The gap aperture means the maximum opening angle; for clarity only the van der Wads surfaccs of the ring substituents (circles) along the maximum opening are shown

\ .c1 3

Figure 2. Schematic illustration of the cone angle a of a metal- bound cyclopentadienyl ligand. The circles represent the van der Waals outlines of the ring substituents; note that a depends on the metal-ring distance

/--

Scheme 2. tert-Butyl-substituted zirconocene dichlorides used in the catalytic investigations

HtBUFigure 3. Illustrating the use of the (half) cone angles (~~12,a,l2) of substituted cyclopentadienyl ligands together with the bending angle of the metallocene (fl) for the calculation of the coordination gap aperture (cga)

'B u &'B"

'Bu

3). The bending angle is taken here as the ring-centroid- cone angle concept has been used for steric measurement metal- ring-centroid angle. of substituted cyclopentadienyl ligands and has also been employed to explain differences in ethene polymerization cga = 360" - a1/2- a2/2- B (1) activities among (C5H4R)2TiC12and (C5H4R)(C5H5)TiCl2/ (a1,a2 = cone angles for different cyclopentadienyl li- Et3AI2CI3and among (C5H4R)2ZrC12/ethylalumoxanecata- gands). lyst~[~~'~~'~~'~].When the cone angle a is known for a sub- Here again, the cone angle can only be determined di- stituted cyclopentadienyl ligand, the coordination gap aper- rectly for the fivefold symmetrically substituted C5ligands. ture (cga) can then be calculated with the help of the bend- However, from the two cone angles a and a' of two differ- ing angle p at zirconium, according to equation 1 (cf. Figure ent fivefold symmetrical ring ligands C5R5 and C5R;, the

1518 Chem. Ber. 1996, 129, 1517-1529 Ethene Polymerization Activity FULL PAPER Table la. Cone angles (a)for (C5Rj.,R:,)Zr systemsLa] 1, assumes free rotatability of the ligands. While this is true for the methyl-substituted zirconocenes, it may not be valid R, R’ n CSR~-~R’~cone angleb] for the tert-butyl substituted analogs[24].Molecular mode- a/” ling calculations give well-formed minima for specific con- formers together with barriers of rotation. To locate the H 0 C5*5 139 minimum energy rotamers the whole hypersurface was Me 5 CgMeg 171 mapped at an artificially lengthened Zr-Cp distance which H, Me 4 CgHMeq 164 just allowed for the passage of the tert-butyl groups upon ring rotation. The minima and their vicinities were then cal- CR= P 0 p5 152 culated at the realistic Zr-ring distance. It was also verified Me,CR‘=P CqMeqP 167 4 that no other minima developed upon decreasing the Zr-Cp contact. Figure 5 provides a view of the total two- 5 “Cg&u5” 224 QU dimensional energy surface and two of the four energy min- H, tBu 1 C5mfBu 156 ima for (C5H3tB~2)2ZrC12,11, with respect to an indepen- 2 CsH3fBu2 173 dent rotation of both rings, and shows the geometries of the minimum energy rotamers. The optimized geometry in 3 CSH~~BU~190 one of the minima corresponds closely to the solid state 4 C5GBu4 207 conformation as observed by an X-ray structure determination[22].The path from one minimum to another has to la]The following geometric parameters were used for the cone angle involve a concerted rotation of both rings. From dynamic determination of the CSR5ligands: d(Zr-Cs ring centroid) = 2.20 for C5H5/2.22for C5Mej/2.27for C,tBu,, radius C5ring = 1.18 for NMR studies of (CSH4tB~)2Zr(s-~i~-butadiene)and (C5HZ- C,H5/1.20 for CSMe5and CStBu5, d(Zr-”P5” ring antroid) = 2.28, 1 ,2,4-SiMe3)2ZrC12[281a barrier to ring rotation of AG* radius “P5”ring = 1.51 (here, these last two distances were taken around 10 kcaVmol may be estimated for 9 and 11 here. from the X-ray structure of (C,Me,P)2ZrC12[17]), CrH = 1.00. C5-Me = 1.50, CS-tBu = 1.52, (C-C)t,, = 1.54 A, C,ing-C- When calculating the gap aperture for the tert-butyl zir- Me-H = Cri,,g-(C-C)lBu = IlO”, van der Wads radii: C = 1.67, conocenes. the hindered ring rotation and the preferential H = 1.32, Me = 2.00, P = 1.85 A note that the methyl group as a whole is assigned a van der Waals radius[l81; for C and H the orientation of the substituents along the maximum opening intermediate value from the range given was taken (J. E. Huheey, E. A. Keiter, R. L. Keiter, Inorganic Chemistry, 4th ed., Harper Collins. New York. 1995. v. 292). The bond distances were averaged Figure 4. Molecular structure of (CSH3-1,3-tBu2)(C5H,)ZrCl2(10) from the X-ray structural data for (CsH5)2ZrC12n91, with the atomic numbering scheme and stereoscopic cell plot (BLA- (C5HMe&ZrCl2[I, (C4Me4P)(C5Hs)ZrC12[71,(C5Me ) C5Hj)- TON-TME, 50% probability euipsoids. and PLIJTON plot, respec- ZrC12[2”] (C5H3rBu)2ZrC121Z’1,(C5H3- 1 ,~-~BU~)~Z~CI~~’(, and ti~ely)[~~], Selected distances (A) and angles (’): Zr-C11 2.440(4), (C,H2-I.~~-tBu,)(C~H3ZrCl, [this workl. - lb] For mixed-substi- Zr-C12 2.441(3), Zr~c,,,~2.452(8)-2.582(6), Zr-ring(Cl-C5)- tu

f-7 cone angle a, for a mixed ligand of the type C5R5.nRA can Lf be approximated according to equation 2[1°1.

a,= (5 - n)/5 ct + n/5. 01’ (2) Hence, from the cone angles of the “pure” CsH5, C5Me5, P5[16],and the hypothetical C5tBu5ligands, the cone angles of the mixed-substituted cyclopentadienyl and the phospholyl ligands can be calculated. It should be noted that the cyclopentadienyl cone angles are dependent on the metal-ligand distance. For zirconium as the apical metal the cone angles are listed in Table la. To have as many structural parameters available as possible, we attempted crystallization of those teut-butyl substituted zirconocenes whose crystal structure was hitherto un- determined. However, crystals suitable for an X-ray structural investigation could only be obtained for (C5H3-l,3- tBu2)(C5H5)ZrC12(10). Figure 4 shows the molecular structure of 10 together with a stereoscopic cell plot. The two tert-butyl groups flank the maximum opening angle spanned by the two ring planes. From the cone and bending angles the coordination gap apertures can now be calculated according to eq. 1. How- ever, there is one further complication: The calculation of an average cone angle according to eq. 2, and its use in eq.

Chern. Ber. 1996, 129, 15 17 - 1 529 1519 FULL PAPER C. Janiak et al.

Figure 5. Energy surface for the independent rotation of both cy- Figure 7 illustrates a fair linear dependency between the clopentadienyl rings in 1 I at a lengthened Zr-Cp distance to allow for passage of the tert-butyl groups. At the actual Zr-Cp distance chemical shifts and the number of alkyl groups for both the two of the four minima found are mapped in more detail. The methyl- and tert-butyl-substituted zirconocene series. When minimum energy rotamers are sketched. The rotation angle, y, cor- the zirconium chemical shifts are based ~n(c~H~)~ZrCl~an responds here to the angle between the single C5-H bond and the mcridional centroid-Zr-centroid plane, with y = 0" being at the increase of about 20 ppm per methyl group or 25 ppm per minimum opening of the ring planes tert-butyl group is apparent. In view of the pronounced val- _.-~~- ence orbital energy dependence of bent metallocenes on the ,,/ ~ - -~ bending angle at the and the inverse pro- portionality of 6("Zr) to the average excitation energy (approximated by the HOMO-LUMO the change in angular parameters might also explain the variation in zirconium chemical To test this assumption, Figure 7 also includes plots of the zirconium chemical shifts versus ,AT-----, the ring-centroid-zirconium-ring-centroid angle. For some of the methyl-substituted systems, including the phospholyl compounds 4 and 5, there is a good linear dependency, but this is not the case for the tert-butyl substituted zirconocenes. Hence, we attribute the difference in chemical shifts mostly to the additive electronic changes induced by the ring substituent~[~I.Photoelectron spectroscopy investigations on permethylated zirconocenes and ferrocenes support such an additive electronic effect on the central metal[3"]. Alkyl groups are electron donating with respect to the cyclopentadienyl ring and the central meta1[30.3'l.Other things being equal. the difference in chemical shift between the methyl- and tert-butyl-substituted zirconocenes allows was taken into account by (surely somewhat arbitrarily) a comparison of the relative electron donor strength of using the cone angle for the cyclopentadienyl ring with one methyl and tert-butyl groups. From 91Zr-NMR we deduce more tert-butyl group. For example, for 8, with the here that a rut-butyl group is 1.25 times as electron donat- CSH4tBuligand, the cone angle for C5H3tBu2was used in- ing as a methyl group. This value is somewhat lower than stead, while for 11 (CSH3rBu2 ligands) the value for that suggested from photoelectron (PE) spectroscopic in- C5H2tBu3 was employed. Table lb contains the coordi- vestigations; there the ratio was about 2[3'].The broad- nation gap apertures for the zirconocenes 1-12. ness of the peaks in the PE spectra leads, however, to a higher uncertainty. We note that we were unable to trace 91Zr-NMR Studies any changes in the polymerization activity (see below) to electronic differences based on the type or number of A recent study suggests a correlation of zirconium-91 alkyl groups. Such electronic variations are apparently in- chemical shifts [6(g'Zr)] and line widths (AvII2)with geo- significant in comparison with the concomitant steric ef- metric parameters, most notably the tilt anglec4](Figure 6). fect s L71. For a more detailed understanding of the electronic varia- bility of the complexes involved in our study, and a possible The correlation between the line width, Av112, of the zir- indicator of the structural flexibility, we investigated the conium NMR signal and the tilt angle, shows no simple 91Zr-NMR parameters and their dependence on alkyl sub- T, dependency (see Table 2). While within 1-6-3 or 4-5 or stitution and coordination geometries. Table 2 lists the 91Zr 10-9-11 the increase in line broadening agrees with the chemical shifts and line widths of the series of zirconocene increase in the experimentally available tilt angle, a com- complexes together with the experimental tilt angles, where parison between the methyl- or tert-butyl-substituted or available from X-ray crystallographic analyses. phospholyl-containing zirconocenes indicates that other effects, e.g. correlation timed4], dominate the line widths. We Figure 6. Schematic representation of the tilt angle, z (dotted line = note that even within a closely related series, the agreement ring normal) between the trends in AvII2 and z might be fortuitous. The ring centroid tilt angles are obtained from solid-state structural data and hence can be influenced by packing effects, while the line widths are measured in solution. Also, rings with different substituents or rings not related by crystallographic symmetry in these bent sandwich molecules will have different tilt angles, e.g. 3.6" [or the C5H3-1,3-tBu2ring and 0.6" for the C5Hs ring in 10, 2.0" and 0.9" for the two independent C5H5 rings in l[lya].

1520 Chem. Ber. 1996, 129, 1517-1529 Ethene Polymerization Activity FULL PAPER Table lb. Coordination gap apertures together with bending angles at zirconium and the sum of cone angles used for the calculation in the zirconocene complexes 1-12 Compound bendin sum of half coordination Ref. to the [bl [GI angle 8 cone angles gap aperture X-ray data P 1" a1i2 4- a2I2 i" cgai" for P 1 (C5H5)2Zc12 129.1 139 92 19 2 (C5J3Me4)(C5H5) zrc12 132 Id] 152 76 3 (c5me4)2ze12 133.1 164 62 7

4 (C4Me4P)(C5H5 )ze12 132.3 153 75 7 5 (C4Me4P)2ZrC12 135.4 167 58 17

6 (C5Me5)(C5H5)ZrC12 130.0 155 75 20 7 (CgMe5)2ZrC12 134 ld1 171 55

[el [el (hind. rot.) (hind. rot.) 8 (CgHqtBu)(CgHg)ZrC12 130 156 74 9 (CSH~~BU)~Z~CI~ 128.7 173 58 21 10 (C-jH3-1,3-tJ3~2)(CgHg)ZrC12129.7 164.5 66 this work 11 (C5H3-1,3-tE3~2)2ZrC12 129.7 190 40 22 [dl 12 (C5H2- 1 ,2,4-tB~3)(C5H5)ZrC12 13 2.5 173 54.5

La] The ring-centroid-zirconium-ring centroid angle. For the difference in the angle between the ring normals from the zirconium center onto the ring planes, see the tilt angle in Table 2. - Lb1 The cone angles a are listed in Table la; a ,a2 are the cone angles for different cyclopentadienyl ligands. - Calculated according to cga = 360" - a1/2 - a& - p (eq. 1). - Where no X-ray data was available, the bending angle was estimated from comparative ZINDOll calculations~']. - Le] The hindered ring rotation and the preferential orientation of the substituents along the maximum opening was taken into account by using the cone angle for the cyclopentadienyl ring with one more tert-butyl group, e.g. for 12 with the CSH2tBu, ligand, half the cone angle for CsHtBu4 was entered as the value of a,/2.

Figure 7. Relation of 91Zr chemical shifts [based on (C5H5)2ZrC12] 1-12. (Note: "Activity" in the following means an "appar- and the number of alkyl groups on the cyclopentadienyl rings (left) ent activity" since differences in activation equilibria of dif- or the centroid-Zr- centroid bending angle, p (right) in (a) methyl- and (b) tert-butyl-substituted zirconocene dichloride complexes: in ferent complexes with MA0 cannot be taken into ac- (b) note that there are two compounds with a total number of two co~nt[~,~~I).We intentionally employed only ethene as a tevt-butyl groups, namely 9 and 10. Parentheses indicate ZINDO/I our estimated bending angles monomer in coinparative studies, as with propene or other a-olefins the stereospecificity of the catalyst strongly affects the acti~ity[~~~~~'].Furthermore, to give comparable catalytic activities, the polymerization conditions with ethene had to be chosen such as to avoid a diffusion-controlled 80 process. Figure 8 illustrates this point by showing the differ- 40 1 ence in ethene uptake over time. The differential ethene ::vi0 consumption (dV/dt) is a direct measure of the activity at 0 2 4 6 8 10 number of methyl groups bending angle, p P time t. In Figure 8 the activity profiles in the ethene poly- (b) merization for the two series of zirconocene complexes are given at high and low zirconocene concentration regimes. Table 3 lists the activity data. The overlapping curves of the .B6 1::vi 60 h 09 activity profiles and the resulting similar activities for many -a 40 m. (48 complexes among both series at the chosen "high" zir- w 20 1 20 0 1;r, conocene concentration of lop5moll-' are due to the rapid 0 1 2 3 4 5 128 129 130 131 132 133 polyethylene precipitation under these conditions. The poly- number of rert-butyl groups bending angle, P merization is truly homogeneous only at the very beginning. With the precipitation of polyethene the active complex becomes more and more embedded in the polymer matrix, Polymerization Studies which represents a transfer to a heterogeneous phase, i.e. The purpose of the polymerization experiments was to a heterogeneous active complex form, and this leads to a check for a correlation between the coordination gap aper- diffusion-controlled reaction. The reaction rate is then con- ture and the catalytic activity of the zirconocene complexes trolled by the rate of diffusion of the monomer through the

Chem. Bee 1996, 129, 1517-1529 1521 FULL PAPER- C. Jdniak et al.

Table 2. 91Zr-NMRdata and ring-tilt angles for the zirconocene Eventually, a zirconium concentration of 2.5 X mol complexes 1 - 12 1-’ together with an Al/Zr ratio of 160000:1 was found to allow comparison of the catalytic activities. The corre- Compd. 6(91Zr)[a1 Av1/2” 6(91Zr)[E1 tilt angle[dl Ref.Iel sponding activity profiles in Figure 8 show a more constant /PPm /ppm (av.) z /” ethene consumption. The smaller amount of polyethene 1 -1 13 260 0 1.3 25 produced under these conditions ensures the existence of a 2 -22 500 81 homogeneous phase over an extended time period. We note 3 54 810 167 3.4 that with the lowering of the zirconocene concentration, the 4 29 1760 142 1.8 AllZr ratio had to be increased overproportionally to offset 5 170 3030 283 1.9 a higher degrec of dissociation upon dilution of the cata- lytically active zirconocene-MA0 comple~[~~~~]. 6 10 340 123 1.7 7 87 140 200 25 Correlation: Polymerization Activity-Coordination Gap Aperture 8 -8 8 1070 25 The relation between the polymerization activity at low 9 -66 2800 47 2.9 4 zirconium concentration and the coordination gap aperture 10 -59 825 54 2.2 is depicted in Figure 9, in separate plots for the methyl- and 11 -1 3 5600 100 4.4 tert-butyl substituted zirconocenes. The methyl-cyclopen- 12 -34 880 79 tadienyl gystems 2, 3, 6 and 7, together with the unsubstituted zirconocene dichloride (1) give rise to a very good Chemical shift from (CSH5*ZrBr2 with F = 0; all samples rnea- sured in CH2CI /CD2C12.- [b 1 Frequency width at half maximum linear dependency. The points for the phospholyl systems, peak height. - h Chemical shift based on (C5H,)2ZrCI,. - Ld] The 4 and 5, fall somewhat below this line, on the side of lower tilt angle T is defined as the an le between the metal-ring-centroid activity for the estimated coordination gap aperture. In vector and the ring P (see Figure 6) and was calculated with the help of the PLATON program[23]from X-ray structural other words, for the activity obtained, the calculated gap data as “90” minus the Zr-Cp,,,-ring plane angle”. When the aperture is too large. As an explanation for this discrepancy two rings are not symmetry related, two tilt angles result - one for each ring - which can be quite different. The average value is listed we suggest the coordination of Lewis acid aluminum moiet- in the table. For the structural data, see the references given in ies to the Lewis base phosphorus donors, schematically Table lb. - References to 91Zr-NMR literature data avaihble shown in Figure 10 for 5. Such an adduct would have a for comparison. higher steric demand, resulting in a smaller gap aperture. (An alternative explanation, which we will not elaborate polymer matrix to the enclosed active center[34]and can no further, for the deviation from simple sterics might be the longer be compared in terms of steric effects. space taken up by the phosphorus “lone pairs”.)

Figure 8. Activity profiles in the cthene polymerization for the two zirconocene series 1-7 (a) and 1, 8-12 (b) activated with MA0 at a zirconium concentration to molar AUZr ratio of rnol-’/4000:1 (upper row) and 2.5 X mol 1-’/160000:1 (lower row). In the lower row the least active complexes are omitted; for further experimental conditions see Table 3

0.80 [zr] = 10-5 moUI 0.70 Al:Zr= 4000:l 0.4 .-- 0.60 E 0.50 - -. ’1 0.3 0.40 2 -* 5 0.30 5 0.2 0.20 0.10 0.1 om 0

0 5 in 15 20 25 30 0 s 10 15 zn 25 30 t [minJ t [min]

1522 Chcm. Ber. 1996, 129, 1517-1529 Ethene Polymerization Activity FULL PAPER

Table 3. Activity and polymer data for ethene polymerizations with Figure 9. Correlation between ethene polymerization activity and zirconocene complexes 1-12 and MA0 as a function oPZr concen- the coordination gap aperture for the methyl-substituted zircono- tration and molar AVZr ratio cone series 1-7 (a) and the tevt-butyl-substituted complexes 8-12 (b); the dashed line in (b) corresponds to the line draw in (a) Compd. Activityta1 Polymer data (a) Zr conc. M, / (q / Mq)rd] M.p.Le1 (b) 800 A (high) B (low) A €3 (B only) 1 37 49 600 12.4 756 134 000 12.7 2 30 79 400 i 2.9 462 (10.5 / 1210 000) 134.1 400 3 30 69 600 / 3.8 165 (8.75 / 995 000) 200 4 37 63 100 / 3.5 262 302 000 / 2.9 50 60 70 80 90 100 30 40 50 60 70 80 90 100 5 11 33 200 18.6 gap aperture ["I gap aperture ["I 14 234 000 14.2 6 30 108 000 / 2.9 373 (13.01 / 1 740 000) 135.6 tions take place at relatively low AUZr ratios up to 25: 1['61. 7 19 19 500 / 4.4 Since these profound changes at the zirconium center at low 35 (9.65 I I 140 000) AI/Zr ratios are not reflected in the 31Pchemical shift, we 1 23 48 700 12.3 suggest that the change in the 3'Presonance at large alumi- 774 100 000 13.0 132.3 8 22 64 200 / 2.6 num excess is instead due to adduct formation. Considering 278 n.a. 138.0 the magnitude and sign of the phosphorus coordination 9 9 110 000 14.2 chemical shift, A(3'P) = 6complex- tjfrrezlIC.IIOLenr = -6.5 15 n.a. 136.3 ppm, similar changes in "P chemical shift on coordination 10 13 107 000 / 3.2 27 n.a. 138.4 have been observed in trimethylaluminum-phosphine com- 11 0.2 n.d. plexes (A range -21 to +13 ppm). With sterically bulkier n.d. - phosphines A was negativer3'1. 12 I3 128 000 / 3.7 In view of the increased steric demand upon aluminum 7 n.a. 134.7 coordination to the phosphorus atoms one might have expected a more pronounced drop in activity for the phospho- Activity in kg PE/(g Zr X h X bar); n.d. = not determined. - ibl T = 70 "C, ethcne prcssure 5 bar, 300 ml toluene, polymerization lyl sytems 4 and 5. Such a large drop is only seen in the time 30 min; A: [Zr] = lops mol l-l, AllZr = 4000: I; B: [Zr] = 2.5 polymerization of propene or 1-hexene. There, the activity X mol I-', AllZr = 160000:l. - ['I M,, = number average molar mass in g mol-'; Q = M,/M,, dispersity; determined by size is 1-2 orders of magnitude lower than those of the corre- exclusion chromatography (SEC), except for the values in parenthe- sponding methylcyclopentadienyl zir~onocenes[~*1.This be- ses. The latter molar mass values were too high for SEC, and were havior may be explained by analogy with nnsa-zirconocene approximated from the Staudinger index q. - n.a. = not available due to the insolubility of the polymer in 1,2,4-trichlorobenzene at complexes with C5ring substituents close to the meridional 135°C. - ld]q = viscosity, Staudinger index in dl g-I, measured in centroid-Zr-centroid plane. These are propene-inactive decalin (0.1% solution) at 135"C, Mq = viscosity averaged molar mass calculated according to the Mark-FIouwing equation M = due to steric interaction of the substituent with the alkyl (T~/K)"~(for comparison pur oses: Mq lies between M,, and group of the olefin, but are still rather active in the polymer- with constants K = 51 X 10-?rnl g-' and a = 0.706 for polyethene ization of the stcrically less demanding ethene[39401. (low pressure in 1,2,4-trichlorobenzene at 135"C, in a range of M from X X 10b to 1.2 X loh g mol-' (see: E. P. Otocka, R. J. Roe, With respect to the electronic perturbation expected M. J. Hellmann, I? M. Muglia, Mucvomoleczile.r 1971, 4, 507.) - upon exchanging a CH group with a phosphorus atom it M.p. = melting point in "C from DSC; the results of the second might appear strange that the phospholyl complexes re- scan (after one heating and cooling cycle) are reported to avoid effects of mechanical and thermal history on the samples, heating semble so much their cyclopentadienyl counterparts. Ab in-

rate 10°C min-'. itio calculations on model L2TiMe + complexes were used to shed some light on this point. For each of the ligands An indication of formation of these P-A1 adducts in an C5H5, C5H4N[24,31,41]and C5H4P, insertion of ethene into equilibrium reaction was provided by 31PNMR. At small the Ti-Me bond was studied by determining the stationary Al/Zr ratios (below 50: 1) the resonance of the phospholyl points for reactants, transition state and products. These complex remains unaffected, but the 3'P-NMR data in calculations were done at the RHF level employing a small Table 4 show a second phosphorus resonance starting above basis set, and were not intended to give quantitatively accu- an AUZr ratio of about 50: 1. From excess aluminum values rate results; however, they should allow a comparison of the above 250:1, only the second signal was observed It was Cp complexes with their heteroatom-substituted analogs. ascertained that this difference in chemical shifts was not For the parent Cp system, a symmetrical qs coordination due to concentration effects. Addition of aluminum alkyls is found as expected. Coordination of the phospholyl group to a zirconocene dichloride complex will also lead to a reac- is also rather symmetrical, but the pyrrolyl complex shows tion of the halide ligands, such as a methylkhloride ex- a significant slippage of the metal atom towards nitrogen, change, and formation of bridged Zr-A1 species. Such reac- probably because of the concentration of negative charge in

Chem Ber: 1996, 129, 1517-1529 I523 FULL PAPER C. Janiak et al.

Figure 10. Schematic drawing of the bisphospholyl zirconocene complex 5, with one or two aluminum moieties from methylalumoxane (MAO) or trimethylaluminum (TMA) coordinated to the phosphorus donors in an equilibrium reaction. Note that reaction of the zirconocene dichloride with a methyl-aluminum species leads foremost to a reaction of the halide ligands. such as methylichloride exchange or abstraction, to give a zirconocenium cation; for simplicity these two ligands have been denoted here as X / /

5a b

Table 4. 3'P-NMR data for the phospholyl-containing zirconocenes Figure 11. Projections of Ti atom on ligand ring for Cp2TiMe+ in combination with various amounts of trimethylaluminum (A), (C5H4P)2TiMe+(B) and (CSf14N)2TiMe' (C); bond lengths (TMA) or methylalumoxane (MAO) in A

~ ~ [a1 A1 Al/Zr Compound 4 Compound 5 PI compd. ratio 8(31P) / ppm"] (Zr conc. / mmol l-1) none - 84.51 (94) 87.17 (43) - 87.26 (3.5) TMA 1:l 84.53 (90) - 2:l 84.52 (82) - 5:l - 87.27 (4) 1O:l 84.53 (59) 87.17 (36) A B C 25:l 84.54 (34) 87.28 (4) (4) 50: 1 - 87.30 Table 5. Energies relative to separated reactants (kcallmol) for in- 1OO:l 87.29/80.70[d1(4) sertion of ethene in L2TiMe+La] 200: 1 87.29/80.66[d1(4) 250: 1 80.80 L Complex T.S.lD' Product 500: 1 80.77 CSH, -3.8 14.6 -2 1.6 1000:1 - 80.44 (4) CSH,P -3.1 15.6 -22.2 80.56 (3) CSH4N 4.0 10.9 -23.0 MA0 1OOO:l - 80.93 (3) In] Barriers calculated at higher levels than RHF and with better basis sets tend to be much lower, sometimes even close to zero. Note added in proof: The bis(tetramethylphospholy1)iron com- Thus, our values should not be compared directly to any experi- plex, (C4Me4P),Fe, was synthesized["] and investigated by NMR mental values. However, comparison of the values for the three spectroscopy in order to verify the P-A1 adduct formation as a ligands should give a reasonable indication of the electronic effects source of the shift of the P resonance upon addition of TMA and of introducing a heteroatom in a metallocene. - rbl T.S. = transi- to exclude the influence of side reactions at the Zr-C1 bonds. The tion state. 31P resonance of the pure iron complex appeared at -60.37 ppm, with TMA at AlfFe ratios of 50, 100 and 150:1, the signal shifted to -54.7 pm (same concentration, toluene solution). - Ib] Molar the slipped pyrrolyl complex is less crowded and has a ratio. - ['PBruker ARX 200, 80.0 MHz for 311', samples in toluene/ [D&oluene. TMA or MA0 was added as a toluene solution to the somewhat lower barrier. zirconocene solutions in [D,]toluene. - Ld] The intensity ratio of Thus, it appears that the phospholyl group is indeed elec- the two signals is about 1 :5. tronically very similar to a Cp group, both in ground-state geometry and in its influence on olefin insertion reactions. that region. Projections of the metal atom onto the ring The reason for this must be the similar electronegativities plane are shown in Figure 11. of carbon (2.5) and phosphorus (2.1): substitution by the The metal-ring interaction does not change significantly more electronegative nitrogen atom (3.0) results in signifi- during olefin insertion: the Cp and phospholyl ligands re- cant changes in structure and reactivity. Of course, one can- main symmetrically q 5-coordinated, and the pyrrolyl group not expect a perfect parallel between Cp and phospholyl retains its slipped coordination mode. This is reflected in chemistry: the phospholyl group has a reactive lone pair the olefin insertion barriers (Table 5): the Cp and phos- which introduces complications not found for the more in- phony1 complexes have nearly the same barriers, whereas ert Cp group. In addition, coordination of e.g. A1 to P

1524 Chem. Bei: 1996,129, 15 17- 1529 Ethene Polymerization Activity FULL PAPER should have some influence on both steric and electronic (ii) In our series of comparisons, the polyethene proper- properties of the complex. ties are affected most dramatically by a variation in tran- The relation between the polymerization activity and the sition metal concentration such that a decrease in zirconium gap aperture for the tert-butyl zirconocenes (Figure 9b) re- concentration gives rise to a large increase in molar mass flects the more complicated approximation of the gap aper- (cf. Table 3) to the extent that the polyethene is no longer ture because of the hindered ring rotation (see above). soluble for SEC measurements. Such substantial molar There is no linear dependency for all data points. When mass increases upon a decrease in zirconocene concen- comparing the data points to the relation obtained from the tration have been observed bef~re[~~x~~].Two possible expla- methyl-substituted series (dashed line in Figure 9b), it could nations are discussed for this not fully understood pheno- be argued that the estimated gap aperture for 8 and 10 is men~n[~~I:(a) a dilution effect favoring the active complex still too large, if one assumes the same relation irrespective form over the inactive precursor or dormant species and of the kind of substituent. Furthermore, it is evident that hence the rate of chain propagation (kp) over that of chain a minimum gap aperture of about 50" is necessary for a termination (kT). It is thereby assumed that chain termi- polymerization activity. At a gap aperture below 50" only nations arise predominantly from "dormant", i.e. tempor- very low activities are observed (compound 11). From this, arily inactive, zirconocene species. (b) A bimolecular chain- the aforementioned deviation of 8 and 10 may be attributed transfer mechanism involving the active complex and a se- to the presence of rotamers having the fert-butyl group(s) cond (active or inactive) zirconocene species is proposed by pointing along the maximum opening where the polymeri- Kaminsky et al.[43]. A decrease in zirconocene concen- zation reaction takes place, so that the actual gap aperture tration would then reduce the rate of the termination reac- for this rotamer becomes smaller than 50". As a tert-butyl tion, kT. The mean degree of polymerization, P,, which is group in this position blocks the path of the C-C bond- proportional to the number average molecular weight Mn, forming reaction, the respective rotamer would be inactive, is given by the ratio of the growth rate to the rate of trans- hence lowering the number of active species. fer, il4, - P, = kp/kT[391. Aside from the activity-gap-aperture correlation further (iii) Within the tert-butyl substituted zirconocene series, remarks concerning the catalyst activity and the polymer the more sterically demanding complexes 9, 11 and 12 parameters can be made: showed rather low polymerization activities. Since it might (i) The polymer dispersity (Q, Table 3) is rather narrow, have been possible that the more sterically demanding com- as expected for metallocene catalysts. The slight increase of plexes were activated more slowly, we carried out a brief Q to 3 - 4, from the theoretical value of 2 €or a Schulz-Flory investigation of the influence of activation or ageing times, distribution, is explained by the polyethene precipitation i.e. the time the zirconocene dichloride and MA0 were al- and subsequent heterogenation of the reaction mixture. lowed to react before pressurizing with ethene. The results Nevertheless, the even broader dispersity of Q= 8.6 for the are summarized in Table 6. Normally, the activation time bisphospholyl complex 5 is remarkable, and is due to a bi- was 10 rnin. Extending this time to 24 h led to considerable modal molar mass distribution as shown in Figure 12. Simi- increase in activity for unsubstituted zirconocene dichloride lar bimodal molar mass distributions with even broader dis- (1) and the mono tot-butyl substituted complex 8. For persities of Q > 10 were observed for the bispentamethyl complexes with two tert-butyl groups, 9 and 10, the rise complex 7 at zirconium concentrations below 5.0 X in activity became very small, while for the tris-tert-butyl mol/l and Al/Zr > 8000: 1. Therefore, at least two types of reactive centers must be present. We suggest that the com- Table 6. Influence of activation times on the polymerization activity plexes with the more bulky tetramethyl- and pentamethyl- of zirconocendMA0 systems ligands are more prone to ligand loss, with the half-sand- PI [a1 wich complex still continuing to be an active polymeri- Compd. Activity after activation time zation catalyst [421. 10 min 3h 24 h

Figure 12. Bimodal molar mass distribution of polyethene obtained 1 23 49 57 with the bisphospholyl zirconocene 5 8 22 - 37 9 9 9 11 1 10 13 - 15 I 12 13 - 5 % I I la]Conditions of activation: In the case of an activation time of 10 min, the reactants were dissolved in 300 ml of toluene, previously thermostatted at 70°C. For the 3 h and 24 h activation times, three- quarters of the MA0 solution was added to the zirconocene dissolved in a small amount of toluene (giving AUZr = 3000: I). and the mixture was allowed to stand for 3 h or 24 h, respectively. The rest of the required MA0 solution (114) had been previously added to the 300 ml portion of toluene for purification purposes. - lbl Acti- vity in kg PE/(g Zr X h X bar). Polymerization conditions: T= 70°C, ethene pressure 5 bar, 300 ml toluene, polymerization time molar mass / g mol-1 30 min; [Zr] = mol I-', AllZr = 4000:l.

Chem. Bm 1996, 129, 1517-1529 1525 C. Janiak et al. FULL PAPER- ____

substituted zirconocene 12 the activity was even decreased. Materials: All complexes were purified by sublimation and the An increase of activity with ageing (5 to 60 min) has been purity was checked by elemental analysis, NMR spectroscopy and reported for (C5H5)2ZrC12/MAO+TMAin ethene polymer- mass spectrometry. The analytical data matched the literature ization, yet a decrease was seen with (C5H&ZrC12/MA0 in values. Analytical data obtained by us and not previously cited in the literature are given below. (C5H5)*ZrC12was purchased from the polymerization of pr~pene[~~]. Merck and used without further purification. Methylalumoxane (MAO) was obtained from Witco (Bergkamen, Germany) as a 10 wt.-%-toluene solution (4.92 wt.-% aluminum, density = 0.9 g/ml. Conclusions average molecular weight of thc MA0 oligomers 900- 1100 glmol). In the ethene polymerization with non-ansa alkyl-substi- Solvents were dried over sodium metal (toluene and benzene), so- tuted cyclopentadienyl and phospholyl zirconocene/MAO dium benzophenone ketyl (pentane and diethyl ether) or potassium systems the catalyst activity under comparable, mostly metal (hexane and THF), followed by distillation and storage under argon. Ethene (BASF AGj was polymerization grade and used homogeneous conditions is a function of steric factors. The without further purification. All experiments were carried out un- order of activity follows the order of steric demand. This der argon using standard Schlenk techniques. The known zir- steric demand can be quantified through the coordination conocene dichlorides were prepared according to literature pro- gap aperture, which was in turn obtained through the ring cedures, or slight modifications thereof, as given below:

cone angle and the zirconium bending angle. In the case of (C,HMq) ( C5H5)ZrC12 (2)['], (C.7FIMe4)2ZrC12 (3)[73451, not freely rotatable rings, such as the lert-butyl substituted (C4Me4P) (C5H.5) Zr Clz (4)C7], (C4 Me4P)2ZrC12 (S)[73'81,

ligands, the influence of specific rotamers must bc con- (C,Me5J (C,H,)ZrCl, (6)[7346], ( CiMer)2ZrCl, (7)J[7,Jq, sidered to account for the activity-gap aperture corre- (C,H4tBu)(C5H5)ZrCl: (8)[481: 'H NMR (C,D6): F =; 1.23 (s, lation. IBu), 5.69 (t. 3J(H,H) = 2.7 Hz, C5-2,3-H), 5.95 (s, C5H5), 6.01 (t, The phospholyl zirconium compounds give Lewis acid/ 3J(H,H) = 2.7 Hz, C5-2,3-H). ~ I3c NMR (C6D6): 6 31.26 base adducts with aluminum species (MA0 or TMA) at (CH3C). 33.37 (CH3C), 112.54, 115.34 (C,-C-2,3), 115.69 (C5HS), 144.15 (CyC-I). high AVZr ratios, in an apparent equilibrium reaction, as shown by 31PNMR. Alteration of the steric situation (C.5H4tB~)2ZrC12(9)[4"1: 'H NMR (C6D6): 6 = 1.29 (s, tBu), through this adduct formation is seen in the activity-gap 5.76 (t, 'J(H.H) = 2.7 Hz, C5-2,3-H), 6.08 (t, 'J(H,H) = 2.7 Hz, aperture correlation, Cy2,3-H). - 13C NMR (C6D6): F = 31.34 (CH3C), 33.49 (CH3C), Zirconium-91 NMR studies show a fairly linear additive 111.71, 115.99 (C5-C-2,3), 143.74 (C5-C-l). - MS (EI, 70 eV, electronic effect of the alkyl groups on the cyclopentadienyl 80'C): m/r (YO)= 402 (21, [MI+'), 387 (33, [M - CH3]'), 367 (3, [M - C1]+). 351 (100, [M HCl - CH3]+), 281 (28, [M - ring or of the phosphorus on the central metal. From this, C5H4tBu]+= [(C5H4tBu)ZrC12]+),265 (22, [(C5H4tBu)ZrCl2 - the relative electronic donor effect of a tert-butyl group ver- CHJ]+), 245 (16, [(CjH4tBu)ZrC12 - HCl]'), 229 (51, sus a methyl substituent is estimated to be 1.25. No influ- [(CSH4fBu)ZrC12- HCI - CH,I+). ence of these electronic effects can be seen, however, in the polymerization activity. (C,H,-I,.~-tBu21(C.7H.~)ZrCI~(10): A solution of 0.77 g (4.31 mmol) of 1,3-C5H4tB~2[5"] in 10 ml of toluene was reacted with an This work was supported by the Deutsche Forschungsgemeirz- equimolar amount of n-BuLi (1.6 M in hexane). After stirring for schaji through grant Ja46613- I, the Funds deer Chemischen Industrie, 30 min, the light-yellow solution was added dropwise to a slurry and the Gesellschafl der Freunde deer TU Berlin. The Polyolefin Di- of C5H5ZrC13(dme)[SL]in 10 ml of toluene. When the addition was vision oj BASF AG, Ludwigshafen is thanked for the donation of complete, the mixture was refluxed for 3 h. The orange toluene ethene, MA0 and the analyses of polymers, Prof. H. Schurnann for solution was decanted from some precipitate and the solvent was his continuous support, and Dr. D. M. Proserpio (University of removed in vacuo. The off-white residuc was then sublimed at Milano) for a copy of the CACAO program. We are grateful to 130- 14O"ClO. 1 Torr to give the product as snow-white fine needles Mrs. A. Miiller for the DSC measurements. (yield 0.7 g, 40%, not optimized). Crystals suitable for X-ray structure analysis were obtained upon slow solvent evaporation from a dichloromethane solution at 2°C. - '€INMR (C,D,): 6 = 1.17 (s, Experimenta1 tBu). 5.94 (d, ,J(H,H) = 2.6 Hz, C5-4,5-H), 6.12 (s: C5Hj),6.18 (t; 4J(H,H) = 2.6 Hz, C5-2-H); (CD2C12):6 = 1.27 (s, tBu), 6.37 (m: Instruments: CHN analysis, Perkin-Elmer Series I1 CHNSlO An- C5-2,4,5-H), 6.48 (s, C5Hj). - I3C NMR (C6D6): 6 = 31.30 alyzer 2400. - NMR spectroscopy, Bruker ARX 200 or ARX 400 (CH3C), 33.84 (CH;C). 112.32 (C5-C-2), 112.42 (Cj-C-4,5), 115.95 ('H and chemical shifts are referenced to TMS via the solvent (C5H5), 144.05 (C5-C-1,3). - MS (EI, 70 eV, 100°C): id: (%) = signal, 31Pto external 85% H3P04,"Zr NMR spectra were meas- 402 (35, 387 (41, [M - CH,]+), 367 (6, [M - CI]'~), 351 ured at ambient temperature against CpzZrC12as cxternal reference [MIt'), (100, [M - HCl - CH,]'), 337 j73, [M - CsH5]+ = [dissolved in an 8.1 mixture of CH2C12/CD2C12]with 6 = -113 [(C5H3tBu2)ZrCl21+1,321 118, [(C5H:,tBu2)ZrC1, - CH,]+}, 301 ppm from CpzZrBr2[2s]).- Mass spectrometry, Varian MAT 311A/ AMD (mass spectrometry peaks given refer to the most abundant (12, [(CSH3tBu2)ZrC12- HCl]'}, 285 117, [(C5H3tBu2)ZrCI2- isotope combination, which contains 90Zr, 35C1, except when two HCI - CH4]+}, 269 122, [(C5H3tBu2)ZrCI2~ HCl - 2 CH,]+}, chlorines are present. In this case, the peak arising from 225 (16; [CsHsZrCi2]+), 57 (23. [tBu]'~'), 41 (25, [tBu - CH,J+*). - C18H26ZrC12(404.53): calcd. C 53.44, H 6.48; found C 53.37, 90Zr35C137Ciand 92Zr35C1is of maximum intensity, as proven by an isotope simulation, but the 90Zr35C12peak is cited to match the H 6.60. fragment losses). - Size exclusion chromatography (SEC, GPC), (C.~H.~-1,3-tBu212ZrCl,(11)[53]: 'H NMR (C6D6): 6 = 1.32 (s, Waters 150 chromatography at 135"C, solvent i ,2,4-trichloroben- tBu), 5.83 [d, 'J(H,H) = 2.6 Hz, Cs-4,5-H], 6.63 [t, 4J(H,H) = 2.6 zene. - DSC, Perkin-Elmer Series 7 Thermal Analysis System. HZ, Cg-2-HJ. - ''C NMR (C6D.5): 6 = 31.15 (CH,C), 34.12

1526 Cheni. Bev. 1996, 129, 1517-1529 Ethene Polymerization Activity FULL PAPER (CH-+C),105.38 (Cj-C-43, 122 79 (C,-C-2), 144.29 (CyC-1,3). - lengths for the two rings. However, the rings were allowed to MS (El, 70 eV, 120°C): mh (%) = 514 (12, [MI*’), 499 (2, [M - “breathe” and to move in space as rigid bodies, which allows ring CH,]+), 479 (2, [M - Cl]’), 463 (2, [M - HCI - CHJ+’), 337 slippage and asymmetry between the two metal -ring interactions. {97, [M - C5H-itBu2]+= [(CsH-irBu2)ZrClz]+}, 321 (13, We never saw any evidence of such deformations in our optmized [(C5H3tBu2)ZrC12 - CH4]+}, 305 (6, [(C5H3tBuJZrC12 - 2 geometries. As a check, we also did a full, unconstrained geometry CHX}, 301 15, [(C5H,tBu2)ZrCl2 - HCI]’), 285 {12, optimization on Cp2TiMe’. This still produced symmetrical q5 [(CsH3tBu2)ZrC12- HC1 - CHd]*}, 269 { 15, [(CSH3tBu2)ZrC12- Cp-Ti bonding; the only difference with the constrained structure HCI - 2 CH,]+}. was a small amount of bending of the ring hydrogens out of the (C~H~-I,2,4-tBu~)(CjHg)ZrC12(12)[5’1: MS (EI, 70 eV, 120°C): ring plane, away from the metal atom. In calculations involving m/= (%I) = 458 (30, [MI”), 443 (19, [M - CHq]+), 423 (4, [M - the pyrrolyl and phospholyl Iigands, we first did a full geometry Cl]+), 407 (49, [M - HCl - CH,]+’), 393 (100, [M - C5HF]’ = optimization on the L2TiCH: derivatives. The ligands were then [(C~H2tBu~)ZrCl~]+l,377 { 1 I, [(CjHl_tBu3)ZrC1, - CH,]+), 341 averaged to local C,, symmetry, and further derivatives were opti- (9, [(C5H2tBu3)ZrClz - CH4 - HCI]+), 325 j16, mized by refining the ligands as rigid bodies. The energy difference [(C5H2tBu3)ZrC12- HCI - 2 CH4]+), 309 f16, [(C5H2?Bu3)ZrCI2 between the fully optimized and the constrained structures is 2-3 - HC1 - 3 CH4]’). 225 (11, [M - C5H2tBu3]+= [C5H5ZrC1,]+), kcal/mol, and the geometrical differences are very small. Table 7 57 (83, [~Bu]’.), 41 (35, [tBu - CHd]’ ’). gves the calculated total energies at optimized geometries. Polymerizations: Polymerization reactions were carried out in a I 1 Buchi-glass autoclave, thermostatted to 70°C and charged with Table 7. Total energies (a.u.) for insertion of ethene into L2TiMe+ 300 ml toluene, MA0 and the transition metal complex. The cala- L L,TiMe+ L,TiMe(C,&)+ L,Ti(Me..C,%)’ L2Ti”Pr+ lyst amount. concentration and AI/Zr ratio are specified in the re- -1344.97215 -1345.02995 spectivc tables After a preactivation time of 10 min the autoclave CsHs -1267.39455 -1345.00156 was pressurized with 5 bar ethene and after 30 min the reaction was Cs&P -1865.34475 -1942.95068 -1942.92091 -1942.98110 CAN -1298.90182 -1376.50916 -1376.48534 -1376.53945 stopped by draining the toluene/polyethene slurry into an acidified - -~ (HCI) waterlmethanol mixture. It was then stirred for 4 h to achieve (C2H4: -77.60099 xu.) complete catalyst-MA0 decomposition and dissolution within the polymer matrix The polymer was separated by filtration, washed with hexane and dried at 80°C. To ensure reproducibility, polyiner- X-ray Structure Determination of 10 izations were carried out at least twice with each zirconium complex and a series of polymerization runs was performed with Crystal Dam. Molecular formula Ci8H26C12Zr,formular yeight charges from the same toluene and MA0 batch To avoid ageing 404.51 glrnol, a = 6.631(3), b = 18.843(9), c = 15.265(5) A, /3 = effects of a series of comparative polymerizations was run 91.3”, Z = 4. dcab = 1.409 gicrn’, monoclinic, P21/n (No. 14). - Data Collectiop: STOE four-circle diffractometei, radiation within a week. Mo-K, (h= 0.71069 A), graphite monochromator, crystal size 0.40 X 0.10 . Modeling Cakulation.~ The semi-empuicai Zl NDOIl geometry X 0.05 mm’, 293 K, o-scan, 4 < 20 < 50°, -1 < h S 7, -22 C optiini~ations~~’]for the estimation of angular structural data were k s 22, -18 < 1s 17, 7303 reflections measured, 3347 indepen- performed with the program HyperChem (Version 3.0, Autodesk dent, ~(Mo-K,) 8.49 cm-I, no absorption correction. - Structurul Tnc., Sausolito. CA 94965, USA). dnalysis and Refinement: Structure solution was performed by di- For the potentlal energy calculations of the ring rotations the rect methods (SHELXS-86[661). Refinement: Full-matrix least- computations were performed within thc extended-Huckel formal- squares on P (SHELXL93[661); all atomic positions except those ism[56]with weighted H,,’s(~’]and the use of the CACAO program for the iert-butyl hydrogens were found and refined (non-hydrogen (Version 4.0)[581.An extended-Huckel molecular orbital (EHMO) atoms with anisotropic temperature factors), 21 8 refined param- program was uscd because it provided a dummy atom topology for eters, final RI = 0 0683, wR2 = 0.1 I41 for 1896 reflections With I the ring centroid and thereby allowed for a controlled rotation of > 20(/), goodness-of-fit on = 1.023[671. the cyclopen tadienyl rings. The atomic parameters for the elements involved in these EHMO calculations were as follows (H,,, c): Zr 5s. -9.87 eV, 1.817; 5p, -6.76 eV, 1.776; 4d, -11.18 eV, 3 835, fi Dedicated to Prof. Dr. Karl-Heznz Rezchert on the occasion of 1SO5 (coefficients for double-( expansion: 0.6224, 0.5782)[591;C1 his 60th birthday. 34 -26.30 eV, 2.183; 3p, -14 20 eV, 1 733rh01; C 2~,-21.4 eV, Li] Ziegler Catalysts (G. Fink, R Muhlhaupt, H. H. Brintzinger, 1.625; 2p, -11.4 eV. 1.625[5hl;H Is, -13.6 eV, 1.3[561.Geometrical Eds.), Springer Verlag, Berlin 1995. For notes concerning the industrial application, see e.g.: Chemr- parameters were fixed as follows: Zr-C = 2.51, Zr-C1 = 2.44, ial & Engineering New.s 1995. October 30, p. 10; September 18, (C-C),, = 1.42, C-H = 1.00, (C-C)tBu = 1.50, CI-Zr-CI = p. 17; September 11, p. 15-20; May 22, p. 34-38. MUJ~I, p. 98”. 7-8. R. Muhlhaupt, B. Rieger, Chimiu 1995, 49, 486-4991, The ab initio calculations were of the all-electron closed-shell or For recent work on metallocene catalysis, see for example: R. H. Grubbs, G W. Coates, Acc: Chem Rec 1996,29, 85-93. J. open-shell Restricted Hartree-Fock type[”], and were carried out P. Mitchell, S. Hajela, S. K. Brookhart, K. I. Hardcastle, L. M. using the CAMESS programr6’l on IBM RS/6000 and Silicon Henhng, J. E. Bercaw,J Am. Chem Soc, 1996,118. 1045-1053. Graphics Crimson workstations. For the ligands we used the STO- K. Patsidis, H. G. Alt, W. Milius, S. J. Palackal, J Organamet 3G minimal basis set[”]; the “reacting” organic fragments (methyl, Chem. I%, SOY, 63-71. M. Arnold, 0. Menschke, J. Knorr, Mucromol Chem Phys 1996, 197, 563-573. L. Reswni, R. L. olefin) were described with the 3-2IG split-valence basis set.[&]For Jones, A. L. Rheingold, G. P. A. Yap, Organometuilics 1996, 15, metal atom (Ti) we used the Dunning split-valence basis[”J. Some 998-1005. T. Yoshida, N. Koga, K. Morokuma, Organometal- geometric constraints (described below) were used for the ligands, IZCS 1996, IS. 766-777. .I. Koivumaki, Pu[yrn. Bull 1994, 36, but the remainder of the molecule was always fully optimized, with- 7-12. C. Psiorz, G. Erker, R. Frohlich, M. Grehl, C/ien.t. Ber. 1995, 128, 357-364. out any symmetry restrictions. M. Buhl, G. Hopp, W. von Phihpsborn, S. Beck, M.-H. Pro- In optimizations of cyclopentad~enylcompounds, local DThsym- senc, U. Rief, H.-H. Brmtzinger, OrganametulZim 1996, 15, metry was enforced for each ring, and equal C-C and C-H bond 778-785.

Chem. Bev. 1996, 129, 1517-1529 1527 FULL PAPER C. Janiak et al.

[’I P C. Mohring, N. J. Coville. J Organomet. C’hem. 1994, 479, [361 A. R. Siedle, R. A. Newmark, J. N. Schroepfer, P. A. Lyon, 1-29. Organomeiallics 1991, 10, 400-404. A. R. Siedlc, R. A. New- C. Janiak, K. C. H. Lange, l? Marquardt, Macromol. Rapid mark, W. M. Lamanna, J. N. Schroepfer, Polyhedron 1990, 9, Comnzun. 1995, 16, 643-650. 301-308. W. Kaminsky, R. Steiger, Polyhedron 1988, 7, 17] C. Janiak, U. Versteeg, K. C. H. Lange, R. Weimann, E. Hahn, 237552381, E S. Dyachkovskii, A. K. Shilova, A. E. Shilova, 1 Organomet. Chem. 1995, 501, 219-234. J. Polyin. Sci,: Part C 1967, 16, 2333-2339. G. Natta, G. Maz- [‘I J. Tian, B. Huang, Macromol. Rapid Cornmun. 1994, 15, zanti, Tetrahedron 1960, 8, 86-100. D. S. Breslow, N. R. New- 923-928. burg, L Am. Chem. Soc. 1959,81, 81-86. G. Natta, l? Pino, G. [’I W. Kaminsky, R. Engehausen, K. Zoumis, W. Spaleck, R. Mazzanti, U. Giannini, J Am. Chem. Soc. 1957, 79, Rohrmann, Macromol. Chem. 1992, 193, 1643- 1651. 29 75 -29 76. [lo] D. White, N. J. Coville, Adv. Organomet. Cliem. 1994, 36, P71 A. R. Barron, J Chum. Soc., Dalton Trans. 1988, 3047-3050. 95- 158. 13’] C. Janiak, K. C. H. Lange, P. Marquardt, R.-P. Kruger, manu- [11] P. C. Mohring, N. J. Coville, J Mol. Catal. 1992, 77, 41-50. script in preparation. [I2] P. C. Mohring, N. Vlachakis, N. E. Grimmer, N. J. Coville, J. [3y1 H.-H. Brintzinger, D. Fischer, R. Muhlhaupt, B. Rieger, R. Organomet. Chem. 1994, 483, 159- 166. Waymouth, Angew. Chem. Int. Ed. Engl. 1995, 34, 1143- 1170. [I3] P. Burger, K. Hortmann, H.-H. Brintzinger, Macromol. Chem. F49 W. A. Herrmann, J. Rohrmann, E. Herdtweck, W. Spaleck, A. Macromol. Symp. 1993, 66, 127- 140. K. Hortmann, H.-H. Winter, Angew. Chenz. Int. Ed. Encl. 1989, 28, 1511-1512. Brintzinger, New J. Chem. 1992, 16, 51 -55. 14’1 For a review on pyrrolyl or azacyclopentadienyl metal com- [I4] C. Janiak, H. Schumann, Adv. Organomet. Chem 1991, 33, plexes, see: C. Janiak, N. Kuhn, Advances in Nitrogen Hetero- 291-393. J. Okuda, Top. Curr: Chem. 1991, 160, 97-145. cycles, Vol 11 (C. J. Moody, ed.), JAI Press, Greenwich, CT, ‘Is1 N. J. Covillc, M. S. Loonat, D. White, D. Carlton, Orgunome- USA, 1996, VOI. 2, 179-210. tallies 1992, 11, 1082-1090. [421 Q. Wang, R. Quyoum, D. J. Gillis, M.-J. Tudoret, D. Jcremic. [I6] Cyclo-P, is not hypothetical but can be stabilizcd in the coordi- B. K. Hunter, M. C. Baird, Organometallics 1996, 15,693-703. nation sphere of transition metals; see 0. J. Scherer, Comments L431 W. Kaminsky, M. Miri, H. Sinn, R. Woldt, Macromol. Chem. Inorg. Chern. 1987, 6, 1-22. M. Baudler, S. Akpapoglou, D. Rapid Commun. 1983, 4, 417-424. W. Kaminsky, K. Kiilper, S. Ouzounis, E Wasgestian, B. Meinigke, H. Budzikiewicz, H. Niedoba, Macrornol. Chem. Macromol. Synzp. 1986, 3, Miinster, Angew. Chem. ht. Ed. Engl. 1988, 27, 280; Angew 377-387. CJiem. 1988, 100, 288; M. Baudler, T. Etzbach, Angew. Chem. J. C. W. Chien, €3.-P. Wang, .I Polym. Sci. A: Polym. Chem. Int. Ed. Engl. 1991, 30, 580; Angew. Chem 1991, 103, 580. 1988, 26, 3089-3102. W. Kaminsky, Macromol. Symp. 1995, lL71 F. Nief, E Mathey, L. Ricard, F. Robert, Organometallic~1988, 97, 79-89. 7, 921-926. 14’1 P. Courtot, R. Pichon, J. Y Salaun, L. Toupet, Cun. J. Chem. [‘*I L. Pauling, The Nature of the Chemical Bond, 3rd ed., Cornell 1991, 69, 661-672. University Press, Ithaca, N.Y., 1960, p. 260-261. [461 P. T. Wolzanski, J. E. Bercaw, Organometallics 1982, 1, [Iy] [Iya] K. Prout, T. S. Cameron, R. A. Forder, S. R. Critchley, B. 793 - 799. Denton, G. V. Rees, Acta Crystullogr., Sect. B 1974, 30, [471 J. M. Manriauez. D. R. McAlister. E. Rosenbere. A. M. Shiller. 2290-2304. - I. A. Ronova, N. V. Alekseev, N. I. Gapot- K. L. Williakon, S. I. Chan, J. E. Bercaw, 1 2m. Chenz. Soc. chenko, Yu. T. Struchkov, Zh. Strukt. Khim. 1970,II, 584-589. 1978. 100. 3079-3083. [’‘I [’‘I R. D. Rogers, M. M. Benning, L. K. Kurihara, K. J. Moriarty, L4*1 P. Renaut, G. Tainturier, B. Gautheron, J Organornet Chem. M. D. Rausch, J Organornet. Chem. 1985,293, 51 -60. 1978, 148, 35-42. iz11 R. A. Howie. G. P. McOuillan. D. W. Thomoson. G. A. Lock. 14’1 M. F. Lappert, C. J. Pickett, P. I. Riley, P. I. W. Yarrow, J Chem. J. Organomci. Chem. 1986, 303, 213-220. Soc., Dalton Trans. 1981, 805-813. LZ2] U. Bohme. H. Lanphof. Z. Krist. 1993. 206. 281-283. R. Riemschneider, Z. Natuvforsch. 1963, B18, 641-645. C. G. Cz3] A. L. Spek, PLATCN-93, PLUTON-92 graphics program; Uni- Vernier. E. W. Casserlv. .I Am. Chem Soc. 1990. 112. versity of Utrecht, The Netherlands. A. L. Spek, Acta Crys- 2808-2809. tullogr., Sect. A 1990, 46, C34. C5l1 E. C. Lund, T. Livinghouse, Orgunometallics 1990, 9, [241 N. Kuhn, G. Henkel, J. Kreutzberg, S. Stubenrauch, C. Janiak, 2426- 2427. J. Organornet. Chem 1993,456,97- 106, and references therein. 15*1 H._-- Sitzmann,- - P. Zhou, G. Wolmershauser, Clzem. Ber 1994, KZ51 R. Benn, A. Rufinska, Angew. Chem. Int. Ed. Engl. 1986, 25, lL/, 3-Y. 861 - 891. [531 1. F. Urazowski, V. I. Ponomaryev, I. E. Nifant’ev, D. A. Le- EZ61 C. G. Howard, G. S. Girolami, G. Wilkinson, M. Thornton- menovskii, j. Organornet. Cliem. 1989, 368, 287-294. Pett, M. B. Hursthouse, J. Am. Chem. Soc. 1984, 106, LS41 S. Lasscrre, J. Derouault, Nouv J. Chim. 1983, 7, 659-665. 203 3 - 2040. [551 A. D. Bacon, M. C. Zerner, Theor. Chim. Acfa 1979,53, 21-54. [271 J. W. Lauher, R. Hoffmann, J. Am. Chem. Soc. 1976, 98, M. C. Zerncr, Reviews in Computational Chemistry, Vol. ZI. (K. 1729- 1742. B. Lipkowitz, D. B. Boyd, Eds.), VCH Weinheim, 1991, p. [28] G. Erker, T. Muhlenbernd, A. Rufinska, R. Benn, Clzein. Ber. 313-365. 1987, 120, 507-519. C. H. Winter, D. A. Dobbs, X.-X. Zhou, [561 R. Hoffmann, 1 Chem. Phys. 1963, 39, 1397-1412. R. 1 Organornef. Cheni. 1991,403, 145-151. Hoffmann, W. N. Lipscomb, 1 Chem. Phys. 1962, 36, 12‘)] 12‘)] U. Bihme, K. H. Thiele, A. Rufinska, Z. anorg. allg. Chem. 2179-2189. R. Hoffmann, W. N. LiDscomb, L Chem. Phys. 1994. 620. 1455-1462. 1962, 37, 2872-2883. r3O] I? G. Gassman, D. W. Macomber, J. W. Hershberger, Organome- P71 J. H. Ammeter, H.-B. Biirgi, J. C. Thibeault, R. Hoffmann, J tallics 1983,2, 1470- 1472. P. G. Gassman, M. R. Callstrom, L Am. Chem. Sac. 1978, 100, 3686-3692. Am. Chem. Soc. 1987, 109, 7875-7876. C. Cauletti, J. C. Green, L5*1 C. Mealli. D. M. Prosermo. J. Chenz Ed. 1990. 67. 399-402. M. R. Kelly, P. Powell, J. v. Tilborg, J. Robbins, J. Smart, 1 [591 K. Tatsumi, A. Nakahura, P. Hofmann, P. Stauffert, R. Electr Spectr. Rel. Phenom. 1980, 19, 327-353. S. Evans, M. Hoffniann, J Am. Chem. SOC.1985, 107, 4440-4451. L. H. Green, B. Jewitt, A. F. Orchard, C. E Pygall, J Chem. L60] R. H. Summerville, R. Hoffmann, 1 Am. Chem. SOC. 1976, Soc., Faraduy Trans. II 1972, 68, 1847-1865. 98, 7240-7254. L3’I C. Janiak, N. Kuhn, R. Gleiter, L Organomet. Chem. 1994, Lo‘] C. C. J. Roothaan, Rev. Mod Php 1951, 23, 69. 475, 223-221. L6’1 M. F. Guest, J. Kendrick, GAMESS Users Manual, SERC Da- L3*1 D. Fischer, S. Jiingling, R. Muhlhaupt, Makromol. Chem., Mac- resbury Laboratory, CCP1/86/1, 1986. M. Dupois, D. Spangler, romol. Symp. 1993,66, 191-202. D. Fischer, Ph. D. Thesis, Uni- J. J. Wendoloski, NRCC Software Catalog, Vol. 1, Program No. versity of Freiburg, Germany, 1992. D. Fischer, R. Muhlhaupt, QGOl (GAMESS), 1980. M. E Guest, R. J. Harrison, J. H. van J Organomet. Chem. 991, 417, C7-Cll. Lenthe, L. C. H. van Corler, Theor. Chim. Artn 1980, 71, 117. [j3]T. Uozumi, K. Soga, Makromol. Cltem. 1992, 193, 823-831. W. J. Hehre. R. F. Stewart. J. A. Poole. 1 Chen?. Phix 1969, 51, [j4]G. Fink, W Zoller, Makromol. Chem. 1981, 182, 3265-3278. 2657. W. J. Hehre, R. Ditchfield, R. E Stcwart, J..A. Pople, J K. Soga, H. Yanaphara; D. Lee. Mukromol. Chem. 1989, 190, Chem. Phys. 1970,52,2769.W. J. Pietro, W. J. Hehre, J CoPnput. 995. K. H. Reichert, K. R. Meyer, IWukrornoL Chem. 1973, Chem. 1983. 4. 241. J. B. Collins. P. v. R. Schlrver. J. S. Binklev. 169, 163-176. J. A. Pople, Chem. Phys, 1976,64,5142. H. Titewaki, S. Hu& r351 B. Rieger, C. Janiak, Angrw. Makromol. Chenz. 1994, 215, naga, J Chem. Phys. 1980, 72, 399. 35-46. [641 J. S. Binkley, J. A. Pople, W. J. Hehre, J Am. Cliein. Soc. 1980,

1528 Chem. Ber. 1996, 129, 1517-1529 Ethene Polymerization Activity - FULL PAPER 102, 939. M. S. Gordon, J. S. Binkley, J. A. Pople, W. J. Pietro, Refinement, Gottingen, 1993; SHELXS-86, Program for Crys- W. J. Hehre, 1 Am. Chem. Soe. 1980, 104, 2797. M. J. Frisch, tal Structure Solution, Gottingen, 1986. J. A. Pople, J. S. Binkley, 1 Chem. Phys. 1984, 80, 3265. K. D. L6’1 Further details of the crystal structure investigations may be Dobbs, W J. Hehre, 1 Comput. Chem. 1986, 7, 359 and 1987, obtained from the Fachinformationszentrum Karlsruhe, D- 8, 861, 880. 76344 Eggenstein-Leopoldshafen (Germany), on quoting the Ch51 T. H. Dunning Jr., P. J. Hay in Modern Theoretical Chemistry, depository number CSD-405685. Vol. 3, H. F. Schaefer ed., Plenum, New York, 1977, p. 1. I961301 Lhh] G. M. Sheldrick, SHELXL-93, Program for Crystal Structure

Chem. Ber: 1996, 129, 1517-1529 I529