Catalytic Asymmetric Addition and Telomerization of Butadiene with Enamine Intermediates

Introduction

The stereoselective construction of C–C bonds is an enduring theme in synthetic organic chemistry.^1–31 As an industrial platform chemical with an annual production of ca. 1.3 × 10⁷ tons, butadiene appears as an ideal starting material for chemical synthesis in the pursuit of atom-economic value-added production.^32,33 To this end, transition-metal catalysts have been explored and carefully tuned to enable selective C–C bond formation reactions of butadiene while suppressing undesired polymeric pathways.^34–51 It is known that butadiene can participate in crotylation,^34–40 1,2- or 1,4-addition,^41–43 and telomerization^41–44 (Scheme 1a). In these reactions, chemoselectivity control has always been an issue and delicate yet judicious selection of ligands and fine-tuning of the additives are normally required. Besides the challenges on chemoselective control, the development of asymmetric catalysis has also been frequently explored, but progress along this line is rather limited. Successful examples are only reported for crotylation by the groups of Krische and Buchwald.^38–40 Asymmetric addition and telomerization with butadiene remains largely underdeveloped (Scheme 1a).

• Download figure
• Download PowerPoint

Mechanistically, the addition to butadiene proceeds via M–H to form a metal-π-allyl intermediate, and the telomerization occurs through oxidative cyclization to form a bis-π-allylpalladium intermediate (Scheme 1).^41–51 The addition of 1,3-dicarbonyl compounds to 1,3-dienes has been known since 1970s with the work of Hata et al.^41,42 However, the reactions yielded mixtures of 1,2-/1,4-addition and telomerization adducts.⁴² In 2004, Hartwig et al.⁵² first reported Pd-catalytic asymmetric intermolecular additions of β-diketones to cyclohexadiene and 2,3-dimethylbutadiene with moderate enantioselectivity. Very recently, Zhou et al.⁵³ and Malcolmson et al.⁵⁴ independently demonstrated asymmetric addition to substituted 1,3-butadiene with Meldrum’s acid derivatives and activated C-pronucleophiles, respectively (Scheme 1a). However, the extension of these catalytic systems to 1,3-butadiene has not been achieved. On the other hand, although the telomerization of butadiene has been developed in academic and industrial research for over 50 years,^45–47 a catalytic enantioselective version remains underdeveloped. All the examined chiral metal catalysis gave unsatisfactory enantioselectivities (<50% ee) (Scheme 1a).^55–59 Herein, we report chiral primary amine/palladium synergistic catalysis for asymmetric C–C bond formation reactions of 1,3-butadiene via enamine intermediates (Scheme 1b). The dual catalysis not only allows for chemoselective tuning between addition and telomerization pathways by simple ligand switch but also enables high enantioselective control for both processes. The reactions could be applied to α-branched aldehydes and β-ketoesters, leading to the formation of acyclic all-carbon quaternary stereocenters with high efficiency and excellent chemo-, regio-, and stereoselectivities.

Results and Discussion

Based on our initial successes on achiral palladium/chiral primary amine catalysis,^60–63 we began by examining the coupling of tert-butyl 2-methyl-3-oxobutanoate 2a and 1,3-butadiene 3a in the presence of chiral primary–tertiary amine catalyst and Pd(OAc)₂. Different phosphine ligands were first examined, and we found that electron-rich bidentate phosphine ligands such as XantPhos and [(oxydi-2,1-phenylene)-bis-(diphenylphosphine)] (DPEPhos) could promote the addition reaction with good chemoselectivity (pathway A, Table 1, entries 1 and 3–5). In these cases, there was only trace or no telomerization adduct 5a detected by ¹H NMR. A bulky tertiary aminocatalyst 1a together with DPEPhos-Pd complex was identified to give the optimal chemoselectivity and enantioselectivity, affording the addition product 4a with 83% yield and 93% ee (Table 1, entry 1). The use of a bulkier primary amine catalyst 1b led to slightly improved enantioselectivity but with compromised activity (Table 1, entry 6 vs 1). In the survey of different palladium salts, it was noted that the use of Pd(C₃H₅)Cp led to considerable formation of a telomerization adduct 5a accompanying 4a (Table 1, entry 7).

The chemoselective telomerization pathway (pathway B) was then separately optimized using Pd(C₃H₅)Cp as the metal catalyst. Considering the fully coordinated nature of the bis-π-allylpalladium species in the telomerization pathway (Scheme 1a), we reasoned the use of monodentate phosphine ligand may facilitate the expected reaction. Previously, monophosphine ligand was shown to mediate a racemic telomerization process via an enol mechanism.^33,38–40 Monophosphine ligand was then screened in the current asymmetric enamine process (Table 1, entries 8 and 10–14). To our delight, the reaction in the presence of tri-p-tolylphosphine was found to exclusively afford the desired telomerization product 5a with 80% yield and 93% ee (Table 1, entry 8). The balance of the electronic and steric effect of phosphine ligand seems to be critical for the activity and chemo- and enantioselectivities. Trialkyl phosphines, such as tricyclohexylphosphine, showed poor yields and rather low enantioselectivities (Table 1, entry 14). The use of electroneutral triphenyl phosphine (entry 10), more electron-rich triaryl phosphine (entry 11), space-demanding ortho-substituted triaryl phosphine (entries 12 and 13), or electron-deficient triaryl phosphines ( Supporting Information Table S1) all led to serious reductions in both activity and enantioselectivity but maintained similar preference to telomerization. Phosphites or phosphoramidites were also examined, showing either poor activity or inferior chemoselectivity (see Supporting Information Table S2).

The use of a bulkier catalyst 1b led to improved enantioselectivity but with sacrifices of both activity and chemoselectivity (Table 1, entry 6 vs 1). In control experiments, the reactions worked well in the absence of aminocatalyst with telomerization product 5a as the major product for both pathways A and B (Table 1, entries 2 and 9). This switch from addition to telomerization (Table 1, entry 1 vs 2 and entry 8 vs 9) suggests a critical role of the aminocatalyst in tuning chemoselectivity besides serving as the primary source of enantioselective control (vide infra).

Substrate scope

With optimized conditions in hand, we then explored the scope of the reactions. Different ester groups with various sizes were tolerated to give the expected allylic adducts for both pathways in good yields and high enantioselectivities (Table 2, entries 1–7 and 18–24). Polar groups such as NHBoc (Boc = t-butyl carbamate) and Cl were well tolerated (entries 6, 7, 23, and 24). Substitution on the α-position of β-ketoesters significantly influenced the reactivity. The ethyl-substituted substrate afforded the telomerization product 5h with 24% yield and 95% ee (Table 2, entry 25) but did not work in the addition pathway (Table 2, entry 8), likely due to steric effect arisen from α-substituent and the space-demanding bidentate-phosphine-palladium complex in the latter process. To our delight, the reaction with cyclic β-ketoesters, such as cyclohexanone, proceeded well to give the corresponding adducts with great control of both chemo- and enantioselectivities for both processes. Variations of ester groups of cyclohexanone retained high yields and enantioselectivities (Table 2, entries 9–11 and 26–29). The nitrogen- and oxygen-containing cyclohexanones could also give the corresponding products with high enantioselectivity, while decreased chemoselectivity was noted in these cases (Table 2, entries 12, 13, 30, and 31). For cyclopentanone, both processes worked well with a slight decrease of chemoselectivity in the telomerization procedure (Table 2, entries 14 and 32). In addition, β-ketoamides were tolerated and the addition pathway proceeded well to give the allylation products with 86–90% yield and 84–87% ee (Table 2, entries 15 and 16), while the telomerization pathway showed low enantioselectivity (Table 2, entry 33). Asymmetric 1,3-diketones, normally considered as challenging substrates in the typical enol process, could be incorporated to give allylation or telomerization products for both pathways A and B; however, the chemoselectivity was rather low in pathway A (Table 2, entries 17 and 34). Both pathways could be conducted in subgram scale with similar performance, highlighting the applicability of the present enantioselective catalysis (Table 2, entries 1 and 18).

We then applied the current catalysis in the late-stage allylic alkylation and telomerization of structurally complexed substrates bearing existing chiral centers. The menthyl esters (Table 2, entries 35 and 37), nopyl (Table 2, entry 36), and even a cholesteryl ester (Table 2, entry 38) could all work smoothly with high reactivity and stereoselectivity.

We next explored the reactions with α-branched aldehydes, for which a highly enantioselective version for both processes has not been reported. In this case, the reactions proceeded to afford telomerization adduct 8a as the major product even under addition-favored conditions with unsatisfied enantioselectivities (Table 3, entries 2 and 3). Further investigation on different phosphine ligands did not improve the chemoselectivity and (p-Tol)₃P still worked preferably for the telomerization process ( Supporting Information Table S3). Different aminocatalysts were further examined to improve the enantioselectivity.

The use of a bulkier tertiary amine 1b/trifluoromethanesulfonic acid (TfOH) favored the enantioselectivity but not the activity and chemoselectivity (Table 3, entries 4 and 5). Changing t-butyl to an adamantyl group led to improved enantioselectivity with good chemoselectivity (Table 3, entry 6 vs 3). Using NHTf₂ instead of TfOH could further increase the ee value to 83% (Table 3, entry 1). Further optimization did not result in additional improvement. It is complicated to tune both chemo- and stereoselectivity by factors from the substrate, the aminocatalyst, and its conjugated acid. A range of α-aryl propanals bearing either electron-donating or -withdrawing groups could be incorporated to afford mainly telomerization adducts 8 with high yields and enantioselectivities (Table 4).

Mechanistic investigation

Control experiments were conducted to gain insight into the reaction mechanism (Scheme 2a). When the bidentate ligand DPEPhos dosage in pathway A was reduced to 5 mol %, the addition was found to be accompanied by telomerization byproduct with 4a/ 5a = 7∶3, but when increasing the loading of monophosphine ligand (p-Me-Ph)₃P to 20 mol % in pathway B, the chemoselectivity deteriorated with addition product appearing ( 5a/ 4a = 4∶1). These could be explained by considering that an electron-rich bidentate coordination is required for the hydropalladation to form I (Scheme 2c), a key intermediate for addition, whereas a delicate yet balanced binding with palladium between monophosphine ligand and butadiene is essential for the telomerization via intermediates IV and V (Scheme 2c). The switch to telomerization pathway in the absence of aminocatalyst 1b/TfOH under pathway A conditions (Table 1, entry 2 vs 1) suggests that the amine-TfOH conjugate could facilitate hydropalladation in the addition pathway by serving as a proton source. On the other hand, such a hydropalladation is disfavored due to the preferential binding with two molecules of butadiene in the context of monophosphine ligand as in the optimized telomerization pathway B.^a

• Download figure
• Download PowerPoint

In line with previous studies,^45–54,64 catalytic cycles for both processes could be proposed as described in Scheme 2c. For pathway A, a Pd-H complex I formed in situ underwent hydrocarbonation with butadiene to afford the key π-allyl-palladium species II, which then coupled with the enamine intermediate III to give the allylation product after hydrolysis. For pathway B, the reaction underwent an oxidative coupling of two coordinated 1,3-butadiene molecules to form the key bis-π-allylpalladium intermediate IV, and subsequent protonation and enamine coupling delivered the telomerization product after hydrolysis (Scheme 2c). Stoichiometric experiments with preformed enamine intermediate III proceeded to give the expected products with comparable enantioselectivity for both pathways, verifying the enamine catalytic nature (Scheme 2b). The observation of diminished chemoselectivity on the telomerization pathway further pinpoints the critical balance and competition between oxidative dimerization (Int. IV) and hydropalladation (Int. II). The presence of a stoichiometric amount of amine/TfOH facilitates the hydropalladation process, hence the formation of addition product 4a in this instance. In accordance with previous studies,^60–63 a steric model could be proposed to account for the observed stereoselectivity.⁶³ Bulkiness from the protonated amino moiety and the space-demanding palladium intermediate ( II or V) both contribute to the stereocontrol.

Conclusion

We have developed tunable asymmetric addition and telomerization of butadiene by synergistic enamine/palladium catalysis. A wide range of α-branched β-ketoesters and aldehydes could be directly coupled with butadiene, affording allylation or telomerization adducts bearing all-carbon quaternary centers in high efficiency and enantioselectivity.

Supporting Information

Supporting Information is available and includes experimental procedures, characterization data, and ¹H NMR, ¹³CNMR, ¹⁹F NMR and high-performance liquid chromatography (HPLC) chromatograms for products (PDF).

Conflict of Interest

The authors declare no competing interests.

Funding Information

This research was made possible as a result of a generous grant from Natural Science Foundation of China.