A Hydrocarbon Soluble, Molecular and “Complete” Al-Cocatalyst for High Temperature Olefin Polymerization

The dinuclear aluminum salt {[iBu2(DMA)Al]2(μ-H)}+[B(C6F5)4]− (AlHAl; DMA = N,N-dimethylaniline) is the prototype of a new class of molecular cocatalysts for catalytic olefin polymerization, its modular nature offering easy avenues for tailoring the activator to specific needs. We report here, as proof of concept, a first variant (s-AlHAl) bearing p-hexadecyl-N,N-dimethylaniline (DMAC16) units, which enhances solubility in aliphatic hydrocarbons. The novel s-AlHAl was used successfully as an activator/scavenger in ethylene/1-hexene copolymerization in a high-temperature solution process.

In the framework of a more general study of the aforementioned activation process [26], we have recently identified a novel Al-based cocatalyst, namely {[iBu 2 (PhNMe 2 )Al] 2 (µ-H)} + [B(C 6 F 5 ) 4 ] − (AlHAl, Figure 1, left) [27]. This unusual homodinuclear cation has a distinctive bridging hydride between two Al atoms. While the two Al centers are coordinatively saturated, AlHAl possesses "latent" Lewis acidity; indeed, we reported indications that it can release the unsaturated [AliBu 2 (PhNMe 2 )] + fragment [27], resembling [AlMe 2 ] + release from methylalumoxane (MAO) [28][29][30][31][32][33]. AlHAl can be viewed as a "tamed" version of the highly unstable [AliBu 2 ] + species [34,35], and yet it is able to competently activate group 4 metal precatalysts (and on top of that to scavenge impurities). Activation of a prototypical ansa-zirconocene precatalyst turned out to occur at [Al]/ [Zr] and [Zr] values as low as 50 and~10 −7 M, respectively [27]; catalyst performance in propene polymerization was comparable to that with MAO activation, even though the latter required 100-fold to 1000-fold larger [Al]/ [Zr]. Moreover and at odds with other wellknown molecular activators (e.g., borate salts of organic Lewis and Brønsted acids) [4,10], AlHAl is able to activate dichloride precatalysts with no need for a separate alkylating and scavenging agent, like e.g., tri-iso-butylaluminum (TIBAL). Last but not least, the ease a separate alkylating and scavenging agent, like e.g., tri-iso-butylaluminum (TIBAL). Last but not least, the ease of synthesis and the modular nature of AlHAl (comprised of Alalkyl groups, neutral N-ligands, a bridging hydride and a borate anion) makes it the prototype of a new cocatalyst class, potentially opening the door to tailored applications. On the other hand, like all ionic activators (and MAO too) AlHAl is poorly soluble in aliphatic hydrocarbons, which represents a major drawback for use in homogeneous phase [16,19,20,24,[36][37][38]. To overcome this problem, we modified the N-donor fragments of AlHAl by installing p-hexadecyl residues on the phenyl rings of aniline. In this paper, we report on the synthesis and characterization of the thus obtained s-AlHAl species (Figure 1, right), its solubility in pentane, and the performance in the activation of a representative ansa-zirconocene. The productivity of the resulting catalyst system in hydrocarbon solution at high temperature was evaluated in ethene/1-hexene copolymerization, comparatively with conventional activators.

Materials and Methods
All manipulations were performed under rigorous exclusion of oxygen and moisture in flame-dried Schlenk-type glassware interfaced to a high-vacuum line (<10 −5 Torr), or in a nitrogen-filled MBraun glovebox (<0.5 ppm O2). Molecular sieves (4Å , MS) were activated for 24 h at ca. 200-230 °C under dynamic vacuum. Benzene-d6 and toluene-d8 were freeze-pumpthaw degassed on a high-vacuum line, dried over Na/K alloy, vacuum-transferred to a dry storage tube with a PTFE valve and stored over activated MS in the glovebox. TTB (Boulder Scientific Co., Longmont, Colorado), AB (Boulder Scientific Co., Longmont, CO, USA), TIBAL (Merck, Milan, Italy), DIBAL-H (Merck, Milan, Italy), 2M HCl in ether (Merck, Milan, Italy), 2,6-di-t-butil-4-metilfenolo (BHT; Merck, Milan, Italy), Li[B(C6F5)4] etherate complex (abcr), DMAC16 (Ricci Chimica, Perugia, Italy), and 1,2-dichlorobenzene (>99.8% isomeric purity; Romil, Cambridge, UK) were purchased and used as received. Ethene (99.95%; Linde, Arluno-MI, Italy) was purified by flowing through a column containing activated 4 Å molecular sieves and an activated Cu catalyst (BASF R0-11G). 1-hexene (99%; Merck, Milan, Italy) was purified by passing it through a mixed-bed column of activated Cu catalyst and 4 Å molecular sieves. Toluene (Romil, Cambridge, UK), pentane (Romil, Cambridge, UK), and ISOPAR-G were dried by using an MBraun SPS-5 solvent purification unit. Dichloromethane (Carlo Erba, Cornaredo, Italy) was dried over calcium hydride, distilled and freeze-pump-thaw degassed on a high-vacuum line. Cat-Zr was synthesized according to reported procedures [39] and kindly donated by the group of Prof. A. Voskoboynikov from Lomonosov Moscow State University, Russia. On the other hand, like all ionic activators (and MAO too) AlHAl is poorly soluble in aliphatic hydrocarbons, which represents a major drawback for use in homogeneous phase [16,19,20,24,[36][37][38]. To overcome this problem, we modified the N-donor fragments of AlHAl by installing p-hexadecyl residues on the phenyl rings of aniline. In this paper, we report on the synthesis and characterization of the thus obtained s-AlHAl species (Figure 1, right), its solubility in pentane, and the performance in the activation of a representative ansa-zirconocene. The productivity of the resulting catalyst system in hydrocarbon solution at high temperature was evaluated in ethene/1-hexene copolymerization, comparatively with conventional activators.

NMR Spectroscopy Experiments
NMR experiments were performed by using a Bruker Avance III HD 400 instrument equipped with a smart probe (400 MHz for 1 H). 1 H NMR spectra were referenced to the residual protons of the deuterated solvent used; 13 C NMR spectra were referenced internally to the D-coupled 13 C resonances of the NMR solvent. To describe the multiplicity of the signals, the following abbreviations are used: s, singlet; bs, broad singlet; d, doublet; bd, broad doublet; dd, doublet of doublets; t, triplet; and m, multiplet.
DOSY NMR measurements were performed at 298 K in pentane by using a coaxial capillary containing acetone-d 6 . The standard double-stimulated echo pulse sequence without spinning was used. The shape of the gradients was rectangular, their duration (d) was 4 ms, and their strength (G) was varied during the experiments. All the spectra were acquired by using 16 K data points, 32 increments, between 16 and 128 scans, a spectral width of 6000 Hz, an acquisition time of 1.3 s, and a relaxation delay of 2 s per transient, and processed by using the inverse Laplace transform (ILT) routine of the Bruker Dynamics Center software suite. Pentane was used as internal standard to account for temperature and viscosity fluctuations; its D t was calibrated by using an external sample of HDO in D 2 O under the same conditions [40]. The D t data were treated as described in the literature to derive the hydrodynamic dimensions [41,42]. The uncertainties were estimated by assuming an error of approximately 5% in the hydrodynamic radius, and 15% in hydrodynamic volume and aggregation number [43,44]; error intervals in K aggr are given at 95% confidence.
The average aggregation number was calculated as the ratio between the average hydrodynamic volume of the aggregate (V H ) and that of the single ion pair (V H 0 ) [45]. The latter was estimated as the sum of the hydrodynamic volume of the cation (V H +0 ) and the anion (V H −0 ) by assuming that where the hydrodynamic volume of the free DMA C16 molecule (V H DMAC16 = 659 ± 99 Å 3 ) and that of TIBAL (V H TIBAL = 381 ± 57 Å 3 ) were estimated by performing ad-hoc DOSY NMR experiments, whereas that of the borate anion (V H −0 = 605 ± 91 Å 3 ) was obtained from the literature [44], resulting in V H 0 = 2432 ± 365 Å 3 . Experimental trends of N vs. C (see Section 3.1) were analyzed as reported in the literature [44] by using three different models for indefinite self-aggregation data: EK (equal K), IK (incremental K), and AK (attenuated K) [46]. The IK gave a better fit than the EK model (as determined by the Euclidean norm of the residuals with respect to the experimental N values, being 0.594 and 9.510, respectively), whereas the AK model was unable to reach convergence.  13

Polymerization Experiments
Ethene/1-hexene copolymerization experiments were performed in a Freeslate Parallel Pressure Reactor setup with 48 reaction cells (PPR48), fully contained in a triple MBraun glovebox operating under nitrogen. The cells, each with a liquid working volume of 6.0 mL, featured an 800 rpm magnetically coupled stirring, and individual online reading/control of temperature, pressure, monomer uptake, and monomer uptake rate. Experiments were carried out according to established experimental protocols [27,39,[48][49][50][51], under the following experimental conditions: ISOPAR-G solvent; T p = 100 • C; p(C 2 H 4 ) = 12 bar (175 psi; monomer fed on demand in a semi-batch fashion); 1-hexene initial concentration in reaction solution of 3.7 v/v% was used (see Section 3.2). Conversion of 1-hexene was kept below 15% to ensure a constant comonomer concentration. All experiments were performed at least in duplicate.
Prior to the execution of a polymerization library, the PPR modules undergo "bakeand-purge" cycles overnight (8 h at 90-140 • C with intermittent dry N 2 flow), to remove any contaminants and leftovers from previous experiments. After cooling to glovebox temperature, the module stir tops are taken off, and the 48 cells are fitted with disposable 10-mL glass inserts (preweighed in a Bohdan Balance Automator, Mettler-Toledo, Milan, Italy) and polyether ether ketone (PEEK) stir paddles. The stir tops are then set back in place, and N 2 in the reactors is replaced with ethene (ambient pressure).
The cells are then loaded with the appropriate amounts of a mixed alkane diluent (ISOPAR-G), containing TIBAL as a scavenger where appropriate, and 1-hexene. The system is then thermostated at the reaction temperature and brought to 12 bar of pressure with ethene. At this point, the catalyst injection sequence is started; aliquots of (a) an ISOPAR-G "chaser", (b) a toluene solution of catalyst, (c) a toluene solution of the proper activator, and (d) a ISOPAR-G "buffer" are uploaded into the needle and subsequently injected into the cell of destination in reverse order, thus starting the reaction. This is left to proceed under stirring (800 rpm) at constant temperature and pressure with feed of ethene on demand until the desired monomer consumption has been reached (for reaction time; see Table S4), and quenched by overpressurizing the cell with 50 psi (3.4 bar) of dry air (preferred over other possible catalyst quenchers because in case of cell or quench line leakage oxygen is promptly detected by the dedicated glovebox sensor).
Once all cells have been quenched, the modules are cooled down to glovebox temperature and vented, the stir tops are removed, and the glass inserts containing the reaction phases are taken out and transferred to a centrifugal evaporator (RVC 2-33 CDplus, Martin Christ, Osterode am Harz, Germany), where all volatiles are removed, and the polymers are thoroughly dried overnight. The dry polymers were robotically weighed while still in the reaction vials, subtracting the prerecorded tare. Reaction yields are double-checked against online monomer conversion measurements. Polymer aliquots are then sent to the characterizations.
All polymers were characterized by means of high-temperature gel permeation chromatography (GPC) and 13 C NMR spectroscopy. GPC curves were recorded with a Freeslate Rapid GPC setup, equipped with a set of two mixed-bed Agilent PLgel 10-µm columns and a Polymer Char IR4 detector. Calibration was performed with the universal method by using 10 monodisperse polystyrene samples (M n between 1.3 and 3700 kDa).

Synthesis and Characterization
s-AlHAl was synthesized by adapting the previously optimized procedure for AlHAl and its mononuclear precursor Al [27], replacing N,N-dimethylaniline (DMA) with its p-hexadecyl substituted analogue (DMA C16 ). Reaction of equimolar amounts of TIBAL, substituted aniline and the corresponding anilinium borate provided access to the mononuclear species s-Al [52], which was then converted into the dinuclear product by addition of one equivalent of di-iso-butylaluminum hydride (DIBAL-H; Scheme 1). to those of AlHAl, except for the additional set of signals of the p-hexadecyl residues. Likewise, the equilibrium constant (K eq ) for the reaction

s-Al + [AliBu 2 H] 3
s-AlHAl + [AliBu 2 H] 2 (1) in toluene-d 8 at 298 K was estimated to be approximately 24, corresponding to a ∆G ≈ −1.9 kcal/mol, implying that s-AlHAl is the largely dominant species in solution ( Figure S3 in Supplementary Materials). These values are comparable within experimental uncertainty with those reported for Al and AlHAl under the same conditions (K eq = 12; ∆G ≈ −1.5 kcal/mol) [27], indicating only marginal effects of the para-alkyl substituents on the relative stability of the mono-and dinuclear species. At the same time, solubility of s-AlHAl in saturated hydrocarbons turned out to be significantly enhanced. Pentane solutions with concentration of (at least) 7 × 10 −2 M can be easily prepared. This value is well adequate, considering that typical catalyst concentrations in industrial reactors are as low as 10 −7 M, and that (as noted before) less than 10 2 equivalents of this novel activator are sufficient for efficient catalyst activation and impurity scavenging [27] (with MAO, which is the preferred cocatalyst in commercial processes, [Al]/[M] ratios between 10 3 and 10 5 are often necessary [9,53]).
Importantly, the mononuclear species s-Al is far less soluble than s-AlHAl; this is convenient from the synthetic point of view because it allows for easy purification of the mononuclear species by washing with pentane (see Materials and Methods). The difference can likely be ascribed to the larger nature of the cation and the more delocalized charge of the dinuclear species [27], facilitating solvation by nonpolar solvents.
The high solubility of the alkyl-substituted cocatalyst also allowed recording NMR spectra in pure pentane at 298 K ( Figure 2). Although the concentrations of s-Al and DIBAL-H could not be reliably estimated under these conditions, dinuclear s-AlHAl appears to be the dominant species in solution, similar to what is observed in toluene-d 8 . Furthermore, diffusion-ordered NMR spectroscopy (DOSY) experiments [54,55] could be carried out to investigate the tendency of s-AlHAl to aggregate in this representative low-polarity solvent [45]. The alkyl substituent on the phenyl ring does not significantly affect the overall structure. The NMR spectra of s-AlHAl (see Supplementary Materials) are analogous to those of AlHAl, except for the additional set of signals of the p-hexadecyl residues. Likewise, the equilibrium constant (Keq) for the reaction in toluene-d8 at 298K was estimated to be approximately 24, corresponding to a ΔG ≈ −1.9 kcal/mol, implying that s-AlHAl is the largely dominant species in solution ( Figure S3 in Supplementary Materials). These values are comparable within experimental uncertainty with those reported for Al and AlHAl under the same conditions (Keq = 12; ΔG ≈ −1.5 kcal/mol) [27], indicating only marginal effects of the para-alkyl substituents on the relative stability of the mono-and dinuclear species. At the same time, solubility of s-AlHAl in saturated hydrocarbons turned out to be significantly enhanced. Pentane solutions with concentration of (at least) 7 × 10 −2 M can be easily prepared. This value is well adequate, considering that typical catalyst concentrations in industrial reactors are as low as 10 −7 M, and that (as noted before) less than 10 2 equivalents of this novel activator are sufficient for efficient catalyst activation and impurity scavenging [27] (with MAO, which is the preferred cocatalyst in commercial processes, [Al]/[M] ratios between 10 3 and 10 5 are often necessary [9,53]).
Importantly, the mononuclear species s-Al is far less soluble than s-AlHAl; this is convenient from the synthetic point of view because it allows for easy purification of the mononuclear species by washing with pentane (see Materials and Methods). The difference can likely be ascribed to the larger nature of the cation and the more delocalized charge of the dinuclear species [27], facilitating solvation by nonpolar solvents.
The high solubility of the alkyl-substituted cocatalyst also allowed recording NMR spectra in pure pentane at 298 K ( Figure 2). Although the concentrations of s-Al and DIBAL-H could not be reliably estimated under these conditions, dinuclear s-AlHAl appears to be the dominant species in solution, similar to what is observed in toluene-d8. Furthermore, diffusion-ordered NMR spectroscopy (DOSY) experiments [54,55] could be carried out to investigate the tendency of s-AlHAl to aggregate in this representative lowpolarity solvent [45].  The trend of average aggregation number (N) vs. concentration (C; Figure 2c) was estimated in the range 0.6-33 mM at 298 K. Experimental datapoints were fitted with the incremental K aggr (IK) model for indefinite self-association [43,44,46,56] to extrapolate the reference equilibrium constant K aggr, IK = 3343 ± 1032 M −1 and the corresponding ∆G aggr, IK = −4.8 ± 0.2 kcal/mol (see also Materials and Methods and Supplementary Materials). These values are comparable to those reported for the conceptually related zirconaziridine [Cp 2 Zr(η 2 -CH 2 N(C 18 H 37 ) 2 )] + [B(C 6 F 5 ) 4 ] − having long N-alkyl substituents (K aggr = 15,589 ± 595 M −1 ; ∆G aggr = −6.0 ± 0.1 kcal/mol in cyclohexane-d 12 , 298 K) [43], and are consistent with the formation of outer sphere ion pairs [4,43,57].

Polymerization Results
The cocatalytic properties of s-AlHAl were tested in ethene/1-hexene c tion at 100 °C in a high-boiling alkane solvent by using rac-Me2Si(2,6-dimet 1-indenyl)2ZrCl2 (Figure 3, Cat-Zr) [58] as a representative ansa-zirconoce precatalyst (see Materials and Methods and Supporting Information for fur  A well-established hydrocarbon-soluble activator, namely [HMeN(C 16 H 33 ) 2 ] + [B(C 6 F 5 ) 4 ] − (RIBS), was used as a benchmark [36,37]. For catalytic activity RIBS alone is not sufficient because it can only serve as alkyl abstractor; therefore, TIBAL was added to the reaction system, thus acting as alkylating agent and scavenger [4,10]. s-AlHAl in principle does not require any additive [27], and it was therefore tested with and without added TIBAL. All polymerization experiments in the presence of TIBAL were performed by using [B]/[Zr] and [Al]/[Zr] ratios of 5 and 5.0 × 10 3 , respectively, which is fairly typical [26].
The results obtained under optimized conditions for each cocatalyst formulation are summarized in Table 1. The productivity obtained with s-AlHAl/TIBAL was appreciably higher than that with RIBS/TIBAL ( up to 10 4 -10 5 may be necessary for optimum performance [9,53]). No significant effects of activator nature and level on average copolymer molar mass and comonomer incorporation were observed.

Conclusions
In this paper, we introduced a novel molecular activator of the AlHAl family bearing long alkyl fragments on the aniline donor so as to enhance solubility in aliphatic hydrocarbons. This s-AlHAl species proved to be a competent cocatalyst in ethene/1-hexene copolymerizations in alkane diluent at high temperature. In combination with a representative ansa-zirconocene, catalyst productivities comparable to those with the better-known RIBS/TIBAL were obtained, with the advantage of working with a single-component activator at much lower [Al]/[Zr] ratios. These results represent a first demonstration that the AlHAl cocatalyst frame can be structurally amplified and tailored for specific applications; further variants are being prepared and screened in our laboratories, as will be reported in due course.  Figure S5: DOSY NMR maps of s-AlHAl in pentane at different concentrations; Figure S6: Ln(I/I 0 ) vs. G 2 plots relative to the DOSY maps in Figure S5; Table S1: DOSY NMR measured self-diffusion coefficient (D t ), and estimated hydrodynamic radius (r H ), hydrodynamic volume (V H ) and average aggregation number (N) for TIBAL, DMA C16 and s-AlHAl at different concentrations in pentane at 298 K; Determination of K eq .