Connection of Stereoselectivity, Regioselectivity, and Molecular

Article Cite This: Macromolecules XXXX, XXX, XXX−XXX

pubs.acs.org/Macromolecules

Connection of Stereoselectivity, Regioselectivity, and Molecular Weight Capability in rac-R′2Si(2-Me-4-R-indenyl)2ZrCl2 Type Catalysts Christian Ehm,*,†,§ Antonio Vittoria,†,§ Georgy P. Goryunov,‡,§ Pavel S. Kulyabin,‡,§ Peter H. M. Budzelaar,†,§ Alexander Z. Voskoboynikov,‡,§ Vincenzo Busico,†,§ Dmitry V. Uborsky,*,‡,§ and Roberta Cipullo*,†,§ †

Dipartimento di Scienze Chimiche, Università di Napoli Federico II, Via Cintia, 80126 Napoli, Italy Department of Chemistry, Lomonosov Moscow State University, 1/3 Leninskie Gory, 119991 Moscow, Russia § DPI, P.O. Box 902, 5600 AX Eindhoven, the Netherlands

Macromolecules Downloaded from pubs.acs.org by UNIV OF WINNIPEG on 10/09/18. For personal use only.

‡

S Supporting Information *

ABSTRACT: A set of 19 silicon-bridged C2-symmetric zirconocenes rac-R′2Si(2-Me-4-R-indenyl)2ZrCl2 of varying steric demand in position 4 were synthesized and screened in propene homopolymerization in a high-throughput experimental setup. The size and accuracy of the experimental data set allow to identify surprisingly good correlations among stereoselectivity, regioselectivity, and molecular weight capability (R2 ≈ 0.8−0.9) over a broad range. We rationalize this trend by assuming that steric tuning in the 4-position affects both preferred insertion and stereoerror formation similarly but leaves other barriers largely unaffected. A quantitative structure−activity relationship based on one single computational descriptor, Δ%VBurusing the difference in the percent of buried volume between the “blocked” and “open” quadrants of the catalyst precursoris established. Provided that a large sphere of 5.0 Å is used, stereoselectivity can be predicted with unprecedented accuracy, i.e., a mean average deviation (MAD) of 0.18 kcal/mol (ΔΔG‡enantio), 0.0007 (σ, probability that the preferred propene enantioface is selected at an active site of given chirality), or 0.3% (mmmm pentads). On the basis of this empirical model, we predicted that the catalyst with R = o-tolyl is an ideal candidate for high stereoselectivity/high MW capability. Ad hoc synthesis and testing of the precursor confirmed the expectations: the catalyst shows the highest stereoselectivity reported so far (σ = 0.9999) for metallocenes at 60 °C, while maintaining a high MW capability (Mw > 1 MDa) and relatively high regioselectivity.

■

INTRODUCTION

concluded that only limited room exists for further improving C2-symmetric bis-indenyl-based ansa-metallocenes,14 in particular as far as molecular weight capability is concerned. The amount of available data regarding the performance of C2-symmetric metallocenes in the literature is enormous, but data from different sources, or sometimes even from the same source,15 are not easily comparable. For example, (i) polymerization conditions (monomer partial pressure, solvent, and polymerization temperature) influence polymer characteristics (regioselectivity, tacticity, and MW);3,4 (ii) activator/ scavenger systems influence primarily activities/productivities and can also limit molecular weights via chain transfer to aluminum;4 and (iii) catalyst heterogenization can influence catalyst performance and polymer properties.16 This situation can hamper rational catalyst tuning/design. With the above-

Group 4 metallocenes initiated the development of molecular olefin polymerization catalysis. Ansa-zirconocenes with a C2symmetric bis-indenyl ligand framework, in particular, were the focus of intensive and elegant research efforts aimed at the implementation of isotactic-selective polypropylene catalysts.1−7 In little more than one decade, several families of complexes featuring various substitution patterns were disclosed with a performance rivaling classical heterogeneous Ziegler−Natta catalysts.3 However, a complicated synthesis, the random distribution of residual stereo- and regiodefects in the polymer chains, and above all an inadequate molecular weight capability at industrially desirable polymerization temperatures (>70 °C, indicatively) altogether limited success in practical application.8,9 Also, due to the more recent boost of non-metallocene (“post-metallocene”) catalysts,10,11 some with superior performance, the interest in those pioneering ansa-metallocenes declined, with only few (albeit remarkable) revivals.12,13 As a matter of fact, Cavallo and Talarico recently © XXXX American Chemical Society

Received: July 18, 2018 Revised: September 13, 2018

A

DOI: 10.1021/acs.macromol.8b01546 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

Figure 1. Precatalysts synthesized and screened in propene homopolymerization. Catalyst M19 was synthesized and tested a posteriori as proof of concept. The synthesis of M3, M5, M12, M16, M18, and M19 is detailed in the Supporting Information. M1 and M2 were kindly donated by SABIC and used as received. Polymer Synthesis and Characterization. Synthesis and characterization of polymers have been performed using a highthroughput workflow, optimized for polyolefins, which has been described in detail before.17 Polymerization experiments were performed in a Freeslate (former Symyx) parallel pressure reactor (PPR) platform in toluene at 60 °C and ppropene of 65 or 95 psi until a desired gaseous monomer consumption was reached (reaction time 2−120 min) using triisobutylaluminum/HNMe2Ph+[B(C6F5)4]− (TIBA/AB) as the scavenger/activator system. 10 μmol of TIBA, 5 mL of solvent, and a 2-fold excess of AB with respect to the precatalyst were used in each cell. The catalyst amount was varied from 10 to 150 nmol, depending on the catalyst. For runs using trityl tetrakis(pentafluorophenyl)borate ([Ph3C+] [B(C6F5)4]−), catalyst amounts were decreased, and the activator was used in 5−10-fold excess. The reactors were quenched with dry air; the polymers (in glass inserts) were collected and dried in a Genevac EZ2-Plus centrifugal evaporator. GPC curves were recorded with a Freeslate Rapid GPC setup, equipped with a set of two mixed-bed Agilent PLgel 10 μm columns and a Polymer Char IR4 detector. Calibration was performed with the universal method, using 10 monodisperse polystyrene samples (Mn between 1.3 and 3700 kDa). 13C NMR spectra were recorded with a 100 MHz Bruker Avance III 400 spectrometer equipped with a 5 mm high-temperature cryoprobe. Polymer samples (∼25 mg) were dissolved at 120 °C in tetrachloroethane-1,2-d2 (0.5 mL) with 0.40 mg mL−1 BHT (2,6-di-tert-butyl-4methylphenol) as stabilizer. DSC curves were obtained with a differential scanning calorimeter (DSC-822 by Mettler Toledo) in a flowing N2 atmosphere at a scanning rate of 10 °C/min from 0 to 200

mentioned limitations of literature comparisons in mind, we embarked on a systematic study, taking advantage of our high throughput experimentation (HTE) catalyst synthesis, screening, and characterization methods combined with computational and QSAR modeling. The ultimate goal is to improve the performance of metallocene catalysts also at higher temperatures relevant for industrial applications. In this initial paper, we focused on variations of the substituent in the 4-position of the indenyl fragments in C2symmetric rac-R′2Si(2-Me-4-R-indenyl)2ZrCl2 type zirconocenes. Nineteen catalysts were synthesized and screened under identical conditions, using a fully automated screening platform (Freeslate PPR48), integrated with a polymer characterization workflow,17,18 including gel permeation chromatography (GPC) and 1H/13C nuclear magnetic resonance (NMR) spectroscopy using a high-temperature cryoprobe. The results were combined in a database which was then used to identify simple quantitative structure−activity relationships (QSAR) and correlations between different catalyst performance indicators.19,20

■

EXPERIMENTAL SECTION

Catalyst Synthesis. The precursor compounds 4-bromo-1methoxy-2-methylindane,21 bis(4-bromo-2-methyl-1H-inden-1-yl)dimethylsilane, bis(4-chloro-2-methyl-1H-inden-1-yl)dimethylsilane,22 and Zr[PhN(CH2)3NPh]Cl2(THF)2,23 metallocenes M6 and M17,24 M4,21 M8,25 M7, M9, M10, and M13,26 and M11, M14, and M1527 were synthesized according to the literature.28 B


Article

Macromolecules

Table 1. Summary of Polymerization of Propene Results with Catalysts M1−M19 Activated with Triisobutylaluminum/ HNMe2Ph+[B(C6F5)4]− (AB) in Toluene at 60 °C and 95 psi Partial Propene Pressurea substituents

NMR

GPC

catalyst

R, bridgeb

mmrrmmb,d

σd

ΔGenantioe

[2,1]c,d

[3,1]c,d

regioTotc,d

ΔGregioe

Mnf

Mwf

PDI

ΔGTerminatione

Tm (°C)

M1 M2 M3 M4 M5 M6 M7 M8 M9 M10 M11 M12 M13 M14

H, Me Ph, Me 2-furyl, Me 2-thienyl, Me Cl, Me Br, Me Me, Me iPr, Me p-F-Ph, Me p-Me-Ph, Me p-tBu-Ph, Me p-CF3-Ph, Me mesityl, Me m,m-tBu2-Ph, Me m,m-F2-Ph, Me C6F5, Me Br, Et Me, Et o-tolyl, Me

1.25(3) 0.12 0.79 0.29(2) 0.39(3) 0.39 0.39(2) 0.37(2) 0.11 0.14 0.14 0.11 0.03 0.06

0.9865(3) 0.9988 0.9917 0.9971(3) 0.9960(3) 0.9960 0.9960(2) 0.9962(2) 0.9989 0.9986 0.9986 0.9989 0.9997 0.9994

2.8 4.5 3.2 3.9 3.7 3.7 3.6 3.7 4.5 4.4 4.4 4.5 5.4 4.8

0.25 0.32 0.90 0.50(2) 0.54(3) 0.82 0.62(2) 0.69 0.29 0.38 0.29 0.32 0.21(2) 0.17

0.03 n.d. n.d n.d. n.d. n.d. 0.04 0.05 n.d. 0.04(4) 0.02(2) n.d. 0.18(3) n.d.

0.29 0.32 0.90(2) 0.50(2) 0.53(3) 0.82 0.67(2) 0.74(2) 0.29 0.40(3) 0.30(2) 0.32 0.40 0.17

3.9 3.8 3.1 3.5 3.5 3.2 3.3 3.3 3.9 3.7 3.9 3.8 3.7 4.2

98 319 126 232 145 136 78 100 316 232 254 293 76 533

189 621 265 474 288 269 162 208 651 551 550 701 169 1139

1.9 2.0 2.1 2.0 2.0 1.9 2.1 2.1 2.1 2.4 2.2 2.4 2.2 2.2

5.2 5.9 5.3 5.7 5.4 5.4 5.0 5.2 5.9 5.7 5.8 5.9 5.0 6.3

145.7 160.1 143.2 153.2 151.7 148.9 148.9 149.7 157.3 157.6 159.3 156.9 155.3 162.4

0.16

0.9984

4.2

0.42

n.d

0.42

3.6

290

678

2.3

5.9

157.2

0.06 0.36 0.39(3) 0.01

0.9994 0.9963 0.9959(2) 0.9999

4.9 3.7 3.7 6.1

0.18 0.78(3) 0.61(2) 0.32

n.d. n.d. 0.05(5) 0.09

0.18 0.79(3) 0.66(5) 0.42

4.2 3.2 3.3 3.6

408 143 91 470

885 281 186 1049

2.2 2.0 2.1 2.2

6.1 5.4 5.1 6.2

160.8 149.0 149.3 158.2

M15 M16 M17 M18 M19 a

[Activator]:[Zr] = 2. For more details (e.g., activity, polymerization time, catalyst loading, etc.) see Tables S1 and S2. bSubstituent on the SiR2 bridge. cIn mol %. dExperimental uncertainty on last significant digit is ±1, unless otherwise indicated in parentheses. eIn kcal/mol. fIn kDa. °C. Polymer melting points (Tm) were collected from the second heating run. Details of polymerization experiments and polymer analysis can be found in the Supporting Information (Tables S1 and S2). Computational Details. All geometries were fully optimized using the Gaussian 09 software package29 in combination with the OPTIMIZE routine of Baker30,31 and the BOpt software package.32 Following the protocol proposed in ref 33, all relevant minima and transition states were fully optimized at the TPSSTPSS level34 of theory employing correlation-consistent polarized valence double-ζ Dunning (DZ) basis sets of cc-pVDZ quality35,36 from the EMSL basis set exchange library.37 The protocol has been successfully used, in combination with M06-2X single point energy corrections to address several polymerization related problems: absolute barrier heights for propagation,38 comonomer reactivity ratios,39,40 metal− carbon bond strengths,41,42 and electronic and steric tuning effects on MW capability.43 The density fitting approximation (resolution of identity, RI) was used throughout.44−47 All calculations were performed at the standard Gaussian 09 quality settings [Scf = Tight and Int(Grid = Fine)]. All structures represent either true minima (as indicated by the absence of imaginary frequencies) or transition states (with exactly one imaginary frequency corresponding to the reaction coordinate). The SambVca 2.0 program was used to calculate Δ%VBur and generate maps of the steric bulk.48

modulation of steric and electronic influence; (III) 4-Ph systems substituted in the para-position by electron-donating [Me (M10) and tBu (M11)] or electron-withdrawing groups [F (M9) or CF3 (M12)] allow for clear identification of electronic trends; direct steric influences can be excluded due to the remoteness of this position; (IV) systems with additional steric crowding close to the active pocket, i.e., mesityl (M13) or 3,5-di-tBu-Ph (M14) substitution; (V) 4-Ph systems with fluorine atoms close to the active pocket, i.e., substitution in the meta-position (M15) or ortho-position (M16, C6F5) allowing insight into steric, electronic and potentially H−F contact effects; (VI) to test the influence of substituents on the silicon bridge, M17 and M18 feature an SiEt2 bridge; all other systems possess an SiMe2 bridge. The performance of M1, M2, and M8 in propene homopolymerization has been described in the literature, although in varying detail.25,50 M11 and M14 have been described by the groups of Linnolathi and Resconi, but in a heterogeneous system.51 M6, M7, M9, M10, M13, M15, M17, M18, and M19 have been described only in the patent literature in little detail.24,26,27 The synthesis of M4 has been described by some of us, but no polymerization performance data are available.21 To our knowledge, M3, M5, M12, and M16 precursors have not been synthesized so far. Rac isomers of metallocenes M12 and M16 were isolated by crystallization of the crude rac/meso mixtures. Rac isomers of metallocenes M3 and M5 were obtained using Jordan’s method.52 Variation of substituent patterns in other positions, i.e., substitution in 2-, 3-, 5-, and 6-position of the indenyl fragment and analysis of cooperative substituent effects, will be the subject of forthcoming publication(s). Polymerization Conditions. All polymerizations were conducted under a standardized set of conditions in toluene. A

■

RESULTS AND DISCUSSION Catalyst Selection and Synthesis. The catalysts were selected to provide a wide variety of substitution patterns, and thereby electronic and steric influences, in the 4-positions of the indenyl groups (see Figure 1) relative to the parent system M1. All catalysts are 2-Me substituted to increase MW.49 In particular, we included the following: (I) different aryl and hetaryl groups, i.e., phenyl (M2), 2-furyl (M3), and 2-thienyl (M4); electronic influences could manifest, imparted by electron-rich aryl fragments; (II) different halogens, i.e., Cl (M5) vs Br (M6) or alkyl, i.e., Me (M7) vs iPr (M8), allowing C


Article

Macromolecules

Figure 2. Bar graphs for stereoselectivity (as ΔΔG‡stereo), regioselectivity (as ΔΔG‡regio), and molecular weight capability (as ΔΔG‡Mn) for the screened catalysts of Figure 1. Polymer melting points (Tm) provided as single yellow dots. Data are taken from Table 1. Catalysts ordered by increasing stereoselectivity.

polymerization temperature of 60 °C (333 K) was chosen. Busico et al. have shown that chain epimerization does not play a role for the parent catalyst M1, if the monomer concentration is high enough (≥2.2 mol/L).53,54 Preliminary studies in our laboratory indicated that Spaleck’s catalyst M2 shows no chain epimerization at even lower propene concentrations. Two propene partial pressures (65 and 95 psi) were tested in the present work to verify that this is the case for all catalysts in this study. Stereoregularity of the polymers is identical within the experimental error (see the Supporting Information), implying that chain epimerization does not affect stereoselectivity appreciably. Triisobutylaluminum/HNMe2Ph+[B(C 6F 5) 4 ]− (TIBA/AB) was chosen as the scavenger/ alkylator/activator system for better control of the polymerization kinetics. All catalysts, except M18, were also tested using TIBA/TTB, i.e., triisobutylaluminum and trityl tetrakis(pentafluorophenyl)borate ([Ph3C+][B(C6F5)4]−), at 95 psi propene partial pressure. Polymerization Results. Table 1 summarizes the main results of our screening. All data are averages of at least duplicate experiments. Catalyst M19 is included in Table 1, but since it was synthesized a posteriori, polymer characteristics will be discussed after rationalizing what drives catalyst performance in these systems. Additional polymerization details can be found in Tables S1 and S2. Figure 2 shows a bar graph representation of stereoerrors, regioerrors, and MW capability for each catalyst and the melting points of all polymers (dots). To bring all error sources (stereoerrors, regioerrors, and chain termination) on the same scale, errors are plotted as Gibbs free energy differences (ΔΔG‡). In the following, we will discuss individual trends in stereoselectivity, regioselectivity, and molecular weight capability. Trends in Stereoselectivity. All activated precatalysts substituted in the 4-position show high isotactic selectivity, regardless of the substituent, with σ values between 0.9917 and 0.9997, which translate to a ΔΔG‡enantio (60 °C) difference of 3.2−5.4 kcal/mol. The unsubstituted catalyst M1 shows a lower stereoselectivity (σ = 0.9865; ΔΔG‡enantio = 2.8 kcal/ mol).50 13C NMR spectra expanded in the methyl region of the PP samples obtained with catalysts M1, M2, M13, and M14 can be found in the Supporting Information (Figures S11 and S12).

Simple substitution of the hydrogen in the 4-position with alkyl or halogen substituents increases σ substantially, but although the substituent size increases in the order Cl (M5) < Br (M6) < Me (M7) < iPr (M8),55 the stereoselectivity remains essentially constant (σ ≈ 0.996). Aromatic substituents show more diversity. The furyl-substituted catalyst M3 shows the “poorest” stereoselectivity (σ = 0.9917) in the whole catalysts set, excluding M1. M4 (thienyl) shows a slightly increased stereoselectivity (σ = 0.9971) over alkyl and halogen substituents, while σ reaches 0.9988 (ΔG‡enantio = 4.5 kcal/mol) in the phenyl-substituted system M2. The increase in stereoselectivity for aryl-substituted systems follows the size of the aromatic system, i.e., furyl < thienyl < phenyl. Change of the bridge from Me2Si to Et2Si has no influence on the stereoselectivity (M6 vs M17 and M7 vs M18). Variation of the substituents on Ph has different effects. Para-substitution on Ph has negligible effects on the stereoselectivity (ΔΔG‡enantio = ±0.1 kcal/mol, compared to M2). Steric bulk, i.e., H (M2) < F (M9) < Me (M10) < CF3 (M12) < tBu (M11), or electronic effects are therefore irrelevant. Conversely, variation of the meta substituents on Ph affects the stereoselectivity, for M15 (meta-F) σ = 0.9984, i.e., ΔG‡enantio = 4.2 kcal/mol, but for M14 (meta-tBu) σ = 0.9994, i.e., ΔG‡enantio = 4.8 kcal/mol. It appears that this does not follow the substituent size H < F ≪ tBu. Introduction of a substituent in the ortho-position on Ph, which points directly into the active pocket, also increases stereoselectivity. M16 (C6F5) reaches σ = 0.9994 while stereoselectivity for M13 (2,4,6-Me3) is σ = 0.9997, ΔG‡enantio = 5.4 kcal/mol. This appears to follow the size of the substituent in the orthoposition H < F < Me. M13, M14, and M16 reach the performance of the best known zirconocenes. The group of Linnolahti has recently reported that rac-Me2Si(2-Me-4-Ph-5-OMe-6-tBu-Ind)2ZrCl2 yields iPP with an isotacticity mmmm of 99.75 (70 °C), corresponding to σ = 0.9995, albeit with a low regioregularity (amount of regioirregular 2,1 units, 1.6%).20 The high performance of Rieger’s ultrarigid metallocenes has been in part explained by a repulsive interaction of the methoxy substituent in the 7-position of the indenyl with the SiMe2 bridge, which lowers the bite angle of the metallocene.12 The ligand framework of M14 is identical to Rieger’s framework D


Article

Macromolecules

trityl tetrakis(pentafluorophenyl)borate ([Ph3C+][B(C6F5)4], TTB) are used (see Table S1). Activation using AB leads to the aniline complex of the activated species; i.e., aniline can compete with olefin coordination. This is not the case for TTB. We retested the catalysts using the latter activator and polymer microstructure turned out to be identical, regardless of which activator was used (see the Supporting Information). Activities given for activation with TTB should be regarded with some caution and likely only correspond to a lower limit, as we did not attempt to optimize performance. Although catalysts M3 (furyl) and M4 (thienyl) possess donor atom substituents (O or S), their performance is not markedly worse, as alkylaluminum binding to the donor atom of a heterocyclic aromatic system is endergonic (see Tables S5 and S6).57 Using a Single Molecular Descriptor To Predict Stereoselectivity. Steric properties of stereoselective olefin polymerization catalysts can be measured using the percentage of buried volume, %VBur, as a molecular descriptor, which essentially measures the steric bulk in the first coordination sphere of the metal.58 Cavallo originally recommended a sphere radius of 3.5 or 4.0 Å for calculation of the %VBur for metallocenes based on a benchmark of metal−NHC bond dissociation energies. However, he also noted that a sphere radius of 3.5 Å is too small to observe noticeable differences in stereoselectivity predictions.59 To calculate %VBur, we used the DFT structures of the dichloride precursors. To partially cover the substituent in the 4-position, i.e., the ipso atom and the ortho atoms, and substituents thereon, which play a crucial role in determining stereoselectivity, a larger sphere is needed (Figure 3). Making

but omits the methoxy group. Although some caution should be used comparing the performance of catalysts under nonidentical conditions, as mentioned in the Introduction, M14 shows increased stereoselectivity compared to Rieger’s catalyst (σ = 0.9995 vs σ ≈ 0.9983 at 60 °C, for both the Zr and Hf derivative),56 implying that the additional 7-MeO substitution might, in fact, be counterproductive with respect to stereoselectivity. Regioselectivity. The iPP produced with M1−M18 shows 0.17−0.90% regioerrors. This translates to a range of ΔΔG‡regio of ≈0.9 kcal/mol for the set of 18 catalysts, i.e., a much narrower range than observed for stereoselectivity (≈ 2.6 kcal/ mol). Trends follow qualitatively those observed for stereoselectivity. Substituents that only lead to a modest increase of stereoselectivity like alkyl (M7, M8, and M18), halogen (M5, M6, and M17), and furyl/thienyl show the largest amounts of regioerrors (0.5−0.9%), while all phenyl-substituted systems produce iPP with lower amounts of regioerrors (0.17−0.42%). The unsubstituted M1 shows only 0.29% regioerrors, and only catalysts with very high stereoselectivity (M14 and M16) exceed this selectivity. M13 is an exception here and shows 0.40% regioerrors despite having a high stereoselectivity. No correlation with electronic properties of the substituent in the 4-position can be observed, as electron-withdrawing substituents like Br (M6, 0.82%) and Cl (M5, 0.53%) produce iPP with less or more regioerrors than electron donating substituents like Me (M7, 0.67%) and iPr (M8, 0.74%). Moreover, the electron-rich arene-substituted systems M3 (0.90%) and M4 (0.50%) can be found toward the low as well as the high end of this series. Substitution on Ph does not yield clear electronic trends either. Regioselectivity appears to increase, if steric bulk of the substituents decreases, looking at the pairs M5/M6 (Br > Cl) and M7/M8 (iPr > Me), but the trend for M3, M4, M2, and M16 (furyl < thienyl < Ph < C6F5) indicates the opposite. Change of the bridge from Me2Si to Et2Si does not affect regioselectivity (M6 vs M17 and M7 vs M18). MW Capability. Also here, we observe that trends in Mw follow qualitatively the trends observed in stereoselectivity; i.e., the higher the stereoselectivity, the higher the molecular weight. M13 (mesityl) is an outlier and produces a short polymer (169 kDa), despite having a very high stereoselectivity. The Me (M7, M18) and iPr (M8) substituted systems and the unsubstituted M1 produce iPP with the lowest molecular weights (160−210 kDa). The Cl (M5) and Br (M6, M17) systems show a somewhat increased Mw (270−290 kDa). M3 (furyl), the smallest aryl-substituted system, shows a similar performance (265 kDa). All other aryl systems show decent to very good (474−1139 kDa) Mw capabilities at 60 °C. Electron-donating substituents (Me, tBu) in the para-Ph position (M10, M11) appear to decrease Mw somewhat (≈550 kDa vs 621 for M2), while electron-withdrawing substituents on Ph (F, CF3, but also C6F5) increase Mw (M2 < M9 < M15 < M12 ≪ M16, 621−885 kDa). M14, which brings bulky tBu groups from the outside near the active pocket, yields iPP with the highest Mw, 1139 kDa. Activity. The focus of this paper is on the microstructure of iPP produced with zirconocenes M1−M19. We used triisobutylaluminum/HNMe2Ph+[B(C6F5)4] (TIBA/AB) as the scavenger/alkylator/activator system for better control of the polymerization kinetics. Resulting activities in this case are low and not comparable to what can be expected of the catalysts when activators like MAO or triisobutylaluminum/

Figure 3. Different sphere sizes used to calculate Δ%VBur. Left 3.5 Å, right 5.0 Å. Only in the latter case, the substituent in the 4-position is partially covered.

use of the high-quality data set in this paper, we decided to benchmark %VBur for metallocenes and found a radius of 5.0 Å to give optimal results (Table S3 and Figure S9).60 The resulting correlation between stereoselectivity (as ΔΔG‡enantio,60 °C = RT(ln σ/(1 − σ)) and Δ%VBur is good (R2 = 0.88, Figure 4) and allows prediction of stereoselectivity with an average mean deviation from experiment (MAD) of 0.18 kcal/mol (ΔΔG‡enantio), 0.0007 (σ), or 0.3 (mmmm%) using eq 1: ΔΔGenantio,60 ° C = 0.4281 × Δ%VBur − 2.5168

(1)

61,62

A leave-one-out cross-validation (LOOCV, see Table S4 and Figure S10 for more details) shows the validity of the E


Article

Macromolecules

Figure 4. Experimental stereoselectivity, as ΔΔG‡enantio,60 °C = RT(ln σ/(1 − σ)), vs Δ%VBur (R2 = 0.88). A sphere size of 5.0 Å was used to calculate Δ%VBur.

Figure 5. Aryl−indenyl torsion angle and buried volume analysis for dichloride precursors (M4, M2, M16, and M13) and lowest re-insertion TS (M3). In the case of M3, the precursor geometry differs significantly from the one found in the active species.67 Torsion angles in deg; buried volume in %. Sphere radius used to calculate buried volume = 5.0 Å. Spheres were generated defining Zr as the center of the sphere, Si−Zr as the zaxis and Si−Zr−CSi as the xz-plane. H atoms were included in the analysis. %VBur = absolute buried volume, Δ%VBur = difference in buried volume between occupied, and “free” quadrants = [(sum of buried volume in occupied quadrants Q1 and Q3 minus buried volume in “empty” quadrants Q2 and Q4)/2].

model. The cross-validated R2 (Q2) = 0.84 is good. A plot of predicted versus experimentally observed stereoselectivity for each iteration of the LOOCV shows a slope close to 1 and goes nearly through the origin (y = 0.9011x + 0.3958, R2 = 0.82; y = 0.9935 and R2 = 0.81 when constrained). To understand why this correlation works so well, let us consider what drives the increase in stereoselectivity in these rac-R′2Si(2-Me-4-R-indenyl)2ZrCl2 systems on the example of the series M3, M4, M2, M16, and M13. Maps of the steric bulk as measured by the buried volume in the four quadrants of the precatalysts are shown in Figure 5. We propose that tuning of the dihedral angle of aryl substituents in the 4-position has a pronounced effect on stereoselectivity.63 As the dihedral angle is increased from 23° to 69°, the parts of the substituent that extend from it also into the “empty” quadrants recede, which increases the dissimilarity between the quadrants (increase in Δ%VBur from 15.6 to 23.2)

and increases chiral recognition. Note that a dihedral angle is not an energy criterion, and a quantitative agreement cannot be expected. Similar to M3 and M4, spherical substituents like Cl, Br, and Me (M6, M5, and M7) also bring considerable steric bulk into the “empty” quadrant, and the resulting stereoselectivity is at the lower end of the range for the catalysts discussed here.64 As can be seen in Table 2, calculated stereoselectivity of insertions into M−Me and M−iPr bonds increases dramatically in the series M3 < M4 < M2 < M16 < M13, which implies that the increasing the difference between the quadrants in these already highly selective systems increases direct ligand−monomer interactions.65,66 Cavallo has shown that smaller spheres can capture the general performance of different catalysts classes, but at the same time, they should not be able to capture the effects of subtle positional tuning. Larger spheres, on the other hand, as F


Article

Macromolecules Table 2. Calculated Stereoselectivity (ΔΔG‡enantio,DFT in kcal/mol) of Propene Insertion into Zr−Me and Zr−iPr Bonds (Level of Theory M06-2X(PCM)/TZ//TPSSTPSS/ DZ, 333 K)

Amassing steric bulk near the equatorial plane of the catalyst should increase all insertion barriers. Experimentally, only a very low productivity is observed. A positive coupling of stereoselectivity increase and regioselectivity/MW capability cannot be expected anymore in such a situation (vide infra). Why Do Stereoselectivity, Regioselectivity, and MW Capability Correlate in 4-Substituted Metallocenes? Stereoselectivity is of course solely based on steric factors. Neither regioselectivity nor MW capability should be solely influenced by steric factors. Resconi and Cavallo found that regioselectivity trends in metallocenes were predominantly based on sterics by comparing metallocenes with different methyl substitution patterns,68,69 but this explanation fails for the broader catalyst set in the present paper (see earlier discussion and the Supporting Information). The question arises why a correlation among σ, [2,1], and MW capability can be observed. A plausible scenario in which a coupling between three independent experimental properties can be expected is shown in Figure 8. The increase of the dissimilarity of the quadrants by gradual recline of the 4-substituent out of the “empty” quadrants into the “blocked” quadrants disfavors TS leading to stereoerrors and also likely favors the preferred insertion pathway. Figure 8 emphasizes this: steric bulk extending into empty quadrants interferes with the incoming monomer, removing this steric bulk will make the preferred insertion easier with respect to competing transition states. TS leading to regioerrors or chain termination after 2,1 insertion on the other hand are not affected as the methyl group of propene is oriented away from the 4-substituents. In such a situation one can expect that all performance indicators (σ, 2,1, and Mw) are positively influenced. The fact that both Mw and regioerror trends span a similar range (≈1.0 kcal/mol) favors this explanation. Trends in stereoselectivity are much more pronounced (>2.0 kcal/mol) over the range where the correlation is valid. Spaleck noted that the lowering of propagation barriers without affecting chain termination barriers might be responsible for increased molecular weights.50 We have recently shown that the increase of Mw in CGC type or McConville type catalyst systems similarly follows reduced barriers for insertion, not increased barriers for chain transfer to monomer.43 Activities/productivities are not necessarily affected.70 Electronic effects on regioselectivity and MW capability appear to be much less pronounced for substituents in the 4-position.

ΔΔG‡enantio,DFT(re‑si) catalyst M13 M16 M2 M4 M3

a

Me

iPr

2.4 1.3 0.4 0.3 −0.1

2.5 1.5 0.7 0.5 −0.8

a

Here, M13 was modeled without the para-Me groups.

used here, appear to capture subtle positional tuning, but at the same time the general applicability is likely lost. We tentatively conclude that adapting the sphere radius for a given narrow catalyst class can pose a way to accurately predict stereoselectivity using a single molecular descriptor. Connection of Stereoselectivity, Regioselectivity, and MW Capability. Figures 6 and 7 show correlation plots of stereo- vs regioselectivity and stereoselectivity vs Mn, respectively. All 4-substituted systems are included. At first sight it might appear surprising that there is a correlation. Stereo- and regioselectivity correlate surprisingly well (R2 = 0.71), considering that both experimental parameters do not usually correlate with each other in metallocene-catalyzed olefin polymerization. The correlation becomes even better when the overly hindered and rigid catalyst M13 is excluded from the set (R2 = 0.93). Stereoselectivity and Mw do not correlate well if all catalysts are included in the plot (R2 = 0.23) but surprisingly well if again M13 is excluded (R2 = 0.80). Because stereoselectivity correlates well with both regioselectivity and Mw, the correlation between the latter two is also good (R2 = 0.80; see Figure S8, M13 excluded). The catalyst with the highest stereoselectivity, M13, is an exception here, and we call it “overtuned”. It does not fit well in the correlations. It differs from all other catalysts in the set for two main reasons. The buried volume maps in Figure 5 show that while the substituent pattern for M13 perfectly increases the dissimilarity between the “open” and “occupied” quadrants, it also amasses steric bulk near the equatorial plane and the center of the catalyst which other catalysts do not. Furthermore, the 2,6-dimethyl substitution pattern locks the ring in place and makes the ligand framework much more rigid.

Figure 6. Regioselectivity (ΔΔG‡regio,60 °C) vs stereoselectivity (ΔΔG‡enantio,60 °C) in kcal/mol. For catalysts M1−M18, blue line, R2 = 0.71; excluding “overtuned” catalyst M13, orange line, R2 = 0.93. G


Article

Macromolecules

Figure 7. MW capability (ΔΔG‡Mn,60 °C) vs stereoselectivity (ΔΔG‡enantio,60 °C) in kcal/mol. For catalysts M1−M18, blue line, R2 = 0.23; excluding “overtuned” catalyst M13, orange line, R2 = 0.80.

Figure 8. Increase of the dissimilarity between the quadrants lowers the preferred insertion TS and increases the lowest TS leading to stereoerrors by similar amounts. TS leading to regioerrors and chain transfer TS are unaffected. As a result, the relative energy differences for TS leading to chain termination or regioerrors increase, but by a smaller amount.

In summary, it appears that steric tuning of the substituents in the 4-position can increase stereoselectivity, regioselectivity, and molecular weight at the same time by making the preferred insertion easier. This effect is likely responsible for the good correlation of all three performance indicators. Overall Catalyst Performance and Performance of M19. The observed correlation between stereoselectivity/ regioselectivity/molecular weight capability means that an increase in one catalyst performance indicator does not necessarily come with trade-offs in the other performance indicators. M14 and M16 deliver the best overall performance of the whole catalyst set; they both give similar stereo- and regioselectivity, but the former gives better molecular weight.

On the other end of the spectrum, M3 and M7 produce one of the shortest and most “flawed” polymers in the test set. The connection of the dihedral angle of 4-aryl substituents, predictability of quadrant steric bulk, and detrimental effect of additional steric bulk in equatorial plane of the catalyst prompted us to test several other catalysts computationally. M19 with a 4-o-tolyl substituent avoids overtuning of the equatorial catalyst plane and possesses an even larger dihedral angle than M13, ensuring separation of steric bulk. While the predicted stereoselectivity remains high (σ > 0.9994), we expected Mw to significantly increase. Ad hoc synthesis and testing of M19 to verify the “model” confirmed the predictions. The catalyst shows the highest stereoselectivity reported so far (σ = 0.9999) for metallocenes at 60 °C, while maintaining a H


Macromolecules

■

high-MW capability (Mw > 1 MDa) and relatively high regioselectivity. The 13C NMR spectrum of an iPP sample prepared with catalyst M19 is shown in Figure S12. Although comparison is hampered by the difference in polymerization conditions, it appears that M19 outperforms Rieger’s metallocenes (Zr/Hf: σ ≈ 0.9984/0.9984 at 60 °C and Mw ≈ 440/ 900 kDa), 12,13 Spaleck’s rac-Me 2 Si(2-Me-4-naphthylindenyl)2ZrCl2 (σ = 0.9982 at 70 °C and Mw = 920 kDa),50 and rac-Me2Si(2-nPr-4-(9-phenanthrylindenyl))2ZrCl2 (σ ≈ 0.9991 at 0 °C and M w = 140 kDa) in terms of stereoselectivity, showing an order of magnitude less stereoerrors.71 The overall performance of the leading catalyst surpasses that of previously known catalysts that benefit from substitution in multiple positions around the indenyl fragment. With respect to the whole set, the performance of these 19 catalysts, differing in the substituent in just a single position, spans a dramatic range (up to 3.3 kcal/mol in ΔΔG‡enantio). We are therefore hopeful that similar studies on different substituent positions (2, 3, 5, 6, and 7-position and bridge) and subsequent combination of the most beneficial substituent patterns will ultimately unlock even better high temperature− high performance catalysts.

■

Article

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.macromol.8b01546. Synthesis of precatalysts M3, M5, M12, M16, M18, and M19; full Gaussian citation, additional polymerization results, full polymer characterization table, regioselectivity/Mw correlation, buried volume analysis (5.0 Å), buried volume analysis−benchmarking, LOOCV; final energies, enthalpies, and Gibbs free energies for all DFT calculations (PDF) DFT structures (XYZ)

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected] (R.C.). *E-mail: [email protected] (D.U.). *E-mail: [email protected] (C.E.). ORCID

Christian Ehm: 0000-0002-2538-5141 Georgy P. Goryunov: 0000-0002-5188-5799 Peter H. M. Budzelaar: 0000-0003-0039-4479 Roberta Cipullo: 0000-0003-3846-1999

CONCLUSIONS AND OUTLOOK

Author Contributions

Cavallo and Talarico have recently concluded in the context of molecular weight tuning that there is not enough space to further improve the ratio between insertion and chain transfer to the monomer, at least for metallocenes.14 This reflects a sentiment in the scientific community that “all has been said” concerning zirconocenes in olefin polymerization. The results presented here indicate the opposite, showing a successful extension of the performance of Rieger’s catalysts in terms of stereoselectivity from ∼0 to 60 °C, which requires an increase of the stereoselectivity of almost 2 kcal/mol. This is a first step on the way to increase the high-temperature tolerance of C2symmetric zirconocenes to levels where they can still produce iPP of commercial value (high melting points) while allowing industry to operate reactors under desirable high-temperature conditions (>100 °C). Good correlations between stereoselectivity, regioselectivity, and MW capability have been identified for 19 C2-symmetric zirconocenes of the type rac-R′2Si(2-Me-4-R-indenyl)2ZrCl2. Trends in all three performance indicators are predominantly governed by steric effects, with the correlations attributed to lowering of the preferred insertion TS. Although shown here only for one catalysts class, the concept should also extend to other classes of olefin polymerization catalysts. Δ%VBur, i.e., the difference in percent buried volume between the “occupied” and “empty” quadrants, has been identified as a suitable computational descriptor allowing prediction of stereoselectivity with high accuracy, provided that the sphere size used for prediction is optimized. The pronounced influence of tuning in the 4-position of the indenyl hints that further improvements might be possible, i.e., by increasing the dihedral angle of aryl substituents without overtuning the catalyst like it was observed for catalyst M13. Catalyst M19 represents a first proof of concept in this regard. Finding optimal substituents via extensive screening of other substituent positions followed by the combination of desired features (cooperative substituent effects) can ultimately lead to further catalyst improvements.

C.E. and A.V. contributed equally. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This research forms part of the research program of DPI, project 800. The authors thank SABIC for the donation of some catalysts.

■

ABBREVIATIONS HTE, high throughput experimentation; QSAR, quantitative structure−activity relationship; GPC, gel permeation chromatography; NMR, nuclear magnetic resonance; PPR, parallel pressure reactors; TIBA, triisobutylaluminum; AB, anilinium borate (HNMe2Ph+ [B(C6F5)4]−); TTB, trityl tetrakis(pentafluorophenyl)borate; MAO, methylaluminoxane; NHC, N-heterocyclic carbene; %VBur, difference in buried volume between occupied and “free” quadrants; MW, molecular weight; σ, probability that the preferred propene enantioface is selected at an active site of given chirality; [2,1], mol % 2,1 regioerrors; [3,1], mol % 3,1 regioerrors.

■

REFERENCES

(1) Kaminsky, W.; Külper, K.; Brintzinger, H. H.; Wild, F. R. W. P. Polymerization of Propene and Butene with a Chiral Zirconocene and Methylalumoxane as Cocatalyst. Angew. Chem., Int. Ed. Engl. 1985, 24 (6), 507−508. (2) Ewen, J. A. Mechanisms of stereochemical control in propylene polymerizations with soluble Group 4B metallocene/methylalumoxane catalysts. J. Am. Chem. Soc. 1984, 106 (21), 6355−6364. (3) Resconi, L.; Cavallo, L.; Fait, A.; Piemontesi, F. Selectivity in Propene Polymerization with Metallocene Catalysts. Chem. Rev. 2000, 100 (4), 1253−1346. (4) Busico, V.; Cipullo, R. Microstructure of polypropylene. Prog. Polym. Sci. 2001, 26 (3), 443−533. (5) Corradini, P.; Barone, V.; Fusco, R.; Guerra, G. Analysis of models for the Ziegler-Natta stereospecific polymerization on the I


Article

Macromolecules basis of non-bonded interactions at the catalytic siteI. The Cossee model. Eur. Polym. J. 1979, 15 (12), 1133−1141. (6) Corradini, P.; Guerra, G.; Cavallo, L. Do New Century Catalysts Unravel the Mechanism of Stereocontrol of Old Ziegler−Natta Catalysts? Acc. Chem. Res. 2004, 37 (4), 231−241. (7) Brintzinger, H. H.; Fischer, D.; Mülhaupt, R.; Rieger, B.; Waymouth, R. M. Stereospecific Olefin Polymerization with Chiral Metallocene Catalysts. Angew. Chem., Int. Ed. Engl. 1995, 34 (11), 1143−1170. (8) Tullo, A. H. Metallocenes rise again. Chem. Eng. News 2010, 88 (42), 10−16. (9) Busico, V. Metal-catalysed olefin polymerisation into the new millennium: a perspective outlook. Dalton Trans 2009, 41, 8794− 8802. (10) Baier, M. C.; Zuideveld, M. A.; Mecking, S. Post-Metallocenes in the Industrial Production of Polyolefins. Angew. Chem., Int. Ed. 2014, 53 (37), 9722−9744. (11) Busico, V. Catalytic Olefin Polymerization is a Mature Field. Isn’t it? Macromol. Chem. Phys. 2007, 208 (1), 26−29. (12) Schöbel, A.; Herdtweck, E.; Parkinson, M.; Rieger, B. UltraRigid Metallocenes for Highly Iso- and Regiospecific Polymerization of Propene: The Search for the Perfect Polypropylene Helix. Chem. Eur. J. 2012, 18 (14), 4174−4178. (13) Machat, M. R.; Lanzinger, D.; Pöthig, A.; Rieger, B. Ultrarigid Indenyl-based Hafnocene Complexes for the Highly Isoselective Polymerization of Propene: Tunable Polymerization Performance Adopting Various Sterically Demanding 4-Aryl Substituents. Organometallics 2017, 36 (2), 399−408. (14) Falivene, L.; Cavallo, L.; Talarico, G. Buried Volume Analysis for Propene Polymerization Catalysis Promoted by Group 4 Metals: A Tool for Molecular Mass Prediction. ACS Catal. 2015, 5, 6815−6822. (15) Rappé, A. K.; Skiff, W. M.; Casewit, C. J. Modeling MetalCatalyzed Olefin Polymerization. Chem. Rev. 2000, 100 (4), 1435− 1456. (16) Severn, J. R.; Chadwick, J. C.; Duchateau, R.; Friederichs, N. Bound but Not Gagged”Immobilizing Single-Site α-Olefin Polymerization Catalysts. Chem. Rev. 2005, 105 (11), 4073−4147. (17) Busico, V.; Cipullo, R.; Mingione, A.; Rongo, L. Accelerating the Research Approach to Ziegler−Natta Catalysts. Ind. Eng. Chem. Res. 2016, 55 (10), 2686−2695. (18) Vittoria, A.; Meppelder, A.; Friederichs, N.; Busico, V.; Cipullo, R. Demystifying Ziegler−Natta Catalysts: The Origin of Stereoselectivity. ACS Catal. 2017, 7, 4509−4518. (19) Cruz, V. L.; Martinez, S.; Ramos, J.; Martinez-Salazar, J. 3DQSAR as a Tool for Understanding and Improving Single-Site Polymerization Catalysts. A Review. Organometallics 2014, 33 (12), 2944−2959. (20) Kuklin, M. S.; Virkkunen, V.; Castro, P. M.; Izmer, V. V.; Kononovich, D. S.; Voskoboynikov, A. Z.; Linnolahti, M. Quantitative structure−property relationships in propene polymerization by zirconocenes with a rac-SiMe2[Ind]2 based ligand framework. J. Mol. Catal. A: Chem. 2016, 412 (C), 39−46. (21) Izmer, V. V.; Lebedev, A. Y.; Nikulin, M. V.; Ryabov, A. N.; Asachenko, A. F.; Lygin, A. V.; Sorokin, D. A.; Voskoboynikov, A. Z. Palladium-Catalyzed Pathways to Aryl-Substituted Indenes: Efficient Synthesis of Ligands and the Respective ansa-Zirconocenes. Organometallics 2006, 25 (5), 1217−1229. (22) Nikulin, M. V.; Tsarev, A. A.; Lygin, A. V.; Ryabov, A. N.; Beletskaya, I. P.; Voskoboinikov, A. Z. Palladium-catalyzed arylation of bis(4-bromo-2-methylinden-1-yl)dimethylsilane and related compounds. Russ. Chem. Bull. 2008, 57 (11), 2298. (23) Zhang, X.; Zhu, Q.; Guzei, I. A.; Jordan, R. F. General Synthesis of Racemic Me2Si-Bridged Bis(indenyl) Zirconocene Complexes. J. Am. Chem. Soc. 2000, 122 (33), 8093−8094. (24) Voskoboynikov, A. Z.; Ryabov, A. N.; Nikulin, M. V.; Lygin, A. V.; Izmer, V. V.; Asachenko, A. F.; Coker, C. L.; Canich, J. A. M. US Pat. US7446216. (25) Spaleck, W.; Antberg, M.; Rohrmann, J.; Winter, A.; Bachmann, B.; Kiprof, P.; Behm, J.; Herrmann, W. A. High Molecular Weight

Polypropylene through Specifically DesigneZirconocene Catalysts. Angew. Chem., Int. Ed. Engl. 1992, 31 (10), 1347−1350. (26) Voskoboynikov, A. Z.; Ryabov, A. N.; Coker, C. L.; Canich, J. A. M. US Pat. US7709670. (27) Voskoboynikov, A. Z.; Nikulin, M. V.; Ryabov, A. N.; Lygin, A. V.; Coker, C. L.; Canich, J. A. M. US Pat. US7910783. (28) Bingel, C.; Goeres, M.; Fraaije, V.; Winter, A. US Pat. US6492539. (29) Gaussian 09, Revision B.1; for the full citation see the Supporting Information. (30) Baker, J. Parallel Quantum Solutions, 2.4 ed.; Fayetteville, AR, 2001. (31) Baker, J. An algorithm for the location of transition states. J. Comput. Chem. 1986, 7 (4), 385−395. (32) Budzelaar, P. H. M. Geometry optimization using generalized, chemically meaningful constraints. J. Comput. Chem. 2007, 28 (13), 2226−2236. (33) Ehm, C.; Budzelaar, P. H. M.; Busico, V. Calculating accurate barriers for olefin insertion and related reactions. J. Organomet. Chem. 2015, 775 (0), 39−49. (34) Tao, J. M.; Perdew, J. P.; Staroverov, V. N.; Scuseria, G. E. Climbing the density functional ladder: Nonempirical metageneralized gradient approximation designed for molecules and solids. Phys. Rev. Lett. 2003, 91 (14), 146401. (35) Balabanov, N. B.; Peterson, K. A. Systematically convergent basis sets for transition metals. I. All-electron correlation consistent basis sets for the 3d elements Sc−Zn. J. Chem. Phys. 2005, 123, 064107. (36) Balabanov, N. B.; Peterson, K. A. Basis set limit electronic excitation energies, ionization potentials, and electron affinities for the 3d transition metal atoms: Coupled cluster and multireference methods. J. Chem. Phys. 2006, 125, 074110. (37) Schuchardt, K. L.; Didier, B. T.; Elsethagen, T.; Sun, L.; Gurumoorthi, V.; Chase, J.; Li, J.; Windus, T. L. Basis Set Exchange: A Community Database for Computational Sciences. J. Chem. Inf. Model. 2007, 47 (3), 1045−1052. (38) Ehm, C.; Cipullo, R.; Budzelaar, P. H. M.; Busico, V. Role(s) of TMA in polymerization. Dalton Trans 2016, 45, 6847−6855. (39) Zaccaria, F.; Ehm, C.; Budzelaar, P. H. M.; Busico, V. Accurate Prediction of Copolymerization Statistics in Molecular Olefin Polymerization Catalysis: The Role of Entropic, Electronic, and Steric Effects in Catalyst Comonomer Affinity. ACS Catal. 2017, 7, 1512−1519. (40) Zaccaria, F.; Cipullo, R.; Budzelaar, P. H. M.; Busico, V.; Ehm, C. Backbone rearrangement during olefin capture as the rate limiting step in molecular olefin polymerization catalysis and its effect on comonomer affinity. J. Polym. Sci., Part A: Polym. Chem. 2017, 55 (17), 2807−2814. (41) Ehm, C.; Budzelaar, P. H. M.; Busico, V. Metal−carbon bond strengths under polymerization conditions: 2,1-insertion as a catalyst stress test. J. Catal. 2017, 351, 146−152. (42) Ehm, C.; Cipullo, R.; Passaro, M.; Zaccaria, F.; Budzelaar, P. H. M.; Busico, V. Chain Transfer to Solvent in Propene Polymerization with Ti Cp-phosphinimide Catalysts: Evidence for Chain Termination via Ti−C Bond Homolysis. ACS Catal. 2016, 6, 7989−7993. (43) Ehm, C.; Budzelaar, P. H. M.; Busico, V. Tuning the Relative Energies of Propagation and Chain Termination Barriers in Polyolefin Catalysis through Electronic and Steric Effects. Eur. J. Inorg. Chem. 2017, 2017, 3343−3349. (44) Whitten, J. L. Coulombic potential energy integrals and approximations. J. Chem. Phys. 1973, 58, 4496. (45) Baerends, E. J.; Ellis, D. E.; Ros, P. Self-consistent molecular HartreeFockSlater calculations I. The computational procedure. Chem. Phys. 1973, 2 (1), 41−51. (46) Feyereisen, M.; Fitzgerald, G.; Komornicki, A. Use of approximate integrals in ab initio theory. An application in MP2 energy calculations. Chem. Phys. Lett. 1993, 208 (5), 359−363. J


Article

Macromolecules

lowest re and si insertion TS for the catalysts shown in Figure 5 reveals that for catalysts M2, M3, M4, and M16 the torsion angle of the aryl fragment does not deviate more than 4° in the lowest re and si insertion TS from the dichloride structure. The difference in torsion angles for the re and si insertion TS in each system is less than 1° for M2, M4, and M16 and 5° for the much smaller furyl fragment in M3. Therefore, deviations expected from restricting the discussion on dichloride precursors are much smaller than the spread in the aryl torsion angles (23°−70°). (64) Spaleck has argued that the increased stereoselectivity of racMe2Si(2-Me-4-naphthyl-Ind)2ZrCl2 “might be due to the presumably high rotational barrier for the substituent”. The dihedral angle for this catalyst is 63°, and it appears that the increased stereoselectivity can be traced back directly to steric changes, not to changes in the rotational barrier height. (65) Talarico, G.; Budzelaar, P. H. M. Analysis of Stereochemistry Control in Homogeneous Olefin Polymerization Catalysis. Organometallics 2014, 33 (21), 5974−5982. (66) The lowest insertion pathway in the present systems is reinsertion in case of R,R-zirconocenes. Talarico and Budzelaar attempted separation of contributions to stereoselectivity for a “plain” rac-bis(R-indenyl)SiMe2ZrCl2 system: (1) the preference of the olefin to insert anti with respect to the growing chain contributes ≈ 2−3 kcal/mol to ΔΔG‡enantio,DFT, (2) interactions of ligand framework and si-oriented propene amount to ≈ 1.5 kcal/mol, and (3) chain orientation contributes ≈1.7 kcal/mol. The balance of these factors depends on the orientation of chain and incoming olefin in the respective transition state. However, for insertions into Zr−iPr bonds or Zr−Me bonds (1) and (3) do not contribute, and it can be reasonably assumed that ΔΔG‡enantio,DFT reflects mostly ligand− monomer interactions. (67) The orientation of the furyl rings in the precursor and the active species differs significantly (see xyz files in the Supporting Information). To not introduce arbitrary errors, the insertion TS was chosen for M3. (68) Toto, M.; Cavallo, L.; Corradini, P.; Moscardi, G.; Resconi, L.; Guerra, G. Influence of π-Ligand Substitutions on the Regiospecificity and Stereospecificity in Isospecific Zirconocenes for Propene Polymerization. A Molecular Mechanics Analysis. Macromolecules 1998, 31 (11), 3431−3438. (69) Correa, A.; Talarico, G.; Cavallo, L. Regiochemistry of propene insertion with group 4 polymerization catalysts from a theoretical perspective. J. Organomet. Chem. 2007, 692 (21), 4519−4527. (70) Olefin insertion transition states are energetically and geometrically early, and the Hammond principle implies that the connected resting state is similarly lowered in energy; i.e., absolute barrier heights for preferred insertion do not necessarily change. (71) Kashiwa, N.; Kojoh, S.; Imuta, J.; Tsutsui, T. In Metalorganic Catalysts for Synthesis and Polymerization; Kaminsky, W., Ed.; Springer-Verlag: Berlin, 1999; p 30.

(47) Vahtras, O.; Almlö f , J.; Feyereisen, M. W. Integral approximations for LCAO-SCF calculations. Chem. Phys. Lett. 1993, 213 (5), 514−518. (48) Falivene, L.; Credendino, R.; Poater, A.; Petta, A.; Serra, L.; Oliva, R.; Scarano, V.; Cavallo, L. SambVca 2. A Web Tool for Analyzing Catalytic Pockets with Topographic Steric Maps. Organometallics 2016, 35 (13), 2286−2293. (49) Jüngling, S.; Mülhaupt, R.; Stehling, U.; Brintzinger, H.-H.; Fischer, D.; Langhauser, F. Propene polymerization using homogeneous MAO-activated metallocene catalysts: Me2 Si(Benz[e]Indenyl)2 ZrCl2/MAO vs. Me2 Si(2-Me-Benz[e]Indenyl) 2ZrCl2 / MAO. J. Polym. Sci., Part A: Polym. Chem. 1995, 33 (8), 1305−1317. (50) Spaleck, W.; Kueber, F.; Winter, A.; Rohrmann, J.; Bachmann, B.; Antberg, M.; Dolle, V.; Paulus, E. F. The Influence of Aromatic Substituents on the Polymerization Behavior of Bridged Zirconocene Catalysts. Organometallics 1994, 13 (3), 954−963. (51) Kuklin, M. S.; Virkkunen, V.; Castro, P. M.; Resconi, L.; Linnolahti, M. Controlling the Microstructure of Isotactic Polypropene by C2-Symmetric Zirconocene Polymerization Catalysts: Influence of Alkyl Substituents on Regio- and Stereocontrol. Eur. J. Inorg. Chem. 2015, 2015, 4420−4428. (52) LoCoco, M. D.; Zhang, X.; Jordan, R. F. Chelate-Controlled Synthesis of Racemic ansa-Zirconocenes. J. Am. Chem. Soc. 2004, 126 (46), 15231−15244. (53) Busico, V.; Brita, D.; Caporaso, L.; Cipullo, R.; Vacatello, M. Interfering Effects of Growing Chain Epimerization on MetalloceneCatalyzed Isotactic Propene Polymerization. Macromolecules 1997, 30 (14), 3971−3977. (54) Busico, V.; Cipullo, R.; Cutillo, F.; Vacatello, M. MetalloceneCatalyzed Propene Polymerization: From Microstructure to Kinetics. 1. C2-Symmetric ansa-Metallocenes and the “Trigger” Hypothesis. Macromolecules 2002, 35 (2), 349−354. (55) Förster, H.; Vögtle, F. Steric Interactions in Organic Chemistry: Spatial Requirements of Substituents. Angew. Chem., Int. Ed. Engl. 1977, 16 (7), 429−441. (56) Rieger does not directly give polymerizaqtion data at 60 °C; mmmm of 99.20% is the average between the reported mmmm at 50 °C (99.40%) and 70 °C (99.00%). (57) Alkylaluminum compounds are used to scavenge heteroatom contaminants that would otherwise poison the catalyst. However, in difference to tetrahydrofuran, for example, heterocyclic aromatic compounds coordinate only very weakly to aluminum, and coordination is endergonic at 60 °C. The additional steric bulk from the ligand framework renders TMA coordination to M3 (precursor and active species) even more endergonic (see the Supporting Information). (58) Viciu, M. S.; Navarro, O.; Germaneau, R. F.; Kelly, R. A.; Sommer, W.; Marion, N.; Stevens, E. D.; Cavallo, L.; Nolan, S. P. Synthetic and Structural Studies of (NHC)Pd(allyl)Cl Complexes (NHC = N-heterocyclic carbene). Organometallics 2004, 23 (7), 1629−1635. (59) Poater, A.; Cavallo, L. Comparing families of olefin polymerization precatalysts using the percentage of buried volume. Dalton Trans 2009, No. 41, 8885−8890. (60) Spheres of 4.5−5.5 Å radius all give a good correlation. Smaller or larger spheres lead to a fast decrease of R2; see the Supporting Information. (61) Cramer, R. D.; Patterson, D. E.; Bunce, J. D. Comparative molecular field analysis (CoMFA). 1. Effect of shape on the binding of steroids to carrier proteins. J. Am. Chem. Soc. 1988, 110, 5959− 5967. (62) Alexander, D. L. J.; Tropsha, A.; Winkler, D. A. Beware of R2: Simple, Unambigious Assessment of the Prediction Accuracy of QSAR and QSPR Models. J. Chem. Inf. Model. 2015, 55, 1316−1322. (63) One of the key assumptions behind the predictive capability of Cavallo’s %VBur parameter is that the active species does not differ much from the precursor. Although one can expect some mobility of the aryl substituents in the 4-position in the precursor, such mobility is lost in the insertion TS. Comparison of precursor structures and K


Connection of Stereoselectivity, Regioselectivity, and Molecular

Recommend Documents