Silyl Tosylate Precursors to Cyclohexyne, 1,2-Cyclohexadiene, and 1,2-Cycloheptadiene

Abstract

Transient strained cyclic intermediates have become valuable intermediates in modern synthetic chemistry. Although silyl triflate precursors to strained intermediates are most often employed, the instability of some silyl triflates warrants the development of alternative precursors. We report the syntheses of silyl tosylate precursors to cyclohexyne, 1,2-cyclohexadiene, and 1,2-cycloheptadiene. The resultant strained intermediates undergo trapping in situ to give cycloaddition products. Additionally, the results of competition experiments between silyl triflates and silyl tosylates are reported.

Graphical Abstract

The chemistry of transient strained cyclic intermediates has been a popular topic of study for over a century.¹ Early efforts in the field established the existence of benzyne (1),² cyclohexyne (2),³ and 1,2-cyclohexadiene (3)⁴ through pioneering studies conducted by Roberts and Wittig in the 1950s and 1960s (Figure 1). Since their discovery, these species, along with their heterocyclic derivatives (e.g., 4), have been employed in a host of synthetic applications spanning natural product synthesis,^1c,d,i,5 heterocycle construction,^1e,g,h,6 and materials chemistry,^1e,7 as exemplified by the syntheses of 5–7.

Figure 1.

Strained cyclic intermediates and selected synthetic applications.

Many of the advanced synthetic applications of strained cyclic intermediates have been enabled by the use of silyl triflates as precursors. Initially developed by Kobayashi as precursors to benzyne (1),⁸ silyl triflates have since become the most commonly employed precursors for accessing arynes, nonaromatic cyclic alkynes, and cyclic allenes.^1,9,10 However, we have encountered difficulties in preparing certain functionalized strained cyclic allene and alkyne precursors due to the instability of the corresponding silyl triflates. This instability can be attributed to the ease of triflate dissociation and cation formation in related systems.^11,12

With the aim of circumventing silyl triflate instability and accessing a wider range of strained intermediates under Kobayashi-type conditions, we sought to develop new precursors to cyclic alkynes and allenes (i.e., 2 and 3, Figure 2). As mentioned above, the most common means to access 2 and 3 is via the corresponding silyl triflates (e.g., 8 and 10, respectively) using fluoride-induced elimination. Encouraged by the success of silyl tosylates as aryne precursors,¹³ we sought to develop silyl tosylate cyclic alkyne and allene precursors 9 and 11, respectively.¹⁴ We hypothesized that the diminished leaving group ability of a tosylate anion relative to a triflate¹⁵ could alleviate difficulties associated with vinyl triflate instability, while retaining sufficient reactivity to form the desired strained intermediates.¹³ These alternative precursors could also allow for new synthetic methods that leverage the differences in reactivity between tosylates and triflates.¹⁶ Furthermore, we hoped that silyl tosylates 2 and 3 would be crystalline,¹⁷ in contrast to silyl triflates, which are often oils. This characteristic could facilitate their purification and use in process chemistry. Herein, we describe the preparation, validation, and synthetic application of the desired silyl tosylates as strained intermediate precursors.

Figure 2.

Silyl triflate (previous) and silyl tosylate (current) precursors to 2 and 3.

Our first objective was to develop synthetic routes to the requisite silyl tosylates, which led to the preparation of 9 and 11 as shown in Figure 3. Following literature procedures,¹⁸ cyclohexanone (12) was converted to silyl ketone 14 in good yield through a sequence involving silyl enol ether formation (12 → 13) and allylic deprotonation/retro-Brook rearrangement (13 → 14). From common intermediate 14, deprotonation under either thermodynamic or kinetic control, followed by quenching with p-toluenesulfonic anhydride, generated silyl tosylates 9 and 11, respectively. Gratifyingly, both alkyne precursor 9 and allene precursor 11 were obtained as crystalline solids and could be prepared on gram-scale.

Figure 3.

Syntheses of silyl tosylates 9 and 11.

As shown in Table 1, we found that silyl tosylate 9 serves as a viable precursor to generate cyclohexyne (2) with in situ trapping. It should be noted that our initial attempts to generate 2 from 9 using CsF (based on literature conditions for the corresponding trimethylsilyl triflate^6b) led to the recovery of unreacted 9. However, the use of the more soluble fluoride source tetrabutylammonium fluoride (TBAF) proved fruitful and enabled the generation of 2 in situ. Trapping with diene 16 furnished oxabicycle 17 via a (4 + 2) cycloaddition (entry 1). Similarly, the use of nitrone 18 as the trapping agent gave rise to isoxazolidine 19 by way of a (3 + 2) cycloaddition (entry 2). In both cases, yields were comparable to those observed using a silyl triflate precursor.^6b,18b Lastly, nucleophilic trapping of 2 with imidazole (20) proved successful, generating vinyl imidazole 21 in moderate yield.¹⁹ These trapping experiments demonstrate that silyl tosylate 9 serves as an effective precursor to 2, rendering 9 a useful intermediate for the synthesis of heterocyclic products.

Table 1.

Silyl Tosylate 9 as a Precursor to Cyclohexyne (2)^a

entry	trapping agent	product	yieldb (lit. yield from 8)c
1			76% (78%)
2			69% (61%)
3			53% (81%)

^aGeneral conditions: silyl tosylate 9 (1.0 equiv, 0.14 mmol), trapping agent (1.5–3.0 equiv), TBAF (5.0 equiv), and THF (0.07 M) heated in a sealed vial under an atmosphere of N₂.

^bYields reflect the average of two isolation experiments.

^cLiterature isolated yields under comparable reaction conditions when using 8 (R = Me or Et).

We also investigated silyl tosylate 11 as a precursor to strained cyclic allene 3 (Table 2). In contrast to our observations in reactions of alkyne precursor 9, CsF could be utilized to induce strained intermediate formation from silyl tosylate 11 under the same conditions reported in the literature for the corresponding silyl triflate 10 (R = Et).^6a We were delighted to find that silyl tosylate 11 could be employed in (4 + 2), (3 + 2), and (2 + 2) cycloadditions to deliver 23, 24, and 26, respectively (entries 1–3). In all cases, yields and diastereomeric ratios were consistent with those reported in the literature for reactions employing silyl triflate 10 (R = Et).^6a,18b,20 The synthesis of isoxazolidine 24 was also carried out on mmol-scale to demonstrate scalability of the reaction.

Table 2.

Silyl Tosylate 11 as a Precursor to 1,2-Cyclohexadiene (3)^a

^aGeneral conditions: silyl tosylate 11 (1.0 equiv, 0.14 mmol), trapping agent (1.0–5.0 equiv), CsF (5.0 equiv), and MeCN (0.1 M) heated in a sealed vial under an atmosphere of N₂.

^bYields reflect the average of two isolation experiments.

^cLiterature isolated yields and diastereomeric ratios under comparable reaction conditions when using 10 (R = Et).

^dDiastereomeric ratios determined by ¹H NMR analysis of the crude reaction mixture.

^eYield determined by ¹H NMR analysis using an external standard.

^fCyclic allene generated from 6,6-dibromobicyclo[3.1.0]hexane.

Having established silyl tosylates as effective substitutes for silyl triflates, we sought to extend this alternative method of strained intermediate generation to address a particular shortcoming in silyl triflate chemistry. As mentioned earlier, silyl triflates can sometimes be unstable due to their pronounced leaving group ability.^11,12 We have observed this type of instability when attempting to synthesize a silyl triflate precursor to 1,2-cycloheptadiene (29) (Figure 4).²¹ Alter natively, silyl tosylate 28, accessible in three steps from 27 (see the Supporting Information for details), could be obtained as a crystalline solid. Treatment of 28 with isobenzofuran 16 under standard conditions for cyclic allene generation and trapping afforded oxabicycle 30 in excellent yield via the intermediacy of cyclic allene 29. This example demonstrates that silyl tosylates can be used to expand the scope of strained intermediates accessible under mild fluoride-based conditions.²²

Figure 4.

Silyl tosylate 28 to access 1,2-cycloheptadiene (29).

Finally, two key experiments were performed to compare the relative reactivity of our silyl tosylates to the corresponding silyl triflates (Figure 5). In the first, equimolar amounts of silyl triflate 8a and silyl tosylate 9, both precursors to cyclohexyne (2), were treated with nitrone 18 under CsF-based reaction conditions. We observed that silyl triflate 8a reacted selectively over silyl tosylate 9 to generate cycloadduct 19. Silyl tosylate 9 did not react under these conditions. An analogous competition experiment was performed using silyl triflate 10a and silyl tosylate 11, both of which serve as precursors to 1,2-cyclohexadiene (3). This led to the efficient formation of 24 and the nearly quantitative retention of silyl tosylate 11. The preferential reactivity of the silyl triflate in both cases can be rationalized based on the relative leaving group abilities of the triflate and tosylate anions.¹⁵ This observed selectivity should prove useful in synthetic applications, analogous to prior studies in which multiple strained intermediates have been generated sequentially to synthesize complex polycyclic products.⁷

Figure 5.

Competition experiments between silyl triflate and silyl tosylate strained intermediate precursors. Yields determined by ¹H NMR analysis with external standard.

In summary, we have developed scalable syntheses of silyl tosylate precursors to the transient strained intermediates cyclohexyne (2), 1,2-cyclohexadiene (3), and 1,2-cycloheptadiene (29). Our synthetic routes to these precursors generate crystalline silyl tosylates, an attribute that could prove useful to process chemists. The silyl tosylate strained intermediate precursors not only replicate the chemistry attained using silyl triflates but also can allow access to strained intermediates inaccessible using known silyl triflate chemistry, as exemplified by silyl tosylate 28. Furthermore, competition experiments demonstrate that silyl triflate precursors to 2 and 3 react chemoselectively in the presence of their silyl tosylate counterparts. This selectivity should prove useful in synthetic design. Collectively, these studies demonstrate the synthetic utility of silyl tosylates as precursors to transient strained intermediates.