Kinetics, Mechanism and Theoretical Studies of Norbornene-Ethylene Alternating Copolymerization Catalyzed by Organopalladium(II) Complexes Bearing Hemilabile α-Amino-pyridine.

Cationic methylpalladium complexes bearing hemilabile bidentate α-amino-pyridines can serve as effective precursors for catalytic alternating copolymerization of norbornene (N) and ethylene (E), under mild conditions. The norbornyl palladium complexes in the formula of {[RHNCH₂(o-C₆H₄N)]Pd(C₇H10Me)(NCMe)}(BF₄) (R = iPr (2a), tBu (2b), Ph (2c), 2,6-Me₂C₆H₃ (2d), 2,6-iPr₂C₆H₃ (2e)) were synthesized via single insertion of norbornene into the corresponding methylpalladium complexes 1a-1e, respectively. Both square planar methyl and norbornyl palladium complexes exhibit facile equilibria of geometrical isomerization, via sterically-controlled amino decoordination-recoordination of amino-pyridine. Kinetic studies of E-insertion, N-insertion of complexes 1 and 2, and the geometric isomerization reactions have been examined by means of VT-NMR, and found in excellent agreement with the results estimated by DFT calculations. The more facile N-insertion in the cis-isomers, and ready geometric isomerization, cooperatively lead to a new mechanism that accounts for the novel catalytic formation of alternating COC.

1. Introduction

<span class="Chemical">Cyclic olefin n class="Chemical">copolymers (<spn>an class="Chemical">COC) have demonstrated wide-range applications in the industries of coating, packaging, medical equipment, etc. [1,2,3,4,5,6], because of their high Tg, high transparency, low dielectric constants, and good biocompatibility and processability [7,8,9,10,11,12,13,14,15,16]. These physical properties are usually controlled by the compositions and the microstructures of the monomers, resulting from the design of the catalysts [17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36]. For instance, with the assistance of methylaluminoxane (MAO) [17,37,38,39,40,41,42,43,44,45,46,47,48,49,50], most of the early-transition-metal metallocene derivatives catalyzed copolymerization to afford random E–N products. The ligand steric hindrance readily impedes the norbornene insertion, and thus the PE segments are usually long. On the other hand, if the copolymerization is catalyzed by the non-metallocene titanium complexes, many examples provide an alternating microstructure for the E–N copolymer products [51,52,53,54,55,56,57,58,59,60,61,62,63].

Although late-transition <span class="Chemical">metaln> catalysts have been employed to generate functionalized <span class="Chemical">polyolefins [64,65,66,67,68,69,70,71,72], the successful examples to produce <span class="Chemical">COC are still limited [73,74,75,76,77,78,79], because of the occurrence of β-H elimination on the alkyl chain. We previously reported our finding of the alternating E–N copolymerization, induced by methylpalladium(II) cations and bearing hemilabile bidentate ligand of α-amino–pyridines without the assistance of MAO (as shown in Scheme 1) [80]. The general features of the copolymerization show Mw~104, PDI~1.5, Tg~100–120 °C, and characteristic N%~50, indicating that the COCs have a predominated alternating E–N microstructure. Furthermore, the reactivity ratios obtained from the linear Fineman–Ross plot show both r1 (kEE/kEN) and r2 (kNN/kNE) are smaller than 1, revealing that both EE- and NN-insertion would be more difficult than EN- or NE-insertion [80]. Synthesis of the alternating copolymer is important for optical applications, since the chemical homogeneous products generate high transparency. The more the percentage compositions of alternating microstructures the copolymers contain, the higher the chemical homogeneity [73].

Scheme 1

Alternating E–N Copolymerization reactions, catalyzed by {[(α-Amino–pyridine)]Pd(Me)(NCMe)]}[BF4].

In order to obtain more detailed information on the structure–reactivity relationship for <span class="Chemical">n class="Chemical">COC formation (in terms of the late transition <spn>an class="Chemical">metal complexe {[(α-Amino–pyridine)]Pd(Me)(NCMe)]}[BF4]), we carefully examined the kinetics in many catalysis-related fundamental steps. Considering E–N copolymerization may be initiated from olefin-coordination, methylpalladium (1) and norbornylpalladium (2) complexes were chosen to study both norbornene and ethylene insertion, as well as their geometric isomerization. The kinetics provide the E–N copolymerization rate constants kEN, kEE, kNN, whilst kNE was given by DFT calculations. The mechanism and the complete energy profiles for the organopalladium catalyzed alternating COC formation were also evaluated. The results establish a comprehensive mechanistic picture that may account for the alternating COC microstructures, caused by the {[(α-Amino–pyridine)]Pd(Me)(NCMe)]}[BF4] system.

2. Results and Discussion

2.1. Synthesis and Characterization of Norbornyl Palladium Complexes

The <span class="Chemical">norbornylpalladiumn> complexes {<span class="Chemical">[RHNCH2(o-C6H4N)]Pd(C7H10Me)(NCMe)}(<span class="Chemical">BF4)(R = Pr (2a), Bu (2b), Ph (2c), 2,6-Me2C6H3 (2d), 2,6-Pr2C6H3 (2e)) were synthesized from the reaction of the corresponding methylpalladium complexes 1a–e and 1.3 molar equivalents of norbornene, to achieve satisfactory yields in MeCN/THF (v/v = 1/8) (Scheme 2). The weakly donating solvents could efficiently suppress successive norbornene-insertion [81].

Scheme 2

Synthesis of Norbornylpalladium Complexes.

Notably, the <n class="Chemical">span class="Chemical">norbornylpalladium complexes, synthesized via single N-insertion into a catalytically active spn>ecies, in a process of <span class="Chemical">polyolefin formation, are rare [81]. Only a few examples had been achieved either by introducing a functionalized <span class="Chemical">alkyl on the olefin [82,83,84,85,86,87,88], or through a neutral palladium species that showed relatively lower affinity to the olefins [89,90,91]. The NMR data measured in CDCl3 show that two diastereomers, exclusively in trans-diastereomers (T-form), are identified for 2a–b, according to the NOE cross-peaks between the norbornyl and ortho-H of pyridine in the NOESY spectra. However, in the cases of 2c–e, the cis-diastereomers (C-form) in low yields are also detected, as listed in Table 1. Herein, T-geometrical isomers are designated according to the trans-disposition between pyridine and MeCN (vide infra); and, the C-form for such ligands that are in the cis configuration (Figure 1). The selectivity of T-form in 2a,b is ascribed to the steric hindrance between the amino substituent and the bulky norbornyl group. Such a factor turns out to be critical to the kinetic and thermodynamic selectivity (vide infra). The formation of small amounts of the C-form in 2c–e is allowed, presumably due to the constraint of free rotation (due to the amino aryl substituents).

Table 1

Relative abundance for the isomers of 1 and 2.

RHNCH2(o-C6H4N)	Me-Pd [8]			MeC7H10-Pd
R		T (%)	C (%)		T (%) (anti/syn)	C (%)
iPr	1a	82	18	2a	100 (30/70)	0
tBu	1b	100	0	2b	100 (27/73)	0
Ph	1c	93	7	2c	96 (41/55)	4
2,6-Me2C6H3	1d	60	40	2d	87 (47/40)	13
2,6-iPr2C6H3	1e	62	38	2e	94 (44/50)	6

Figure 1

Isomers of complexes 1a–e and 2a–e.

The major diastereomer for 2a,b is in syn-configuration, in which the <n class="Chemical">span class="Chemical">methyl on <spn>an class="Chemical">norbornyl and the amino substituent are at the same side, with respect to the coordination plane (Figure 1). The NOESY spectra also show the NOE cross-peaks between these two alkyl groups. The pyridinyl o-H signal of anti-configuration for 2a,b is found more downfield than in the syn-configuration (2a: anti/syn = 8.42/8.32 ppm; 2b: anti/syn = 8.39/8.29 ppm). Accordingly, the relative yields of syn/anti-products for 2c–2e are evaluated.

The crystallographic analyses provide unequivocal evidence for the molecular structures of these geometrical isomers. The single crystals of <span class="Chemical">T-2bn> and C-2e were grown from co-solvents of <span class="Chemical">Et2O/<span class="Chemical">MeCN/CH2Cl2, and their ORTEP drawings are depicted in Figure 2. The collection of C-2e, which is of the kinetic product from N-insertion into 1e, is probably due to ready decomposition of the thermodynamically stable T-2e, via C-H activation at Pr2C6H3 [92,93]. The bond lengths of Pd-N1(sp2), Pd-N2(sp3), and Pd-N3(sp) of 2b,e are comparable with the data of 1a–e [80]. All structural parameters collected in Table S1 are in agreement with the literature data [89,90,91].

Figure 2

ORTEP drawings of (A) {[BuHNCH2(o-C6H4N)]Pd[(C7H10)Me](NCMe)}(BF4) (T-2b); (B) {[(2,6-Pr2C6H3)HNCH2(o-C6H4N)]Pd[(C7H10)Me](NCMe)}(BF4)(C-2e); all H-atoms are omitted for clarity.

2.2. Geometrical Isomerisation of Norbornylpalladium Complex

Dissolving the crystal of C-2e in <span class="Chemical">CDCl3n> in ambient conditions resulted in transformation into T-2e, with the T/C ratio of 94/6. The equilibrium constant (Keq) is estimated as 15.7 (T over C). The value of ΔG0 at 298 K is thus −1.6 kcal/mol, calculated from Keq. The time-resolved spectra of such an isomerization reaction were monitored by variable-temperature <span class="Chemical">1H-NMR technique. The changes of pyridine(o-H), NH, and the <span class="Chemical">norbornyl endo-CH2 signals show the formation of two diastereomers (Figure 3). The chemical shifts for the endo-CH2 of the norbornyl group in C-2e are much more upfield than in T-2e, presumably caused by the ring current from the amino aryl. Kinetic analysis showed that this isomerization reaction gives first-order rate constants 5.48 × 10−5–5.69 × 10−3 s−1, at 248–273 K. The differences of activation enthalpy (ΔH‡isom), Gibbs energy (ΔG‡isom), and entropy (ΔS‡isom), derived from the linear Eyring relationship, are 21.5 kcal/mol, 18.7 kcal/mol, and 9.5 cal/mol·K, respectively (Figure S5).

Figure 3

Time-resolved 1H-NMR spectra for geometrical isomerization of C-2e to T-2e, at 253 K in CDCl3.

The small positive value of ΔS‡isom suggests that the geometrical isomerization of 2e may undergo intramolecular <span class="Chemical">aminen> decoordination–recoordination, along with <span class="Chemical">amine inversion, leading to enantioisomeric exchange (Scheme 3) (although the solvent-assisted associative pathway may also be possible) [94]. The steric repulsion between the amino substituent and the <span class="Chemical">norbornyl group significantly destabilizes the ground state of C-2e, and drives the amine group dissociation. Such a mechanism is also supported by the diastereotopic H-exchange of backbone methylene, in prior work of similar derivatives [90].

Scheme 3

Isomerization reaction mechanism of C-2e.

2.3. Kinetics of Norbornene Insertion Reactions in Organopalladium(II) Complexes

Treatment of 1a–e with excess <span class="Chemical">norbornenen> gave rise to the formation of polyads of <span class="Chemical">norbornyl units, via successive <span class="Chemical">norbornene insertions. To our surprise, all products are in T-form, as displayed in Scheme 4. Besides, the product of tetra-ads (n = 3) was absent upon the reaction course. If the reactions were carried out at 228–263 °C, the products 2a–e were respectively produced first, and the multi-N-inserted products 3a–e (n = 1), such as diads and triads 3a’–e’ (n = 2), were ensuing. In addition, the norbornene-coordinated species and free MeCN were not observed, indicating that the norbornene coordination, instead of migratory insertion, may be rate-determining (Scheme 4). The fact suggests that this case is very different from the reactions of alternating E–N copolymerization, catalyzed by bis(pyrrolide–imine)titanium, in which the olefin-insertion steps are rate-determining [23].

Scheme 4

Reactions of successive N-Insertion in 1a–e.

The kinetic studies on the reactions of the first N-insertion for 1a–e were further ex<n class="Chemical">span class="Chemical">amined with use of <spn>an class="Chemical">1H VT-NMR in a pseudo-first-order manner. As listed in Table 2, entries 1–5 show the effect of the amino substituents on the reactivity at 263 K. In entries 1–3, only the rate constants for the T-form (kobs) were evaluated, since the C-isomer was consumed under the conditions before the 1st measurement. An attempt to record kobs in the condition of lower norbornene concentration failed, because the reaction would proceed too slowly to be monitored. Complex 1b is significantly slower toward norbornene insertion than the runs of other analogous complexes. The observation of the smaller kobs value of 1b is probably because of the bulky Bu substituent on the amine group, which would hinder the norbornene coordination, and thus retard the norbornene-insertion rate.

Table 2

Kinetic data for norbornene insertion reactions of {[RHNCH2(o-C6H4N)]Pd(Me)(NCMe)}(BF4)(R = Pr (1a), Bu (1b), Ph (1c), 2,6-Me2C6H3 (1d), 2,6-Pr2C6H3 (1e)) in CDCl3.

Entry	Cat	Temp. (K)	[Cat] (×10−3 M)	[N] a/[Cat]	kobs (10−4 s−1)
					C	T
1	1a	263	5.3	28.7	- b	8.69
2	1b	263	24.9	82.2	- c	2.40
3	1c	263	2.0	23.2	- b	51.60
4	1d	263	1.1	16.7	134.00	21.80
5	1e	263	3.4	10.8	- b	5.47
6	1e	263	3.4	24.2	- b	6.19
7	1e	263	3.4	62.6	- b	10.10
8	1e	263	3.4	134.5	- b	12.90
9	1e	263	3.4	260.9	- b	23.80
10	1e	263	1.7	21.0	110.00	-
11	1e	258	3.3	31.1	109.00	2.35
12	1e	253	3.4	26.3	41.30	1.29
13	1e	243	3.3	28.0	16.80	-
14	1e	238	3.4	27.2	9.68	-
15	1e	233	3.3	20.1	7.11	-
16	1e	233	3.3	33.7	10.80	-
17	1e	233	3.3	40.5	12.90	-
18	1e	228	3.4	31.3	2.50	-

a Determined by 1H-NMR, using tetramethylsilane as an internal standard; b the C-isomer was consumed under the conditions before the 1st measurement; c no C-isomer.

It is worth noting that in entry 4 the <span class="Species">kobsn> is 6.1 fold faster than <span class="Species">kobs. Similarly, the <span class="Species">kobs are 46 and 32 folds of kobs in entries 11 and 12, respectively. Moreover, if the reactions were carried out in temperatures lower than 243 K (entries 13–18), only the consumption of the C-isomer was observed. The facts indicate that in such low temperatures, no T-to-C isomerization occurs, and the reaction of norbornene-insertion only takes place on the C-isomer. This is an unprecedented example, of which one geometrical isomer in square-planar configuration shows substantially more facile reactivity than another toward the fundamental reactions of norbornene insertion [85,95].

To shed further light on the kinetics of the <span class="Chemical">norbornenen> insertion, we chose 1e to determine the rate laws. The plots of the pseudo-first-order rate constants, against the concentration of <span class="Chemical">norbornene ([N]) in Figure S24, show that <span class="Species">kobs is directly proportional to [N] (Entries 15–17), while kobs[N] relationship is linear with a non-zero intercept, which represents the rate constant of T-to-C isomerization, kisom. (Entries 5–9). Therefore, the rate law for the norbornene-insertion reactions can be written as Equations (1) and (2), indicating that in C-1e the rate of N-insertion is much faster than that of isomerization (whereas in T-1e, the two reactions compete with each other).

Indeed, the rate constants kn class="Chemical">N- (Entry 10) and kN- at 263 K were evaluated as 3.06 × 10−1 and 2.14 × 10−3 M−1s−1, and the rate constants of T-to-C and C-to-T isomerization (kisom and kisom−1) were estimated as 4.60 × 10−4 and 8.48 × 10−4 s−1 (Data were obtained from the intercept of the kobs-[N] linear and the equilibrium constant of isomerization (KC/T = 0.54 at 263 K)). Furthermore, ΔH‡N and ΔS‡N for the N-insertion reaction of C-1e were obtained from the Eyring plot, as 13.8 kcal/mol and −8.9 cal/mol·K, respn>ectively (entries 11–14 and 18, Figure S25). The small negative value of ΔS‡N agrees with an intermediate A-pathway [96]. One may assert that the rate-determining step of <span class="Chemical">norbornene-coordination tends to be critically dependent on the steric hindrance between the amino substituent and the norbornene. Such a stress may be released via either geometrical isomerization or migratory N-insertion.

If the N-insertion is much more favored in the C-form, the problem then arises: why were no Pn class="Chemical">N products observed? A reasonable assumption is that the steric hindrance would be increased by the increasing numbers of inserted <span class="Chemical">norbornene, inhibiting the coordination step. Accordingly, we investigated the 2nd and the 3rd insertion rates for 1a at 263 K, and found they are ranked in the order of 1st > 2nd ~3rd >> 4th (Figure S26 and Table S9).

The <span class="Chemical">norbornenen>-insertion reactions of 1 and 2 are worthy to note in several aspects: (1) the N-coordination step is rate-determining, and the N-insertion reactions are favored in the cis isomers; (2) the products favor the trans isomers, and the C–T isomerizations take place owing to the steric effects in the hemilabile ligands; and (3) the geometrical isomerization is capable of winning the competition over the successive N-insertion, thus facilitating the alternating N–E-insertion.

2.4. Kinetics of Ethylene Insertion Reactions in Organopalladium(II) Complexes

In the presence of excess <span class="Chemical">ethylenen>, <span class="Chemical">MeCN in 1a–e were readily replaced by <span class="Chemical">ethylene to form T/C-{[RHNCH2(o-C6H4N)]Pd(Me)(C2H4)}(BF4) (1a’–e’), followed by the E-insertion and an ensuing E-coordination to give T/C-{[RHNCH2(o-C6H4N)]Pd(C3H7)(C2H4)}BF4) (4a’–e’), as shown in Scheme 5.

Scheme 5

E-Insertion Reactions in complexes 1a–e.

Consequently, <n class="Chemical">span class="Chemical">ethyl(ethylene)palladium in the form of T/<spn>an class="Chemical">C-{[RHNCH2(o-C6H4N)]Pd(Et)(C2H4)}-(BF4) (5a’–e’), along with the propene and butenes of chain-transfer products, were detected [97]. T-{[RHNCH2(o-C6H4N)]Pd(Et)(MeCN)}(BF4) (5a–c), with respective yields of 29%, 50%, and 24%, were also obtained. The absence of the C-5a–c is presumably due to the conformational steric hindrance between the amino substituent and the ethyl ligand [80,98]. T-forms are of the more favored configurations in 1a’–d’ and 5a’–e’, again, ascribed to the steric effect (Table 3). The E-coordinated complexes appear to undergo geometrical isomerization more facile than that in 2a–e, according to the broadened or coalescent Pd-alkyl signals for 1b’–e’, 5a’–b’, and 5d’–e’ found at 263 K.

Table 3

The yields of E-coordinated Pd(II) species, from the reactions of 1 and ethylene in CDCl3.

RHNCH2(o-C6H4N)	E-Coord. Complexes
R		T (%)	C (%)		T (%)	C (%)
iPr	1a’	58	42	5a’	59	41
tBu	1b’	63	37	5b’	58	42
Ph	1c’	100	0	5c’	100	0
2,6-Me2C6H3	1d’	59	41	5d’	64	36
2,6-iPr2C6H3	1e’	38	62	5e’	67	33

The kinetic data in Table 4 were collected by monitoring the Me-n class="Chemical">Pd signals of 1a–e under pseudo-first-order conditions, with excess of [E]. The values of kE were measured from the transformations of 1a’–e’ to 4a’–e’. Unlike in the processes of N-insertion, the migratory E-insertion in 1’ is comparable between two geometrical isomers. On the other hand, the differences of kE are limited in terms of ligand variation (Entries 1–4 and 6).

Table 4

Kinetic data for the pseudo-first-order kE of ethylene insertion reaction of 1a–e in CDCl3.

Entry	Temp. (K)	Cat	[Cat] (10−3 M)	[E] a/[Cat]	kE (10−4 s−1)
					C	T	Total b
1	263	1a	5.3	18.3	c	c	6.44
2	263	1b	3.6	21.3	d	8.05	-
3	263	1c	4.4	25.9	e	1.93	-
4	263	1d	3.2	17.7	4.74	4.72	4.28
5	268	1e	3.4	21.7	13.10	14.30	13.60
6	263	1e	3.3	14.3	4.36	3.80	4.14
7	263	1e	3.3	25.0	4.46	4.19	4.37
8	263	1e	3.3	51.9	4.91	4.27	4.71
9	258	1e	3.4	49.4	3.40	3.20	3.37
10	253	1e	3.4	32.5	1.16	1.17	1.16
11	243	1e	3.3	27.2	0.41	0.41	0.41

a Determined by 1H-NMR spectra using tetramethylsilane as an internal standard; b the rate is estimated from the sum of decrease in intensity of C- and T- isomers in 1H-NMR spectra; c the signals of Pd-Me for C- and T- isomers are overlapped; d all the signals are overlapped; and e lack of C- isomer.

In entries 6–8, the values of kE for 1e appear to be independent on [E], suggesting the rate-determining step ought to be the step of E-insertion [93,99,100,101,102]. Moreover, the Eyring analyses provide the activation parameters of the E-insertion in 1e, as 17.3 kcal/mol for ΔH‡E and −7.4 cal/mol·K for ΔS‡E, that are comparable to those for n class="Chemical">N-insertion in C-1e. That is probably also an accidental factor for the occurrence of the E–N alternating <span class="Chemical">copolymerization.

In summary for the E-insertion reactions of 1a–e, the E-coordination to 1, as well as the geometrical isomerization of the <n class="Chemical">span class="Chemical">ethylene-coordinated spn>ecies, has to be ready. The successive E-insertions are quenched by β-elimination, but may be accelerated at high [E]. It is assumed that a hydrido<span class="Chemical">palladium species may be formed by chain transfer, and serves as the actual catalytically active resting state. Overall, the E-insertions are less critical than the N-insertions in <span class="Chemical">{[RHNCH2(o-C6H4N)]Pd(Me)(NCMe)]}[BF4] catalyzed formation of alternating COC.

2.5. DFT Approach for E–N Copolymerization

Kinetic studies for N-insertion and E-insertion reactions provide the evaluations of kn class="Chemical">N and kE, respectively, which can model the kEN, kNN, and kEE for E–N copolymerization. Nevertheless, the <span class="Chemical">1H-NMR spectra for the E-insertion reactions of 2a–e are too complicated to have successful kinetic analysis. In order to gain the missing information for kNE, and to have a comprehensive mechanistic picture for such alternating E–N copolymerization, the gas-phase DFT studies were applied (The chirality of amine is fixed as R form not only for easier comparison but also due to the enantiomers giving almost the same energies in preliminary calculations. The energies for syn- and anti-isomers of norbornyl complex, showing similar energies, were also calculated from energy scan of rotating the Pd–C bond on norbornyl ligand). The complex 1e was chosen for the calculations of the reaction energy profile, in which the fundamental steps were concerned. Considering the replacement of E for MeCN in organopalladium(II) complexes is feasible and complete (as shown in Scheme 5), the E-coordinated species 6/6’ and 8/8’, as illustrated in Chart 1 below, are proposed as the resting states for the calculations.

Chart 1

Structures of resting states. (R = 2,6-Pr2C6H3).

2.5.1. C–T Isomerization

As aforementioned, the isomerization reaction of 1e’ is more facile than those of 1e and 2e, suggesting that the isomerizations with E-coordinate and <n class="Chemical">span class="Chemical">MeCN-coordinate organopalldium species might not be the same. We calculated several kinds of pathways such as A-type [103,104,105], D-type [106], <spn>an class="Chemical">amine-dissociation, Berry pseudorotation [107,108], and direct geometry change [109,110,111,112,113], and found that the last one is the most energy-favored. Rotation of the dihedral angle of Npy–Nam–Pd–CMe affords pseudo penta-coordinate transition states TS(6–6’) and TS(6–6’)’, with barriers of 14.0 and 18.1 kcal/mol, respectively (Scheme 6). TS(6–6’) has a weak δ-agostic interaction of the Pr group, with the distance Pd–H = 2.5286 Å, shorter than the sum of Van der Waals radii (2.72 Å).

Scheme 6

C–T isomerization via distorted tetrahedral species for norbornyl complexes (E + ZPE, kcal/mol). (R = 2,6-Pr2C6H3).

On the other hand, the <span class="Chemical">aminen> decoordination–recoordination pathway has a barrier of 21.7 kcal/mol, in agreement with the case of 2e. The results show C–T isomerization via pseudo penta-coordinate geometry is much lower in energy than via the <span class="Chemical">amine-dissociation pathway. Accordingly, the isomerization mechanisms of the E-bound and <span class="Chemical">MeCN-bond complexes are different. The former may undergo an unusual geometry change, and the later may likely go through amine dissociation. The difference successfully accounts for the experimental results that isomerization of E-bound species are more rapid than the MeCN-bound ones. Likewise, the norboryl species 8 requires 15.1 kcal/mol to overcome activation energy to form 8’, via a similar pathway of geometry change.

2.5.2. Ethylene Insertion into (C2H4)Pd-Me

The calculations for <span class="Chemical">ethylenen> insertion reactions started from the E-bound <span class="Chemical">methyl <span class="Chemical">palladium species 6 and 6’. As depicted in red in Figure 4, the endothermic insertion steps proceed with activation energies of 18.4 and 17.2 kcal/mol, respectively, to give γ-agostic species 10 and 10’. The former data is in excellent agreement with the kinetic data for 1e’ (17.3 kcal/mol). IRC scan led to local minima 6 and 6’, and no in-plane π-complex could be found. The mechanism of ethylene insertion is also supported by the previous studies [93,99,100,101,102,114,115,116].

Figure 4

Energy profile for isomerization, ethylene, and norbornene propagation reaction of 6 and 6’ (E + ZPE, kcal/mol). (R = 2,6-Pr2C6H3).

2.5.3. Norbornene Insertion into (C2H4)Pd-Me.

The <span class="Chemical">norbornenen> insertion reactions of 6 and 6’ start to approach Pd from the side, above the coordination plane of 6/6’. The transition states TS(6–11)/TS(6’–11’) were found to be in the <span class="Chemical">TBP structure. IRC calculations showed that TS(6–11) and TS(6’–11’) would respectively form the N-bound π-products 11 and 11’, first by the <span class="Chemical">olefin replacement. The activation barriers are 19.6 (via T-form) and 13.7 (via C-form) kcal/mol.

Although both substitution reactions are endothermic by 3.9 kcal/mol, the conformation of <n class="Chemical">span class="Chemical">norbornene in 11 and 11’ are different. In 11, the bridge-head CH2 moiety tilts away from <spn>an class="Chemical">pyridine, while the CH2 directs toward the pyridine in 11’. In consequence, insertion directly takes place from 11 to form the γ-agostic product 9 by 10.2 kcal/mol barrier. In contrast, 11’ was turned to 12’ by rotating norbornene, then norbornene insertion followed to give 9’. The rotational motion and insertion barriers are calculated as 1.7 and 8.1 kcal/mol, respectively. The reactions from 9 to 11 and 9’ to 11’ are thus exothermic by 4.8 and 7.4 kcal/mol, respectively (Figure 4).

Overall, the rate-determining step of the <span class="Chemical">norbornenen> insertions should be the substitution of <span class="Chemical">ethylene by <span class="Chemical">norbornene, via a transition state of N-association (TS(6–11)/TS(6’–11’)) from the E-bound species. The reaction via C-form requires a smaller energy barrier, and thus proceeds faster, consistent with the kinetic results of 1e. Again, the activation energy between 6’ and TS(6’–11’), 14.1 kcal/mol, meets the kinetic result of C-1e, 13.8 kcal/mol very well.

The subsequent <span class="Chemical">ethylenen> coordination <span class="Species">to 9’ is calculated as an exothermic reaction by 12.8 kcal/mol, and the corresponding activation energy is 5.2 kcal/mol (Scheme 7). In contrast, the energy difference and activation energy for <span class="Chemical">norbornene coordination to 9’ were calculated as −2.8 and 14.0 kcal/mol. Apparently, the process of ethylene coordination to 9’ is substantially favored after the N-insertion reaction, thereby supporting the alternating copolymerization.

Scheme 7

Coordination pathway of ethylene or norbornene for γ-agostic intermediate 9’ (E + ZPE, kcal/mol). (R = 2,6-Pr2C6H3).

2.5.4. Ethylene and Norbornene Insertion into (C2H4)Pd-C7H10Me

The energy profiles for <span class="Chemical">ethylenen> and <span class="Chemical">norbornene insertion processes from 8 and 8’ are illustrated in Figure 5. The transition structure TS(8’–13’) in the cis-form is 2.0 kcal/mol lower than TS(8–13), thus the activation barrier is 3.4 kcal/mol lower, supporting the kinetic results (red line in Figure 5). As a consequence, one may propose the thermodynamically more stable isomer 8 converts to 8’, followed by <span class="Chemical">ethylene insertion to form 13’, confirming the alternating COC formation.

Figure 5

Energy profile for isomerization, ethylene and norbornene propagation reaction of 8 and 8’ (E + ZPE, kcal/mol). (R = 2,6-Pr2C6H3).

The N-insertion reactions from 8 and 8’ are found to go through the pathways similar to those started from 6 and 6’. However, the activation energies of n class="Chemical">N-coordinating steps are 26.2 (via T-form) and 22.4 (via C-form) kcal/mol, much higher than the cases for 6 and 6’. The process for the C-form still has a smaller energy barrier, although not enough to make competition with the E-insertion. The successive N-insertions appear to achieve the N<span class="Chemical">N-diads in the racemic form, and also agree with the experimental results.

2.5.5. Mechanism of Alternating E–N Copolymerization

The structure-reactivity relationship of alternating E–N <n class="Chemical">span class="Chemical">copolymerization may be illustrated as the catalytic cycle, shown in Scheme 8.

Scheme 8

Alternating Catalytic Cycle of E–N Copolymerization. (Solid line: R = 2,6-Me2C6H3, 2,6-Pr2C6H3; dash line: R = Pr, Bu, Ph).

The resting states of T-I and C-I, representing the E-last-inserted intermediates, are the result of chain transfer. In the major cycle, the C-I reacting with <span class="Chemical">norbornenen> is favored to yield T-II, which prefers to isomerize to C-II. The C-II undergoes facile E-insertion to give T-III, which prefers to isomerize to C-III, then T-IV and C-IV, and so on. The <span class="Chemical">E-N alternating configurations thus may be achieved.

The rapid isomerization between T-I and C-I equilibrates, but is less likely to go to C-II from T-I. In the equilibrium between T-II and C-II, the former one is predominant because of the steric hindrance between the <span class="Chemical">norbornyln> and amino moieties. The T-II can alternatively convert to C-III too, via E-insertion. In the case of T-III, instead of isomerizes to C-III, it may undergo E-insertion to yield C-IV. Such pathways will also achieve E–N alternating configurations.

The successive N-insertion would have a chance to take place in the C-II n class="Chemical">species to give racemic norbornene diads, particularly with high <span class="Chemical">norbornene feeding. The T-II species unlikely proceeds via such a process, due to its disfavored steric effect. The successive E-insertion may also take place in the C-III. Such results were indeed observed under the conditions of high pressure of ethylene and low feeding with norbornene.

One may notice that the equilibria of geometrical isomerizations allow the catalysts to choose the more facile pathway of insertion. For the N-last-inserted n class="Chemical">species such as II and IV, the cis-isomers that undergo E-insertion will be predominant. And, for the E-last-inserted species as I and III, N-insertion may occur in either isomer. The control factor is mainly attributed to the steric effect.

Of course, n class="Chemical">competitions are always available when the kinetic conditions vary upon the reaction course. The formation of N-diads or even N-triad, and E-polyads or branches, had been indeed observed by <span class="Chemical">13C-NMR. With the N-feeding in the region of 6–79 mol%, the produced COCs containing norbornene units in 42.0–59.6% and the alternating compositions of E–N are calculated as 70.4–94.2% (Table S1) [80]. One may conclude that the alternating copolymerization is of a more general situation.

Comparing with the catalytic E–n class="Chemical">N <span class="Chemical">copolymerization using <span class="Chemical">diimino palladium catalysts, in which 5–80 mol % norbornene-feeding resulted in the norbornene contents are ranged in 9–60% in the COCs, the hemilabile unsymmetric amino–pyridine ligands in this study provide unprecedented examples that can afford selective isomerization-controlling olefin-insertions, and lead to high norbornene content as well as a high percentage of alternating E–N microstructure in the COCs.

The activity data, <span class="Chemical">norbornenen> contents, and <span class="Chemical">norbornene blocks, with alternating percentages in the <span class="Chemical">COC products catalyzed by 1a–e, are collected in Table 5. All data for 1d and 1e are generally comparable. Complexes 1a,b,c show substantially lower activity; however, slightly less or nearly comparable results in the norbornene contents and alternating compositions were acquired compared to those of 1d,e. The unusually high percentages of N-block for 1c may be blamed on its relatively large kN and kNN.

Table 5

Norbornene blocks of E–N copolymers catalyzed by 1a–1e. a

Cat	Act. b	NBcoc c (mol %)	Single c (mol %)	Diads c (mol %)	Triads c (mol %)	Alter. c (mol %)
1a	14	44.0	33.7	6.9	3.4	76.6
1b	5	42.9	39.0	3.9	0	81.9
1c	6	47.5	26.4	12.4	8.7	71.0
1d	27	48.1	34.8	6.9	6.4	80.8
1e	28	49.0	38.7	6.1	4.2	86.3

a Reaction conditions: 0.06 mmol of catalysts, 21 bar of ethylene, 50 mL of CH2Cl2, 30 min, and room temperature; b Activity = kg (COC)mol−1(Pd) h−1. Determined by 13C-NMR.

We accordingly ascribe that the activity depends on the initialization of n class="Chemical">N-coordination, as well as the chain transfer or the termination steps. The N-block and alternating compositions depend on the propagation steps, due to N- and E-insertions. Apparently, the fundamental kinetics follow the relationship: kEN > kNE >> kNN, kEE, as well as the concentrations of the monomers. In addition, the critical isomerization equilibria should never be overlooked by viewing the distinct C/T ratios for the resting complexes 5’ and the precursors 1 and 2.

3. Materials and Methods

3.1. Materials

Commercially available reagents were purchased and used without further purification unless otherwise indicated. <n class="Chemical">span class="Chemical">THF, <spn>an class="Chemical">diethyl ether, dichloromethane, hexane and toluene were dried by passage through an M. Braun solvent purification system (MB-SPS) prior to use. Acetonitrile was distilled over anhydrous CaH2. All manipulations of air-sensitive material were performed under a nitrogen atmosphere either in a glove box or by standard Schlenk techniques. The preparation of [(N^N)PdMe(NCMe)][BF4] (1a–e) was described in our previous work [80].

3.2. Measurements

The NMR n class="Chemical">spectra were measured on a Bruker DPX-400, AVIII-400 or DMX-500 spectrometer. The corresponding frequencies for 13C-NMR spectra were 100.625 and 125.753 MHz, respn>ectively. Values upfield of <span class="Chemical">1H- and 13C- data were given in δ (ppm) relative to chloroform in CDCl3 (7.24, CHCl3; 77.0, CHCl3) or to benzene in d6-benzene (7.15, C6H6; 128.7, C6H6). To have good integration data, the 13C NMR spectra of copolymers were obtained at 100.625 MHz in CDCl3 or C6D6 using inverse gated proton decoupling with 30 degree pluse and 3 s delay between the pluses. The NMR probe temperature for kinetic analysis was calibrated by using methanol standard (183–273 K) before the measurement. Assignments are based on 1H- and 13C-NMR spectroscopy with COSY, NOESY and HSQC techniques. High-resolution mass spectrometric analyses were collected on a WATERS LCT Premier XE spectrometer. Elemental analysis was done on a Thermo Scientific FLASH 2000 CHNS analyzer. Gel permeation chromatography (GPC) was performed in toluene at 40 °C using a Kratos model spectroflow 400 equipped with PL-mixed D exclusion limit 400 k columns, and the polystyrene calibration curve was used for analyses. Differential scanning calorimetry was measured under a continuous nitrogen purge (20 mL/min) on a Perkin-Elmer Pyris 6 DSC instrument. The data were gathered on the secondary heating cycle using a heating and cooling scan rate of 10 °C/min.

4. Conclusions

This article reports detailed kinetics and a comprehensive mechanistic study for the catalytic alternating <n class="Chemical">span class="Chemical">copolymerization of <span class="Chemical">norbornene and ethylene, with use of the precursors (methylpalladium(II) cations (1a–e) bearing hemilabile bidentate of α-amino–pyridines).

The kinetic studies evidence that the key resting states are likely to be the <span class="Chemical">ethyl(ethylene)palladiumn> complexes (5a’–e’). The kinetic data of insertion reactions for 1a–e reveal that the catalytic cycle is predominated by the N-insertion reactions, in which the N-coordination is rate-determining. The N-insertion in the C-isomer is significantly more facial (kN-/kN- = 143 at 263 K), and the feasible geometrical isomerization is mainly controlled by the steric effects, cooperatively achieving the alternating N–E copolmerization.

The DFT calculations not only show consistent results with the kinetic measurements, but also afford a n class="Chemical">complimentary mechanistic understanding. The calculations for C-8’ also provide the favored E-insertion. The successive N-insertion in C-8’ will lead to racemic N-block.

This study thus is good example, of which the structure-reactivity relationship may be fine-tuned by steric control. The hemilabile unsymmetric bidentate ligand of α-amino–<n class="Chemical">span class="Chemical">pyridine on the square planar organo<spn>an class="Chemical">palladium(II) cations demonstrate reactivities of convenient geometric isomerization and isomer-differentiated kinetic selectivity towards fundamental olefin-insertion. Such an unusual combination achieves the catalysis of COC-formation, with a high percentage of alternating microstructure.