Catalytic Performance of Cobalt(II) Polyethylene Catalysts with Sterically Hindered Dibenzopyranyl Substituents Studied by Experimental and MLR Methods

Given the great importance of cobalt catalysts supported by benchmark bis(imino)pyridine in the (oligo)polymerization, a series of dibenzopyran-incorporated symmetrical 2,6-bis(imino) pyridyl cobalt complexes (Co1–Co5) are designed and prepared using a one-pot template approach. The structures of the resulting complexes are well characterized by a number of techniques. After activation with either methylaluminoxane (MAO) or modified MAO (MMAO), the complexes Co1–Co4 are highly active for ethylene polymerization with a maximum activity of up to 7.36 × 106 g (PE) mol−1 (Co) h−1 and produced highly linear polyethylene with narrow molecular weight distributions, while Co5 is completely inactive under the standard conditions. Particularly, complex Co3 affords polyethylene with high molecular weights of 85.02 and 79.85 kg mol−1 in the presence of MAO and MMAO, respectively. The 1H and 13C NMR spectroscopy revealed the existence of vinyl end groups in the resulting polyethylene, highlighting the predominant involvement of the β-H elimination reaction in the chain-termination process. To investigate the mechanism underlying the variation of catalytic activities as a function of substituents, multiple linear regression (MLR) analysis was performed, showing the key role of open cone angle (θ) and effective net charge (Q) on catalytic activity.


Introduction
Concerted efforts have been devoted in last two decades to exploring the propensity of bis(imino)pyridine metal complexes' precatalysts towards ethylene (oligo)polymerization. This effort was driven by their distinct advantages of low cost, ease of preparation, and high polymerization activity [1][2][3][4][5]. Indeed, it is a continuous topic of interest to enhance the scope of these precatalysts in terms of polymerization activity, thermal stability, and the molecular weight of the resulting polyethylene, either through adjusting the steric and electronic substitution pattern of the benchmark bis(imino)pyridine framework or designing a completely new ligand framework [6][7][8][9][10][11][12][13][14][15]. Tremendous studies have been performed by adjusting the imine-N and imine-C substituents, as well as the focal pyridine unit, to examine their influence on the catalytic activity for ethylene polymerization [16][17][18][19][20][21][22][23][24][25][26][27][28]. In particular, more focus was placed on the incorporation of the bulky substituents at the axial sites of the chelate ring in order to improve the chain propagation versus the chain transfer reactions at elevated reaction temperatures [17,19,28].
The benzhydryl substituent is one of the substituents of the bis(imino)pyridine cobalt complexes that have been investigated to explore the extent of the complex catalyst space. This class of catalyst system experienced difficulties in depicting high catalytic activity and thermal stability simultaneously. However, these cobalt catalysts can produce characteristically linear and high-density polyethylene with a high molecular weight and broad to narrow distributions [9][10][11][12]. Furthermore, different substitutions at the para-position of N-aryl rings of cobalt complex systems yield different products ranging from oligomers to polymers. To this end, many works regarding the modification of substitutions at the para-position of N-aryl rings in bis(imino)pyridyl cobalt complexes have been reported [29][30][31][32][33]. These modified cobalt complexes result in a drastic change of catalytic activity and molecular weight of the produced product, attributed to the steric and electronic effect of the induced species. For example, Wu and coworkers studied the influence of electronic and steric effects at the para-position [7]. Their results exhibited no significant change in activity on the basis of electronic perturbations. Equally, variation in the steric substituents makes a notable contribution to enhancing the molecular weights of the resulting polymer. Likewise, the Li group noticed a large difference in catalytic performance with variations in the steric hindrance at the para-position of the N-aryl units (H < Br < iPr) [32]. It is worth mentioning that the precatalysts with a sterically enhanced bulky group such as electron-donating (iPr) and electron-withdrawing (Br) substituents at the para-position of the imines usually display superior activities and molecular weights compared to precatalysts with a hydrogen substituent.
With this development, our group studied the influence of different steric groups on the bis(imino)pyridine metal precatalysts framework. Firstly, catalysts with an ortho-benzhydryl substituted group as shown in Chart 1A [9] exhibited high polymerization activity up to 9.87 × 10 6 g (PE) mol −1 (Co) h −1 , and the molecular weight of resulting polyethylene fall in the range of 100-330 kg mol −1 with narrow molecular weight distributions. However, the incorporation of a benzhydryl substituent at only one ortho-position provided a slightly lower activity of 2.85 × 10 6 g (PE) mol −1 (Co) h −1 , in addition to producing a lower-molecular-weight polyethylene in the range of 53.81-57.56 kg mol −1 (Chart 1B) [11]. Moreover, introducing a benzhydryl substituent at the para-position within the same N-aryl group results in high activity (1.81 × 10 7 g (PE) mol −1 (Co) h −1 ) and molecular weight (118 kg mol −1 ) of the resultant polyethylene exhibiting the unimodal behavior (Chart 1C) [12]. Similarly, the successful incorporation of dibenzocycloheptyl groups at the orthoand para-positions of N-aryl rings attached to the imine nitrogen atom in the generic bis(imino)pyridine yielded polyethylene with high activity up to 1.00 × 10 7 g (PE) mol −1 (Co) h −1 and thermally stability at 60 • C. This report depicted the high thermal stability for unsymmetrical 2,6-bis(imino)pyridylcobalt (II) chloride precatalysts (Chart 1D) [29]. The activity of D is lower than that of C, because of the steric hindrance caused by the bulky dibenzocycloheptyl group, which inhibits the incoming monomer to approach the central metal atom. In addition to that, our research group reported a series of symmetrical cobalt complexes with a dibenzocycloheptyl substituent at the para position of the N-aryl rings. These cobalt complexes yielded highly linear polyethylene with catalytic activity up to 1.21 × 10 7 g (PE) mol −1 (Co) h −1 and narrow molecular distribution (Chart 1E) [33]. With a view to further enhance the polymerization scope of these cobalt precatalysts toward ethylene polymerization, we were interested in functionalizing the bis(imino)pyridine with dibenzopyran bulky group as the para-substituent on the N-aryl group (Chart 1F). For this purpose, a series of symmetrical 2,6-bis(imino) pyridyl cobalt complexes with a para-substituted dibenzopyran group are prepared. The synthesized cobalt complexes are fully characterized, and extensive ethylene polymerization is performed for investigating the impact of structural changes on catalytic performance. The effect of a co-catalyst, temperature, and pressure on the catalytic activity is examined during the polymerization runs. In addition, the properties of polymers, including molecular weight, polydispersity, melting temperature, and microstructure, are well described. To gain a better understanding of the variation mechanism regarding the catalytic activity as a function of substituents, multiple linear regression (MLR) analysis is carried out to unravel the role of electronic and steric effects, thus providing a guide for the future design of novel catalysts with enhanced catalytic activity.

Synthesis and Characterization of Co1-Co5
As shown in Scheme 1, various 4-dibenzopyranol-2,6-dialkyllanilines were prepared from treating corresponding dialkylanilines with dibenzopyranol. The title cobalt complexes (Co1-Co5) were obtained in the one-pot reactions of 2,6-diacetylpyridine, 4-dibenzopyranol-2,6-dialkyllanilines, and CoCl 2 refluxed with acetic acid. The cobalt complexes were precipitated with an excess of diethyl ether and isolated as an air-stable green powder in high yields. The representative complexes Co1 and Co4 were the subject of single-crystal X-ray diffraction analysis. The FT-IR spectra of complexes (Co1-Co5) showed signals located in the region from 1602 to 1652 cm −1 ascribed to the stretching vibration of C=N imine bonds. This information demonstrated the active coordination between the central metal and the donor atoms. The elemental analysis of all the complexes exhibited their consistency with the molecular formula. Moreover, the molecular structures of Co1 and Co4 were well described using single-crystal X-ray diffraction analysis as shown in Figures 1 and 2.   Single crystals of Co1 and Co4 subjected to X-ray determination were obtained by dissolving them slowly in n-hexane, which was further treated with the solution of the corresponding complex in dichloromethane. The crystal structures of Co1 and Co4 are shown in Figures 1 and 2, and the selected values of bond length and bond angles are listed in Table 1. The structures of Co1 and Co4 are fundamentally similar, because both have a five-coordinate structure system with the cobalt center surrounded by two chloride atoms and three nitrogen atoms from the bis(arylimino)pyridine. In both cases, a tridentate ligand with imino and pyridine moieties is basically planar; the N-aryl groups are inclined almost perpendicularly to the bis(imino)pyridine plane. The geometry can be explained well by the tau value (τ 5 ) calculated by the equation (β − α)/60, where β is the largest, and α is the second-largest angle in the coordination sphere [29]. According to the τ 5 values (0.21 for Co1 and 0.07 for Co4), the crystal structures of the complexes described the bonding of a single cobalt center with bis(arylimino)pyridine ligands and two chlorine atoms leading to distorted square-pyramidal and perfect square-pyramidal geometry, respectively.
The basal plan includes all the nitrogen atoms (N1, N2, and N3) and one chlorine atom Cl(1), while the second chlorine atom Cl(2) forms the apical position. The dihedral angles of Co1 and Co4 are 82.55 • and 77.87 • , respectively, where the cobalt atoms exist above the N1, N2, and N3 planes. These results can be found in previously reported studies [29,[33][34][35]. Different substitutions of aryl rings connected with imine nitrogen atoms on both sides cause the change in steric hindrance around their atoms leading to a slight change in bond lengths and bond angles. The bond lengths for Co-N imine atoms are 2.223(3) Å and 2.225(3) Å in the Co1 complex, while these bond length values are a little bit higher in Co4 complex, i.e., 2.239(4) Å and 2.246(4) Å. The fluctuation of bond lengths in each complex is due to the presence of different substitutions on N-aryl rings. Moreover, further description reveals that the Co-N pyridine bond is comparatively stronger than the Co-N imine bonds for Co1 and Co4 complex systems. The length N(3)-C(8) bonds in both complexes describe C=N.

Ethylene Polymerization
The microstructure of the resultant polyethylene may sometimes vary due to the variation of co-catalysts. Primarily, different alkylaluminium co-catalysts were tried to check their catalytic activity, and the results indicated that methylaluminoxane (MAO) or modified methylaluminoxane (MMAO) are more effective co-catalysts than other alkyl aluminium reagents [36,37]. Co1 was chosen as a precatalyst for the initial screening and finding of the standard conditions in order to examine the catalytic performance of the cobalt precatalysts toward ethylene polymerization. Different reaction conditions such as temperature, Al/Co ratio, and time were steadily varied under an ethylene pressure of 10 atm using methylaluminoxane (MAO) or modified methylaluminoxane (MMAO) as cocatalysts. The set of optimal reaction conditions was extended to predict the catalytic performance of the rest of the cobalt precatalysts Co2-Co5. Furthermore, Co1 was also screened at an ethylene pressure of 5 and 1 atm. Polymer characterization was carried out using different techniques. The molecular weight (M w ) and molecular weight distribution (M w /M n ) of the resultant polyethylene were determined using gel permeation chromatography, and the melting temperature was measured using differential scanning calorimetry (DSC). The microstructure analysis of resultant polyethylene samples was performed using high-temperature NMR ( 1 H/ 13 C) spectroscopy.

Catalytic Evaluation Using Co1 with MAO as a Cocatalyst
Ethylene polymerization was carried out by optimizing Co1 in toluene under the ethylene pressure of 10 atm, and all the results are listed in Table 2. Considering different conditions, the first polymerization run was performed at different temperatures ranging from 30 to 70 • C. During all the running at different temperatures, the Al/Co ratio was constant with the reaction time of 30 min ( Table 2, entries 1-5, runs 1). By increasing the temperature from 30 • C onwards, the highest activity was observed at 40 • C and then suddenly started to decrease until there was no activity at 70 • C. This behavior exhibits the sensitivity of catalytic activity towards high temperatures. This may be attributed to the low solubility of ethylene at high temperatures and the deactivation ion of cobalt's active species. On the contrary, the molecular weight of produced polyethylene decreases from 5.02 kg mol −1 to 2.29 kg mol −1 with an increase in temperature. This decline may happen due to a high chain-transfer rate and low ethylene concentration at elevated temperatures [36]. Moreover, Figure 3a depicts the unimodal behavior of GPC curves at different temperatures, and Figure 3b shows the clear graphical variation of activity and molecular weight with temperatures corresponding to entries 1-5 of Table 2.  Furthermore, the variations of the molar ratio of Al/Co were tried from 2000 to 3500 at 40 • C and 30 min (Table 2, entries 2, 6-10). The highest value of the activity (7.36 × 10 6 g (PE) mol −1 (Co) h −1 ) was observed at an Al/Co of 2500. Additionally, the polymerization activity was decreased by increasing the concentration of MAO. The molecular weight of the resultant PE ranges from 1.42-2.27 kg mol −1 , as shown in Figure 4. The effect of reaction time from 5 to 60 min on the polymerization process was examined at optimum conditions, i.e., temperature at 40 • C and Al/Co ratio of 2500 (Table 2, entries 7, [11][12][13][14]. The slow decrease of activity from 14.93 × 10 6 g (PE) mol −1 (Co) h −1 (5 min) to 5.37 × 10 6 g (PE) mol −1 (Co) h −1 (60 min) is due to some deactivation of the active species over longer time spans of reactions. It is unlikely that the molecular weight of PE showed a slight change with time, and it ranged from 2.42-3.12 kg mol −1 ; this may happen due to a chain transfer and the β-H elimination processes. The slight increase in molecular weight of polyethylene with reaction times is plotted in the form of GPC curves as shown in Figure 5a. The unimodal molecular weight distribution was observed for a resultant product with a melting point ranging from 121.9-124.8 • C. The plots of activity and the molecular weight of polyethylene as a function of the reaction time are given in Figure 5b. These dependencies exhibit similarities with related catalytic systems [10][11][12][13]. To investigate the effect of ethylene pressure on the catalytic activity and molecular weight of the resultant polyethylene, another polymerization run was performed ( Table 2, entries 15 and 16) at optimized reaction conditions (T = 40 • C, Al/Co = 2500, t = 30 min) at 1 and 5 atm, respectively. There is a strong influence of pressure on both activity and molecular weight as shown in Figure S1 of Supplementary Materials. The GPC trace exhibits the unimodal curves of the resultant polymer with the melting point of 124.8 • C and 121.9 • C at 1 and 5 atm, respectively.

Catalytic Evaluation Using Co1 with MMAO as a Cocatalyst
The ethylene polymerization process for Co1/MMAO was conducted in the same manner as the Co/MAO system. Following the same procedure, the molar ratio of Al/Co was fixed at 2000, and the ethylene polymerization was performed in toluene by adjusting the temperature from 30 • C to 70 • C under 10 atm pressure and over 30 min reaction time ( Table 3, entries 1-5). The highest value of activity is achieved at 40 • C. Furthermore, there is a gradual decrease in activity and molecular weight with increasing temperature, and this trend corresponds to the results obtained by the MAO cocatalyst system ( Figure 6).  By fixing the optimum temperature at 40 • C, the second series of ethylene polymerization was performed by changing the Al/Co ratio from 1000 to 2500 (Table 3, entries 2, 6-10).
The polymerization activities varied from 0.02-5.46 × 10 6 g (PE) mol −1 (Co) h −1 with the highest active peak at the molar ratio of 1750. There is no regularity of the variation of molecular weight as depicted by GPC traces, as shown in Figure 7, with the PDI value of resultant PEs around 2.0. The reaction time was prolonged from 5 to 60 min in the third series of ethylene polymerization (Table 3, entries 11-14, run 3). By increasing the reaction time, the catalytic activity decreased from 16.93 to 3.33 × 10 6 g (PE) mol −1 (Co) h −1 , which could possibly be due to the deactivation of active species. Figure 8 shows the GPC trace of the resultant PEs, and, clearly, there is an inverse relationship between molecular weight and activity as a function of reaction time. There is a significant effect of ethylene pressure on the catalytic activity and molecular weight of the produced polyethylene as shown in Figure S2 in the Supplementary Materials, exhibiting the similarity of the trend in comparison to those observed for the Co1/MAO system.

Catalytic Evaluation Using Co1-Co5 with MAO and MMAO as the Co-Catalysts
After optimizing the reaction conditions based on Co1/MAO system, the behavior of the remaining cobalt precatalysts Co2-Co5 toward ethylene polymerization was investigated under optimized reaction conditions, i.e., the reaction temperature of 40 • C, the Al/Co ratio of 2500, ethylene pressure of 10 atm, and the run time of 30 min as listed in entries 2-5 of Table 4. All the complexes exhibit good activity with values ranging from 1.36 to 7.36 × 10 6 g(PE) mol −1 (Co) h −1 towards the ethylene polymerization, except for the Co5 complex, which is probably due to the different electronic negativity of the ortho-substituents within the N-aryl group. The steric effect has a significant influence on catalytic activities. The hindrance of ortho-substituent due to bulky groups decreases the activity of the resultant polymer in the following order Co1 (Me, Me) > Co4 (Me, Et) > Co2 (Et, Et) > Co3 (iPr, iPr). Thus, the highest value of activity 7.36 × 10 6 g(PE) mol −1 (Co) h −1 obtained by the Co1 (Me, Me) precatalyst bearing the least bulky substituents. However, Co3 (i-Pr, i-Pr) displayed the least value of activity 1.36 × 10 6 g(PE) mol −1 (Co) h −1 . The iso-propyl substituent at the ortho-positions of N-aryl rings hinders the incoming ethylene monomers and affects the coordination, leading to a decrease of activity. Meanwhile, the substituent in the ortho-position of N-aryl group strongly influences the molecular weight of the resulting polyethylene. The highest value of molecular weight 85.02 kg mol −1 for Co3 corresponds to the bulkiest iso-propyl N-aryl group. This shows that bulky substituents protect the active sites and suppress the chain transfer. Similar behavior was observed in previous reports [9][10][11][12]33]. Moreover, the melting point of the polyethylene produced by the Co3 precatalyst is 134.9 • C, which is the highest value among all cobalt complexes. This could be attributed to the polymer's high crystallinity. Figure 9a shows that all the PEs obtained from Co1, Co2, and Co4 possessed narrow distributions (PDI ≈ 2.0), and their GPC curves show unimodal distribution possessing a single-site catalyst system. The GPC trace of PE from Co3 demonstrated the bimodal distribution, which is due to the multisite active species during the polymerization process. The values of the activities and molecular weight of different cobalt precatalysts at the optimized reaction conditions are given in Figure 9b.  For the MMAO cocatalyst, ethylene polymerization was conducted at the optimized conditions, i.e., reaction temperature of 40 • C, Al/Co ratio of 1750, ethylene pressure of 10 atm, and reaction time of 30 min for the remaining cobalt complexes with the decreasing order of activity Co1 (Me, Me) > Co4 (Me, Et) > Co2 (Et, Et) > Co3 (i-Pr, i-Pr). All the cobalt complexes bearing polymers are unimodal, describing the single-site active species, except Co3 as shown in Figure 10a. The trend of activity and the molecular weight Co/MMAO system are similar to Co/MAO. Co3 produces polyethylene with a very high molecular weight of 79.85 kg mol −1 ; the molecular weights of the remaining cobalt complexes range from 1.17 to 4.38 kg mol −1 . This high value for Co3 depicts that steric hindrance of the ortho-substituent of the N-aryl group dominatingly affect the molecular weight of polymer describing the bulky groups are responsible for suppressing the chain transfer. Additionally, the Co3 precatalyst exhibits the lowest catalytic activity as listed in Table 4 and displays a broader and bimodal molecular weight distribution corresponding to Figure 10. Comparatively, MAO generates polyethylene of a much higher molecular weight than the system containing MMAO, shown as entries 3 and 8 of Table 4.  (Table 4, entries 6-9).

Microstructural Properties of Polyethylene
A variety of polyethylene samples exhibit high melting temperatures (T m ) from 112.5 • C to 124.5 • C under different conditions, indicating that the microstructure is linear because of the absence of branching content. The detailed examination of product microstructure was completed using high-temperature 1 H and 13 C NMR spectroscopy for the Co1/MMAO sample obtained at optimum conditions mentioned in entry 9 of Table 3. The existence of PEs with vinyl (CH 2 =CH 2 -) end groups was assured from both 1 H NMR (H a and H b ) and 13 C NMR peaks at C a and C b at δ 139.50 and 114.38 ppm, respectively, in Figures 11 and 12. It is impossible to calculate the exact ratio of saturated to unsaturated PEs (H a to H g ), due to the peak of the end methyl group overlapping partly with the main backbone (-CH 2 -). However, the ratio of H a to H g is close to 1:3, indicating the presence of saturated PEs components along with high T m values of 123.7 • C and 124.5 • C, respectively. A comparatively lower T m is related to unsaturated PEs components. Methyl groups at the saturated end of the macromolecules were observed at 14.22 ppm (C g ) and 32.24, 22.92, and 18.33 ppm (C d , C e , and C f ), respectively.  The detailed examination of product microstructure was also completed using hightemperature 1 H and 13 C-NMR spectroscopy for the Co1/MAO sample obtained at optimum conditions listed in entry 7 of Table 2. The produced polymers have high values of T m , which range from 121.9 • C to 124.8 • C at different conditions ascribed to the highly linear structures. The signals corresponding to the vinyl (CH 2 =CH 2 -) monomers were observed in both spectra, exhibiting linear polyethylene, as displayed in Figures S3 and S4 of the Supplementary Materials.
Compared with previous reported systems in Chart 1, the catalytic activities of the Co1-Co4 precatalysts in the present study are relatively higher than those of 1B but lower than other categories. As for the molecular weight of produced polyethylene, it shows that the present cobalt catalytic system has a higher M w than 1E and 1B, but lower than 1A, 1C, and 1D. Moreover, the current catalytic system exhibits a slightly broader molecular weight distribution (1.84-5.52) relative to the reported analogues, depicting a narrow molecular weight distribution (~2.0). Comparatively, the cobalt complex systems possess relatively good catalytic performance, which has potential application in the practical fields.

Interpretation of Catalytic Activity by MLR Analysis
In the present work, calculations were performed to explore the effect of the substituents on the catalytic activities. The variation of experimental activities for all the cobalt complexes in both cocatalysts, MAO and MMAO, present the same trend, so the results from the MAO runs were selected as representative in order to explore the relationship between the structures of complexes and experimental activities. First, to validate the calculated parameters, the geometries of the complexes Co1 and Co4 were optimized and compared to their experimental crystal structures. The comparisons of bond lengths and bond angle between calculated and experimental results are listed in Tables S1 and S2 of the Supplementary Materials. Although the structures at the quartet have higher values of optimized energy (∆E), it is evident that the standardized deviations (δ) for bond length and bond angle at the quartet states are smaller than those at the doublet states. Therefore, we chose the calculation results of complexes Co1-Co4 at the quartet state, which is in correspondence with our previous report [38].
On the basis of optimized structures, the values of descriptors were calculated and are listed in Table 5. Herein, seven descriptors are considered based on our previous studies [38][39][40], including effective net charge (Q eff ), HOMO-LUMO energy gaps (∆ε 1 , ∆ε 2 ), energy difference (∆E), Hammett constant (F), bite angle (β), and open cone angle (θ). The detail definition and calculation procedure of each descriptor are provided in Supplementary Materials. For Co1-Co3 complexes in which the ortho-substituents R 1 /R 2 vary from methyl through ethyl to isopropyl, the values of Q eff decrease substantially from 0.527 to 0.493 due to the increase in electron-donating ability of corresponding alkyl groups. Since the ethylene polymerization is electrophilic reaction, the lower net charge on the central metal may cause lower reaction activity. As a result, the experimentally observed activities of these three precatalysts decrease from 7.36 through 4.56 to 1.36 × 10 6 g (PE) mol −1 (Co) h −1 (Table 5). Due to the increase in steric hindrance of the alkyl substituents, the value θ decreases from 139.83 • to 115.98 • for Co1-Co3, along with the decrease of catalytic activities. Furthermore, the energy difference (∆E), HOMO-LUMO energy gap (∆ε 1 ), and (∆ε 2 ) values present the increasing trend with the decrease of the experimentally determined catalytic activity. There is no discernible correlation between the bite angle (β) and the experimental activity. Regarding Co4, in which both methyl-and ethyl-substituent occupy the ortho-position within the N-aryl group, its effective net-charge value is greater than that of Co2 but lower than that of the Co1 complex. Meanwhile, due to the increase of hindrance compared to methyl counterpart, its value of θ is lower than that of Co1 and greater than that of Co2. In order to calculate the effect of these substituents in terms of the electronic and steric effect on the catalytic activity, a multiple linear regression (MLR) analysis is used to explore the relationship between structures with experimental activities [35]. Linear fitting is performed individually for each single descriptor to check its impact on catalytic activities, as shown in Figure S3. Among all the descriptors, it is clear that an open cone angle (θ) and effective net charge (Q eff ) exhibit the highest correlations with R 2 values of 0.978 and 0.902, respectively. Therefore, the comparisons between the experimental and the calculated activities of the Co complexes using the combination of θ and Q eff show good correlation, as displayed in Figure 13.
The calculated contribution of each descriptor indicates that the open cone angle (θ) is the dominant factor with the contribution value of 62.53%, revealing that the effect of substituents on the catalytic activity is primarily attributable to the steric effect. Higher values of θ are relatively favorable to catalytic activity.

General Consideration
All manipulation of water and/or moisture sensitive compounds was performed under a nitrogen atmosphere using a standard high vacuum Schlenk or in a conventional nitrogen-filled glovebox. Before usage, the solvents (toluene and THF) were refluxed for 8-12 h over a suitable drying agent and distilled under a nitrogen environment. Cocatalysts MAO (1.46 M solution in toluene) and MMAO (2.00 M in heptane) were purchased from the Akzo Nobel Corporation. High-purity ethylene was purchased from the Beijing Yansan Petrochemical Company. Other reagents were purchased from Aldrich, Acros, or local suppliers. Crystal data were collected on a MM007HF single-crystal diffractometer with Confocal-monochromatized Mo Kα radiation (λ = 0.71073 Å) at 173(2) K. 1 H NMR spectra (except for PE) were recorded with a Bruker DMX 400 MHz instrument at room temperature using TMS as an internal standard, while the NMR spectra of the polyethylenes were recorded on a Bruker DMX 300 MHz instrument at 100 • C in deuterated dichlorobenzene (C 6 D 4 Cl 2 ) or 1,1,2,2-tetrachloroethane-d 2 (C 2 D 2 Cl 4 ) as a solvent with TMS as an internal standard. A PerkinElmer System 2000 FT-IR spectrometer was used to obtain infrared spectra. A Flash EA 1112 micro-analyzer was used for elemental analysis. The molecular weight (M w ) and molecular weight distribution (M w /M n ) of polyethylenes were measured on a PL-GPC220 equipment running at 150 • C with 1,2,4-trichlorobenzene as the solvent. The melting temperature (T m ) of polyethylenes was determined using the second scanning run of a PerkinElmer TA-Q2000 DSC analyzer in a nitrogen atmosphere. A sample (4.0-6.0 mg) was heated to 150 • C at a rate of 20 • C/min, held for 5 min at 150 • C to eliminate the thermal history, and then cooled to −20 • C at a rate of 20 • C min −1 . Before testing the 13 C NMR spectra of the polyethylenes, they were dissolved to 80 to 100 mg of polyethylene in 1,1,2,2-tetrachloroethane-d 2 (2 mL) with TMS as an internal standard using a heat gun. The inverse gated 13 C spectra were then acquired on a Bruker DMX 300 spectrometer at A 250 mL round-bottomed flask was loaded with 2,6-dimethylaniline (1.45 g, 12 mmol) and dibenzopyranol (3 g, 15 mmol). The mixture was stirred for 30 min under nitrogen at 130 • C, and then the catalytic amount of ZnCl 2 in HCl solution was added dropwise. The reaction mixture was further heated for 2 h, and, upon cooling, the solid was dissolved in DCM (200 mL). It was filtered and washed twice with the saturated solution of NH 4 Cl and NaCl. The organic part was collected, dried over anhydrous Na 2 SO 4 , and concentrated at reduced pressure. The product was recrystallized from DCM using n-heptane to result in a white crystalline solid (1.25 g, 34%). TLC and NMR spectra confirmed that the resultant product is 2,6-dimethyl-4-(dibenzopyranyl)aniline. 1   The cobalt complex Co1 was obtained using the one-pot synthesis method. The Schlenk tube was loaded with 2,6 diacetyl pyridine (0.03 g, 0.20 mmol), 2,6-dimethyl-4-(dibenzopyranyl)aniline (0.20 g, 0.60 mmol) and CoCl 2 (0.02 g, 0.20 mmol) refluxed with acetic acid (30 mL). The reaction continued for 6h at 120 • C, and the volatiles were evaporated to result in a concentrated solution using the pump. The cobalt complex was precipitated using diethyl ether (20 mL

X-ray Crystallographic Studies
Two X-ray-quality crystals of complexes Co1 and Co4 were grown by the layering of n-hexane into the solution of the corresponding complex in dichloromethane at an ambient temperature. The X-ray studies were performed on an XtaLAB Synergy-R HyPix diffractometer with mirror monochromatic Cu-Kα radiation (λ = 1.54184 Å) at 170.01 or 169.99 K; cell parameters were obtained by global refinement of the position of all collected reflections. The intensities of the Lorentz and polarization effects were corrected, and an empirical absorption was applied. The structures were resolved by the direct methods and refined by full matrix least-squares on F 2 . All hydrogen atoms were placed in calculated positions. The structure's solution and refinement were conducted using SHELXT-97 [41,42].The free solvent molecules were squeezed with PLATON software. The details of the X-ray structure determinations data and structural refinements for Co1 and Co4 are provided in Table 6.

Ethylene Polymerization
Ethylene polymerization is conducted in a stainless-steel autoclave (250 mL) equipped with a temperature-and pressure-control system and a mechanical stirrer. The autoclave is initially evacuated and then filled with nitrogen gas. This process is repeated three times after the final evacuation ethylene is introduced. A solution of the corresponding complex (1.5 µmol) in freshly distilled toluene (25 mL) is injected into the autoclave. Another batch of freshly distilled toluene (25 mL) is added, and then the required amount of a cocatalyst (MAO, or MMAO) is injected. After adding another batch of toluene (50 mL), the autoclave is pressurized immediately with ethylene (10 atm), and the content is stirred at a rate of 400 rpm. Upon completion, the stirring is stopped, and the pressure is slowly released. The reaction is quenched with 10% hydrochloric acid in ethanol, and the polymer is washed with ethanol, filtered, and dried under reduced pressure at 40 • C. Finally, the product is weighed. A Schlenk tube is used instead of an autoclave for ethylene polymerization at 1 atm, following a similar procedure.

Conclusions
The successful incorporation of the 4-dibenzopyranly group at the para-position of the N-aryl ring of 2,6-bis(imino)pyridyl ligands was performed, and cobalt(II) complexes were synthesized from this ligand framework. The produced cobalt complexes were further characterized by FT-IR, elemental analysis, and X-ray diffraction analysis. The molecular structures of Co1 and Co4 depicted the distorted square-pyramidal geometry. Upon activation with MAO or MMAO, good activities were exhibited by these cobalt complex precatalysts up to the value of 7.36 × 10 6 g (PE) mol −1 (Co) h −1 . The molecular weight and polydispersity indices of resultant polyethylenes are highly dependent on the substitutions of the ligand framework and reaction parameters such as Al/Co molar ratio, reaction temperature, and reaction time. The 1 H-and 13 C-NMR spectroscopy depicted the microstructure of the product as the linear polyethylene with the presence of vinyl (CH 2 =CH 2 -) end groups. Moreover, multiple linear regression (MLR) analysis was performed to investigate the effect of different substituents, showing the dominant role of the open cone angle and effective net charge on catalytic activity. Meanwhile, the open cone angle presents a relatively higher correlation and bigger contribution with catalytic activity, indicating its dominant role on activity. Compared with previously reported analogues, the present cobalt complexes possess good catalytic performance, such as high catalytic activity and molecular weight, as well slight broader dispersity, showing the potential utility of the dibenzopyranyl group as substituent.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/molecules27175455/s1, Figure S1: GPC curves of the obtained polyethylene (a); activity and M w as a function of ethylene pressure, atm (b) for the Co1/MAO system at optimized condition (Table 2, entries 7, 15-16). Figure S2: GPC curves of the obtained polyethylene (a); activity and Mw as a function of ethylene pressure, atm (b) for the Co1/MAO system at optimized condition (Table 5, entries 9, 15-16). Figure S3: The 1 H-NMR spectrum of the polyethylene obtained with Co1/MAO (Table 2, entry 7). Figure S4: The 13 C-NMR spectrum of the polyethylene obtained with Co1/MAO (Table 2, entry 7). Figure S5: Definition of open cone angle (θ) of complex. Table S1: Comparisons of bond lengths and bond angles between the calculated geometry and experimental crystal data for complex Co1 along with the standard deviation (δ) and energy variation (∆E) values at various spin states. Table S2: Comparisons of bond lengths and bond angles between the calculated geometry and experimental crystal data for complex Co4 along with the standard deviation (δ) and energy variation (∆E) values at various spin states. Table S3: The values of correlation coefficient (R 2 ) for Co complexes by the combinations of two and single descriptors. Ref. citation of [43][44][45][46][47][48][49][50].