Controlled Rh-Catalyzed Mono and Double Decarbonylation of Alkynyl α-Diones to Form Conjugated Ynones and Disubstituted Alkynes

Abstract

A Rh-catalyzed controlled decarbonylation of alkynyl α-diones is described. By using different ligand and solvent combinations, mono and double decarbonylation can be selectively achieved to give conjugated ynones and disubstituted alkynes, respectively. A fundamental study on catalytic activation of un-strained C−C bonds under non-oxidative conditions is presented.

Graphical abstract

Transition metal-catalyzed C−C activation has emerged as a useful tool to carry out transformations that can be difficult or impossible under traditional methods.¹ While novel transformations and synthetic applications have been elegantly demonstrated in strained systems or using a permanent/temporary directing group, far fewer examples are available for catalytic activation and subsequent functionalization of non-strained C−C bonds without the aid of chelation.^2,1j,1o Inspired by the broad applicability of the C-C≡N bond activation,³ our laboratory has been interested in the activation of analogous C-C≡C bonds.⁴ Recently, we reported the Rh-catalyzed decarbonylation of diynones and mono-ynones enabled by using bidentate phosphine ligands with a large bite angle (Scheme 1, eqs 1 and 2).⁵ DFT calculations supported an initial cleavage of the alkynyl acyl bond for these transformations. Given the unique nature of this bond cleavage, it is natural to wonder if such a mode of activation can be extended to other systems, such as 1,2-diketones. Perhaps, a more intriguing question is whether a controlled activation of the vicinal di-carbonyls, e.g. mono vs. double decarbonylation, can be realized in a catalytic fashion. As an exploratory study, here we describe the development of a controlled Rh-catalyzed mono and double decarbonylation of alkynyl α-diketones to form conjugated ynones and disubstituted alkynes (Scheme 1, eq 3).

Scheme 1.

Decarbonylation of Alkynyl Ketones

In 1974, Kaneda and coworkers reported the first examples of Rh-catalyzed decarbonylations of 1,3- and 1,2-diketones; however, low yields and only single decarbonylation were observed.^2d We have recently shown that mono decarbonylation of isatins (cyclic 1,2-diones) followed by alkyne insertion can be catalyzed by a Rh (I) catalyst, albeit requiring the use of a directing group.⁶ To the best of our knowledge, double decarbonylation of 1,2-diketones remains an elusive transformation;⁷ moreover, C-C activation of an yn-α-dione system has not been studied previously. Hence, these reasons motivated us to investigate a controlled catalytic decarbonylation of alkynyl α-diketones. The challenges are twofold: 1) compared to other ketones, yn-α-diones exhibit significantly higher electrophilicity and can undergo facile cycloaddition⁸ and intramolecular cyclization⁹; thus, substrate stability and chemoselectivity are not trivial issues; 2) conjugated ynones, the product of mono-decarbonylation of yn-α-diones, were previously demonstrated to undergo facile decarbonylation to give 1,2-disubstituted alkynes,^5b thus, discovering a new catalytic system that can selectively halt at the mono-decarbonylation stage can be another concern.

To investigate the reactivity of alkynyl diones, 1,4-diphenylbut-3-yne-1,2-dione (1a) was employed as the model substrate, and was first subjected to the previous decarbonylation conditions with 5 mol % [Rh(COD)Cl]₂ and 12 mol % Xantphos in ethylbenzene (Table 1, entry 1). While the doubly decarbonylated product 3a can be isolated in 58% yield, the mono-decarbonylated ynone 2a can only be observed in a trace amount. Nevertheless, this preliminary result was encouraging regarding the reactivity of the yn-dione functional group; next we sought to control the selectivity for the formation of 2a over 3a.

Table 1.

Selected Optimization Studies^a

entry	ligand	solvent (M)	temperature (°C)	time(h)	2a/yield (%)b	3a/yield (%)b
1	Xantphos	ethylbenzene (0.1)	136	48	trace	58
2	Xantphos	chlorobenzene (0.1)	131	48	64	21
3	Xantphos	toluene (0.1)c	130	48	33 (40)	–
4	dppb	toluene (0.1)c	130	48	66 (70)	–
5	dppe	toluene (0.1)c	130	2.5	63	–
6	dppe	toluene (0.05)c	130	2.5	69	–
7	dppe	toluene (0.01)c	130	18	77	–
8	dppe	toluene (0.02)d	110	8	92	–
9	Xantphos	ethylbenzene (0.02)d	136	48	20	75
10	Xantphos	ethylbenzene (0.05)	136	40	–	89

^aThe reactions were run with alkynyl dione 1a (0.10 mmol) under a positive pressure of argon.

^bIsolated yields; number in parenthesis is percent conversion of starting material.

^cReaction run in a sealed vial. The heating bath temperature is 130 °C.

^dSubstrate added via syringe pump over 5 h in the dark.

Lowering the temperature to 131 °C (reflux under PhCl) yielded a mixture of 2a and 3a (Table 1, entry 2). Running the reaction in a sealed vial (versus an open system under a positive pressure of argon) in toluene effectively shut down production of alkyne 3a; however, conversion to ynone 2a remained low (Table 1, entry 3). A screen of ligands revealed that ligands with a smaller bite angle accelerated the reaction rate, with dppe giving the 63% yield in only 2.5 h (Table 1, entry 5), although the exact reason remains unclear. The major byproducts from this reaction were found to be a mixture of structurally unidentifiable dimeric compounds, which was likely caused by undesired intermolecular reactions. Consequently, the reaction concentration was adjusted to avoid dimerizations. Ultimately, it was found that running the reaction at 0.02 M in toluene, with slow addition of the substrate, provided conjugated ynone 2a in 92% yield after 8 hours (Table 1, entry 8).^10,11

With the optimal conditions for single decarbonylation in hand, we next optimized the process for double decarbonylation. When slow addition and diluted conditions were applied with the Xantphos and ethylbenzene system, the yield of 3a was increased, albeit with forming 20% of 2a (Table 1, entry 9). Increasing the concentration to 0.05 M ultimately furnished 3a exclusively in 89% yield (Table 1, entry 10).¹²

With the optimized conditions in hand, the scope of mono decarbonylation was investigated (Scheme 2). A variety of aryl groups containing electron-donating or -withdrawing groups are compatible for this transformation. Lower yields were obtained when an electron-donating group is present on the arene (e.g. 2b and 2c), which is largely attributed to the light sensitivity of the substrates. In general, the reactions were run in the dark to minimize such an issue. Alkyl groups were also tolerated. While giving a low yield (24%) under the dppe conditions, the t-butyl ketone 1f gave 76% yield with the Xantphos ligand and under reflux of PhEt (vide infra). Methyl ketone 1g proceeded with a much lower yield (2g, 17%), which was likely due to the poor stability of the starting material. A range of substituents on the alkyne part, including alkyl, aryl and alkenyl groups, were well tolerated giving the corresponding ynones in good yields without significant electronic bias (2h–2n). Gratifyingly, a complex ethynyl estradiol-derived substrate 1n furnished the mono decarbonylation in 90% yield.

Scheme 2. Mono Decarbonylation of Alkynyl Diones^a.

^a Reactions were run on a 0.10 mmol scale; all yields are isolated yields. ^b Number in parenthesis is percent conversion of starting material. ^c The reaction was run with Xantphos as the ligand in refluxing PhEt.

The scope of double decarbonylation was also explored (Scheme 3). A variety of aryl substituents were tolerated giving good to excellent yields (3a–3e, 53–94%).¹³ While the methyl substituted ketone is certainly more challenging due to the aforementioned stability issue, 32% yield of the corresponding alkyne 3g was nevertheless isolated. Not surprisingly, the bulky t-Bu ketone substrate 1f only gave mono decarbonylation under the standard double decarbonylation conditions (vide supra). The alkyne substituents were also examined. Aryl, alkenyl and alkyl groups were all found suitable for this transformation. While the t-butyl-substituted alkyne 1n only offered the mono decarbonylation, the analogous OMe substituted substrate 1m and the estradiol derivative 1o smoothly provided the double decarbonylation products, suggesting an interesting electronic or coordination effect.

Scheme 3. Double Decarbonylation of Alkynyl Diones^a.

^a Reactions were run on a 0.10 mmol scale; all yields are isolated yields; 1f and 1n gave mono-decarbonylated products only.

Finally, encouraged by the success of C-C activation of α-diketones, we wondered whether this transformation could be generalized to α-keto esters. To our delight, subjecting the alkynyl α-keto ester 4 to our standard Xantphos conditions indeed gave the decarbonylated product, ynoate 5 in 58% yield (Scheme 4). To the best of our knowledge, this represents the first example of catalytic C-C activation of α-keto esters. Further investigation of this topic is ongoing.

Scheme 4.

Decarbonylation of an α–Keto Ester

In summary, an unusual reactivity of alkynyl α-diones is disclosed. Utilizing different rhodium-catalysis conditions, mono and double decarbonylative C-C bond formation can be realized in a controlled fashion. These reactions were operated under pH and redox neutral conditions. Both alkyl and aryl groups were tolerated, suggesting a promising substrate scope. In addition, an α-keto ester substrate also successfully underwent decarbonylation to give an ynoate. The new and fundamental reactivity discovered in this work should help to advance our understanding of the transition metal-mediated activation of unstrained C-C bonds, which is expected to open the door to new avenues for developing alkynyl transfer reactions.^5b Mechanistic studies and further expansion of the reaction scope are currently under investigation.