Preparation of Pincer Hafnium Complexes for Olefin Polymerization

Pincer-type [Cnaphthyl, Npyridine, Namido]HfMe2 complex is a flagship among the post-metallocene catalysts. In this work, various pincer-type Hf-complexes were prepared for olefin polymerization. Pincer-type [Namido, Npyridine, Namido]HfMe2 complexes were prepared by reacting in situ generated HfMe4 with the corresponding ligand precursors, and the structure of a complex bearing 2,6-Et2C6H3Namido moieties was confirmed by X-ray crystallography. When the ligand precursors of [(CH3)R2Si-C5H3N-C(H)PhN(H)Ar (R = Me or Ph, Ar = 2,6-diisopropylphenyl) were treated with in situ generated HfMe4, pincer-type [Csilylmethyl, Npyridine, Namido]HfMe2 complexes were afforded by formation of Hf-CH2Si bond. Pincer-type [Cnaphthyl, Sthiophene, Namido]HfMe2 complex, where the pyridine moiety in the flagship catalyst was replaced with a thiophene unit, was not generated when the corresponding ligand precursor was treated with HfMe4. Instead, the [Sthiophene, Namido]HfMe3-type complex was obtained with no formation of the Hf-Cnaphthyl bond. A series of pincer-type [Cnaphthyl, Npyridine, Nalkylamido]HfMe2 complexes was prepared where the arylamido moiety in the flagship catalyst was replaced with alkylamido moieties (alkyl = iPr, cyclohexyl, tBu, adamantyl). Structures of the complexes bearing isopropylamido and adamantylamido moieties were confirmed by X-ray crystallography. Most of the complexes cleanly generated the desired ion-pair complexes when treated with an equivalent amount of [(C18H37)2N(H)Me]+[B(C6F5)4]−, which showed negligible activity in olefin polymerization. Some complexes bearing bulky substituents showed moderate activities, even though the desired ion-pair complexes were not cleanly afforded.


Introduction
Transition metal pincer complexes have been prepared to discover applications in various areas, especially in organometallic catalysis [1,2]. The tridentate chelating pincer ligand binds a metal in a meridional fashion to form a coplanar structure with the metal at the center. The ligand-metal interaction is tight and inflexible, which confers high stability. For homogeneous olefin polymerization, the initial zirconium (Zr)-based metallocene catalysts were followed by titanium (Ti)-based half-metallocenes and subsequently post-metallocenes with non-cyclopentadienyl ligands [3][4][5][6][7][8][9]. Among the post-metallocenes that have been developed, a pincer-type [C naphthyl , N pyridine , N amido ]HfMe 2 complex is a flagship catalyst. The complex was discovered in the early 2000s through high-throughput screening and high-throughput screening and has since been extensively explored and applied in a commercial process (I in Figure 1) [10,11]. The pincer-type Hf complex I is able to incorporate a large amount of α-olefin in ethylene/α-olefin copolymerizations [12], and is capable of controlling the tacticity of propylene polymerization to produce isotactic polypropylene [13][14][15]. An unique characteristic of I is that the β-elimination process, an intrinsic chain transfer reaction that inevitably occurs in the olefin polymerizations performed with the conventional Zr-based metallocene and Ti-based half metallocene catalysts, is completely prevented [14][15][16]. With these merits, it is possible not only to grow a polyolefin (PO) chain from a Hf-site in a living fashion but also to grow PO chains uniformly from diethylzinc (Et2Zn) deliberately added in excess as a chain transfer agent. The latter is termed coordinative chain transfer polymerization (CCTP) [17][18][19]. The CCTP technique is judiciously utilized in the commercial production of olefin block copolymers at the Dow Chemical Company [10,[20][21][22]. By performing anionic polymerization of styrene in a one-pot reaction after CCTP with I, it is also possible to synthesize polyolefin-polystyrene block copolymers [23][24][25]. In this context, many studies have been performed to detail I and to improve the catalytic performance by modifying the skeleton of I [26][27][28][29][30][31][32][33][34][35]. With an aim to develop upgraded catalyst for I, we prepared various pincer-type Hf-complexes. The results are presented herein.

Results and Discussion
Pincer-type [N amido , N pyridine , N amido ]HfMe2 complexes were prepared (3 and 4 in Scheme 1). Bis(amino)pyridine compounds 1 and 2 were prepared according to the reported method involving the stepwise methylation of the corresponding bis(imino)pyridine compounds with AlMe3 [ 36 ]. Zr and Y complexes have been successfully prepared using 1 [36], but the synthesis of its Hf analogue has not been reported. When 1 and 2 were treated with HfMe4 generated by the treatment of HfCl4 with 4 equiv methylmagnesium bromide (MeMgBr) at -30 °C, the desired pincer-type Hf-complexes 3 and 4 were cleanly generated [37]. In 1 H NMR spectra of 3 and 4, Hf(CH3)2 signals were observed as a singlet at 0.36 and 0.10 ppm, respectively. The structure of 4 was unambiguously confirmed by X-ray crystallography. Metallation of the bis(imino)pyridine compound containing 2,6-Me2C6H3N(H)-moieties was unsuccessful under the same reaction conditions.

Results and Discussion
Pincer-type [N amido , N pyridine , N amido ]HfMe 2 complexes were prepared (3 and 4 in Scheme 1). Bis(amino)pyridine compounds 1 and 2 were prepared according to the reported method involving the stepwise methylation of the corresponding bis(imino)pyridine compounds with AlMe 3 [36]. Zr and Y complexes have been successfully prepared using 1 [36], but the synthesis of its Hf analogue has not been reported. When 1 and 2 were treated with HfMe 4 generated by the treatment of HfCl 4 with 4 equiv methylmagnesium bromide (MeMgBr) at -30 • C, the desired pincer-type Hf-complexes 3 and 4 were cleanly generated [37]. In 1 H NMR spectra of 3 and 4, Hf(CH 3 ) 2 signals were observed as a singlet at 0.36 and 0.10 ppm, respectively. The structure of 4 was unambiguously confirmed by X-ray crystallography. Metallation of the bis(imino)pyridine compound containing 2,6-Me 2 C 6 H 3 N(H)-moieties was unsuccessful under the same reaction conditions. high-throughput screening and has since been extensively explored and applied in a commercial process (I in Figure 1) [10,11]. The pincer-type Hf complex I is able to incorporate a large amount of α-olefin in ethylene/α-olefin copolymerizations [12], and is capable of controlling the tacticity of propylene polymerization to produce isotactic polypropylene [13][14][15]. An unique characteristic of I is that the β-elimination process, an intrinsic chain transfer reaction that inevitably occurs in the olefin polymerizations performed with the conventional Zr-based metallocene and Ti-based half metallocene catalysts, is completely prevented [14][15][16]. With these merits, it is possible not only to grow a polyolefin (PO) chain from a Hf-site in a living fashion but also to grow PO chains uniformly from diethylzinc (Et2Zn) deliberately added in excess as a chain transfer agent. The latter is termed coordinative chain transfer polymerization (CCTP) [17][18][19]. The CCTP technique is judiciously utilized in the commercial production of olefin block copolymers at the Dow Chemical Company [10,[20][21][22]. By performing anionic polymerization of styrene in a one-pot reaction after CCTP with I, it is also possible to synthesize polyolefin-polystyrene block copolymers [23][24][25]. In this context, many studies have been performed to detail I and to improve the catalytic performance by modifying the skeleton of I [26][27][28][29][30][31][32][33][34][35]. With an aim to develop upgraded catalyst for I, we prepared various pincer-type Hf-complexes. The results are presented herein.

Results and Discussion
Pincer-type [N amido , N pyridine , N amido ]HfMe2 complexes were prepared (3 and 4 in Scheme 1). Bis(amino)pyridine compounds 1 and 2 were prepared according to the reported method involving the stepwise methylation of the corresponding bis(imino)pyridine compounds with AlMe3 [ 36 ]. Zr and Y complexes have been successfully prepared using 1 [36], but the synthesis of its Hf analogue has not been reported. When 1 and 2 were treated with HfMe4 generated by the treatment of HfCl4 with 4 equiv methylmagnesium bromide (MeMgBr) at -30 °C, the desired pincer-type Hf-complexes 3 and 4 were cleanly generated [37]. In 1 H NMR spectra of 3 and 4, Hf(CH3)2 signals were observed as a singlet at 0.36 and 0.10 ppm, respectively. The structure of 4 was unambiguously confirmed by X-ray crystallography. Metallation of the bis(imino)pyridine compound containing 2,6-Me2C6H3N(H)-moieties was unsuccessful under the same reaction conditions.  The prototype complex I characteristically contains a Hf-C(aryl) bond. We prepared the related pincer-type [C silylmethyl , N pyridine , N amido ]HfMe 2 complexes containing the Hf-CH 2 Si bond instead of the Hf-C(aryl) bond (Scheme 2). 2-Br-6-(R 2 R'Si)-pyridines (R = Me, Ph, iPr; R' = Me, iPr; 5-7) were prepared from 2,6-dibromopyridine by the treatment of nBuLi and subsequently Me 3 SiCl, Ph 2 (Me)SiCl, or iPr 3 Si(OSO 2 CF 3 ) [38,39]. Compounds 5-7 were treated with 2 equiv tBuLi to generate 2-Li-6-(R 2 R'Si)-pyridines, which were reacted with imine compound PhC(H)=N(2,6-iPr 2 C 6 H 3 ) to generate pyridines substituted with the R 2 R'Si-group and (2,6-iPr 2 C 6 H 3 )N(H)C(Ph)(H)-group at the 2-and 6-positions (8)(9)(10). Treatment of 8 and 9 with in situ generated HfMe 4 afforded the targeted pincer-type [C silylmethyl , N pyridine , N amido ]HfMe 2 complexes containing the Hf-CH 2 Si bond (11 and 12). The 1 H NMR spectrum of 11 distinguished the two diastereotopic protons on HfCH 2 Si moiety at 1.31 and 0.24 ppm as doublets with a large geminal coupling constant (J = 12.6 Hz) ( Figure S7 in Supporting Information). The two methyl groups attached on Si and the two methyl groups attached on Hf were also diastereotopic, respectively, and four singlet methyl signals were observed at 0.84, 0.42, 0.37, and 0.27 ppm. The same signal pattern was observed in the 1 H NMR spectrum of 12. In contrast, a totally different pattern was observed in the 1 H NMR spectrum of the product derived from 10, which had an iPr 3 Si-substituent. Analysis of the spectrum indicated that the Si(Me) 2 C-Hf bond was not formed, while [N pyridine , N amido ]HfMe 3 complex 13 was generated ( Figure S9). The σ-bond metathesis via agostic interaction of SiC-H bond might be a process for the formation of SiC-Hf bond. In the case of iPr 3 Si-substituent, steric hindrance might hamper the agostic interaction not to afford the desired pincer-type complex. The prototype complex I characteristically contains a Hf-C(aryl) bond. We prepared the related pincer-type [C silylmethyl , N pyridine , N amido ]HfMe2 complexes containing the Hf-CH2Si bond instead of the Hf-C(aryl) bond (Scheme 2). 2-Br-6-(R2R'Si)-pyridines (R = Me, Ph, iPr; R' = Me, iPr; 5-7) were prepared from 2,6-dibromopyridine by the treatment of nBuLi and subsequently Me3SiCl, Ph2(Me)SiCl, or iPr3Si(OSO2CF3) [38,39]. Compounds 5-7 were treated with 2 equiv tBuLi to generate 2-Li-6-(R2R'Si)-pyridines, which were reacted with imine compound PhC(H)=N(2,6-iPr2C6H3) to generate pyridines substituted with the R2R'Si-group and (2,6-iPr2C6H3)N(H)C(Ph)(H)-group at the 2-and 6-positions (8)(9)(10). Treatment of 8 and 9 with in situ generated HfMe4 afforded the targeted pincer-type [C silylmethyl , N pyridine , N amido ]HfMe2 complexes containing the Hf-CH2Si bond (11 and 12). The 1 H NMR spectrum of 11 distinguished the two diastereotopic protons on HfCH2Si moiety at 1.31 and 0.24 ppm as doublets with a large geminal coupling constant (J = 12.6 Hz) ( Figure S7 in Supporting Information). The two methyl groups attached on Si and the two methyl groups attached on Hf were also diastereotopic, respectively, and four singlet methyl signals were observed at 0.84, 0.42, 0.37, and 0.27 ppm. The same signal pattern was observed in the 1 H NMR spectrum of 12. In contrast, a totally different pattern was observed in the 1 H NMR spectrum of the product derived from 10, which had an iPr3Si-substituent. Analysis of the spectrum indicated that the Si(Me)2C-Hf bond was not formed, while [N pyridine , N amido ]HfMe3 complex 13 was generated ( Figure S9). The σ-bond metathesis via agostic interaction of SiC-H bond might be a process for the formation of SiC-Hf bond. In the case of iPr3Si-substituent, steric hindrance might hamper the agostic interaction not to afford the desired pincer-type complex. Synthesis of pincer-type [C naphthyl , S thiophene , N amido ]HfMe2 complex was attempted by replacing the pyridine moiety in I with thiophene (Scheme 3). The Suzuki-coupling reaction of 1-naphthylboronic acid with the imine compound constructed with 5-bromo-2-thiophencarboxaldehyde and 2,6-iPr2C6H3NH2 generated the thiophene compound bearing imine (2,6-iPr2C6H3N=C(H)-) and naphthyl moieties (14). In the synthesis of I, 2-iPrC6H4Li facilely attacked the imine group to afford the desired ligand precursor. In the case of the reaction between 2-iPrC6H4Li and thiophene analogue 14, the desired ligand precursor was not obtained. However, nBuLi readily reacted with 14 to produce the desired thiophene compound 15 substituted with naphthyl and (2,6-iPr2C6H3)N(H)C(nBu)(H)-group at the 2-and 5-positions. When 15 was reacted with HfMe4, the Hf-C aryl bond was not formed, which failing to generate the desired [C naphthyl , S thiophene , N amido ]HfMe2 complex. Instead, the [N amido , S thiophene ]HfMe3 complex 16 was cleanly obtained. S-C-C ipso (naphthyl) angle might be too wide to generate the desired pincer-type complex via formation of Hf-C(naphthyl) bond. A single signal assigned to Hf(CH3)3 was observed as a singlet at 0.35 ppm in the 1 H NMR spectrum ( Figure S12). Synthesis of pincer-type [C naphthyl , S thiophene , N amido ]HfMe 2 complex was attempted by replacing the pyridine moiety in I with thiophene (Scheme 3). The Suzuki-coupling reaction of 1-naphthylboronic acid with the imine compound constructed with 5-bromo-2-thiophencarboxaldehyde and 2,6-iPr 2 C 6 H 3 NH 2 generated the thiophene compound bearing imine (2,6-iPr 2 C 6 H 3 N=C(H)-) and naphthyl moieties (14). In the synthesis of I, 2-iPrC 6 H 4 Li facilely attacked the imine group to afford the desired ligand precursor. In the case of the reaction between 2-iPrC 6 H 4 Li and thiophene analogue 14, the desired ligand precursor was not obtained. However, nBuLi readily reacted with 14 to produce the desired thiophene compound 15 substituted with naphthyl and (2,6-iPr 2 C 6 H 3 )N(H)C(nBu)(H)-group at the 2-and 5-positions. When 15 was reacted with HfMe 4 , the Hf-C aryl bond was not formed, which failing to generate the desired [C naphthyl , S thiophene , N amido ]HfMe 2 complex. Instead, the [N amido , S thiophene ]HfMe 3 complex 16 was cleanly obtained. S-C-C ipso (naphthyl) angle might be too wide to generate the desired pincer-type complex via formation of Hf-C(naphthyl) bond. A single signal assigned to Hf(CH 3 ) 3 was observed as a singlet at 0.35 ppm in the 1 H NMR spectrum ( Figure S12). The prototype complex I was discovered through the high-throughput screening and a variety of derivatives were prepared for screening [11,40,41]. The starting material for I (6-bromo-2-pyridinecarboxaldehyde) is expensive. The naphthyl group is introduced by the Suzuki-coupling reaction with naphthylboronic acid, and the aldehyde group is converted by condensation with aniline derivatives to imine. This group is reactive with 2-isopropylphenyllithium. The imine bond that forms with alkylamine is easily hydrolyzed in the presence of moisture. Therefore, compounds prepared using alkylamine were not included in the screening. In this work, we prepared the derivatives of I containing various alkylamido moieties instead of the arylamido moiety in I. The synthetic scheme was different from that developed for I and the starting material, 2,6-dibromopyridine, was relatively inexpensive (Scheme 4). 2-Bromo-6-naphthylpyridine 17, which was prepared through the Suzuki-coupling reaction of 2,6-dibromopyridine and 1-naphthylboronic acid [42], was treated with 2 equiv tBuLi to generate 2-lithio-6-naphthylpyridine, which was subsequently reacted with the imines generated through the condensation of benzaldehyde and various alkylamines. The resulting alkylamine compounds 18-21 were purified by the conventional column chromatography using silica gel. When 18-21 were treated with HfMe4, the desired pincer-type [C naphthyl , N pyridine , N alkylamido ]HfMe2 complexes 22-25 were cleanly generated. 1 H and 13 C NMR spectra agreed with the structures (Figures S17-S20) and the structure of 22 and 25 were unambiguously confirmed by X-ray crystallography.

X-ray Crystallographic Studies
The molecular structure of pincer-type [N amido , N pyridine , N amido ]HfMe2 complexes 4 was confirmed by X-ray crystallography ( Figure 2). The geometry around the Hf-center can be defined as a distorted trigonal bipyramid with a basal plane formed by pyridine-N(1), methyl-C(32), and methyl-C(33), with the axial sites occupied with amido N(2) and N(3) atoms. The sum of the bond angles of C(32)-Hf-N(1), C(33)-Hf-N(1), and C(33)-Hf-C(32) is 360°, indicating that the Hf atom is The prototype complex I was discovered through the high-throughput screening and a variety of derivatives were prepared for screening [11,40,41]. The starting material for I (6-bromo-2pyridinecarboxaldehyde) is expensive. The naphthyl group is introduced by the Suzuki-coupling reaction with naphthylboronic acid, and the aldehyde group is converted by condensation with aniline derivatives to imine. This group is reactive with 2-isopropylphenyllithium. The imine bond that forms with alkylamine is easily hydrolyzed in the presence of moisture. Therefore, compounds prepared using alkylamine were not included in the screening. In this work, we prepared the derivatives of I containing various alkylamido moieties instead of the arylamido moiety in I. The synthetic scheme was different from that developed for I and the starting material, 2,6-dibromopyridine, was relatively inexpensive (Scheme 4). 2-Bromo-6-naphthylpyridine 17, which was prepared through the Suzuki-coupling reaction of 2,6-dibromopyridine and 1-naphthylboronic acid [42], was treated with 2 equiv tBuLi to generate 2-lithio-6-naphthylpyridine, which was subsequently reacted with the imines generated through the condensation of benzaldehyde and various alkylamines. The resulting alkylamine compounds 18-21 were purified by the conventional column chromatography using silica gel. When 18-21 were treated with HfMe 4 , the desired pincer-type [C naphthyl , N pyridine , N alkylamido ]HfMe 2 complexes 22-25 were cleanly generated. 1 H and 13 C NMR spectra agreed with the structures (Figures S17-S20) and the structure of 22 and 25 were unambiguously confirmed by X-ray crystallography. The prototype complex I was discovered through the high-throughput screening and a variety of derivatives were prepared for screening [11,40,41]. The starting material for I (6-bromo-2-pyridinecarboxaldehyde) is expensive. The naphthyl group is introduced by the Suzuki-coupling reaction with naphthylboronic acid, and the aldehyde group is converted by condensation with aniline derivatives to imine. This group is reactive with 2-isopropylphenyllithium. The imine bond that forms with alkylamine is easily hydrolyzed in the presence of moisture. Therefore, compounds prepared using alkylamine were not included in the screening. In this work, we prepared the derivatives of I containing various alkylamido moieties instead of the arylamido moiety in I. The synthetic scheme was different from that developed for I and the starting material, 2,6-dibromopyridine, was relatively inexpensive (Scheme 4). 2-Bromo-6-naphthylpyridine 17, which was prepared through the Suzuki-coupling reaction of 2,6-dibromopyridine and 1-naphthylboronic acid [42], was treated with 2 equiv tBuLi to generate 2-lithio-6-naphthylpyridine, which was subsequently reacted with the imines generated through the condensation of benzaldehyde and various alkylamines. The resulting alkylamine compounds 18-21 were purified by the conventional column chromatography using silica gel. When 18-21 were treated with HfMe4, the desired pincer-type [C naphthyl , N pyridine , N alkylamido ]HfMe2 complexes 22-25 were cleanly generated. 1 H and 13 C NMR spectra agreed with the structures (Figures S17-S20) and the structure of 22 and 25 were unambiguously confirmed by X-ray crystallography.

X-ray Crystallographic Studies
The molecular structure of pincer-type [N amido , N pyridine , N amido ]HfMe2 complexes 4 was confirmed by X-ray crystallography ( Figure 2). The geometry around the Hf-center can be defined as a distorted trigonal bipyramid with a basal plane formed by pyridine-N(1), methyl-C(32), and methyl-C(33), with the axial sites occupied with amido N(2) and N(3) atoms. The sum of the bond angles of C(32)-Hf-N(1), C(33)-Hf-N(1), and C(33)-Hf-C(32) is 360°, indicating that the Hf atom is

X-ray Crystallographic Studies
The molecular structure of pincer-type [N amido , N pyridine , N amido ]HfMe 2 complexes 4 was confirmed by X-ray crystallography ( Figure 2). The geometry around the Hf-center can be defined as a distorted trigonal bipyramid with a basal plane formed by pyridine-N(1), methyl-C(32), and methyl-C(33), with the axial sites occupied with amido N (2)  angle of 180 • expected for the ideal trigonal bipyramidal structure. The Hf atom is not situated in a plane formed by the chelating ligand framework (i.e., a plane formed by N(2), C(6), C(1), N(1), C(5), C (19), and N(3) atoms), but rather is situated slightly above the plane (0.46 Å). The sum of the bond angles around the amido N(2) and N(3) atoms is 360 • , respectively, indicating that both N atoms adopt an sp 2 hybridization for π-donation from N to Hf-center. Hf-N amido (i.e., Hf-N(2) and Hf-N(3)) distances are significantly shorter than that of Hf-N pyridine (i.e., Hf-N(1)) (2.08 vs. 2.27 Å).
Molecules 2019, 24, x 5 of 18 perfectly situated in the basal plane. The N(2)-Hf-N(3) angle is 137.10(6) o , which deviated from the angle of 180° expected for the ideal trigonal bipyramidal structure. The Hf atom is not situated in a plane formed by the chelating ligand framework (i.e., a plane formed by N(2), C(6), C(1), N(1), C(5), C (19), and N(3) atoms), but rather is situated slightly above the plane (0.46 Å). The sum of the bond angles around the amido N(2) and N(3) atoms is 360°, respectively, indicating that both N atoms adopt an sp 2 hybridization for π-donation from N to Hf-center. Hf-N amido (i.e., Hf-N(2) and Hf-N(3)) distances are significantly shorter than that of Hf-N pyridine (i.e., Hf-N(1)) (2.08 vs 2.27 Å). The molecular structure of [C naphthyl , N pyridine , N amido ]HfMe2 complex bearing isopropylamido moiety (22) was confirmed by X-ray crystallography (Figure 3a). Geometry around the Hf-center was defined as a distorted trigonal bipyramid with a basal plane formed by pyridine-N(1), methyl-C(26), and methyl-C(27) atoms. The Hf atom is situated in a plane formed by the chelating ligand framework (i.e., a plane formed by N(2), C(16), C(15), N(1), C(11), C(10), and C(1) atoms). Either the naphthalene or the pyridine ring is slightly tilted from the plane formed by the chelating ligand framework (9.5(4) o and 9.6(4) o , respectively). Amido-N(2) atom adopts sp 2 hybridization for π-donation and, accordingly, the CH(iPr) atom is almost coplanar with the plane formed by the chelating ligand framework (C(15)-C(16)-N(2)-C(23) torsional angle, 171°). The Hf-C aryl distance is slightly longer than Hf-C methyl distances (2.29 vs 2.20 or 2.22 Å). Molecular structure of 25 bearing the adamantylamido moiety was also confirmed by X-ray crystallography (Figure 3b). In this structure, N-C-C-N chelating ligand framework and pyridine ring form a plane with the Hf-center, while the naphthalene ring is rather severely tilted from the plane (20.0°). While the Hf-N pyridine distance is almost the same with that in 22, the Hf-N amido and Hf-C naphthyl distances are longer by 0.01 to 0.03 Å than the corresponding distances in 22. The molecular structure of [C naphthyl , N pyridine , N amido ]HfMe 2 complex bearing isopropylamido moiety (22) was confirmed by X-ray crystallography (Figure 3a). Geometry around the Hf-center was defined as a distorted trigonal bipyramid with a basal plane formed by pyridine-N(1), methyl-C(26), and methyl-C(27) atoms. The Hf atom is situated in a plane formed by the chelating ligand framework (i.e., a plane formed by N(2), C(16), C(15), N(1), C(11), C(10), and C(1) atoms). Either the naphthalene or the pyridine ring is slightly tilted from the plane formed by the chelating ligand framework (9.5(4) • and 9.6(4) • , respectively). Amido-N(2) atom adopts sp 2 hybridization for π-donation and, accordingly, the CH(iPr) atom is almost coplanar with the plane formed by the chelating ligand framework (C(15)-C(16)-N(2)-C(23) torsional angle, 171 • ). The Hf-C aryl distance is slightly longer than Hf-C methyl distances (2.29 vs. 2.20 or 2.22 Å). Molecular structure of 25 bearing the adamantylamido moiety was also confirmed by X-ray crystallography (Figure 3b). In this structure, N-C-C-N chelating ligand framework and pyridine ring form a plane with the Hf-center, while the naphthalene ring is rather severely tilted from the plane (20.0 • ). While the Hf-N pyridine distance is almost the same with that in 22, the Hf-N amido and Hf-C naphthyl distances are longer by 0.01 to 0.03 Å than the corresponding distances in 22.

Activation Reactions
Activation reaction of the prototype complex I is complex [43,44]. which generated the desired ion-pair complex, which is stable in benzene [21].  (Figure 4). The major set of signals was unambiguously assigned to the desired ion-pair complex 26 (Scheme 5). In the 1 H NMR spectrum of 3, all four isopropyl groups were equivalent and a single Me2CH-resornace was observed at 3.65 ppm as a septet. However, in the 1 H NMR spectrum of the desired ion-pair complex 26, Me2CH-resonances were split as a pair of signals at 3.14 and 2.69 ppm, which was attributed to the persistent coordination of (C18H37)2NMe to the Hf-center. With the persistent coordination of

Activation Reactions
Activation reaction of the prototype complex I is complex [43,44]. Reaction with B(C 6 F 5 ) 3 results in decomposition through a process involving C 6 F 5 -transfers. Reaction with [Ph 3 4 ] − , which generated the desired ion-pair complex, which is stable in benzene [21].
When 3 was treated with [(C 18 H 37 ) 2 N(H)Me] + [B(C 6 F 5 ) 4 ] − in C 6 D 6 , methide abstraction occurred as was evident from the observation of the methane signal at 0.16 ppm in the 1 H NMR spectrum. In the 1 H NMR spectrum collected at the early stage of the reaction, two sets of signals corresponding to the ligand framework were observed in a ratio of approximately 10:1 ( Figure 4). The major set of signals was unambiguously assigned to the desired ion-pair complex 26 (Scheme 5). In the 1 H NMR spectrum of 3, all four isopropyl groups were equivalent and a single Me 2 CH-resornace was observed at 3.65 ppm as a septet. However, in the 1 H NMR spectrum of the desired ion-pair complex 26, Me 2 CH-resonances were split as a pair of signals at 3.14 and 2.69 ppm, which was attributed to the persistent coordination of (C 18 H 37 ) 2 NMe to the Hf-center. With the persistent coordination of amine, the two protons on an α-methylene of (C 18 H 37 ) 2 NMe are diastereotopic to each other, even though the two α-methylene carbons are equivalent. This was reflected as separately observed NCH 2 resonances at 2.39 and 2.13 ppm with a triplet-doublet coupling pattern (J = 13 and 4.2 Hz). Due to the persistent coordination of amine, the NCH 2 signals as well as the NCH 3 signal were sharp. With time, the minor set of signals became more prominent as the set of signals mainly observed at the initial stage became depressed and disappeared by 6 h. The spectrum was assigned to complex 28 generated from the ion-pair complex 26 through C-H bond activation of a H-CH 2 (Me)CH-group (i.e., isopropyl group). This unwanted reaction was also observed in the activation reaction of the prototype complex I with [Ph(Me) 2 N-H] + [B(C 6 F 5 ) 4 ] − . In the 1 H NMR spectrum of 28, three Me 2 CH-resonances were observed at 3.67, 3.22, and 2.39 ppm as a septet, while the resonance of the isopropyl group that was engaged in the C-H bond activation process (i.e., Hf-CH 2 (Me)CH-signal) was observed at 2.84 ppm as a broad multiplet (Figure 3). In the case of 4, which contained ethyl substituents instead of isopropyl, the targeted ion-pair complex 27 was also cleanly generated by the action of [(C 18 H 37 ) 2 N(H)Me] + [B(C 6 F 5 ) 4 ] − through methide abstraction. In contrast with 3, the activated complex was stable in C 6 D 6 with no further reaction. initial stage became depressed and disappeared by 6 h. The spectrum was assigned to complex 28 generated from the ion-pair complex 26 through C-H bond activation of a H-CH2(Me)CH-group (i.e., isopropyl group). This unwanted reaction was also observed in the activation reaction of the prototype complex I with [Ph(Me)2N-H] + [B(C6F5)4] − . In the 1 H NMR spectrum of 28, three Me2CH-resonances were observed at 3.67, 3.22, and 2.39 ppm as a septet, while the resonance of the isopropyl group that was engaged in the C-H bond activation process (i.e., Hf-CH2(Me)CH-signal) was observed at 2.84 ppm as a broad multiplet (Figure 3). In the case of 4, which contained ethyl substituents instead of isopropyl, the targeted ion-pair complex 27 was also cleanly generated by the action of [(C18H37)2N(H)Me] + [B(C6F5)4] − through methide abstraction. In contrast with 3, the activated complex was stable in C6D6 with no further reaction. Due to presence of chiral centers at the ligand framework as well as the Hf-center, two sets of signals were observed. The reaction rate of 11 was slow, requiring approximately 3 h for complete methide abstraction. At this stage of the complete reaction, the ratio of the two diastereomers was approximately 1:1, which slowly changed and ultimately became 1.0:0.60 after an overnight reaction. In the case of 12, the reaction rate was rapid and efficiently generated two diastereomers at a ratio of 1.0:0.50 within 30 min. Upon an overnight reaction, the ratio also slowly changed to become 1.0:0.85. The change of the ratio indicated that the site epimerization occurred, even though the rate was slow.

Activation Reactions
Activation reaction of the prototype complex I is complex [43,44]. which generated the desired ion-pair complex, which is stable in benzene [21].  (Figure 4). The major set of signals was unambiguously assigned to the desired ion-pair complex 26 (Scheme 5). In the 1 H NMR spectrum of 3, all four isopropyl groups were equivalent and a single Me2CH-resornace was observed at 3.65 ppm as a septet. However, in the 1 H NMR spectrum of the desired ion-pair complex 26, Me2CH-resonances were split as a pair of signals at 3.14 and 2.69 ppm, which was attributed to the persistent coordination of (C18H37)2NMe to the Hf-center. With the persistent coordination of with concomitant generation of methane. Due to presence of chiral centers at the ligand framework as well as the Hf-center, two sets of signals were observed. The reaction rate of 11 was slow, requiring approximately 3 h for complete methide abstraction. At this stage of the complete reaction, the ratio of the two diastereomers was approximately 1:1, which slowly changed and ultimately became 1.0:0.60 after an overnight reaction. In the case of 12, the reaction rate was rapid and efficiently generated two diastereomers at a ratio of 1.0:0.50 within 30 min. Upon an overnight reaction, the ratio also slowly changed to become 1.0:0.85. The change of the ratio indicated that the site epimerization occurred, even though the rate was slow.  4 ] − to generate methane. However, the 1 H NMR signals were too complicated to be unambiguously assigned ( Figure S24). In contrast, [N amido , S thiophene ]HfMe 3 -type complex 16 reacted slowly with [(C 18 H 37 ) 2 N(H)Me] + [B(C 6 F 5 ) 4 ] − and a set of signals eventually appeared after 3 days. These were assigned to 31 ( Figure S25). Two Hf-CH 3 signals were observed at 0.60 and 0.41 ppm. Reaction of 22 and 23 bearing secondary amido moieties cleanly afforded the targeted ion-pair complexes 32 and 33, for which two sets of signals were observed, respectively, in 1 H NMR spectra, due to the presence of chiral centers on both the Hf-center and at the ligand framework ( Figures S26 and S27). Reaction of 24 and 25 bearing tertiary amido moieties did not cleanly afford the targeted ion-pair complexes ( Figures S28 and S29).

Polymerization Studies
The targeted ion-pair complex 27 derived from [N amido , N pyridine , N amido ]HfMe 2 bearing ethyl substituents (4) exhibited moderate activity in ethylene/propylene copolymerization. The activity was approximately 1/7th that of the prototype Dow catalyst I (entry 1 vs. 4 in Table 1). All other complexes (i.e., 3, 11, 12, 16, and 22-23) that showed clean 1 H NMR signals in the activation reaction with [(C 18 H 37 ) 2 N(H)Me] + [B(C 6 F 5 ) 4 ] − exhibited negligible activities. Complex 13 and 25 bearing bulky iPr 3 Si-and adamantyl group, which did not show clean 1 H NMR signals in the activation reaction, exhibited moderate activities (approximately 1/10th that of I; entries 2 and 3). Comonomer incorporation ability was also inferior to that of I; 6.7, 0, and 5.1 mol% propylene was incorporated with 4, 13, and 25, respectively, while the propylene content was very high (56 mol%) with I under the identical polymerization conditions. [N pyridine , N amido ]HfMe 3 complex 13 bearing bulky iPr 3 Si-group generated the relatively high-molecular-weight polymer (M n , 430 kDa). The prototype Hf catalyst I was exceptional in terms of the activity and α-olefin incorporation capability. We also reported various types of Hf-complexes ([N,P]Hf(CH 2 Ph) 3 , [N,P,N]HfMe 2 , and [N,N]Hf(CH 2 Ph) 3 -type) with tetrahydroquinoline and tetrahydrophenanthroline framework, which were also inferior to I in terms of both activity and α-olefin incorporation capability [45][46][47]. The Hf-C bonding character was significantly ionic when compared to those of Zr-C and Ti-C bonding, and steric congestion around Hf-center might be crucial for the high activity [48]. When steric congestion is not significant, the ionic Hf-C bond becomes strong and the insertion of olefin through the Hf-C bond may be less favorable, leading to lowered activity.

General Remarks
All manipulations were performed under an inert atmosphere using standard glove box and the Schlenk technique. Toluene, hexane, diethyl ether, THF and C 6 D 6 were distilled from benzophenone ketyl. Methylcyclohexane used for the polymerization reactions was purchased from TCI and was purified over a Na/K alloy. Ethylene was purified by contact with molecular sieves and copper at a pressure of 50 bar. 1 H NMR (600 MHz) and 13 C NMR (150 MHz) spectra were recorded using an ECZ 600 apparatus (JEOL, Tokyo, Japan). Compounds 5 [39], 7 [38], 17 [42], 2-iPrC 6 H 4 Li [41], and PhC(H)=NCH(CH 3 ) 2 were prepared according to previously reported procedures and conditions [49].

Complex 3
MeMgBr (0.400 mL, 1.20 mmol, 3.0 M solution in diethyl ether) was added dropwise at −78 • C to a solution of HfCl 4 (93.5 mg, 0.292 mmol) in toluene (2 mL). The resulting solution was stirred at −40 to −35 • C for 1 h to precipitate a white solid. After cooling to −78 • C, a solution of 1 (0.100 g, 0.195 mmol) in toluene was added dropwise. The resulting mixture was stirred at −40 to −35 • C for 2 h and then warmed slowly to room temperature. After stirring overnight, all volatiles were removed using a vacuum line. Toluene (10 mL) was added to extract the product. The extract was collected through filtration over Celite. After the solvent was removed, the residue was triturated with hexane to obtain a pink solid (0.099 g, 70%). The isolated product was contaminated with some amount of chloromethyl-Hf analog, which was converted to the desired dimethyl-Hf complex by treatment with MeLi in toluene. 1