Reexamining the Evidence for Proton Transfers in Ethylene Polymerization

LETTER LETTER

Reexamining the evidence for proton transfers in ethylene polymerization

Delley et al. (1) propose a new proton trans- Si(OH)Cr–R sites would be extremely rare fer mechanism to explain the mysterious and therefore unobservable (1). To ascribe initiation mechanism (2) of the Phillips po- the predicted vanishingly small abundance to ’ lymerization catalyst. The authors mecha- DFT error (typically approximately ±5–10 nism invokes dormant (D) sites and active kcal/mol for B3LYP) would invoke unchar- Si(OH)CrR sites, as summarized in Fig. 1. acteristically large errors, approximately 17 A combination of transition state theory, kcal/mol = k T ln[109] at temperatures of ’ B Delley et al. s (1) density functional theory catalyst operation (4). When ab initio calcu- (DFT) results, and standard microkinetic lations for a hypothesized mechanism result modeling techniques (3) for the mechanism in such large discrepancies, it is useful to con- of Fig. 1 yield several testable predictions. sider alternative mechanisms and interpreta- DFT is not sufficiently accurate to predict tions of data. When polyethylene samples are absolute rates, but selectivities, molecular sufficiently thick, the polymer itself shows weights, and relative abundances can often −1 be predicted more accurately because they bands at 3,605 and 3,640 cm (5), which are likely two-phonon combinations of CH2 asym- involve ratios of rates. Relative abundances − metric stretching (2,883/2,919 cm 1) and rock- are particularly important because the evi- − Fig. 2. Transmission IR spectra of polyethylene thin films: ing modes (721/731 cm 1) (5). We and others dence for proton transfer hinges upon the (a) made by a supported catalyst Cr[CH(SiMe3)2]3/SiO2, have observed these peaks after removal of after extraction of the catalyst with HF; and (b)madebya assignment of infrared (IR) peaks at 3,640 = −1 homogeneous catalyst, [(C6H3Me2)N C(Me)C(NC6H3Me2) – Phillips catalyst and in polyethylene made with 2 1 and 3,605 cm to Si(OH)Cr R intermediates O-κ -N,O]Ni(η -CH2CMeCH2)(PMe3), activated by B(C6F5)3. (1). According to the proposed mechanism, different catalysts (Fig. 2), as well as in paraf- The Inset shows the spectrum of deuterium-labeled poly- the total abundance of these species is fins of noncatalytic origin. ethylene. Another major issue involves the pro- θ = θ + θ + θ + ... posed termination mechanism. The proton SiðOHÞCr−R 1 2 3 transfer termination pathway is effectively steps (1). Furthermore, the authors dismiss + ð + Þ termination by β-H transfer, citing a large = kT kP 1 K PC2H4 the microscopic reverse of the initiation . barrier to product diene release (see figure kT + kPð1 + K PC2H4Þ + kT ðk−I + kPÞ=kI pathway (apart from vinyl vs. alkyl chain S8 in ref. 1). However, ethylene may assist differences and steric effects of the polymer). in diene displacement after β-H termina- π Because the proton transfer step is unfavor- Although the authors included a -bound tion. This alternative termination pathway able and slow, the authors’ computed free en- ethylene ligand to facilitate initiation, they is important because it would regenerate θ ≈ −9 π ergies predict SiðOHÞCr−R 10 , meaning the omitted -bound ethylene in the termination an Si(OH)CrH site that is ready to insert ethylene, and thus create a faster polymerization cycle. However, the dormant (D) state would still act as an off-cycle trap for vast majority of Cr sites. In summary, Delley et al.’s(1)calcula- tions do not support their interpretation ofthenewIRpeaksasevidenceforthe Si(OH)Cr–R site. Moreover, the IR peaks are indistinguishable from peaks that are intrinsic to polyethylene. Further studies of alternative pathways and site geome- tries are needed to determine whether Si(OH)Cr–R sites can catalyze ethylene polymerization.

Author contributions: B.P. and S.L.S. designed research; B.P., S.L.S., A.F., Y.W., and A.E.S. performed research; and B.P. and S.L.S. wrote the paper. Fig. 1. Initiation (I), propagation (P), and termination (T) steps in the mechanism of Delley et al. (1). The authors report standard free-energy changes and barriers for all steps except for the ethylene binding con- The authors declare no conflict of interest. stant K, which lies in the range 0.1/atm < K < 10/atm, based on quantities from their paper and Supporting 1To whom correspondence should be addressed. Email: baronp@ Information. engineering.ucsb.edu.

www.pnas.org/cgi/doi/10.1073/pnas.1422589112 PNAS Early Edition | 1of2 Downloaded by guest on September 25, 2021 ACKNOWLEDGMENTS. The authors thank the Depart- bDepartment of Chemistry and Biochemistry, and cDepartment of Chemistry and Biochemistry, ment of Energy Basic Energy Sciences for support under Grant DE-FG02-03ER15467. University of California, Santa Barbara, CA 93106; Florida State University, Tallahassee, FL 32306

Baron Petersa,b,1, Susannah L. Scotta,b, a b 1 Delley MF, et al. (2014) Proton transfers are key elementary steps 3 Keii T (2004) Heterogeneous Kinetics: Theory of Ziegler-Natta- Anthony Fong , Youhong Wang , in ethylene polymerization on isolated chromium(III) silicates. Proc Kaminsky Polymerization (Springer, Berlin). c and Albert E. Stiegman Natl Acad Sci USA 111(32):11624–11629. 4 Harvey J (2006) On the accuracy of density functional theory in 2 – aDepartment of Chemical Engineering, University McDaniel M (2010) A review of the Phillips supported chromium catalyst transition metal chemistry. Ann Rep Prog Chem C 102:203 226. and its commercial use for ethylene polymerization. Advances in Catalysis 5 Nielsen J, Woollett A (1957) Vibrational spectra of polyethylenes of California, Santa Barbara, CA 93106; eds Gates BC, Knozinger H (Academic, New York), Vol 53, pp 123–606. and related substances. J Chem Phys 26(6):1391–1400.

2of2 | www.pnas.org/cgi/doi/10.1073/pnas.1422589112 Peters et al. Downloaded by guest on September 25, 2021