The Three-Component Synthesis of 4-Sulfonyl-1,2,3-triazoles via a Sequential Aerobic Copper-Catalyzed Sulfonylation and Dimroth Cyclization

4-Sulfonyl-1,2,3-triazole scaffolds possess promising bioactivities and applications as anion binders. However, these structures remain relatively unexplored and efficient synthetic procedures for their synthesis remain desirable. A practical room-temperature, aerobic copper-catalyzed three-component reaction of aromatic ketones, sodium sulfinates, and azides is reported. This procedure allows for facile access to 4-sulfonyl-1,5-disubstituted-1,2,3-triazoles in yields ranging from 34 to 89%. The reaction proceeds via a sequential aerobic copper(II)chloride-catalyzed oxidative sulfonylation and the Dimroth azide–enolate cycloaddition.


Scheme 1. Synthetic routes towards 4-sulfonyl-triazoles.
In light of the above-mentioned limitations, we envisioned a multicomponent 4-sulfonyl-1,2,3-triazole synthesis that starts from readily available aromatic ketones, sodium sulfinates, and organic azides, via a one-pot oxidative sulfonylation and Dimroth cyclization sequence, as shown in Scheme 1. In order to increase the sustainability of the envisioned procedure, we wanted to avoid the use of stoichiometric oxidants such as hypervalent iodine [25], peroxides [41], copper salts [42,43], and silver salts [44,45] that are commonly used in oxidative coupling reactions, in favor of air as the terminal oxidant, which is environmentally benign and results only in formation of water as a side product [43,46,47]. Over the past decade, many catalysts have been developed that allow for aerobic/O 2 oxidative coupling of diverse substrates, with the significant focus being on palladium [48][49][50][51] and other noble metal catalysts [43,50,[52][53][54][55]. However, copper catalysts are particularly interesting when compared to noble metal catalysts. Next to generally being inexpensive, readily available, and of low toxicity, they have great potential for catalyzing a broad range of reactions. Catalysis by natural copper-oxidases can serve as an example [46]. This is epitomized by the wide array of aerobic copper-catalyzed oxidative coupling reactions that have been reported to date [46,56,57].

Results and Discussion
As starting point for our three-component aerobic oxidative sulfonylation/Dimroth sequence, we utilized the reaction conditions reported by Lan et al. for the CuBr 2 -catalysed synthesis of α-alkyl ketosulfones, as shown in Table 1 and Table S1 [26]. While the original ligand-free procedure furnished the ketosulfones from α-unsubstituted acetophenone only in a low yield of 20%, it was expected that subsequent transformation of the in situ formed ketosulfone into the corresponding 1,2,3-triazole would result in a significantly improved yield. Pleasingly, when applying these reaction conditions, the corresponding 4-tosyl triazole 4a was obtained in 54% yield (Entry 1). Different copper salts, including CuI, Cu(OAc) 2 , Cu(OTf) 2 , and CuCl 2 were screened (Entries 1-5), and CuCl 2 was found to be superior, furnishing 4a in an isolated yield of 71%, and 72% upon repetition of the experiment (Entry 2). Next, different bases and solvents were evaluated, as well as a reduction in the equivalents of base (Entries 6-15, and Tables S16-S18). However, any deviation from the standard conditions (Entry 2) resulted in a decreased yield. The amine base may perform a dual role, acting both as base and ligand. Pyridine and Et 3 N (Entry 6 and 7) presumably are unreactive because they are both weaker bases and ligands than DBU, thereby not sufficiently deprotonating the enol and stabilizing the copper complex. Surprisingly, DBN results in a markedly lower yield than DBU (Entry 8), 52% versus 72%, which may be the result of it being too strongly coordinated to the copper and thereby hindering the formation of the copper-enolate. K 2 CO 3 and KOtBu presumably are unreactive since they are both less soluble in DMSO and weaker ligands (Entry 9 and 10). Presumably, DMSO is superior since it is both a polar coordinating solvent as well as a mild oxidant. Subsequently, different tertiary amine ligands were screened, including TMEDA, 2,2bipyridine, 1,10-phenanthroline, and neocuproine (Entries [16][17][18][19]. Primary and secondary amine ligands were excluded due to the risk of α-amination [58], and phosphine ligands were excluded due to their susceptibility to oxidative degradation [51]. Out of the evaluated amine ligands, TMEDA proved to be the superior with a yield of 81% (Entry 16). Finally, the catalyst loading and solvent volume were varied and an optimal loading of 10 mol% CuCl 2 /TMEDA and 2 mL volume of DMSO was found, resulting in a yield of 89% (Entry 28). As final control experiments, reactions were set up under the optimized conditions in either absence of CuCl 2 (Entry 29) or under argon atmosphere (Entry 30), which resulted in no reaction and in less than 5% of triazole being formed.
With the optimized conditions in hand, we set out to explore the substrate scope for this reaction and investigated various acetophenone derivatives, sodium sulfinates, and organic azides, as shown in Scheme 2. For the reaction of acetophenone derivatives 1a-j with sodium p-toluene sulfinate 2a and phenyl azide 3a, the yields of triazoles 4a-j varied from high to low. The reactions of electron-deficient acetophenones progressed at similar or faster rates yet resulted in lower yields. The 4-tosyl triazoles derived from p-trifluoromethyl substituted 4b, p-fluoro substituted 4c, and o-bromo substituted 4d acetophenone were obtained in yields of 55 (4b), 73 (4c), and 64% (4d). The reduced yields compared to 4a can be in part explained by the occurrence of Regitz diazo transfer, as observed previously for cyclic sulfonyl ketones [24] and evidenced by the observation of nitrogen evolution from the reaction mixture and the presence of a minor quantity of aniline in the crude mixture, as determined by GC/MS. However, no other product from this side-reaction could be isolated. The presence of donating substituents resulted in a reduced reaction rate and concomitantly reduced yields. The p-methyl and p-methoxy substituted triazoles 4e and 4f were obtained in yields of 51% and 34% after 48 h. Next, the influence of steric hindrance was investigated and the expected negative correlation between steric hindrance and product yields was observed. The o-methyl and 1-naphthyl triazoles 4g and 4h give reduced yields in comparison to their less sterically hindered counterparts, the p-methyl and 2-naphthyl triazoles 4e and 4i, 36% (4g) versus 51% (4e) and 60% (4h) versus 71% (4i). Several aliphatic ketones including acetone, pinacolone, and trifluoromethyl acetone were screened without success, as expected. A plausible reason for the limitations of the reaction scope for ketones is given in the final paragraph, after discussing the reaction mechanism.
Pleasingly, 4-methyl sulfinate triazole 4j was furnished in a good yield of 85%. However, in the case of the p-chlorophenyl sulfonyl triazole 4k, a yield of only 51% was obtained. Finally, the scope in azide was evaluated and both for phenyl azides with electron-withdrawing and donating groups, good yields of 74% (4l) and 64% (4m) were obtained. Regrettably, sterically hindered azides were unreactive and no triazole 4n formed. Benzyl azide performed far less effectively and furnished the triazole 4o in 18% yield, along with 45% of sulfonyl ketone 5a.
A plausible reason for the limitations of the reaction scope for azides is given in the final paragraph, after discussing the reaction mechanism.
In order to gain more insight into the mechanism, several control experiments were performed. In the absence of CuCl 2 no reaction occurred, and under argon atmosphere less than 5% of product formed, as shown in Table S1 (Entry 32-33), which shows that the copper salt is required for catalyzing the reaction and oxygen is needed for catalytic turnover. Addition of four equivalents of TEMPO completely inhibited the reaction, as shown in Scheme 3a. From the control reaction in the presence of 1,1-diphenylethylene (DPE) 6, a radical trapping reagent, the sulfonyltriazole 4a, could be isolated in a reduced yield of 39% and the yield of the radical trapping product 7 was rather low, 4%, as shown in Scheme 3b. These results indicate involvement of both free radicals and alternative mechanisms, such as via oxidative chlorination or via an organometallic intermediate. Phenacyl chloride 8 is a possible reaction intermediate, considering that the halogenation of the carbonyl α-position by stoichiometric copper halides, as well as diverse aerobic copper-catalyzed oxidative halogenations, have been reported [46,59,60]. From the reaction of 8 under standard conditions in presence and absence of the copper catalyst, triazole 4a was obtained in yields of 11% and 45%, as shown in Scheme 4c. This indicates that 8 may be an intermediate in the oxidative sulfonylation, although, it shows that there must be alternative operative pathways. However, it should be noted that 8 was not isolated from or observed in the reaction mixture. Additionally, the reported copper-catalyzed oxidative halogenation reactions generally take place under (Lewis) acidic conditions, which raise the Cu(II) reduction potential and consequently promote single-electron transfer SET reactivity.
Conversely, basic conditions and stronger ligands stabilize both Cu(II) and Cu(III), or lower the Cu(II) reduction potential, which reduces SET reactivity and favors the formation of organometallic intermediates [46,61]. The intermediacy of the sulfonyl ketone 5 is supported by the observation of the NMR characteristic peak in the crude reaction mixture at reduced reaction times. When the oxidative sulfonylation was performed in absence of azide good yields of phenacyl sulfones 5a and 5b were obtained, as shown in Scheme 3d, which proves its intermediacy. Advantageously, this shows that the presented methodology can also be used for the synthesis of sulfonyl ketones in good yield.
The reason why the reaction scope is limited to aromatic ketones can be explained by the reaction mechanisms shown in Scheme 4, in which the first step involves the formation of a (copper-) enolate. The observation that the reaction times increase for more electron-rich aromatic ketones indicates that this deprotonation may be the rate limiting step. This fact would explain why acetone and pinacolone are unreactive since these are less acidic than acetophenone by at least 1.8 orders of magnitude. On the other hand, 1,1,1,-trifluoroacetone, while more acidic, may be unreactive due to its high tendency to form hydrates with water [68]. The reason for the reduced yields of 1,2,3-triazole 4o from the reaction with benzyl azide are intrinsic to the Dimroth azide-enolate cycloaddition, which was defined by L'abbé in 1971 as "the condensation of organic azides with active methylene compounds in the presence of an equimolar amount of organic or inorganic base leading to highly substituted 1,2,3-triazoles in a regioselective manner" [69,70]. The reaction mechanism involves either a stepwise concerted [3+2] cycloaddition or a stepwise addition of the enolate to the azide with formation of an N1-triazenyl ion intermediate. The rate of this cycloaddition depends on the stabilization of this ion [71]. For this reason, aromatic azides with electron-donating groups are expected to react more slowly than those with withdrawing substituents, and benzyl azides and aliphatic azides are expected to react more slowly than their aromatic counterparts.

Conclusions
In conclusion, we have developed an efficient Cu-catalyzed three-component synthesis of 4-sulfonyl triazoles from aromatic ketones, azides, and sodium sulfinates, with air oxygen as terminal oxidant, operating at room temperature. This reaction involves a sequential oxidative sulfonylation of aromatic ketones/Dimroth cyclization. Preliminary mechanistic investigations indicate that both sulfonyl free radicals and organometallic Cu(III)-intermediates are involved.