Mechanistic Studies of Ethylene and α-Olefin Co-oligomerization Catalyzed by Chromium-PNP Complexes

Abstract

To explore the possibility of producing a narrow distribution of mid- to long-chain hydrocarbons from ethylene as a chemical feedstock, co-oligomerization of ethylene and linear α-olefins (LAOs) was investigated, using a previously reported chromium complex, [CrCl₃(PNP^OMe)] (1, where PNP^OMe = N,N-bis(bis(o-methoxyphenyl)phosphino)methylamine). Activation of 1 by treatment with modified methylaluminoxane (MMAO) in the presence of ethylene and 1-hexene afforded mostly C₆ and C₁₀ alkene products. The identities of the C₁₀ isomers, assigned by detailed gas chromatographic and mass spectrometric analyses, strongly support a mechanism that involves five- and seven-membered metallacyclic intermediates comprising of ethylene and LAO units. Using 1-heptene as a mechanistic probe, it was established that 1-hexene formation from ethylene is competitive with formation of ethylene/LAO co-trimers, and that co-trimers derived from one ethylene and two LAO molecules are also generated. Complex 1/MMAO is also capable of converting 1-hexene to C₁₂ dimers and C₁₈ trimers, albeit with poor efficiency. The mechanistic implications of these studies are discussed and compared to previous reports of olefin co-trimerization.

INTRODUCTION

Ethylene, the most widely produced organic compound in the world,^1,2 is attractive as a cheap and abundant feedstock for the manufacturing of value-added products including linear α-olefins (LAOs), a class of commodity chemicals that are used in industries such as synthetic polymers, detergents, plasticizers, and lubricants, and have a global demand of over four million tons per year.³ LAOs are produced primarily by catalytic oligomerization of ethylene via linear chain growth (Cossee-Arlman mechanism).⁴ Such reactions produce a non-selective Schulz-Flory distribution of oligomers, requiring additional procedures to separate the desired products. More recently, selective oligomerization of ethylene to 1-hexene and/or 1-octene has been reported,^5–8 first using chromium/2-ethylhexanoate^9,10 and subsequently with complexes containing a diverse array of transition metal ions and ligand architectures.⁵ These reactions most likely proceed through a metallacycle ring expansion mechanism, in which oxidative coupling of two ethylene molecules at an electrophilic metal center affords a metallacyclopentane intermediate, which undergoes ethylene insertion to give a metallacycloheptane that eliminates 1-hexene (Scheme 1).^11–16 The elimination step may involve either a β-hydride elimination/reduction elimination sequence or a concerted 3,7-hydride shift; it is difficult to distinguish between these alternatives,¹³ but the latter is favored by recent theoretical studies.^15–17 Whichever mechanism operates, the elimination is expected on geometric grounds to be much more favorable, relative to further insertion, for the seven-membered rather than for the five-membered metallacycle, accounting for the observed selectivity. With some catalysts, and especially at higher ethylene pressures, a second insertion becomes favored over elimination, yielding 1-octene. This chemistry has also been extended to homo-trimerization of LAOs^18,19 and other olefinic substrates.^20–22

Scheme 1.

Proposed catalytic cycle for the selective trimerization of ethylene to 1-hexene

Significant effort has been devoted to understanding the chemistry of selective ethylene oligomerization,^5–7 including issues such as catalyst initiation, effect of activators, metal oxidation states, and reaction mechanism. One interesting feature of ethylene trimerization is the formation of some C₁₀ alkenes in addition to 1-hexene, suggested to be secondary products from the coupling of two ethylenes with 1-hexene. Co-oligomerization of ethylene and an α-olefin, such as propylene,¹³ 1-octene,²³ and styrene,²⁰ by homogeneous chromium catalysts has been demonstrated previously. “Product recycling,” wherein co-oligomerization of ethylene with the primary α-olefin produced competes effectively with homo-oligomerization, could constitute a selective route to hydrocarbons larger than C₆ or C₈, using only ethylene as starting material.24 Here we explore the feasibility of such a process by examining the efficiency of 1-hexene incorporation in co-trimerization, using the previously reported chromium complex, [CrCl₃(PNP^OMe)] (1, where PNP^OMe = N,N-bis(bis(o-methoxyphenyl)phosphino)methylamine, Chart 1) as the catalyst precursor.^25,26

Chart 1.

Structure of the chromium pre-catalyst [CrCl₃(PNP^OMe)] (1)

RESULTS

Co-oligomerization of ethylene and 1-hexene

To establish a benchmark for the present studies, compound 1 was treated with modified methylaluminoxane (MMAO) in chlorobenzene under an atmosphere of ethylene. Analysis of the reaction mixture after 1 h by gas chromatography (GC) (Figure S1A) revealed that C₆ (92 wt %) and C₁₀ (8 wt %) alkenes were the primary products (Table 1, entry 1), with a productivity of 571 ± 14 g/g Cr·h; 1-hexene was the only detectable C₆ isomer and only trace amounts (<10 mg) of polyethylene were obtained.

Table 1.

Product distribution of olefin homo- and co-trimerization^a

Entry	Monomer(s)	Productivitye (g/g Cr/h)	Products (wt %)
			C6f	C10	C11	C12	C13	C14	C15	C16	C18
1	ethyleneb	571 ± 14	92	8	–	–	–	–	–	–	–
2	ethyleneb/1-hexenec	926 ± 93a	nd	94	–	1	–	4	–	–	0.2
3	ethyleneb/1-heptene,d	863 ± 97	51	2	43	–	0.5	0.6	0.8	1.2	–
4	1-hexenec	7.6 ± 0.3	–	–	–	27	–	–	–	–	73

^aReaction condition: complex 1 (12 μmol), monomer(s), and MMAO (250 mg, ~55 equiv, solution in isohexanes) diluted to 10 mL with chlorobenzene, room temperature.

^bEthylene (1 atm).

^c1-Hexene (35 mmol).

^d1-Heptene (24 mmol).

^eThe standard deviations were determined from reactions performed in triplicate.

^fThe C₆ fraction contains >99% 1-hexene.

To favor the production of C₁₀ olefins and to obtain sufficient material for characterization, the same reaction was performed with an excess of added 1-hexene; the ratio of 1-hexene:ethylene in solution, under the reaction conditions employed, based on NMR spectroscopic measurement, was approximately 1:1. Quantification of the olefins heavier than C₆ by GC showed that C₁₀ olefins were the major products (94 wt %), followed by much smaller amounts of C₁₂ (1 wt %), C₁₄ (4 wt %), and C₁₈ (0.2 wt %) olefins (Figure S1B). Because 1-hexene was used as a starting material, the amount that was formed during the reaction could not be determined; the productivity based on the other products was 926 ± 93 g/g Cr·h (Table 1, entry 2).

The C₁₀ olefins were isolated by fractional distillation of the crude reaction mixture. GC analysis showed three major species with retention times of 36.98, 37.19, and 37.38 min, and several minor ones between 37.75–38.07 min (Figure 1, black trace). By comparison to authentic samples of the nine isomers expected from co-trimerization through a metallacycle mechanism (Scheme 2), it is apparent that 3-propyl-1-heptene (P, red trace) corresponds to the peak at 36.98 min, 4-ethyl-1-octene (O, orange trace) corresponds to the peak at 37.19 min, 5-methyl-1-nonene (N, yellow trace) corresponds to the peak at 37.38 min, 2-butyl-1-hexene (M, green trace) corresponds to the peak at 37.75 min, and the linear C₁₀ olefins (H-L; light blue, brown, pink, purple, and blue traces) correspond to the peaks between 37.78–38.07 min. The relative amounts of C₁₀ olefins were determined from the integrated areas of the peaks in the gas chromatogram: linear C₁₀ olefins H–L, 3%; M, 5%; N, 47%; O, 18%; and P, 27% (Scheme 2).

Figure 1.

Gas chromatogram of the C₁₀ alkenes (black, bottom trace) obtained by fractional distillation from the reaction of 1/MMAO/ethylene/1-hexene; the products appear at 36.98, 37.19, 37.38, 37.75, 37.86, 38.07 min. The C₁₀ olefin standards are shown directly above (from top to bottom, RT = retention time in min): cis-5-decene (H, light blue; 38.08), cis-4-decene (I, brown; 38.07), 1-decene (J, pink; 37.84), trans-4-decene (K, purple; 37.86), trans-5-decene (L, blue; 37.90), 2-butyl-1-hexene (M, green; 37.76), 5-methyl-1-nonene (N, yellow; 37.44), 4-ethyl-1-octene (O, orange; 37.22), 3-propyl-1-heptene (P, red; 37.02). The peak marked with an asterisk (*) is the trans-4-decene impurity present in the 3-propyl-1-heptene standard (see SI).

Scheme 2.

C₆ and C₁₀ alkenes formed via co-trimerization of ethylene and 1-hexene and their hydrogenation products^a

^aThe percentages shown in green correspond to the relative amounts of C₁₀ alkenes present in the reaction product. Similarly, the percentages shown in blue correpond to the relative amounts of C₁₀ alkanes present after the alkenes were subjected to hydrogention. The discrepancies (up to ±5%) between the total amount of each group of alkene isomers observed and the amounts of hydrogenation products obtained are attributed to experimental error.

To obtain further support for the assignments, a sample of the C₁₀ olefin mixture was hydrogenated and analyzed by GC. As shown in the black trace in Figure 2, the gas chromatogram of the saturated C₁₀ products displays two prominent peaks at 36.79 and 36.90 min, and a significantly smaller one at 37.72 min. Comparing the retention times of the peaks in this trace to those of the expected products (Scheme 2) indicate that indeed, they correspond to 2-ethyloctane (S, red trace), 5-methylnonane (R, orange trace), and decane (Q, green trace), respectively, with a relative ratio of 40:56:4 (Figure 3). (Although the expected ratio based on the distribution of C₁₀ alkenes is 45:52:3, this discrepancy is attributable to experimental error, such as non-ideal peak integration, slight differences in the GC response factors for the different isomers, or incomplete hydrogenation.) Consistent with the chemical formula C₁₀H₂₂, mass spectral analysis reveals that all three species in the hydrogenated sample display a peak with an m/z value of 142. In addition, each compound exhibits a unique fragmentation pattern in its mass spectrum that matches well with that of the species to which it was assigned (Figure 3).

Figure 2.

Gas chromatogram of the C₁₀ alkanes (black, bottom trace) obtained from hydrogenation of the C₁₀ alkene mixture (see Figure 1); the products appear at 36.79, 36.90, and 37.72 min. The C₁₀ alkane standards are shown directly above (from top to bottom, RT = retention time in min): decane (Q, green; 37.28), 5-methylnonane (R, orange; 36.86), and 2-ethyloctane (S, red; 36.71).

Figure 3.

Electron-impact (EI) ionization mass spectra (black) of the peaks that appear in the C₁₀ alkane sample (Figure 2) at A) 36.79 min, B) 36.90 min, and C) 37.72 min. For comparison, the mass spectra for authentic samples of 4-ethyloctane (pink), 5-methylnonane (blue), and decane (green) are shown directly above that for A, B, and C, respectively. RT = retention time.

Homo-oligomerization of 1-hexene

To test the feasibility of LAO homo-oligomerization by the Cr-PNP^OMe system, reactions were conducted using 1/MMAO in the presence of 1-hexene without added ethylene. GC-MS analysis of this reaction mixture showed that both C₁₂ and C₁₈ olefins were formed, in a ratio of 27:73 wt % (Figure S2; Table 1, entry 4). Although several isomers of each chain length were obtained, no attempt was made to ascertain their identities. The overall productivity was quite low, 7.6 ± 0.3 g/g Cr·h.

Co-oligomerization of ethylene and 1-heptene

An analogous reaction was carried out with ethylene and 1-heptene as co-monomers, in approximately 4:3 ratio (as determined by NMR spectroscopic measurements). GC analysis of the crude mixture showed that 1-hexene and C₁₁ olefins constituted approximately 51 and 43 wt %, respectively, of the total products (Figure S1C; Table 1, entry 3). Minor amounts of C₁₀ (2 wt %), C₁₃ (0.5 wt %), C₁₄ (0.6 wt %), C₁₅ (0.8 wt %), and C₁₆ (1.2 wt %) olefins were also observed. The overall productivity was 863±97 g/g Cr·h.

DISCUSSION

The chromium (III) complex CrCl₃(PNP^OMe) (1) was selected as the pre-catalyst for the present studies because of its high selectivity for ethylene trimerization.^25–27 That selectivity has been attributed, in part, to the interaction of a pendant methoxy moiety of the PNP^OMe ligand with the chromium center;⁷ when these methoxy groups are absent, as in the [CrCl₃(PNP)]₂ variant (where PNP = N,N-bis(diphenylphosphino)methylamine), both ethylene tri- and tetramerization occur.^28–30 Although the methoxyaryl functionalities are not essential to obtaining a high ratio of 1-hexene versus 1-octene,³¹ only the PNP^OMe ligand forms a catalyst that produces 1-hexene with >99% trimerization selectivity.

The ability of 1 to facilitate selective trimerization is indicated by the high yield of C₁₀ olefins, ~94 wt % of the quantifiable products (i.e., those other than C₆) obtained when ethylene/1-hexene were used as co-monomers (Table 1, entry 2). Because identification of the reaction products would provide insight into their possible mechanism of formation, studies were undertaken to characterize the C₁₀ species. As shown in Scheme 2, there are three possible metallacycloheptanes from co-trimerization of two ethylenes with one 1-hexene, and two possible routes to each of them. Unsubstituted chromacyclopentane A (from oxidative coupling of two ethylene molecules with the Cr center) can undergo insertion of ethylene to give D (i.e. homo-trimerization), or either 2,1- or 1,2-insertion of 1-hexene, leading to chromacycloheptanes E and F respectively. Alternatively chromacyclopentanes B and C, formed by oxidative coupling of ethylene and 1-hexene, could undergo subsequent insertion of another ethylene to yield chromacycloheptanes E and G or F and G, respectively.

Elimination of olefin from the various chromacycloheptanes, by either the stepwise or concerted route (see above), would afford the products shown in Scheme 2. Only linear decenes can result from E, with the position of the double bond depending on the relative preference for hydride elimination (or transfer) from substituted vs. unsubstituted and endo- vs. exocyclic C_β positions. Likewise, F should give rise to either 2-butyl-1-hexene (M) or 5-methyl-1-nonene (N), and G, to 4-ethyl-1-octene (O) or 3-propyl-1-heptene (P). As shown in Figure 1, GC analysis shows that all of the branched, and some or all of the unbranched, expected C₁₀ olefins are indeed formed, supporting the proposed mechanism.

Detailed analysis of the C₁₀ olefin distribution indicates several important characteristics of the [Cr-PNP^OMe] intermediates. First, since only 3–4% of the C₁₀ olefins are linear, their precursor intermediate E must be significantly disfavored. E can be generated either by 2,1-insertion of 1-hexene into chromacyclopentane A or ethylene insertion into the unsubstituted side of chromacyclopentane B; both sequences must therefore be of low probability. In each case there are several possible interpretations. A is certainly formed, since homo-trimerization of ethylene is taking place (see below); hence insertion of olefin into A exhibits a strong preference for ethylene over LAO and/or for 1,2-over 2,1-insertion. In previous studies using a Cr-PNP^OMe/biphenyldiyl complex as a model for the chromacyclopentane intermediate,¹³ it was found that ethylene inserts into the five-membered metallacycle more than twenty times faster than LAOs, and the latter insert exclusively in 1,2-mode, consistent with the current observations. In contrast, in the co-trimerization of ethylene and styrene by 1/MAO the major products (~80%) were those derived from 2,1-insertion of styrene into A, and no 1,2-insertion products were observed.²⁰ A preference for 2,1-insertion of styrene is common, both in the polymer literature^32–34 and elsewhere,^35–38 although 1,2-insertion can be achieved using sterically-hindered catalysts.³⁹

It seems quite unlikely that insertion of ethylene into the substituted position of B, to give G, should be favored over insertion into the unsubstituted one, to give E. A much more reasonable explanation is that formation of B itself is relatively unfavorable, presumably for steric reasons, and that isomer C is strongly favored in the co-oxidative coupling of 1-hexene with ethylene.

The relative overall rates of homo- and co-trimerization may be inferred to be approximately the same, from the observation that the reaction of ethylene with 1-heptene (at approximately 4:3 ratio in solution) gives close to 50% each of 1-hexene and heavier products (Table 1). Since there is more than one available route to each intermediate, it is not possible to assess the relative contributions of all the pathways. However, as the above model studies show, insertion of LAO is significantly less favored than that of ethylene. Hence a reasonable interpretation would be that A and C are formed at about the same rate, with little or no B, and that both A and C react further almost entirely by insertion of ethylene rather than LAO. A smaller amount (~4%) of C₁₄ and even less (~0.2%) C₁₈ is formed in the reaction of ethylene with 1-hexene; these could result from sequential secondary reactions of C₁₀ and C₁₄ with two ethylenes. Another possible route to C₁₄ would be insertion of C₆ into C; that is, a C₂/LAO/LAO co-trimerization; results from the ethylene/1-heptene co-oligomerization experiment suggest that route does contribute (see below). In a related study on co-oligomerization of ethylene with 1-octene using a bis(pyridine)carbene complex of Cr, insertion of LAO into the analog of intermediate C was indeed found to contribute to the products,²³ although that system exhibits several other significant differences from the present one: there the C₁₂ products (the result of co-trimerization of two ethylenes with one 1-octene) did not predominate in the product distribution, and the most abundant C₁₂ species identified was 2-butyl-1-octene, the analog of M, which is not the major C₁₀ product here.

Insertion of ethylene into C can take place either on the side near to or away from the β-substituent to give G and F respectively. Since the sums of the products (M + N) and (O + P) are not too different, it appears that there is relatively little preference in that step. On the other hand, it is notable that N (47%) is obtained in a significantly higher yield than M (5%), suggesting that hydrogen transfer, whether it occurs in a stepwise or concerted manner,^15–17 is less favored from the tertiary substituted β position in F than from the secondary, unsubstituted one. Again, this pattern does not appear to be general for all ethylene trimerization catalysts. For example, ethylene trimerization by a titanium-phenoxyimine complex, believed to involve a similar metallacycle mechanism, afforded M as the major (~90%) C₁₀ olefin product.⁴⁰ The more similar yields of O and P (18% and 27%, respectively) imply that proximity to a substituent in the γ position has a much smaller effect on hydride transfer preference.

Although the formation of C₁₀ olefins from the co-trimerization of one 1-hexene and two ethylenes is unambiguous, routes to the observed C₁₂, C₁₄, and C₁₈ products (Table 1, entry 2) are less obvious. The various possible combinations are shown in Table 2; the C₁₄ and C₁₈ olefins are most simply explained as secondary and tertiary co-trimerization products of two ethylene molecules with the C₁₀ and C₁₄ olefins generated during the reaction. Co-trimerization of one ethylene with two 1-hexenes is an alternate route to C₁₄, as noted above and discussed further below.

Table 2.

Possible combinations of ethylene (C₂), 1-hexene (C₆), and 1-heptene (C₇) dimers and trimers that make up the observed product distributions^a

Ethylene	Ethylene/1-Hexene	Ethylene/1-Hepteneb
C2 + C2 + C2 = C6	C2 + C2 + C2 = C6	C2 + C2 + C2 = C6
C2 + C2 + C6 = C10	C2 + C2 + C6 = C10	C2 + C2 + C6 = C10
	C6 + C6 = C12	C2 + C2 + C7 = C11
	C2 + C2 + C10 = C14	C6 + C7 = C13
	C2 + C6 + C6 = C14	C7 + C7 = C14
	C2 + C2 + C14 = C18	C2 + C2 + C10 = C14
	C6 + C6 + C6 = C18	C2 + C6 + C6 = C14
		C2 + C6 + C7 = C15
		C2 + C7 + C7 = C16

^aEach combination must satisfy the following criteria: 1) homo-oligomerization of ethylene form trimers, 2) co-oligomerization of ethylene/LAO forms trimers, 3) homo-oligomerization of LAO forms dimers and trimers, 4) dimers do not form between LAO and ethylene, and 5) tetramers of alkenes do not form.

^bUse of 1-heptene allowed the determination of trimers formed from two molecules of LAOs and one molecule of ethylene.

Because co-dimerization of ethylene/LAO does not occur, as evidenced by the absence of any C₈ products from ethylene and 1-hexene, C₁₂ cannot be the coupling product of ethylene and C₁₀. A more reasonable route is through homo-dimerization of 1-hexene; 1/MMAO was found to homo-oligomerize 1-hexene, at a productivity about 1% of that of reactions involving ethylene (Table 1, Entry 4; Figure S2). The yield of C₁₂ in the ethylene/1-hexene co-trimerization, about 1% of all products heavier than C₆, is roughly consistent with that relative reactivity; some of the C₁₈ products might also arise by that route. In comparison, chromium complexes supported by triazacyclohexane ligands (Cr-NNN) can oligomerize LAO with high efficiency, up to ~98% conversion;^18,19 the bis(pyridine)carbene Cr complex is also effective for homo-dimerization of LAOs.²³

Some additional inferences may be obtained from the reaction of ethylene/1-heptene with 1/MMAO, which affords C₆, C₁₀, C₁₁, C₁₃, C₁₄, C₁₅, and C₁₆ species (Table 1, entry 3). Clearly C₆ and C₁₁ arise via homo-trimerization of ethylene and co-trimerization of two ethylenes and one 1-heptene respectively; because they are formed in comparable amounts (51 and 43 wt %), the reaction rates for homo-trimerization and co-trimerization by 1 must be similar. (Although no attempts were made to measure actual reaction kinetics in the present work, it has been the subject of other reports in the literature.^41,42) The remaining identified products comprise only a small percentage of the total. The only obvious pathway to C₁₃ is by co-dimerization of C₆ and C₇; a small amount of LAO dimerization was indicated by appearance of C₁₂ in the ethylene/1-hexene reaction, but since much more C₇ than C₆ is present at any time here, it is surprising that the amounts of C₁₄ and C₁₃ are nearly the same —especially given that there are additional possible routes to C₁₄ isomers, co-trimerization of C₂/C₂/C₁₀ (not likely to be very significant, as the amount of C₁₀ is relatively low) or C₂/C₆/C₆. In contrast to the situation for ethylene/1-hexene, where the two sequences C₂/C₂/C₁₀ and C₂/C₆/C₆ both give C₁₄ and hence their relative contributions cannot be determined, the corresponding sequences here would be C₂/C₂/C₁₁ and C₂/C₇/C₇, giving C₁₅ and C₁₆ respectively. Again, comparable amounts of the two are obtained, implying that some co-trimerization of the form C₂/LAO/LAO does take place.

CONCLUSIONS

The following generalizations appear to best account for the observations in co-oligomerization of ethylene and LAO by the chromium-PNP^OMe complex: 1) the initial intermediates are chromacyclopentanes formed by oxidative coupling of either two ethylenes, or one ethylene and one LAO, at the Cr center; 2) the rate constants for formation of those two species are similar; 3) the isomer of the mixed chromacyclopentane with the alkyl chain in the β position is strongly favored; 4) insertion of ethylene into the 5-membered ring is strongly favored over insertion of LAO; 5) insertion occurs preferentially on the side of the ring away from the substituent; 6) hydride transfer occurs preferentially from a secondary over a tertiary carbon site; 7) homo-di- (and -tri-) merization of LAOs does take place, but at a rate much lower than that of any reactions involving ethylene; and 8) co-dimerization of ethylene with LAO is not observed. While points 7) and 8) seem at first to be at odds, they need not be: it may be that the disubstituted chromacyclopentane is sufficiently inhibited towards a further insertion that formation of dimer (presumably by slow β-hydrogen elimination/reductive elimination) competes effectively with further growth, whereas for un- or monosubstituted rings (intermediates A–C in Scheme 2) ethylene insertion dominates over elimination. Only small amounts of the products that would result from insertion of ethylene into a disubstituted ring (C₁₆ from ethylene/1-heptene; probably some of the C₁₄ from ethylene/1-hexene) are observed.

Comparison of 1 to other olefin trimerization catalysts in the literature (see Discussion section) shows clearly that these generalizations do not universally apply to reactions proceeding via the metallacycle-based mechanism. There is a wide range of reactivity behaviors and consequent product distribution profiles, including preferences for ethylene vs. LAO in both oxidative coupling and insertion steps, relative rates of growth by insertion vs. elimination by hydride transfer, and regiochemistry of insertion and hydride transfer. Ongoing work is aimed at gaining the understanding the factors influencing those preferences, at a quantitative as well as qualitative level, that will be needed to make this chemistry a viable alternative to traditional non-selective oligomerization as a route from ethylene to C₁₀₊ hydrocarbons.

EXPERIMENTAL SECTION

General Considerations

Reagents obtained from commercial suppliers were used as received. The chromium pre-catalyst [CrCl₃(PNP^OMe)] (1) was synthesized according to a literature procedure.12 The following gas chromatography standards were obtained from the sources shown in parentheses: 4-ethyl-1-octene (Novel Chemical Solutions), 2-butyl-1-hexene (ChemSampCo), trans-5-decene (Aldrich), cis-5-decene, (TCI), trans-4-decene (ChemSampCo), 1-decene (Aldrich), and decane (Aldrich). The compounds 5-methylnonane and 4-ethyloctane were prepared by hydrogenation of 2-butyl-1-hexene and 4-ethyl-1-octene, respectively, using H₂ over Pd/C. Syntheses of 3-propyl-1-heptene, 5-methyl-1-nonene, and cis-4-decene are provided in the Supporting Information. The modified methylaluminoxane activator (MMAO-C4 solution in isohexanes, 7 wt % Al; referred to as MMAO in the text) was obtained from Albemarle. MMAO-C4 was the activator of choice because it is available as a homogeneous solution that could be transferred quantitatively in repeated trials and gave the most reproducible data. Chlorobenzene was purged with argon and dried over calcium hydride before use. All ethylene oligomerization trials were performed using a high vacuum Schlenk manifold. Polymer-grade ethylene gas (>99.9% purity) was purified by passage through columns containing activated molecular sieves and Ridox, an oxygen-scavenger.

Gas Chromatographic Analyses

Gas chromatography-mass spectrometry (GC-MS) was performed using an Agilent 6890N system with an HP-5MS capillary column (30 m length, 0.25 mm diameter, and 0.5 μm film) that is equipped with an Agilent 5973N mass selective detector. Gas chromatography (GC) was performed using an Agilent 6890N instrument with a flame ionization detector (FID). Routine runs were performed using a DB-1 capillary column (10 m length, 0.10 mm diameter, 0.40 μm film) with the following heating program: hold at 40°C for 3 min, ramp temperature at 50°C/min to 290°C and then hold for 3 min (total run time = 13 min). For detailed C10 isomer analyses, an Agilent GS-GasPro capillary column (30 m, 0.32 mm diameter) was used, with the following heating program: hold at 90°C for 30 min, ramp temperature at 50°C/min to 260°C, and then hold for 10 min (total run time = 43 min).

The amount of products in each oligomerization trial was determined from the integrated areas of the peaks observed in the gas chromatogram, using the integrated area of a biphenyl internal standard as a reference. The integrated area for each peak was corrected using the appropriate response factor, which was determined experimentally. The response factors for all isomers containing the same number of carbon atoms were assumed to be the same.

To determine the identities of the C10 alkene isomers, the C10 fraction was separated from the crude reaction mixture by fractional distillation under atmospheric pressure. The GC retention times of the C10 products were compared to those of authentic samples that were either prepared independently or obtained from commercial sources. These assignments were further confirmed by GC analysis of the C10 sample after it had been hydrogenated.

Procedure for Olefin Oligomerization Trials

Homo-trimerization of ethylene

Inside a nitrogen-filled glovebox, [CrCl₃(PNP^OMe)] (8.0 mg, 12 μmol) was suspended in 10 mL of chlorobenzene, along with a stir bar, in a 250 mL round bottom flask equipped with a 180° needle valve. The flask with sealed, brought outside of the glovebox, and then attached to a high vacuum manifold. The reaction flask was cooled to −78°C in a dry ice/acetone bath, degassed, and then warmed to room temperature. The solution was stirred under an atmosphere of ethylene and a solution of MMAO (250 mg diluted to 1 mL in chlorobenzene) was added using an airtight syringe through a septum that was placed over the needle valve adapter, resulting in the formation of a pale yellow-green homogeneous mixture. Ethylene consumption was monitored using a U-tube mercury monometer during the course of the reaction, and the ethylene pressure was maintained between 700–790 Torr. (It should be noted that gradual catalyst decomposition was observed during the course of the reaction; the underlying cause for such behavior has not yet been determined.) After 1 h, the reaction with quenched with HCl(aq) and stirred for 1 min, resulting in a colorless solution. Solid biphenyl was added (50 mg, 325 μmol) as an internal standard. A 2.0 mL aliquot of the organic layer was filtered through a short silica plug and then analyzed by GC.

Co-oligomerization of ethylene/1-hexene

The same procedure was used as described for homotrimerization of ethylene except that 1-hexene (2.98 g, 35.4 mmol) was added to the reaction flask in addition to the chromium pre-catalyst before diluting to 10 mL with chlorobenzene.

Co-oligomerization of ethylene/1-heptene

The same procedure was used as described for homotrimerization of ethylene except that 1-heptene (2.32 g, 24 mmol) was added to the reaction flask in addition to the chromium pre-catalyst before diluting to 10 mL with chlorobenzene. Less 1-heptene was added to the reaction mixture, compared to the amount of 1-hexene used in the ethylene/1-hexene trials, because its increased lipophilicity was found to cause clumping of the chromium complex at high concentration when activated with MMAO-C4.

Homo-oligomerization of 1-hexene

Inside a nitrogen-filled glovebox, [CrCl₃(PNP^OMe)] (8.0 mg, 12 μmol) and 1-hexene (2.98 g, 35 mmol) were diluted to 10 mL with chlorobenzene. The mixture was treated with MMAO-C4 (250 mg) and then stirred at room temperature for 1 h. Aqueous HCl was added to quench the reaction and then biphenyl (50 mg, 325 μmol) was added to the organic layer as an internal GC standard. A 2.0 mL aliquot of the organic phase was filtered through a short silica plug and then analyzed by GC.