Adjusting Ortho-Cycloalkyl Ring Size in a Cycloheptyl-Fused N,N,N-Iron Catalyst as Means to Control Catalytic Activity and Polyethylene Properties

Abstract

Five examples of bis(arylimino)tetrahydrocyclohepta[b]pyridine dichloroiron(II) complex, [2-{(Ar)N=CMe}-9-{N(Ar)}C₁₀H₁₀N]FeCl₂ (Ar = 2-(C₅H₉)-4,6-(CHPh₂)₂C₆H₂ Fe1, 2-(C₆H₁₁)-4,6-(CHPh₂)₂C₆H₂ Fe2, 2-(C₈H₁₅)-4,6-(CHPh₂)₂C₆H₂ Fe3, 2-(C₁₂H₂₃)-4,6-(CHPh₂)₂C₆H₂ Fe4, and 2,6-(C₅H₉)2-4-(CHPh₂)C₆H₂ Fe5), incorporating ortho-pairings based on either benzhydryl/cycloalkyl (ring sizes ranging from 5 to 12) or cyclopentyl/cyclopentyl groups, have been prepared in reasonable yield by employing a simple one-pot template strategy. Each complex was characterized by FT-IR spectroscopy, elemental analysis, and for Fe3 and Fe5 by single crystal X-ray diffraction; pseudo-square pyramidal geometries are a feature of their coordination spheres. On treatment of Fe1–Fe5 with modified methylaluminoxane (MMAO) or methylaluminoxane (MAO), a range in catalytic activities for ethylene polymerization were observed with benzhydryl/cyclopentyl-containing Fe1/MMAO achieving the maximum level of 15.3 × 10⁶ g PE mol⁻¹ (Fe) h⁻¹ at an operating temperature of 70 °C. As a key trend, the activity was found to drop as the ortho-cycloalkyl ring size increased: Fe1_C5H9/CHPh2~Fe5_C5H9/C5H9 > Fe2_C6H11/CHPh2 > Fe3_C8H15/CHPh2 > Fe4_C12H23/CHPh2. Furthermore, strictly linear polyethylenes (T_m > 126 °C) were formed with molecular weights again dependent on the ortho-cycloalkyl ring size (up to 55.6 kg mol⁻¹ for Fe1/MAO); narrow dispersities were a characteristic of all the polymers (M_w/M_n range: 2.3–4.7), highlighting the well-controlled nature of these polymerizations.

1. Introduction

The transition metal-catalyzed polymerization of ethylene is one of the most important carbon–carbon bond-forming reactions and moreover, it is widely used in the chemical industry [1,2]. Late transition metal complexes, and in particular those based on iron and cobalt [3,4,5,6,7,8], have emerged as effective catalysts in ethylene polymerization [9,10,11,12,13]; others involving the first row d-block centers such as nickel [14,15], chromium [16,17], and vanadium [18] have also shown great potential. While catalytic activity represents a key attribute of the catalyst, there is nowadays a drive toward systems that can operate effectively at high temperature whilst still producing high molecular weight polyethylene with a narrow molecular weight distribution. In this regard, variations to the classic bis(imino)pyridine-iron(II) halide structure [9,10,11,12,13] (A, Scheme 1) in the form of modifications to the ortho-/para-substitution pattern (i.e., steric and electronic effects) or even more dramatic structural changes to the N,N,N-ligand core itself have seen improvements in the catalytic performance and molecular weight of the polyolefin.

Of note, iron catalysts bearing carbocyclic-fused bis(imino)pyridine ligands have shown a capacity to display optimal productivity at temperatures in the range of 50–80 °C. In particular, bis(imino)pyridines bearing doubly or singly fused seven-membered rings have provided robust ligand frameworks for a range of iron(II) precatalysts (B–E, Scheme 1) [19,20,21,22,23,24]. In terms of catalytic performance, doubly fused B [19] bearing relatively small ortho-alkyl substituents displayed excellent activity (up to 10⁷ g PE mol⁻¹ (Fe) h⁻¹) with the molecular weight of the resulting polyethylenes falling in the range of 3.5–96.9 kg mol⁻¹ albeit with broad dispersity. By comparison, the installation of larger benzhydryl groups (CHPh₂) to the ortho-sites in C [20] led to similar high catalytic activities but afforded lower molecular weight polyethylene that exhibited narrower dispersity. For the singly fused iron(II) precatalysts D and E (Scheme 1), a noticeable increase in thermal stability was noted, especially for E in which the ortho-positions of the N-aryl group were substituted with cycloalkyl and hydrogen/methyl [22].

In this work, we revisit the singly-fused bis(arylimino)tetrahydrocyclohepta[b]pyridine skeleton shown in iron-containing D and E, with a view to introducing N-aryl ortho-pairings based on benzhydryl/cyclopentyl, benzhydryl/cyclohexyl, benzhydryl/cyclooctyl and benzhydryl/cyclododecyl; for comparative purposes, we also report the ortho-combination cyclopentyl/cyclopentyl (Scheme 1), as it is a common feature to all iron complexes to be synthesized is the presence of a para-benzhydryl group. To explore the impact of the size of ortho-cycloalkyl ring on catalyst activity, thermal stability as well as various polymer properties, a comprehensive polymerization study is presented involving changes to the type/amount of co-catalyst, temperature, pressure, and run time. Additionally, the synthetic and characterization data for all new iron complexes are disclosed.

2. Results

2.1. Materials and Methods

The synthesis and handling of air- and moisture-sensitive compounds were carried out under an atmosphere of nitrogen using standard Schlenk techniques. Prior to use, toluene was heated under reflux over sodium/benzophenone and then distilled. Methylaluminoxane (MAO) (1.46 M solution in toluene) and modified methylaluminoxane (MMAO) (1.93 M solution in n-heptane) were purchased from Albemarle Corporation. High-purity ethylene was purchased from Beijing Yanshan Petrochemical Co. and used as received. Other reagents were purchased from Aldrich or local suppliers. IR spectra were recorded using a PerkinElmer System 2000 FT-IR spectrometer. Elemental analysis was carried out with a Flash EA 1112 microanalyzer. Molecular weights and molecular weight distributions of the polymers were determined with an Agilent PL 220 GPC instrument operating at 150 °C with 1,2,4-trichlorobenzene as solvent. Melting temperatures of the polyethylenes were measured from the second scanning run on a Perkin-Elmer DSC-7 differential scanning calorimeter (DSC) under a nitrogen atmosphere. ¹H and ¹³C NMR spectra of the polyethylenes were recorded using a Bruker DMX 300 MHz instrument at 100 °C in deuterated 1,1,2,2-tetrachloroethane with tetramethylsilane (TMS) as an internal standard. The compounds 2-acetyl-5,6,7,8-tetrahydrocyclohepta[b]pyridin-9-one [25] and the 2-cycloalkylanilines [26,27,28] have been prepared using the literature methods.

2.2. [2-{(Ar)N=CMe}-9-{N(Ar)}C₁₀H₁₀N]FeCl₂ (Fe1–Fe5)

Ar = 2-(C₅H₉)-4,6-(CHPh₂)₂C₆H₂ Fe1. A mixture of 2-acetyl-5,6,7,8-tetrahydrocyclohepta[b]pyridin-9-one (0.061 g, 0.30 mmol), 2-cyclopentyl-4,6-dibenzhydrylaniline (0.140 g, 0.66 mmol), and iron(II) chloride tetrahydrate (0.057 g, 0.29 mmol) in glacial acetic acid (20 mL) was stirred and heated under reflux for 3 h. Once cooled to ambient temperature, the solvent was partially removed under reduced pressure, and an excess of diethyl ether was added to induce precipitation. Then, the precipitate was collected, washed with diethyl ether (4 × 15 mL), and dried to afford Fe1 as a green powder (0.197 g, 53%). FT-IR (cm⁻¹): 3025 (w), 2946 (m), 2865 (w), 1598 (m, ν(C=N)), 1569 (m), 1492 (s), 1448 (s), 1030 (m), 840 (w), 744 (m), 698 (s), 613 (m). Anal. Calcd for C₈₆H₇₉Cl₂FeN₃ (1281.34): C, 80.61; H, 6.21; N, 3.28. Found: C, 80.44; H, 6.16; N, 3.45%.

Ar = 2-(C₆H₁₁)-4,6-(CHPh₂)₂C₆H₂ Fe2. By employing the same procedure as that described for Fe1 but with 2-cyclohexyl-4,6-dibenzhydrylaniline as the aniline, Fe2 was isolated as a green powder (0.220 g, 58%). FT-IR (cm⁻¹): 3025 (w), 2927 (m), 1599 (m, ν(C=N)), 1572 (m), 1493 (s), 1447 (s), 1261 (m), 1030 (m), 744 (m), 698 (s), 614 (m). Anal. Calcd for C₈₈H₈₃Cl₂FeN₃ (1309.40): C, 80.72; H, 6.39; N, 3.21. Found: C, 80.48; H, 6.33; N, 3.44%.

Ar = 2-(C₈H₁₅)-4,6-(CHPh₂)₂C₆H₂ Fe3. By employing the same procedure as that described for Fe1 but with 2-cyclooctyl-4,6-dibenzhydrylaniline as the aniline, Fe3 was isolated as a green powder (0.246 g, 62%). FT-IR (cm⁻¹): 3025 (w), 2916 (m), 2854 (m), 1599 (m, ν(C=N)), 1572 (s), 1493 (s), 1446 (s), 1261 (m), 1030 (m), 744 (m), 697 (s), 613 (m). Anal. Calcd for C₉₂H₉₁Cl₂FeN₃ (1365.51): C, 80.92; H, 6.72; N, 3.08. Found: C, 80.71; H, 6.90; N, 3.12%.

Ar = 2-(C₁₂H₂₃)-4,6-(CHPh₂)₂C₆H₂ Fe4. By employing the same procedure as that described for Fe1 but with 2-cyclododecyl-4,6-dibenzhydrylaniline as the aniline, Fe4 was isolated as a green powder (0.171 g, 40%). FT-IR (cm⁻¹): 3025 (w), 2924 (m), 2856 (w), 1601 (m, ν(C=N)), 1571 (s), 1494 (s), 1030 (m), 744 (m), 699 (s), 614 (m). Anal. Calcd for C₁₀₀H₁₀₇Cl₂FeN₃ (1477.72): C, 81.28; H, 7.30; N, 2.84. Found: C, 81.38; H, 7.22; N, 3.04%.

Ar = 2,6-(C₅H₉)₂-4-(CHPh₂)C₆H₂ Fe5. By employing the same procedure as that described for Fe1 but with 2,6-dicyclopentyl-4-benzhydrylaniline as the aniline, Fe5 was isolated as a green powder (0.173 g, 55%). FT-IR (cm⁻¹): 3025 (w), 2948 (m), 2864 (m), 1598 (m, ν(C=N)), 1569 (s), 1491 (s), 1029 (m), 742 (m), 700 (s), 619 (m). Anal. Calcd for C₇₀H₇₅Cl₂FeN₃ (1085.14): C, 77.48; H, 6.97; N, 3.87. Found: C, 77.76; H, 6.80; N, 3.52%.

2.3. General Procedure for Ethylene Polymerization

A stainless steel autoclave (250 mL) containing an ethylene pressure control system, a mechanical stirrer, and a temperature controller was used to conduct the polymerization runs. The autoclave was placed under reduced pressure and backfilled with ethylene (×3). On reaching the required temperature, the iron complex (2 μmol), pre-dissolved in toluene (25 mL), was injected into the autoclave under an ethylene atmosphere (ca. 1 atm) followed by the addition of more toluene (25 mL). Subsequently, the specified quantity of co-catalyst (MMAO or MAO) was added by syringe, and finally, more was toluene introduced to bring the total volume of solvent to 100 mL. The apparatus was immediately pressurized to an ethylene pressure of 5 or 10 atm, and the stirring started. Following the set reaction time (5–60 min), the autoclave was allowed to cool to ambient temperature and the ethylene pressure was vented. The reaction mixture was quenched with 10% HCl in ethanol and the resulting polyethylene was filtered, washed with ethanol, and then finally dried at 60 °C until of constant weight.

2.4. X-ray Diffraction Studies

X-ray quality crystals of Fe3 and Fe5 were obtained by diffusing hexane onto a dichloromethane solution containing the corresponding complex. A single crystal of each was selected and mounted on a Bruker SMART CCD diffractometer incorporating a graphite-monochromated Mo–Kα radiation (λ = 0.71073 Å) source and a nitrogen cold stream (−100 °C). The data were corrected for Lorentz and polarization effects (SAINT) and semiempirical absorption corrections based on equivalent reflections were applied (SADABS). The structures were solved by direct methods and refined by full-matrix least-squares on F². Hydrogen atoms were placed in calculated positions. Structure solution and structure refinement were carried out by using the SHELXT (Sheldrick, 2018) [29]. The structural disorder exhibited by the cyclopentyl and dichloromethane solvent molecules in Fe5 was also processed by the SHELXL (Sheldrick, 2018) [30]. The X-ray structure determination and refinement details are collected in

Table 1. Crystallographic data and structure refinement for Fe3 and Fe5.

-	Fe3	Fe5
Crystal color	blue	gray
Empirical formula	C92H91Cl2FeN3	2 C70H75Cl2FeN3·3CH2Cl2
Formula weight	1363.41	2424.93
T (K)	172(2)	173(2)
Wavelength (Å)	0.71073	0.71073
Crystal system	Tetragonal	Orthorhombic
Space group	I41/a	P212121
a/Å	32.4190(2)	18.8805(6)
b/Å	32.4190(2)	19.1069(5)
c/Å	31.8280(3)	35.1775(10)
α/°	90	90
β/°	90	90
γ/°	90	90
Volume/Å3	33,451.0(5)	12,690.2(6)
Z	16	4
ρcalcg/cm3	1.083	1.269
μ/mm−1	2.359	0.492
F(000)	11,552.0	5112
Crystal size/mm3	0.15 × 0.1 × 0.05	0.184 × 0.181 × 0.062
Θ range (°)	6.7 to 151.132	3.032 to 50
Limiting indices	−34 ≤ h ≤ 36	−22 ≤ h ≤ 22
	−35 ≤ k ≤ 40	−22 ≤ k ≤ 22
	−39 ≤ l ≤ 37	−41 ≤ l ≤ 39
No. of rflns collected	65,052	76,355
No. unique rflns [R(int)]	16,088(0.0368)	21,892(0.1219)
Completeness to θ (%)	96.2	99.8
Goodness of fit on F2	1.024	1.024
Final R indices [I > 2σ(I)]	R1 = 0.0531	R1 = 0.0953
	wR2 = 0.1370	wR2 = 0.2468
R indices (all data)	R1 = 0.0623	R1 = 0.1449
	wR2 = 0.1425	wR2 = 0.2910
Largest diff peak and hole (e Å−3)	0.99/−0.35	0.92/−0.56

.

3. Results and Discussion

3.1. Synthesis and Characterization of Fe1–Fe5

A one-pot strategy was utilized to synthesize the bis(arylimino)tetrahydrocyclohepta[b]pyridine dichloroiron(II) complexes, [2-{(Ar)N=CMe}-9-{N(Ar)}C₁₀H₁₀N]FeCl₂ (Ar = 2-(C₅H₉)-4,6-(CHPh₂)₂C₆H₂ Fe1, 2-(C₆H₁₁)-4,6-(CHPh₂)₂C₆H₂ Fe2, 2-(C₈H₁₅)-4,6-(CHPh₂)₂C₆H₂ Fe3, 2-(C₁₂H₂₃)-4,6-(CHPh₂)₂C₆H₂ Fe4, and 2,6-(C₅H₉)₂-4-(CHPh₂)C₆H₂ Fe5). Typically, by combining 2-acetyl-5,6,7,8-tetrahydrocyclohepta[b]pyridin-9-one, the respective aniline, and iron(II) chloride tetrahydrate in acetic acid under reflux for several hours, the target ferrous complexes could be isolated as green solids on work-up in reasonable yield (Scheme 2). All complexes have been characterized by FT-IR spectroscopy, elemental analysis, and for Fe3 and Fe5, by X-ray diffraction.

Single crystals of Fe3 and Fe5 of suitable quality for the X-ray determinations were obtained by diffusing hexane onto a dichloromethane solution containing the complex. X-ray crystallographic data for Fe3 and Fe5 can been found as Supplementary Materials from the Cambridge Crystallographic Data Centre (CCDC): 2021622 (Fe3), 2020110 (Fe5). Following the structure refinement of Fe5, two independent molecules (A and B) were evident within the asymmetric unit with only minimal differences apparent between them. Views of Fe3 and Fe5 (molecule A) are depicted in Figure 1 and Figure 2; selected bond lengths and angles are tabulated in

Table 2. Selected bond lengths and angles for Fe3 and Fe5 (molecule A).

-	Fe3	Fe5
-	Bond lengths (Å)	-
Fe(1)-N(1)	2.2256(18)	2.287(8)
Fe(1)-N(2)	2.0826(17)	2.132(8)
Fe(1)-N(3)	2.1972(17)	2.292(8)
Fe(1)-Cl(1)	2.3250(6)	2.316(4)
Fe(1)-Cl(2)	2.2456(6)	2.282(4)
-	Bond Angles (deg)	-
N(1)-Fe(1)-N(2)	73.48(6)	73.5(3)
N(1)-Fe(1)-N(3)	141.04(6)	144.8(3)
N(2)-Fe(1)-N(3)	73.23(6)	72.1(3)
N(1)-Fe(1)-Cl(2)	97.70(5)	102.0(2)
N(2)-Fe(1)-Cl(2)	152.55(6)	135.5(3)
N(3)-Fe(1)-Cl(2)	102.58(5)	97.2(2)
N(1)-Fe(1)-Cl(1)	104.23(5)	97.0(2)
N(2)-Fe(1)-Cl(1)	92.60(5)	113.0(2)
N(3)-Fe(1)-Cl(1)	96.92(5)	103.1(3)
Cl(2)-Fe(1)-Cl(1)	114.84(2)	111.51(13)

. The structures of Fe3 and Fe5 are similar differing only in the N-aryl group substitution pattern (viz. 2-cyclooctyl-4,6-dibenzhydrylphenyl Fe3, 2,6-dicyclopentyl-4-benzhydrylphenyl Fe5) and will be described together. Each structure contains an iron center coordinated to two chloride atoms and three nitrogen donors from the chelating N,N,N-ligand to form a geometry best identified as a distorted square pyramidal [17,19,20,21,22,23,24]. The square base of the pyramid is filled by the N1, N2, N3, and Cl2 while Cl1 occupies the apical position. The iron atom is located above the basal plane by 0.516 Å for Fe3 and 0.623 Å for Fe5, which is reminiscent of that seen in a number of structurally related analogs [22,23]. Of the three iron–nitrogen distances, the central Fe-N_pyridine bond length [2.0826(17) Å for Fe3 and 2.132(8) Å for Fe5] is the shortest, likely reflecting the constraints of the pincer ligand and the stronger donor ability of the pyridine nitrogen. The Fe-N_imine distances though longer are comparable despite their inequivalent environments [2.2256(18), 2.1972(17) Å (Fe3), 2.287(8), 2.292(8) Å (Fe5)]. The planes of the N-aryl groups are positioned almost perpendicularly to their adjacent imine vectors as is evidenced by the dihedral angles (71.1°, 82.5° Fe3; 82.2°, 86.5° Fe5), the first angle in each pair being slightly less on account of steric hindrance imposed by the neighboring pyridine-fused carbocyclic ring. As expected, the saturated section of the fused carbocycle (C7–C8–C9–C10) shows some deviation from planarity, owing to the sp³-hybridization of these four carbon atoms. The ortho-substituted cycloalkyl groups display boat-chair/tub (Fe3_cyclooctyl) and envelope (Fe5_cyclopentyl) configurations.

The microanalytical data for all five complexes were consistent with compositions based on the general formula LFeCl₂. In addition, their IR spectra revealed ν(C=N) imine-stretching frequencies at around 1600 cm⁻¹, wavenumbers that are typical for that seen in similar N,N,N-iron complexes [3,4,5,6,7,8,19,20,21,22,23,24,31,32,33,34].

3.2. Ethylene Polymerization

In previous studies, precatalysts A–E (Scheme 1) exhibited their optimum performance for ethylene polymerization when activated with either modified methylaluminoxane (MMAO) or methylaluminoxane (MAO) [31,32,33,35]. Hence, this work is concerned with using both of these co-catalysts with Fe2 chosen as the test precatalyst to permit an optimization of the reaction parameters. Changes to the run temperature, Al:Fe molar ratio, run time, and pressure will be independently undertaken for both Fe2/MMAO and Fe2/MAO before extending the corresponding set of optimized conditions to the remaining precatalysts, Fe1, Fe3, Fe4, and Fe5 [19,20,21,22,23,24,28,31,32,33,34]. Typically, the polymerization runs will be performed in toluene at an ethylene pressure of 10 bar over a 30 min run time; the full set of data are collected in

Table 3. Catalytic evaluation of Fe1–Fe5 using modified methylaluminoxane (MMAO) as co-catalyst ^a.

Entry	Precat.	Al:Fe	T/°C	t/min	Activity b	Mwc	Mw/Mnc	Tmd/°C
1	Fe2	2000	50	30	4.2	7.0	2.7	127.4
2	Fe2	2000	60	30	7.9	9.4	3.0	128.6
3	Fe2	2000	70	30	11.2	12.2	3.1	130.3
4	Fe2	2000	80	30	8.8	7.2	2.3	127.5
5	Fe2	2000	90	30	7.5	9.3	2.3	126.4
6	Fe2	1000	70	30	7.4	31.1	4.5	132.6
7	Fe2	1500	70	30	9.4	21.8	3.7	131.9
8	Fe2	2500	70	30	9.0	12.5	3.0	130.1
9	Fe2	3000	70	30	7.8	9.4	2.8	128.4
10	Fe2	2000	70	5	23.8	10.9	2.6	129.3
11	Fe2	2000	70	15	15.3	14.5	3.3	129.9
12	Fe2	2000	70	45	9.9	41.9	4.7	132.2
13	Fe2	2000	70	60	9.0	64.8	4.0	132.5
14 e	Fe2	2000	70	30	6.4	9.2	2.8	127.9
15	Fe1	2000	70	30	15.3	27.2	4.2	130.4
16	Fe3	2000	70	30	6.4	14.1	2.9	129.7
17	Fe4	2000	70	30	0.5	4.6	2.8	125.7
18	Fe5	2000	70	30	14.6	37.5	4.0	128.7

^a Conditions: iron precatalyst (2.0 μmol), ethylene pressure (10 atm), toluene (100 mL). ^b Activity: 10⁶ g PE per mol Fe per h. ^c M_w in kg per mol. M_w and M_w/M_n measured by gel permeation chromatography (GPC). ^d Measured using differential scanning calorimeter (DSC). ^e ethylene pressure (5 atm).

and

Table 4. Catalytic evaluation of Fe1–Fe5 using MAO as co-catalyst ^a.

Entry	Precat	Al:Fe	T/°C	t/min	Activity b	Mwc	Mw/Mnc	Tmd/°C
1	Fe2	2000	30	30	2.7	45.8	2.7	135.0
2	Fe2	2000	40	30	2.9	90.6	3.0	132.6
3	Fe2	2000	50	30	3.1	32.0	3.1	134.6
4	Fe2	2000	60	30	3.3	103.5	2.3	131.9
5	Fe2	2000	70	30	1.2	25.9	2.3	133.8
6	Fe2	2000	80	30	1.0	24.7	4.5	131.1
7	Fe2	1000	60	30	1.0	186.8	3.7	135.1
8	Fe2	1500	60	30	2.9	41.6	3.0	132.2
9	Fe2	2500	60	30	4.1	35.5	2.8	131.7
10	Fe2	3000	60	30	2.7	13.6	2.6	130.3
11	Fe2	2500	60	5	7.2	88.9	3.3	133.1
12	Fe2	2500	60	15	5.6	52.8	4.7	132.3
13	Fe2	2500	60	45	2.9	44.5	4.0	132.5
14	Fe2	2500	60	60	2.4	71.2	2.8	134.2
15 e	Fe2	2500	60	30	1.2	28.8	3.2	133.0
16	Fe1	2500	60	30	5.3	55.6	2.9	131.8
17	Fe3	2500	60	30	1.6	11.2	2.8	131.3
18	Fe4	2500	60	30	0.3	5.1	4.0	129.6
19	Fe5	2500	60	30	6.0	25.6	2.7	131.0

^a Conditions: iron precatalyst (2.0 μmol), ethylene pressure (10 atm), toluene (100 mL). ^b Activity: 10⁶ g PE per mol Fe per h. ^c M_w in kg per mol. M_w and M_w/M_n measured by GPC. ^d Measured by DSC. ^e ethylene pressure (5 atm).

. In all cases, differential scanning calorimetry (DSC) and gel permeation chromatography (GPC) are used to measure various polymer properties (M_w, M_w/M_n and T_m), while high-temperature NMR spectroscopy is employed for selected samples. As a matter of course, gas chromatography (GC) will be employed to check for any oligomeric products present in the polymerization solutions.

3.2.1. Catalytic Evaluation Using Fe1–Fe5/MMAO

In order to determine the optimal run temperature, Fe2/MMAO was screened at temperatures between 50 and 90 °C with the Al:Fe molar ratio fixed at 2000:1 (entries 1–5, Table 3). The highest activity of 11.2 × 10⁶ g PE mol⁻¹ (Fe) h⁻¹ was observed at 70 °C. In terms of the molecular weight, all the polyethylenes displayed values between 7.0 and 12.2 kg mol⁻¹ with temperature proving an influential factor. For example, the molecular weight of the polyethylene initially increased as the temperature of the run was raised from 50 to 70 °C and then decreased with a further rise in the reaction temperature; similar behavior has been reported for the ortho-cycloalkyl-containing E [22] (Scheme 1). It would seem the ortho-cyclohexyl groups in Fe2 can protect the active iron species at temperatures up to 70 °C before the rate of chain transfer over chain propagation becomes more important at higher temperature. Across the temperature range explored, all the polyethylenes exhibited narrow unimodal distributions (M_w/M_n range = 2.3–3.1) indicative of single-site active species.

Then, the influence of the Al:Fe molar ratio was investigated using Fe2/MMAO (entries 3, 6–9, Table 3). Specifically, this ratio was increased from 1000:1 to 3000:1, resulting in a peak in activity being noted with 2000 molar equivalents of co-catalyst. Interestingly, all these runs maintained good activities (>7.4 × 10⁶ g PE mol⁻¹ (Fe) h⁻¹) even with lesser amounts of MMAO. With respect to the molecular weight of the polymer, this reached a maximum of 31.1 kg mol⁻¹ when the Al:Fe molar ratio was 1000:1. As the ratio was raised, the molecular weight of the polyethylene decreased (Figure 3), which indicated that the rate of chain transfer increased; although uncertain at this stage, a process involving transfer of the polymer chain from the active iron catalyst to an alkyl-aluminum species seems likely [20,31,33,34,36,37].

With regard to the activity/time profile, the polymerization runs using Fe2/MMAO were quenched after pre-determined run times, typically 5, 15, 30, 45, and 60 min (entries 3 and 10–13, Table 3); the run temperature was kept at 70 °C and the molar ratio of Al to Fe was kept at 2000:1. An uppermost activity of 23.8 × 10⁶ g PE mol⁻¹ (Fe) h⁻¹ was seen after 5 min (entries 10, Table 3) before a gradual decrease in catalytic activity was seen on extending the reaction time. Nonetheless, even after 60 min, the catalyst still maintained a credible activity (9.0 × 10⁶ g PE mol⁻¹ (Fe) h⁻¹), highlighting the sizable lifetime of this catalyst. On lowering the ethylene pressure from 10 to 5 atm (entries 3 and 14, Table 3), the catalytic activity showed a dramatic drop, while the polyethylene generated at the two different pressures possessed similar molecular weights and molecular weight distributions, which is a finding that is consistent with related iron analogs [22].

Based on the optimal polymerization conditions established for Fe2/MMAO, the remaining precatalysts Fe1 and Fe3–Fe5 were then investigated using MMAO as co-catalyst at 70 °C. A wide range in activities were observed between 0.5 × 10⁶ g PE mol⁻¹ (Fe) h⁻¹ and 15.3 × 10⁶ g PE mol⁻¹ (Fe) h⁻¹ (entries 3 and 15–18, Table 3). In terms of their relative performance, their activities were found to fall in the order: Fe1_C5H9/CHPh2~Fe5_C5H9/C5H9 > Fe2_C6H11/CHPh2 > Fe3_C8H15/CHPh2 > Fe4_C12H23/CHPh2 (Figure 4). Two key points emerge from the inspection of this order. Firstly, the cyclopentyl-containing precatalysts, Fe1 and Fe5, displayed the highest activities with the benzhydryl-containing Fe1 marginally higher. Secondly, the ortho-cycloalkyl ring size is of crucial importance to catalytic activity with the value observed using Fe1 far exceeding that seen using Fe2, Fe3, and Fe4. Furthermore, Fe1 gave the highest molecular weight polyethylene (27.2 kg mol⁻¹) of this series, which suggests that the cyclopentyl systems were more conducive to chain propagation and to the suppression of chain transfer. As for the dispersities of the polyethylene, Fe1–Fe5/MMAO all generated materials with M_w/M_n values that fell in the range 2.8–4.2, which is unlike the broad ranges often observed with structurally related iron analogs [19]. It would seem this class of catalyst has a predilection toward single site-like behavior.

3.2.2. Catalytic Evaluation Using Fe1–Fe5/MAO

To allow a comparison with the MMAO study, we also studied the impact of using methylaluminoxane (MAO) as the activator for all five iron precatalysts. Complex Fe2 was once again employed as the test precatalyst to optimize the conditions; the results of the evaluation are tabulated in Table 4. By maintaining the Al:Fe molar ratio at 2000:1, the highest activity of 3.3 × 10⁶ g PE mol⁻¹ (Fe) h⁻¹ was achieved at 60 °C when investigated over the 30 to 80 °C temperature range (cf. 70 °C with Fe2/MMAO) (entry 4, Table 4). Although this level of performance was lower than that displayed using Fe2/MMAO (11.2 × 10⁶ g PE mol⁻¹ (Fe) h⁻¹), the molecular weight of the polyethylene reached 103.5 kg mol⁻¹, which is more than eight times that seen with MMAO. It is unclear as to the reason behind this molecular weight enhancement but it could plausibly be due to the greater stability of the active catalyst toward propagation over chain transfer at this lower temperature. Indeed, when the run temperature was raised to 70 °C, not only did the activity drop, but the molecular weight of the polymer significantly declined as well (entry 5, Table 4 and Figure 5).

With the reaction temperature maintained at 60 °C, a series of polymerization runs were conducted using Fe2/MAO with the Al:Fe molar ratio systematically raised from 1000:1 to 3000:1 (entries 4 and 7–10, Table 4). A peak activity of 4.1 × 10⁶ g PE mol⁻¹ (Fe) h⁻¹ was observed with an Al:Fe molar ratio of 2500:1 (entry 9, Table 4). On increasing the molar ratio above 2500:1, a decline in activity was seen which is consistent with more chain transfer (Figure 6) [19,22,33,38].

As with Fe2/MMAO, the activity dropped as the run time was extended though not as noticeably (entries 9 and 11–14, Table 4). Hence, at the five minute mark, a value of 7.2 × 10⁶ g of PE mol⁻¹ (Fe) h⁻¹ was achieved that only decreased to 2.4 × 10⁶ g of PE mol⁻¹ (Fe) h⁻¹ after 60 min (entry 11, Table 4, and entry 14, Table 4). It is apparent that this catalyst though less active than Fe2/MMAO maintained a more uniform activity/time profile. As with Fe2/MMAO, a short induction period is needed to reach peak performance with Fe2/MAO; similar observations have been noted elsewhere for iron and cobalt catalysts [39].

With the optimal polymerization conditions identified for Fe2/MAO (viz., an Al:Fe molar ratio = 2500:1, run temperature = 60 °C, run time = 30 min), the four other iron precatalysts, Fe1 and Fe3–Fe5, were also evaluated using these set of conditions; their results are displayed alongside Fe2 in Table 4 (entries 9 and 16–19). Examination of the data reveals that these iron precatalysts exhibited a range in activities from 0.3 to 6.0 × 10⁶ g of PE mol⁻¹ (Fe) h⁻¹ which is narrower than that seen with MMAO. In terms of their relative performance, their catalytic activities fell in the order: Fe5_C5H9/C5H9~Fe1_C5H9/CHPh2 > Fe2_C6H11/CHPh2 > Fe3_C8H15/CHPh2 > Fe4_C12H23/CHPh2. This order essentially mirrors that seen with MMAO, although closer inspection reveals Fe1 to be slightly more active than Fe5 (the opposite was seen with MMAO). Once again, the catalytic activity is shown to drop as the size of the ortho-cycloalkyl ring was increased from 5 to 12 (Figure 7). One possible explanation for this observation may relate to steric protection imparted by the larger cycloalkyl rings on the active iron center, thereby impeding ethylene coordination and insertion. In addition, the molecular weight of the polyethylene was found to drop from 55.6 to 5.1 kg mol⁻¹ as the ortho-cycloalkyl increased, reaching the minimum value for the most sterically hindered Fe4 (Figure 7). Nonetheless, the dispersities remained narrow for all five catalysts (M_w/M_n range: 2.7–4.0).

3.3. Microstructural Properties of the Polyethylene

As highlighted in Table 3 and Table 4, the melting points displayed by the polyethylenes fell between 126.4 and 135.1 °C, which is characteristic of linear polyethylene [32,35,37]. To confirm this assertion and gather more information about the end group composition and likely chain transfer pathways, samples of polyethylene obtained using Fe2/MAO and Fe2/MMAO have been analyzed by ¹H and ¹³C NMR spectroscopy (recorded in tetrachloroethane-d₂ at 100 °C).

As shown in Figure 8, the ¹H NMR spectrum of the polymer generated using Fe2/MMAO (M_w = 12.2 kg mol⁻¹; entry 3, Table 3) revealed resonances typical of a –C(H_b)=C(H_a)₂ end group, which was backed up by the ¹³C NMR spectrum with the corresponding vinylic carbon signals visible at δ 116.5 and 139.0 [19,40,41,42]. Such an observation would imply that β-hydrogen elimination is a major chain transfer pathway. However, on inspection of the ratio of the integrals for the vinylic H_a to methyl H_g protons (δ 0.97) in the ¹H NMR spectrum, a slight excess of the expected 2:3 ratio was evident, signifying the co-existence of some fully saturated polyethylene. This finding would suggest that a termination mechanism involving chain transfer to AlMe₃ (but not Al(i-Bu₃)), and its aluminum derivatives found in MMAO solutions, also plays a minor role [20,22]. In addition, carbon signals corresponding to saturated n-butyl end chain ends were clearly visible in the lower frequency region of the ¹³C NMR spectrum (d, e, f and g, Figure 8).

In addition, a sample of the higher molecular weight (M_w = 35.3 kg mol⁻¹) polyethylene prepared using Fe2/MAO at 60 °C (entry 9, Table 4) was characterized by ¹³C NMR spectroscopy. High-intensity singles observed around δ 29.44 (see Figure 9) were the only signals detectable which can be assigned to the –(CH₂)_n– repeat unit in accord with high linearity of the material. No clear evidence was found to support the existence of unsaturated nor saturated chain ends presumably due to the high molecular weight of this sample.

4. Conclusions

The bis(arylimino)tetrahydrocyclohepta[b]pyridine dichloroiron(II) complexes, Fe1–Fe4, comprising N-aryl groups appended with benzhydryl/cycloalkyl ortho-substituents have been successfully synthesized; in addition, Fe5 based on a cyclopentyl/cyclopentyl ortho-pairing is also disclosed. All five complexes have been fully characterized including by single crystal X-ray diffraction in the cases of Fe3 and Fe5. On their treatment of Fe1–Fe5 with either MAO or MMAO, a range in catalytic activities was exhibited with cyclopentyl-containing Fe1 and Fe5 at the top end for both activators. Indeed, Fe1/MMAO achieved exceptionally high activity at a temperature of 70 °C (1.53 × 10⁷ g PE per mol Fe per h) underlining the appreciable thermostability of this iron catalytic system. Furthermore, the activities were found to drop as the ortho-cycloalkyl ring size increased with the most sterically encumbered cyclododecyl-containing Fe4 displaying only modest activity. In a similar fashion, the molecular weights for the polymers were also found to decline as the ring size increased with cyclopentyl-containing Fe1/MAO producing the highest molecular weight polyethylene of the MAO-activated series (55.6 kg mol⁻¹). Strictly linear polyethylene (T_m > 126 °C) with narrow distribution (M_w/M_n range: 2.3–4.7) were features of all the polymers generated in this study, the latter highlighting the good control of these polymerizations. Despite the high molecular weights, end group analysis revealed evidence for vinyl-terminated polymers supporting the role of β-hydride elimination as a key chain transfer pathway.