Fluorine- and Trifluoromethyl-Substituted Iminopyridinenickel(II) Complexes Immobilized into Fluorotetrasilicic Mica Interlayers as Ethylene Oligomerization Catalysts

Abstract

Heterogeneous catalysts comprising immobilized nickel(II) complexes bearing a fluorine- or trifluoromethyl-substituted iminopyridine ligand (X_n-C₆H_5–n-N=C (CH₃)-C₅H₅N, X = F or CF₃) in fluorotetrasilicic mica interlayers were prepared by reacting Ni²⁺-exchange fluorotetrasilicic mica with the appropriate ligand. Upon activating the precatalyst with triethylaluminum or triisobutylaluminum, the generated active species showed catalytic activity for ethylene oligomerization, yielding low-molecular-weight polyethylene (PE), ethylene oligomers, and wax-like PE. The oligomer distribution almost agreed with what we expected according to the Schultz–Flory distribution. However, the amount of solid products was much higher than the theoretical value, indicating that at least two active species were formed, i.e., the oligomer and low-molecular-weight PE. The precatalyst with a 2,4-F₂C₆H₃ group on the imino nitrogen atom activated by triethylaluminum showed the highest catalytic activity for ethylene oligomerization (408 g-C₂ g-cat⁻¹ h⁻¹), with selectivities to the liquid and solid products of 51.0% and 11.5%, respectively, with the rest of the product corresponding to wax-like PE. Meanwhile, the highest selectivity to the liquid product (66.7% at 233 g-C₂ g-cat⁻¹ h⁻¹) was obtained using the precatalyst with a 2-FPh group on the imino nitrogen atom activated by triisobutylaluminum.

1. Introduction

Ethylene oligomerization is an essential industrial process for preparing raw materials of various chemicals and base oils of lubricants [1]. Numerous late transition metal complexes have been designed and applied for ethylene oligomerization owing to their high catalytic activity upon activation with methylalumoxane (MAO) [2,3,4,5,6,7,8]—in particular, catalysts consisting of a Ni complex bearing a bidentate or tridentate ligand and MAO exhibit high catalytic activity for ethylene polymerization/oligomerization. Therefore, modifications of the ligand backbone in Ni-based catalysts have been widely investigated, resulting in the development of remarkable ligands [3,4,5,6,7,8,9,10,11]. For instance, iminopyridines were reported as useful bidentate ligands for Ni-based polymerization catalysts, and their complexes showed moderate catalytic activity for ethylene polymerization producing characteristic polyethylene (PE) with short branches upon activation with MAO [12,13,14,15,16,17,18,19,20,21,22]. However, the molecular weight of the PE obtained from iminopyridine-bearing Ni catalysts is relatively low owing to the low steric bulk around the metal center. Therefore, these complexes are suitable for producing ethylene oligomers [23,24,25,26,27,28,29], which are used as an intermediate of oxo alcohols, lubricants, and other chemicals. In addition, the oligomerization of bioethylene produced from bioethanol to afford hydrocarbons with carbon numbers 8–16 has received much attention because these hydrocarbons form environmentally friendly jet fuel [30,31,32].

As a recent development in ethylene oligomerization, heterogeneous catalysts consisting of nickel complexes and metal–organic frameworks (MOFs) were investigated by Arrozi et al. [33]. These catalyst systems, combined with Et₂AlCl, afforded only C₄ and C₆ products with high catalytic activity. Wei et al. also developed notable heterogeneous catalysts composed of nickel complexes and covalent organic frameworks with a porous structure [34]. Although these catalysts exhibited reproducibility with moderate activity, the major products were C₄ and C₆. In addition, these reports did not discuss the detailed composition of the C₆ and C₈ fractions.

Many previous studies on ethylene oligomerization have focused on the production of linear α-olefins, which are valuable as comonomers for linear low-density polyethylene and as feedstocks for various industrial chemicals. In contrast, a certain amount of branched hydrocarbons are desirable for bio-jet fuel applications as they provide good fluidity. Olefins with an inner double bond are also advantageous because they can be converted into branched hydrocarbons through coupling reactions catalyzed by acid.

Recently, we developed an effective procedure for the preparation of heterogeneous catalysts comprising late transition-metal complexes immobilized into the interlayer space of layered clay minerals via the direct reaction of metal ion-exchanged clay minerals (Mⁿ⁺–clay, Mⁿ⁺ = Fe³⁺, Co²⁺, and Ni²⁺) with appropriate ligands such as bis(imino)pyridine [35], acetylimnopyridine [36], α-diimine [37,38,39,40,41,42], and quinolineimine [43]. The prepared precatalysts could be readily activated by conventional aluminum compounds such as triethylaluminum (TEA), triisobutylaluminum (TIBA), and MAO, showing high catalytic activity for ethylene polymerization/oligomerization with excellent chemical processability, that is, with morphology control and no fouling.

In this study, we prepared catalysts via immobilizing iminopyridinenickel(II) complexes into the interlayers of fluorotetrasilicic mica by reacting iminopyridine ligands with F- or CF₃ groups on the iminophenyl ring (L = 1–6) with Ni²⁺-exchanged fluorotetrasilicic mica (Figure 1). Additionally, precatalysts containing unsubstituted, CH₃-substituted, and PhO-substituted ligands (L = 7, 8, and 9, respectively) were also prepared via a one-pot synthesis method originally developed by our research group. This approach is particularly useful for preparing precatalysts with an unstable complex bearing a less sterically hindered substituent. The catalytic performance of the prepared heterogeneous catalyst systems in ethylene oligomerization was evaluated, with special attention paid to the product distribution—particularly the contents of branched olefins, as well as linear terminal and internal olefins.

2. Results and Discussion

2.1. Characterization of Precatalysts

Figure 2 shows the X-ray diffraction (XRD) profiles of the prepared precatalysts. The basal spacing of dried Ni²⁺–exchanged mica (hereinafter referred to as Ni²⁺–mica) was 1.0 nm, corresponding to the thickness of one layer. After treating the dried Ni²⁺–mica with acetonitrile (CH₃CN), the basal spacing changed to 1.2 nm, indicating that CH₃CN was intercalated into the interlayers of the mica and oriented along the mica sheets. When Ni²⁺–mica was treated with the iminopyridine ligands in CH₃CN, the peak assigned to the basal spacing in the XRD profile shifted to low diffraction angles due to ligand intercalation. The interlayer spacing between two layers (basal spacing–layer thickness) ranged between 0.24 and 0.27 nm. These values are smaller than the expected size of an iminopyridinenickel(II) complex, probably owing to distortion of the Ni complex in the interlayer due to the strong Coulomb force between Ni²⁺ ions and the negative charged mica sheets. Moreover, the phenyl ring on the iminonitrogen atom in the ligand was probably tilted toward the mica sheet.

The thermogravimetry-differential thermal analysis (TG-DTA) results of CH₃CN-treated Ni²⁺–mica and 3/Ni²⁺–mica were compared to study the decomposition of the Ni(II) complex. Although the precatalyst used for the oligomerization (

Table 1. Ethylene oligomerization using PI/Ni²⁺–mica precatalyst in the presence of aluminum compound ¹.

Entry	L	Time	R3Al	Activity 2	Selectivity/wt%		Selectivity in Alkenes 4/mol%			MB 5	α 6
		/h			Solid	Liquid 3	SLAO	SIO	SBO	/%
1	1	1	TEA	333	9.2	55.3	43.7	46.6	9.7	64.7	0.72
2	2	1		249	16.0	46.7	42.7	48.2	9.1	62.9	0.78
3	3	1		408	11.5	51.0	49.6	43.5	6.9	62.9	0.73
4	4	1		175	19.4	41.1	49.6	42.7	7.7	60.9	0.77
5	5	1		312	40.5	10.5	34.4	44.7	20.9	50.9	0.81
6	6	1		210	16.2	46.8	61.3	32.0	6.7	62.9	0.73
7 7	7	1.25		236	14.3	50.6	43.6	47.8	8.6	64.8	0.74
8 7	8	1.25		229	59.8	7.5	31.4	48.9	19.7	67.3	0.85
9 8	9	1		80	20.8	14.9	56.3	nd 9	nd 9	56.3	0.69
10	1	2	TIBA	233	8.4	66.7	44.0	46.6	9.4	75.4	0.68
11	2	2		195	19.2	54.8	43.0	47.9	9.1	74.2	0.69
12	3	2		241	15.8	52.2	49.9	43.0	7.1	68.6	0.78
13	4	2		168	19.1	48.1	51.4	40.7	7.9	67.6	0.78
14	5	1		419	40.3	15.8	40.0	42.7	17.4	56.1	0.85
15	6	1		296	20.1	57.8	44.9	45.5	9.6	78.1	0.72
16	3	1	MAO	~0	–	–	–	–	–	–	–

¹ Reaction conditions: Temperature = 50 °C, C₂ pressure 0.40 MPa, n-hexane (solvent) = 50 mL, R₃Al = 0.60 mmol, L/Ni²⁺–mica = 200 μmol g⁻¹, catalyst weight = 5.0 mg. ² Activity (g-C₂ g-cat⁻¹ h⁻¹) was determined based on ethylene consumption and the amount of used catalyst. ³ Total amounts from C₄ to C₁₈. ⁴ S_LAO, S_IO, and S_BO indicate the selectivity to linear terminal alkenes, linear internal alkenes, and branched alkenes, respectively. ⁵ MB = Material balance = 100 × (liquid product weight + solid product weight)/(total C₂ consumption by weight). ⁶ Schultz–Flory constant determined based on mole fraction of C₈ to C₁₆ alkenes. ⁷ Prepared using the one-pot synthesis method. ⁸ C₂ pressure 0.70 MPa. ⁹ Not determined.

, entry 3) was prepared using 200 μmol of ligand per g of Ni²⁺–mica, the 3/Ni²⁺–mica precatalyst for the TG-DTA measurement was prepared using 800 μmol of ligand per g of Ni²⁺–mica to ensure that the weight loss due to decomposition of the complex was appropriately detected. The results are shown in Appendix B (Figure A1). When the ligands reacted with Ni²⁺ ions located in the mica interlayers to form coordination complexes, the theoretical ligand content in the resulting precatalyst was estimated to be 15.7 wt.% (800 μmol g-cat⁻¹), assuming that all the ligands participated in the complex formation. In the TG–DTA profile, the sample of 3/Ni²⁺–mica exhibited an additional weight loss of 5.0 wt.% in the temperature range of 350–800 °C compared with CH₃CN-treated Ni²⁺–mica. This additional weight loss is attributed to the thermal decomposition of the coordinated ligands. The observed weight loss corresponds to 215 μmol of L3 per gram of catalyst (molar mass of L3 = 232 g-mol⁻¹), indicating that approximately 215 μmol of nickel complexes were actually formed within the mica interlayers.

Based on the FT-IR spectrum of 3/Ni²⁺–mica, we confirmed the formation of nickel complexes in the mica interlayers. In our previous work, the absorption peak corresponding to the C=N stretching vibration, which appeared when a C=N bond was coordinated to a nickel center in the mica interlayers, was typically observed around 1590–1600 cm⁻¹ [36,42,43]. In the present spectrum, this characteristic peak was clearly observed at 1601 cm⁻¹, indicating that the desired nickel complexes were successfully formed in the mica interlayers. Additional peaks corresponding to the bending vibration (1370 cm⁻¹) attributed to a CH₃ group and the stretching vibrations (1446, 1501, and 1573 cm⁻¹) assigned to the C=C and C=N bonds of the aromatic ring were also observed.

2.2. Catalyst Performance in Ethylene Oligomerization

2.2.1. Effects of the Ligand Structure on the Catalyst Performance

Table 1 shows the results of the ethylene oligomerization using the 1–9/Ni²⁺–mica precatalysts in the presence of TEA, TIBA, or MAO.

The precatalysts with L7-8 backbones prepared using the one-pot synthesis method were also evaluated. The catalysts having an F-substituted phenylimino ligand showed moderate activity for the formation of ethylene oligomers compared with the F-unsubstituted analog (Table 1, entries 7 and 8). The ethylene oligomerization afforded liquid and solid products and wax-like oligomer. The selectivities shown in Table 1 indicate the weight fraction of the liquid and solid products relative to the consumed ethylene. Because the wax-like oligomer with a carbon number over 20 was soluble in n-heptane (solvent), the amount of dissolved wax-like oligomers could be determined as the difference between 100% and the material balance. The selectivity to linear terminal alkenes (linear α-olefins) is expressed as “S_LAO.” Apart from linear α-olefins, linear internal olefins (IO) were formed via migration of the Ni²⁺ center bonded to the terminal carbon of the chain to the β-position followed by abstraction of a γ-hydrogen on the chain. Moreover, branched olefins (BO) were formed by inserting another ethylene molecule into the Ni²⁺–alkyl bond after the chain walking. These mechanisms are summarized in Scheme 1. The a value is the Schulz–Flory distribution constant, which was calculated according to the amount of produced linear α-olefins, and correlates with the content of oligomers with longer chains, with higher a value indicating higher content of such oligomers.

Among the 1–4/Ni²⁺–mica precatalysts having F-substituted iminopyridine ligands (F_nC₆H_5-n-N=C(CH₃)-C₅H₄N, n = 1–3), the 3/Ni²⁺–mica precatalyst (n = 2) showed the highest activity for ethylene oligomerization upon activation with a conventional aluminum compound. When the number of F atoms (n) on the iminophenyl ligand increased from 2 to 3, the activity decreased (Table 1, entries 3 and 12 vs. entries 4 and 13). Because oligomerization catalysts consisting of a transition-metal complex and an aluminum compound are typically used only once, catalyst lifetime is an important factor for improving productivity.

Figure 3a shows the activity profiles of the catalysts based on the ethylene consumption. A higher consumption rate was observed when the oligomerization was performed using the 1/Ni²⁺–mica and 3/Ni²⁺–mica precatalysts with an o-F substituent (Table 1, entries 1 and 3). According to the results shown in entries 7 and 8, Table 1, the bulkiness of the substituent at the o-position on the iminophenyl ring was not crucial for the catalytic activity. Therefore, the electronic effect of the o-F substituent was most likely the main factor determining the increase in activity. However, the activity based on the product yield depended on several factors, such as the activation efficiency, propagation rate, and deactivation rate.

The catalysts bearing the more electron-withdrawing group CF₃ in the iminophenyl group exhibited high activities (Table 1, entries 5, 6, 13, and 14). The catalyst having an o-CF₃ group showed higher activity than that with an m-CF₃ group, irrespective of the type of activator (Table 1, entries 5 and 6 or entries 14 and 15). The o-CF₃ group affected the product distribution. Thus, the amount of solid products was 40.5% when using the catalyst having the o-CF₃ group, whereas that obtained with the catalyst having the m-CF₃ group was 16.2%. The a value observed in entry 5 (0.81) was significantly higher than that in entry 6 (0.73), indicating that an o-CF₃ group suppressed the elimination of the growing chain.

The catalyst prepared with the o-CH₃-substituted ligand (Entry 8) exhibited moderate activity. In contrast, the precatalyst containing an o-PhO substituent (Entry 9) showed lower activity and produced a smaller fraction of liquid products. The selectivity toward liquid and solid products depended strongly on the steric bulk of the o-substituents on the iminophenyl ring of the ligand. Precatalysts bearing o-CH₃ or o-CF₃ groups predominantly afforded solid products (Table 1, Entries 5 and 8). When the catalyst with an o-PhO substituent was employed, although the proportion of solid products was relatively low (20.8 wt.%), the combined fraction of solid and wax-like products reached 74.5 wt.%, indicating that chain propagation was the dominant reaction pathway.

Conversely, for the precatalyst bearing the CF₃ substituent at the m-position, the selectivity to solid products decreased and liquid products increased (Table 1, entry 6). Meanwhile, no significant change in the selectivity to liquid products was observed when H atom was replaced with F at the o-position of the iminophenyl ring (Table 1, entry 7 vs. entries 1 and 3). These results indicated that the selectivity to the liquid products depended on the steric bulk of the substituent at the o-position of the iminophenyl ring.

Careful observation of the ethylene consumption profiles provided useful information. In all cases, an induction period was observed during oligomerization, which was shortened with increasing the number of the F-substituents on the iminophenyl group. The o-F substituent seemed to increase the reactivity of the complex toward the aluminum compound, resulting in a faster formation of active species. In contrast, 4/Ni²⁺–mica having 2,4,6-F₃-substituted phenyl group showed low activity, which can be attributed to the faster deactivation of this catalyst compared with other catalysts. Meanwhile, no clear difference was observed between the a values of the 1-4/Ni²⁺–mica precatalysts, suggesting similar propagation rates. The same trend was previously observed in the polymerization using F-substituted α-diimine Ni(II) complexes immobilized into mica interlayers [37].

2.2.2. Detailed Analysis of Alkene Products

In the oligomerization, the terminal and internal linear olefins were formed along with the small amounts of branched olefins. In a typical run (Table 1, entry 3), S_LAO was 49.6 mol%, and the mole fractions of internal and branched olefins were 43.5 and 6.9 mol%, respectively. In the C₈ fraction, the major internal olefin products were 2- and 3-octenes, which were formed via the migration of the C₈ chain on the Ni center from the α-carbon to the β-carbon before elimination. Unfortunately, the branched olefins could not be identified via gas chromatography–mass spectrometry (GC-MS) because of the lack of standard samples. However, according to the proposed mechanism shown in Scheme 1, the branched C₈ products were most likely 3-methyl-1-heptene and 5-methyl-1-heptene. The mole fraction of C₈ products is summarized in

Table 2. Product distribution in C₈ fraction ¹.

Entry	L	Selectivity in C8 Alkenes/mol%
		1O	(E)2O	(Z)2O	(E)3O	(Z)3O	3M1H	5M1H
1	1	43.1	15.1	28.1	5.4	5.8	1.6	1.0
2	2	41.5	16.5	27.7	6.9	4.5	1.7	1.2
3	3	48.2	14.4	26.3	4.5	4.8	1.1	0.6
4	4	50.3	14.3	25.4	4.5	3.6	1.1	0.9
5	5	32.6	9.8	31.1	8.3	9.5	6.4	2.3
6	6	60.4	11.5	21.7	2.9	2.2	0.8	0.5
7	7	39.0	17.3	26.7	5.7	6.0	1.5	3.8
8	8	27.9	12.9	31.9	8.1	11.1	5.2	3.0
9	1	43.5	15.1	28.0	5.9	4.8	1.6	1.1
10	2	42.2	16.0	27.8	6.6	4.6	1.8	1.1
11	3	49.3	14.2	26.6	3.9	3.9	1.2	0.9
12	4	51.4	14.2	24.9	4.3	3.5	1.2	0.5
13	5	37.7	8.8	26.4	9.4	11.3	4.8	1.7
14	6	44.3	15.4	27.3	5.4	4.7	1.8	1.1

¹ 1O = 1-octene, (E)2O = (E)-2-octene, (Z)2O = (Z)-2-octene, (E)3O = (E)-3-octene, (Z)3O = (Z)-3-octene, 3M1H = 3-methyl-1-heptene, 5M1H = 5-methyl-1-heptene.

. The primary C₈ alkene product was 1-octene, and 2-octenes were favorably formed over 3-octenes in all reactions. According to the chain walking mechanism, these product distributions agreed with the expected ones. The (Z)/(E) ratio of 2-octenes reveals that the formation of the (Z) isomer was more favorable than that of the (E) isomer due to steric limitation. This trend was observed in all linear 2-alkenes.

Scheme 1. Possible formation route of linear and branched alkenes.

Scheme 1. Possible formation route of linear and branched alkenes.

The product distributions summarized in Table 2 (entries 1–4) exhibited nearly identical trends: over 90% of the product were linear terminal and internal olefins (1-, 2-, and 3-octenes) with only a small fraction of branched isomers (3-methyl-1-heptene and 5-methyl-1-heptene), when oligomerization was conducted using 1–4/Ni²⁺–mica, regardless of the aluminum compound employed. A similarly high selectivity toward linear olefins was also observed for oligomerization using 6/Ni²⁺–mica bearing an m-CF₃-substituted iminopyridine ligand combined with TEA; however, this catalyst exhibited a notably higher selectivity for 1-octene (Entry 6). Although the origin of this specific behavior remains unclear at this stage, the absence of a comparable trend for 7/Ni²⁺–mica containing an unsubstituted ligand suggests that the m-CF₃ substituent plays a crucial role in inducing this effect.

In contrast, 5/Ni²⁺–mica incorporating an o-CF₃-substituted iminopyridine ligand afforded higher selectivity toward branched C₈ olefins, namely 3-methyl-1-heptene and 5-methyl-1-heptene (Table 2, Entry 5). The propagation-to-elimination ratio (α) increased, likely due to the steric bulk exerted by the o-substituent on the iminophenyl group. A similar product distribution was observed when 8/Ni²⁺–mica, bearing a ligand with an electron-donating o-CH₃ group, was used for oligomerization (Table 2, Entry 8).

Furthermore, when catalysts containing ligands with relatively bulky o-substituents were employed (Table 1, Entries 5, 8, and 13), the proportion of solid products increased as that of liquid products decreased. Because branched olefins are generated through ethylene insertion following chain walking, such insertion must occur prior to chain termination. The o-substituent appears to suppress chain elimination during the chain-walking process via steric interference. This suppression extends the chain length, thereby promoting the formation of solid or wax-like products while decreasing the yield of liquid olefins.

2.2.3. Effects of Oligomerization Temperature and Ethylene Pressure on Activity and Product Distribution

Figure 4 shows the profiles of the C₂ consumption rate when the catalyst consisting of 3/Ni²⁺–mica and either TEA or TIBA was employed. Both aluminum compounds effectively activated the precatalyst, yet a distinct difference was observed in the induction period. Specifically, TIBA resulted in a considerably longer induction period. The duration of this period shortened as the oligomerization temperature increased, regardless of the type of aluminum compound. At 40 °C, a prolonged induction period and a slow activation profile were evident, particularly for the TIBA-activated system. In contrast, at 70 °C, the precatalyst was readily activated by the aluminum compound, but the maximum consumption rate (peak height) decreased. After reaching this peak, the catalyst gradually deactivated over time, and the rate of deactivation accelerated with increasing temperature. The time required to reach the peak is presumed to depend on the balance between activation and deactivation rates. The deactivation rate observed during the oligomerization with TIBA was slower than that with TEA, suggesting that the aluminum compound participated in the deactivation process.

The prolonged induction period observed here can be interpreted based on the activation mechanism previously proposed for Cp₂ZrCl₂/Mg²⁺–mica catalysts [44]. In the initial stage of the activation, the aluminum compound primarily reacts with the complex located at the edges of the stacked mica layers. Subsequently, the mica sheets are gradually exfoliated as alkyl chains grow on the formed active species. Finally, the sites situated within the inner layers become exposed and play activation of the complexes. Therefore, the extended induction period observed under conditions employing a weaker activator or lower temperature can be attributed to the sluggish initial activation of complexes at the layer edge. A comparable induction behavior was also reported in our previous studies using the mica-based precatalysts [36,37,39,40].

Table 3. Effects of temperature and ethylene pressure on catalyst performance of 3/Ni²⁺–mica ¹.

Entry	T	PC2	Time	Activity 2	Selectivity/wt%		Selectivity in Alkenes 4/mol%			MB 5	α 6
	/°C	/MPa	/h		Solid	Liquid 3	SLAO	SIO	SBO	/%
17	40	0.4	1.5	388	34.9	37.5	63.5	31.2	5.4	72.4	0.79
3	50	0.4	1.0	408	11.5	51.0	49.6	43.5	6.9	62.9	0.73
18	60	0.4	1.0	258	3.1	76.7	36.6	54.6	8.8	81.0	0.64
19	70	0.4	1.0	171	1.1	74.0	27.3	60.2	12.6	77.0	0.61
20	50	0.2	1.0	242	5.3	67.2	34.8	55.6	9.5	72.7	0.66
21	50	0.7	1.0	581	24.3	49.1	60.0	34.8	5.3	73.5	0.71
22 7	50	0.4	2.0	237	17.6	58.5	51.7	41.1	7.2	76.2	0.75
23 8	50	0.4	1.5	283	16.3	48.9	53.8	40.0	6.3	65.2	0.72
24 9	50	0.4	1.5	121	16.0	53.0	54.0	40.3	5.6	69.0	0.75

¹ Reaction conditions: T = reaction temperature, P_C2 = C₂ pressure, n-hexane (solvent) = 50 mL, Et₃Al =0.30 mmol, catalyst weight = 5.0 mg, 3/Ni²⁺–mica = 200 μmol g⁻¹. ² g-product g-cat⁻¹ h⁻¹. ³ Total amounts from C₄ to C₁₈. ⁴ S_LAO, S_IO, and S_BO indicate the selectivity to linear terminal alkenes, linear internal alkenes, and branched alkenes, respectively. ⁵ Material balance = 100 × (liquid product weight + solid product weight)/(total C₂ consumption by weight). ⁶ Schultz–Flory constant determined based on mole fraction of C₈ to C₁₆ alkenes. ⁷ Prepared by the one-pot synthesis method (3/Ni²⁺–mica = 200 μmol g⁻¹). ⁸ Prepared by the one-pot synthesis method (3/Ni²⁺–mica = 800 μmol g⁻¹). ⁹ Prepared by the one-pot synthesis method (3/Ni²⁺–mica = 1400 μmol g⁻¹).

summarizes the results of oligomerization using the catalyst composed of 3/Ni²⁺–mica and TEA. As the oligomerization temperature increased (Entries 17 to 19), the catalytic activity and selectivity to solid products decreased, whereas the selectivity to the liquid products increased. This trend can be rationalized by the decrease in the α value with increasing temperature. Because the activation energy for the chain-transfer reaction is higher than that for chain propagation, the chain-transfer process—leading to the formation of shorter-chain olefins—became more dominant at elevated temperatures. Furthermore, the selectivities to internal olefins (S_IO) and branched olefins (S_BO) increased, while that to 1-olefins (S_LAO) decreased as the temperature rose. This shift can be attributed to the promotion of chain walking through β-hydrogen elimination under higher-temperature conditions.

The results presented in Entries 3, 20, and 21 demonstrate the effects of ethylene pressure on catalytic activity and product selectivity. As the ethylene pressure increased, the activity rose from 242 g-product g-cat⁻¹ h⁻¹ at 0.2 MPa to 581 g-product g-cat⁻¹ h⁻¹ at 0.7 MPa. The proportion of solid products also increased from 5.3 wt.% to 24.3 wt.% with increasing ethylene pressure, owing to the higher ethylene concentration in the reaction medium. In contrast, the chain-propagation reaction became faster than the chain-walking process as the ethylene concentration increased, resulting in a decrease in the selectivity to branched olefins.

The influence of the ligand-to-nickel molar ratio on catalyst performance was also investigated using 3/Ni²⁺–mica catalysts prepared via the one-pot synthesis method (entries 22–24). As the ligand-to-nickel ratio increased, the catalytic activities markedly decreased—from 237 g-product g-cat⁻¹ h⁻¹ at 200 μmol g-Ni²⁺–mica⁻¹ to 121 g-product g-cat⁻¹ h⁻¹ at 1400 μmol g-Ni²⁺–mica⁻¹. This trend is likely attributed to the preferential formation of bis-liganded nickel complexes under higher ligand concentrations. Such species are known to be inactive or only weakly active toward ethylene oligomerization, particularly when the ligand is sterically less hindered.

Overall, the reaction conditions significantly affected the catalyst performance, including both activity and product distribution. To efficiently produce liquid oligomers rich in branched olefins, higher reaction temperatures and lower ethylene pressure were favorable. However, these conditions simultaneously led to a decrease in catalytic activity. Although the activity would normally be expected to increase with rising temperature, an opposite trend was observed, likely due to the thermal instability of the active species.

Figure 5 shows the Schultz–Flory distribution plots of the carbon number vs. the logarithm of the mass fraction corresponding to the conditions described in entries 3 and 19, Table 3. The plots were constructed according to the equation

C_n = α² (1 − α)ⁿ⁻¹

(1)

where n is the carbon number in the oligomer chain and C_n is the mole fraction of the carbon in the oligomer with a carbon number of n. A linear correlation was obtained between ln C_n and n, with the slope being experimentally determined and representing the ratio of the rate constant for chain propagation to that for chain migration. Thus, higher α values indicate that the chain propagation is dominant. Although both plots showed an almost linear correlation, they gradually diverged from the line at higher carbon numbers. These results suggest that diverse active species with different α values were formed from the precatalyst. When the oligomerization temperature was increased, the α value obtained from the slope increased considerably. In alkene oligomerization using Ziegler-type catalysts, a temperature increase is favorable for generating shorter-chain products because the activation energy for chain migration is higher than that for chain propagation.

Table A1. Mole fraction of each carbon number in liquid products.

Entry	Mole Fraction in Liquid Products (C4-C18)/mol%
	C4	C6	C8	C10	C12	C14	C16–18
1	31.4	24.4	14.0	9.6	7.1	5.2	8.4
2	28.4	21.4	13.7	10.4	8.0	6.6	11.6
3	28.9	22.8	14.0	11.6	8.2	5.7	8.8
4	29.0	21.0	13.7	10.3	8.1	6.3	11.6
5	22.1	18.9	14.6	10.3	10.3	8.8	15.1
6	28.9	22.5	14.9	10.1	7.7	6.1	9.8
7	42.8	22.5	12.2	7.8	5.4	4.0	5.5
8	28.5	20.0	16.1	11.6	9.0	6.2	8.6
9	42.4	22.6	12.1	8.1	5.4	3.7	5.6
10	45.9	20.3	11.2	7.9	5.4	3.9	5.6
11	35.6	19.3	12.4	9.1	7.2	5.6	10.8
12	38.1	18.4	11.8	8.9	6.8	5.7	10.2
13	56.4	11.4	7.0	6.0	6.0	4.5	8.8
14	42.4	20.3	11.6	8.3	5.8	4.5	7.0
16	26.3	22.1	13.5	10.7	7.8	6.6	13.0
17	42.2	24.4	12.6	7.3	5.2	3.9	4.4
18	52.8	23.0	9.5	5.0	3.9	2.8	2.9
19	40.6	24.0	12.4	7.9	5.7	4.0	5.3
20	26.9	24.8	15.4	11.1	7.9	5.4	8.6

(Appendix A) indicates the relationship between the mole fraction of the generated oligomer and the carbon numbers. At oligomerization temperatures of 40 °C, 50 °C, 60 °C, and 70 °C, the mole fractions of C₄ products were 26.3%, 28.9%, 42.2%, and 52.8%, respectively (Table 3, entries 17, 3, 18, and 19, respectively). Meanwhile, the mole fractions of C_6-18 components were 13.0%, 8.8%, 4.4%, and 2.9%, respectively. These trends were in good agreement with the theory as mentioned above.

Meanwhile, the divergence from the linear correlation in the plot obtained at 70 °C was more pronounced than that obtained at 50 °C, which contradicts the results expected according to the difference in activation energy as discussed earlier. This can be attributed to the active species for generating solid products exhibiting higher thermal stability than those for generating liquid products.

2.3. Precatalyst Prepared Using One–Pot Synthesis Method

Compared with the results obtained using the standard method (Table 3, entry 3), the activity of the catalysts prepared using the one-pot synthesis method was low (Table 3, entries 21–23), but the selectivities to the solid and liquid products for both type of the catalysts were nearly equal. No apparent differences were observed in the selectivity to alkenes, which suggests that the same active species were formed for the two preparation methods. Comparative experiments are being planned to investigate the influence of ligand structure on the catalytic performance of catalysts prepared by one-pot and conventional methods.

The one-pot synthesis method has some advantages, for example, no ligand preparation and purification are required. We have already applied this method to prepare heterogeneous catalysts with various ligands [40,43].

2.4. Comparison with Homogeneous Catalyst System

The catalytic performance of 3/Ni²⁺–mica was compared with that of a homogeneous catalyst system consisting of the 3–NiCl₂ complex and modified methylalumoxane (MMAO). The results are summarized in Table S2 of the Supporting Information. The nickel-based activity of 3/Ni²⁺–mica was estimated based on the assumption that all ligands used in the catalyst preparation (200 μmol g-Ni²⁺–mica⁻¹) reacted with nickel ions in the interlayer spaces. This assumption likely underestimates the actual catalytic activity per nickel complex, because not all ligands may have reacted with the nickel ions. A noteworthy feature of this system is that 3/Ni²⁺–mica was readily activated by TEA (or TIBA) and exhibited high catalytic activity without requiring the expensive cocatalyst MAO. The catalytic activity of 3/Ni²⁺–mica combined with TEA (4090 kg-product mol-Ni⁻¹ h⁻¹) was substantially higher than that of the homogeneous 3–NiCl₂/MMAO system (130 kg-product mol-Ni⁻¹ h⁻¹).

The homogeneous catalyst, however, showed a clear advantage in product selectivity: the proportion of solid products obtained with the homogeneous system was markedly lower than that obtained with the 3/Ni²⁺–mica catalyst, reflecting the formation of more uniform active species in the homogeneous medium. In contrast, the mica sheet in the heterogeneous catalyst likely acts as large physical barriers surrounding the active species. Moreover, the nickel complexes are electrostatically attracted to the negatively charged mica layers, which may induce structural distortion of the complexes and lead to the generation of multiple types of active species.

The lifetimes of the active species in both the homogeneous and heterogeneous catalyst systems were evaluated, and the results are presented in Figure S8 of the Supporting Information. These results clearly illustrate the distinct characteristics of the two systems. The homogeneous catalyst rapidly generated active species and reached its maximum activity within several minutes after activation. In contrast, for the 3/Ni²⁺–mica/TEA catalyst system, the activity gradually increased to its maximum through a moderate activation process. After reaching the peak, deactivation proceeded slowly over more than 3 h. The long lifetime of the catalyst observed in the heterogeneous 3/Ni²⁺–mica catalyst system represents one of its major advantages over the homogeneous counterpart.

3. Materials and Methods

3.1. Precatalyst Preparation

Fluorotetrasilicic mica (Na⁺–mica, ME-100) was purchased from Katakura & Co-op Agri Corporation (Tokyo, Japan). The ion-exchange reaction of Na⁺–mica was performed using an aqueous solution of Ni(NO₃)₂·6H₂O according to a previously reported procedure [36]. After the reaction, the recovered solid was calcined at 200 °C for 4 h in air and then dried at the same temperature for 4 h under reduced pressure to obtain Ni²⁺–mica. The Ni²⁺ content and surface area of Ni²⁺–mica were determined via X-ray fluorescence (Malvern Panalytical Ltd., Tokyo, Japan, AXIOUS) and the BET method (MicrotracBEL Corp., Osaka, Japan, BELSORP Mini II), respectively. These data are summarized in Table S1 in the Supporting Information.

The iminopyridine ligands (L = 1–6) were prepared from 2-acetylpyridine and substituted aniline using a previously reported method [13] and via ¹H nuclear magnetic resonance spectroscopy (Bruker Japan, Yokohama, Japan, Avance III 300) and mass spectrometry (JEOL Ltd., Tokyo, Japan, AM 500). Detailed are provided in the Supporting Information.

Ligands 1–6 and 9 (200 μmol per g of dried Ni²⁺–mica) were allowed to react with Ni²⁺–mica at 70 °C for 120 h in CH₃CN. Precatalysts with ligand amounts of 800 and 1400 μmol g-mica⁻¹ were also prepared using one-pot synthesis method to confirm the effect of the ligand/Ni²⁺ ratio on the activity and the product distribution. After the reaction, the solution was removed using a syringe, and the solid fraction was successively washed with CH₃CN, toluene, and n-hexane and dried at ambient temperature for 4 h under reduced pressure to obtain the 1–6 and 9/Ni²⁺–mica precatalysts. To clarify the effects of the ligands with F-containing substituents on the catalyst performance, precatalysts containing ligands having a 2-methylphenyl and a phenyl group on the nitrogen atom in the imino group (L = 7 and 8, respectively) were also prepared using the one-pot synthesis method [38]. Details of the one-pot synthesis method are provided in Appendix B.2, along with the XRD profiles of the obtained precatalysts.

For the XRD characterization (Rigaku Corporation, Tokyo, Japan, Ultima III), the precatalyst was dried at 110 °C for 1 h and then mixed with dry liquid paraffin, applied to a glass specimen holder, and covered with a polyester film in a glove box. The diffractometer was operated under following conditions: voltage, 40 kV; current, 40 mA; scan rate, 1.0° min⁻¹; range, 3–15°; sampling interval, 0.01°.

The amount of complex formed in the precatalyst was estimated on the basis of TG–DTA (Shimazu Corporation, Kyoto, Japan, DTG-60) of the precatalysts. The measurement was performed under N₂ flow (100 mL min⁻¹) at a temperature range from room temperature to 800 °C and a heating rate of 10 °C min⁻¹.

To confirm the formation of nickel complexes in the mica interlayers, 3/Ni²⁺–mica was characterized by FT-IR spectroscopy (Jasco Corporation, Tokyo, Japan, FT/IR 4100 equipped with an MCT detector), and the resulting spectrum is shown in Figure A3 in Appendix B.3. The sample for the FT-IR measurement was prepared by pressing a mixture of 40 mg of 3/Ni²⁺–mica and 20 mg of SiO₂ (binder) into a 10 mm-diameter wafer inside a glove box. The measurement conditions were as follows: resolution, 2 cm⁻¹; number of accumulations, 128.

3.2. Procedure of Ethylene Oligomerization

The ethylene oligomerization was performed at 40–70 °C and 0.2–0.7 MPa (gauge pressure) for 1–2 h using a 120 mL stainless steel autoclave equipped with a magnetic stirrer. Into the autoclave, n-hexane (as the solvent, 50 mL), the precatalyst (typically, 5.0 mg) as a toluene slurry, and the aluminum compound as the activator were added under N₂ atmosphere. The autoclave was purged with ethylene and then placed into a water bath maintained at the desired temperature. Ethylene was continuously supplied at a constant pressure during oligomerization.

After the oligomerization, a sample was taken out from the autoclave to analyze the products (butenes) in the gas phase, and the autoclave was then cooled in an ice bath. The liquid phase products (C₄-C₁₈ hydrocarbons) were quantitatively analyzed using a gas chromatograph in the presence of tridecane as an internal standard. The analysis conditions were as follows: methyl silicon column, 0.25 mm (internal diameter) × 60 m (length); initial conditions, 40 °C for 15 min; heating rate, 10 °C min⁻¹; final conditions, 220 °C for 10 min. The powdery products (low-molecular-weight PE) were recovered by filtration. The obtained PE powder was dried at 40 °C overnight and then weighed to determine the activity.

4. Conclusions

A series of iminopyridine Ni(II) complexes immobilized into fluorotetrasilicic mica interlayers were prepared as ethylene oligomerization catalysts via the reaction of Ni²⁺–mica and an appropriate iminopyridine ligand. All precatalysts were readily activated using conventional alkyl aluminum compounds such as TEA and TIBA. The products of the ethylene oligomerization reaction were ethylene oligomers consisting of a liquid fraction (C₄-C₁₈) and PE (wax-like PE and powdery PE). The selectivity to the liquid products changed by modifying the substituent on the iminophenyl group. Specifically, the catalyst with an o-methyl substituent on the iminophenyl group afforded the highest selectivity to wax-like PE. In contrast, the catalyst with a relatively smaller substituent, such as F and H at the o-position of the iminophenyl group predominantly produced ethylene oligomers.