Zirconium-catalysed hydrosilylation of esters and depolymerisation of polyester plastic waste

Abstract

Schwartz's reagent, Cp₂Zr(H)Cl, has traditionally been used as a stoichiometric reagent for the reduction of unsaturated organic molecules. Recently, methods to use Cp₂Zr(H)Cl as a catalyst have been developed through turnover of the [Zr]–X intermediates, formed upon reaction with an organic substrate, by hydride reagents. Herein, we report the development of a new catalytic pathway for the reduction of esters that uses the bench-stable silane Me(OMe)₂SiH (DMMS) as a mild stoichiometric reductant and Cp₂Zr(H)Cl as the catalyst. This system exploits the regeneration of [Zr]–H through the σ-bond metathesis of [Zr]–OR with [Si]–H to achieve catalyst turnover. These reaction conditions tolerate a range of reducible functional groups (e.g. alkyne, alkene, and nitro) and give high yields of the corresponding alcohols (up to 91% isolated yield). This methodology also proved fruitful for the reductive depolymerisation of a variety of polyesters found in household plastic waste.

---

The monohydride Zr(VI) complex Schwartz's reagent (Cp₂Zr(H)Cl) (1) was first reported by Wailes and Weigold in the 1970s.^1–3 Complex 1 has been shown to be a versatile stoichiometric reducing agent for unsaturated organic compounds.⁴ This reactivity is enabled by the availability of an empty valence orbital on the zirconium centre that allows the coordination of the substrate, followed by reduction via a hydrozirconation pathway. Stoichiometric hydrozirconation using complex 1 was demonstrated for alkenes,⁵ alkynes,³ amides,^6,7 nitriles,⁸ imines,⁹ isocyanates,¹⁰ phosphine oxides,¹¹ and ketones.¹² In 1996, Majoral and Skowronska developed the first stoichiometric application of complex 1 for the reductive cleavage of ester groups in lactones and cyclic anhydrides.¹³ They obtained the corresponding diol derivatives in excellent yields at room temperature after 1 h (Scheme 1). Although these results were limited in terms of scope and mechanistic investigation, they highlighted the efficiency of Schwartz's reagent (1) as a reducing agent for ester functionalities.

Scheme 1 Limited previous examples of ester reduction mediated/catalysed by [Zr]–H.

Only quite recently, there has been a growing interest in using zirconocene complexes as catalysts combined with hydride donors for the reduction of carbonyl groups. In 2019 and 2021, Lecourt et al. described the reduction of primary acetates using Cp₂ZrCl₂ as the catalyst and diisobutylaluminium hydride (DIBAL-H) as the reductant.^14,15 However, this required an excess of pyrophoric reagents, a high catalyst loading (30 mol%) and cryogenic temperatures (−20 to −30 °C). In 2022, Bayeh-Romero et al. reported a single example of the reduction of an ester (a lactone) by the in situ preparation of the catalyst 1 from Cp₂ZrCl₂ and Me(OMe)₂SiH (DMMS) in the presence of Et₂NH.¹⁶ The desired diol was obtained in a moderate yield (55%) even after an extended reaction time (24 h).

Recently, we proved that Cp₂Zr(H)Cl (1) could also catalyse the partial reduction of secondary amides to imines.¹⁷ This reaction proceeded with high chemoselectivity, avoiding over reduction to amine products. The reactivity was enabled by the exquisite inherent selectivity of the catalyst and the use of hydrosilanes as mild reductants. Given that amides with their poor carbonyl electrophilicity are more difficult to reduce than esters and inspired by the previous examples of Zr-catalysed/mediated ester reduction, we herein describe the first general and practical method for the [Zr]–H catalysed hydrosilylation of linear monomeric esters into their alcohol derivatives (Scheme 2). To demonstrate the robustness of our methodology, we expanded the substrate scope to include a range of polyester plastic materials. By overcoming the inherent low reactivity of these polymer materials due to their lack of solubility, inhomogeneity and relatively low purity, we provide an example of chemical upcycling by transforming waste materials into value-added monomer products.^18–20

Scheme 2 Development of a general procedure for Zr-catalysed ester hydrosilylation.

Based on the work of Bayeh-Romero et al. and our previous work on the reduction of secondary amides,^16,17 initial conditions were applied to the reduction of ethyl 4-methylbenzoate 2a, using 10 mol% of Cp₂Zr(H)Cl (1) with 2.1 equiv. of DMMS in THF-d₈ at 80 °C (Table 1, entry 1). After only 1 h, the corresponding 4-methylbenzyl silylated ether 3a^Si was obtained in a yield of >95%. Reducing the catalyst loading to 5 mol% required an extended reaction time of 2 h to reach a similar yield in 3a. The conditions using 1 mol% of catalyst 1 led to the reaction stalling at 81% yield after 18 h (Table 1, entries 2 and 3). To study the effect of varying the reductant, we tested using 5 mol% of catalyst 1 with different hydrosilanes as reductants (Table 1, entries 4 to 7). Polymethylhydrosiloxane (PMHS), Et₃SiH and Ph₃SiH only gave trace yields of 3a, whereas (EtO)₃SiH gave >95% yield after 5 h. This highlights the importance of having a more Lewis acidic silicon centre on the hydrosilane, as noted previously for the reduction of secondary amides.¹⁷ A solvent screen showed a rate increase in C₆D₆, with >95% yield of 3a already after 1 h (Table 1, entry 8). Similar reactivity to that in THF-d₈ was observed in CD₂Cl₂ and in anisole (Table 1, entries 2, 9 and 10), which is regarded as a non-toxic and biodegradable solvent according to the CHEM21 guide.²¹ The influence of the temperature was investigated and full conversion was achieved at 25 °C after 12 h and at 60 °C after 2 h (Table 1, entries 11 and 12). Exchanging DMMS for pinacolborane (HBpin) resulted in the efficient reduction of ester 2a after 2 h at 60 °C, demonstrating that reductants based on boron are also applicable to this system, although only sluggish reactivity was observed using catecholborane (HBCat) (Table 1, entries 13 and 14). The reaction did not proceed in the absence of the catalyst while in the absence of hydrosilane, the 5 mol% of catalyst 1 reacted quantitatively with 2a to give the alkoxyzirconocene products in trace yield (see ESI†).

Table 1 Optimisation of reaction conditions for the reduction of the model ester 2a

Entry	X (mol%)	Solvent	E–H	Time (h)	Yield^a (%) 3a
a Yields measured by ^1H NMR against mesitylene as an internal standard. n.d. = not detected. b Reaction conducted at 25 °C. c Reaction conducted at 60 °C.
1	10	THF-d8	DMMS	1	>95
2	5	THF-d8	DMMS	1	86
				2	>95
3	1	THF-d8	DMMS	1	48
				18	81
4	5	THF-d8	PMHS	7	<5
5	5	THF-d8	Et3SiH	7	n.d.
6	5	THF-d8	Ph3SiH	7	n.d.
7	5	THF-d8	(EtO)3SiH	1	31
				2	65
				5	>95
8	5	C6D6	DMMS	1	>95
9	5	CD2Cl2	DMMS	1	72
				2	>95
10	5	Anisole	DMMS	1	85
				2	>95
11^b	5	C6D6	DMMS	1	30
				12	>95
12^c	5	C6D6	DMMS	1	70
				2	>95
13^c	5	C6D6	HBpin	2	>95
14	5	C6D6	HBCat	2	8

---

To investigate the mechanism of this catalytic system, we first studied the elementary steps by a series of stoichiometric reactions (Fig. 1). The 1 : 1 reaction of complex 1 with ester 2a in C₆D₆ resulted in almost 50% conversion (47%) of the ester 2a to the alkoxyzirconocene products 3a^Zr and 4a^Zr after 2 h at 80 °C. This ratio in 3a^Zr and 4a^Zr did not change upon heating for a further 16 h and at this stage, complex 1 was fully consumed. The addition of 1.1 equiv. of DMMS to this mixture gave full conversion of ester 2a with the formation of the two alkoxyzirconocene products 3a^Zr (36%) and 4a^Zr (21%) and the two silylethers 3a^Si (61%) and 4a^Si (78%) after 1 h at 80 °C. The ratio between these products did not change significantly upon heating for a further 25 h. The potential over reduced products (p-xylene and ethane) resulting from exhaustive deoxygenation were not detected during this experiment.

Fig. 1 Time-step analysis by ¹H NMR spectroscopy of the stepwise stoichiometric reduction of 2a by complex 1 and DMMS.

As only 1.1 equiv. instead of 2 equiv. of DMMS were used, the yields in 3a^Si and 4a^Si cannot reach 100% and thus the presence of both zirconium- and silicon-alkoxides is expected. In line with the relative oxophilicities of Si and Zr,²² the complexes 3a^Zr and 4a^Zr are converted to the silyl ethers 3a^Si and 4a^Si by σ-bond metathesis of [Zr]–OR with [Si]–H with concomitant formation of Cp₂Zr(H)Cl (1). The newly formed complex 1 then reacts with the remaining ∼50% of 2a. The distinct percentages in 3a^Zr/Si and 4a^Zr/Si mentioned above evidence that the metathesis reactions of the two Zr-alkoxide complexes 3a^Zr and 4a^Zr with the hydrosilane do not proceed at the same rate. The conversion from 4a^Zr to 4a^Si is faster than 3a^Zr to 3a^Si due to the greater affinity for the silicon atom of the more electron rich alkoxide EtO⁻. Buchwald and co-workers reported the use of Cp₂TiCl₂ in combination with hydrosilanes and n-BuLi as an activator for the selective reduction of esters.^23,24 A ‘self-activating’ system was also developed using Ti(OiPr)₄ as the catalyst which was proposed to proceed by σ-bond metathesis of the Ti–OiPr bond with the Si–H bond of the silane to generate a reactive Ti–H species.²⁵ However, these methods can require high catalyst loadings (stoichiometric in some cases) and suffer from the potential generation of the flammable and toxic gas SiH₄ depending on the hydrosilane used.^25,26 We have now shown that Schwartz's reagent exhibits similar reactivity to these Ti^IV catalysts and related d⁰ catalysts in the hydrosilylation of esters.

Based on these results and the previous work in this area, we can propose a catalytic cycle (Scheme 3) which is similar to that proposed by Buchwald and co-workers for the catalytic hydrosilylation of esters with Ti(OiPr)₄.²⁵ After the addition of the Zr–H bond of the catalyst 1 to the carbonyl C=O bond of the ester 2, the hemiacetal intermediate (Int-1) (not detected) degrades rapidly into the zirconium alkoxide 4^Zr and aldehyde via a β-alkoxide elimination in a similar fashion to classical stoichiometric ester reduction reagents (e.g. LiAlH₄). Under the catalytic conditions, the aldehyde (not detected) is immediately reduced by complex 1 to the Zr alkoxide complex 3^Zr.¹⁶ The two alkoxyzirconocene complexes 3^Zr and 4^Zr perform a σ-bond metathesis with the hydrosilane to form the desired silyl ethers 3^Si and 4^Si, respectively, with concomitant regeneration of the catalyst 1.^16,17

Scheme 3 Proposed catalytic cycle for the reduction of esters 2 by complex 1.

Using the optimised conditions defined above for ethyl 4-methylbenzoate 2a (Table 1, entries 2 and 8), the generality of this system was explored through a substrate scope (Scheme 4). The benzoate esters 2a and 2b and methyl phenylacetate 2c were cleanly reduced to the corresponding alcohols. Iodo- and amino-substituted esters 2d and 2e were well tolerated in this reaction with no observed hydrodehalogenation nor N–H silylation, respectively. The keto-substituted ester 2f was reduced to the diol by addition of 3.1 equiv. of DMMS. An ester containing a nitro group 2g required heating at lower temperature (40 °C) for an extended reaction time (16 h) to control the chemoselectivity. Alkynes and alkenes 2h and 2i were not reduced under these conditions with only selective ester reduction observed for these examples, although it was necessary to perform the reduction of 2h at a lower temperature of 40 °C for 16 h to ensure good chemoselectivity. ε-Caprolactone 2j was efficiently reduced to the corresponding diol. The electron-poor ester 2k was converted to the desired products but required an extended reaction time of 16 h to reach full conversion. The sterically congested ester 2l was also efficiently transformed to the alcohol products.

Scheme 4 Substrate scope for the hydrosilylation of monomeric esters 2 catalysed by complex 1. The yields of the silyl ethers were measured by ¹H NMR spectroscopy by integration against mesitylene as an internal standard (reactions conducted in C₆D₆). The yields of the alcohols are isolated yields from reactions in THF after basic workup (see ESI†). ^aReactions conducted at 40 °C.

Given that this catalytic hydrosilylation system could reduce monomeric esters to alcohols in high efficiency, we were encouraged to apply it to the reductive depolymerisation of polyesters found in household plastic waste.^27,28 Within this growing research field, our group previously reported the catalytic reductive depolymerisation of a variety of polyesters with the strongly Lewis acidic catalysts B(C₆F₅)₃ and [Ir(POCOP)H(THF)][B(C₆F₅)₄] (POCOP = 1,3-(tBu₂PO)₂C₆H₃) using hydrosilanes as hydride donors.^29,30 Recently, Fernandes et al. performed the hydrosilylation of polyesters from commercial and waste samples using MoO₂Cl₂(H₂O)₂ and Zn(OAc)₂ as catalysts.^31,32 Related to this hydrosilylation reactivity, we have demonstrated that an f-element complex (La(N(SiMe₃)₂)₃) could act as a catalyst to depolymerise polyesters and polycarbonates in the presence of hydroboranes at low catalyst loadings (1 mol%).³³ However, this catalyst was inactive for the catalytic hydrosilylation of polyesters.

Therefore, we tested the above optimized conditions on three commonly used polyesters, the biodegradable synthetic polycaprolactone (PCL), two samples of polylactic acid (PLA) from blue and white 3D printer filaments and polyethylene terephthalate (PET) from an Evian© bottle (Scheme 5). Excellent yields of the desired disilyl ethers from PCL and PET were obtained with 94% of 3j^Si and >95% of 3n^Si and 4n^Si after 4 h and 22 h, respectively, at 80 °C. Interestingly, this system evidenced a high tolerance toward additives contained in the white and blue PLA filaments, giving 3m^Si as the sole product with distinct yields due to the differing additive content (in 75% and >95% respectively, after 4 h at 80 °C). In all cases, the pure diol products 3j, 3m and 3n could be readily isolated by hydrolysis of the O–Si bond under basic conditions. Hydroboranes (HBpin and HBCat) were tested in this system as boron-based reductants but were poorly reactive in comparison with DMMS (see ESI†). In the interest of improving the sustainability of this process, anisole was tested as a solvent for the depolymerisation of the three plastics. Similar yields of the corresponding disilyl ether products were reached (81–84%) but required extended reaction times (38–54 h) (see ESI†). Contrary to the hydrosilylation conditions with B(C₆F₅)₃,³⁰ Ir(III),²⁹ Zn(II)³¹ and Mo(VI)³² catalysts which displayed poor chemoselectivity and were sensitive to the reaction conditions, the Cp₂Zr(H)Cl (1) catalyst evidenced high selectivity toward the disilyl ether derivatives without formation of over-reduced products (alkanes or monosilyl ethers), which is reminiscent of the hydroboration conditions using a La(III) catalyst.³³ Surprisingly, compared to the La(III)/HBpin system, the PCL depolymerisation required harsher conditions (4 h at 80 °C vs. 1 h at r.t.) and under similar experimental conditions, higher yields were obtained with PET (>95% after 22 h at 80 °C vs. 73% after 24 h at 100 °C).

Scheme 5 Reductive depolymerisation of polyesters catalysed by complex 1. The yields of silyl ethers were measured by ¹H NMR spectroscopy by integration against mesitylene as an internal standard (reactions conducted in THF-d₈). The yields of alcohols are isolated yields from reactions in THF after basic workup (see ESI†).

Conclusions

In conclusion, we have developed a robust and efficient catalytic system based on the Earth-abundant metal complex Cp₂Zr(H)Cl (1) with a mild and stable reductant (DMMS) for the hydrosilylation of ester compounds and materials into their alcohol derivatives. In the case of monomeric substrates, we have demonstrated excellent functional group tolerance (alkene, alkyne, amine, etc.) coupled with excellent chemoselectivity to form alcohol derivatives exclusively. With this catalytic system, a range of polyesters obtained from commercial and household waste samples (PCL, PLA and PET) were also successfully depolymerised into their monomeric diols in high yields. Mechanistic studies suggest that catalyst turnover is achieved via the metathesis of Zr–O and Si–H σ-bonds. This work highlights the use of the classical organometallic reagent Cp₂Zr(H)Cl as a powerful catalyst for selective reduction reactions. New catalytic transformations using this complex are currently under investigation in our laboratory.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

For financial support we acknowledge CEA, CNRS, University of Paris-Saclay, European Research Council (ERC consolidator grant agreement no. 818260) and H2020 Sun-to-X project (grant agreement no. 883264). T. C. thanks the Fondation Louis D.–Institut de France for its support.

References

1. P. C. Wailes and H. Weigold, J. Organomet. Chem., 1970, 24, 405–411.
2. P. C. Wailes and H. Weigold, J. Organomet. Chem., 1970, 24, 413–417.
3. P. C. Wailes, H. Weigold and A. P. Bell, J. Organomet. Chem., 1971, 27, 373–378.
4. M. M. Więcław and S. Stecko, Eur. J. Org. Chem., 2018, 6601–6623.
5. D. W. Hart and J. Schwartz, J. Am. Chem. Soc., 1974, 96, 8115–8116.
6. D. J. A. Schedler, A. G. Godfrey and B. Ganem, Tetrahedron Lett., 1993, 34, 5035–5038.
7. D. J. A. Schedler, J. Li and B. Ganem, J. Org. Chem., 1996, 61, 4115–4119.
8. G. Erker, W. Frömberg, J. L. Atwood and W. E. Hunter, Angew. Chem., Int. Ed. Engl., 1984, 23, 68–69.
9. D. Vargová, B. Mudráková, I. Némethová and R. Šebesta, Eur. J. Org. Chem., 2019, 7606–7612.
10. V. Pace, K. de la Vega-Hernández, E. Urban and T. Langer, Org. Lett., 2016, 18, 2750–2753.
11. M. Zablocka, B. Delest, A. Igau, A. Skowronska and J.-P. Majoral, Tetrahedron Lett., 1997, 38, 5997–6000.
12. E. Cesarotti, A. Chiesa, S. Maffi and R. Ugo, Inorg. Chim. Acta, 1982, 64, L207–L208.
13. N. Cénac, M. Zablocka, A. Igau, J.-P. Majoral and A. Skowronska, J. Org. Chem., 1996, 61, 796–798.
14. M. Gavel, T. Courant, A. Y. P. Joosten and T. Lecourt, Org. Lett., 2019, 21, 1948–1952.
15. T. Courant, M. Gavel, R. M. Q. Renard, V. Gandon, A. Y. P. Joosten and T. Lecourt, J. Org. Chem., 2021, 86, 9280–9288.
16. R. A. Kehner, M. C. Hewitt and L. Bayeh-Romero, ACS Catal., 2022, 1758–1763.
17. L. J. Donnelly, J.-C. Berthet and T. Cantat, Angew. Chem., Int. Ed., 2022, 61, e202206170.
18. L. D. Ellis, N. A. Rorrer, K. P. Sullivan, M. Otto, J. E. McGeehan, Y. Román-Leshkov, N. Wierckx and G. T. Beckham, Nat. Catal., 2021, 4, 539–556.
19. Q. Hou, M. Zhen, H. Qian, Y. Nie, X. Bai, T. Xia, M. Laiq Ur Rehman, Q. Li and M. Ju, Cell Rep. Phys. Sci., 2021, 2, 100514.
20. J. Payne and M. D. Jones, ChemSusChem, 2021, 14, 4041–4070.
21. D. Prat, A. Wells, J. Hayler, H. Sneddon, C. R. McElroy, S. Abou-Shehada and P. J. Dunn, Green Chem., 2016, 18, 288–296.
22. K. P. Kepp, Inorg. Chem., 2016, 55, 9461–9470.
23. S. C. Berk, K. A. Kreutzer and S. L. Buchwald, J. Am. Chem. Soc., 1991, 13, 5093–5095.
24. K. J. Barr, S. C. Berk and S. L. Buchwald, J. Org. Chem., 1994, 59, 4323–4326.
25. M. T. Reding and S. L. Buchwald, J. Org. Chem., 1995, 60, 7884–7890.
26. S. C. Berk and S. L. Buchwald, J. Org. Chem., 1992, 57, 3753–3755.
27. C. Wang and O. El-Sepelgy, Curr. Opin. Green, 2021, 100547.
28. A. C. Fernandes, Green Chem., 2021, 23, 7330–7360.
29. L. Monsigny, J.-C. Berthet and T. Cantat, ACS Sustainable Chem. Eng., 2018, 6, 10481–10488.
30. E. Feghali and T. Cantat, ChemSusChem, 2015, 8, 980–984.
31. A. C. Fernandes, ChemSusChem, 2021, 14, 4228–4233.
32. B. F. S. Nunes, M. C. Oliveira and A. C. Fernandes, Green Chem., 2020, 22, 2419–2425.
33. M. Kobylarski, J.-C. Berthet and T. Cantat, Chem. Commun., 2022, 58, 2830–2833.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d2gc02186b

‡ These authors contributed equally to this work.

---