Aza-Wol ﬀ rearrangement of N - ﬂ uoroalkyl triazoles to ketenimines †

N -Fluoroalkylated 1,2,3-triazoles underwent a microwave-heating-assisted ring opening, nitrogen molecule elimination and concomitant group rearrangement to form isolable N - ﬂ uoroalkylketenimines. This reagent-free process is characterized by a wide scope and high e ﬃ ciency and provides a new route to unexplored N - ﬂ uoroalkyl compounds. The reaction mechanism was investigated by a combination of mechanistic and computational studies. [2 + 2] cycloaddition of ketenimines with alkynes or alkenes a ﬀ orded novel cyclobutenimines and cyclobutanimines, respectively. Addition of oxygen, sulfur-and nitrogen nucleophiles to ketenimines gave new N - ﬂ uoroalkyl imidates, thioimidates and amidines.


Introduction
The Wolff rearrangement of α-diazo carbonyl compounds to ketenes is one of the most significant transformations that has contributed to the growth of chemical synthesis in the 20th century (Scheme 1A). 1 The ketenes generated in this way are highly useful substrates with a wide range of synthetic applications. 2etenimines, aza-analogues of ketenes, are a related class of compounds with cumulated double bonds and a broad range of applications in organic chemistry. 3Although they are generally more stable than ketenes, most ketenimines are not isolable materials and therefore need to be formed in situ as reactive intermediates.For example, highly reactive N-sulfonylketenimines are generated by copper-catalyzed azide alkyne cycloaddition (CuAAC) of N-sulfonyl azides and alkynes and subsequent spontaneous denitrogenative rearrangement of a copper-triazole intermediate (Scheme 1B, left).They can serve as three atom (C-C-N) synthons in the syntheses of amides, [4][5][6] amidines, 7 imidates, 8 four-, five-, six-and sevenmembered heterocycles, amino acids, and many other biologically and pharmaceutically valuable compounds (see Scheme 1B, right for selected examples). 9he formation of ketenimines by CuAAC is limited to electron-deficient sulfonyl/phosphoryl azides; alkyl or aryl azides pre-ferably form triazole rings. 10Even N-fluoroalkyl azides with a strong electron-withdrawing N-CF 2 R groups react with terminal alkynes in a CuAAC fashion to form N-fluoroalkyl triazoles, [11][12][13][14] substrates which have been showcased as being useful for the synthesis of N-fluoroalkyl azoles (e.g.][21][22] The incorporation of fluorinated groups into small molecules is a widely used approach to the modification of their pharmacological as well as pharmacokinetic properties. 23hereas C-, and O-fluoroalkyl compounds have been studied extensively, N-fluoroalkyl substrates have received more attention only recently. 24That their potential is unexplored is caused mainly by the lack of efficient methods to synthesize these compounds using readily available starting materials.Introducing a CF 3 group directly into nitrogen functionalities is a greatly challenging and substrate-specific process. 25owever, an approach based on N-CF 3 synthons has recently emerged as an alternative and atom-economical route.The Schoenebeck group introduced versatile N-CF 3 carbamoyl fluorides as building blocks for the synthesis of highly attractive tertiary N-CF 3 amides, 26 hydrazines, 27 ureas, 28,29 formamides 30 and other compounds. 31Synthesis of tertiary N-CF 3 amides was also recently reported by Toste and Wilson from carboxylic acid derivatives and isothiocyanates in the presence of AgF. 32nother synthon, the N-CF 3 nitrilium ion, was introduced by Xu and Wang for the synthesis of N-CF 3 azoles and imido derivatives. 33,34Despite the great progress in recent years, most of the strategies for this are limited to N-CF 3 compounds and methods for the synthesis of a wide range of N-fluoroalkyl compounds remain underdeveloped.
Here, we report a highly efficient, atom-economical, waste-

Results and discussion
While exploring the thermal stability of N-pentafluoroethyl-4phenyl-1,2,3-triazole (1a) by differential scanning calorimetry (DSC), thermogravimetric analysis (TG) and gas chromatography with mass detection (GC/MS), we noticed decomposition at temperatures of around 140-160 °C in solution or without solvent and also decomposition in the gas phase (GC inlet temp.: 250 °C) with a mass loss of m/z 28.Detailed NMR and IR analyses of preparative reactions revealed the quantitative formation of N-( pentafluoroethyl)-2-phenylethen-1-imine (2a), which was fully characterized (Scheme 2, see the ESI †).RFA analysis ruled out any possible participation of the copper catalyst carried over from the synthesis of 1a with CuAAC.Various solvents were tested including chloroform, toluene, THF, acetone and polar protic solvents, but the best results were obtained using DCE (see the ESI † for details).
Conventional heating is also possible but leads to 2a in a longer reaction time and unidentified side products.
Subsequent investigation revealed that the denitrogenative rearrangement of N-fluoroalkyl-1,2,3-triazoles 1 to N-fluoroalkylketenimines 2 is of a wide scope and, in a vast majority of cases, highly efficient (Table 1).
Triazoles with various combinations of aryl, substituted aryl, heteroaryl, alkyl, substituted alkyl or cycloalkyl groups in position four, hydrogen, Cl, Br, I or an alkyl group in position five and trifluoromethyl (-CF 3 ), pentafluoroethyl (-CF 2 CF 3 ), bromotetrafluoroethyl (-CF 2 CF 2 Br), substituted tetrafluoroethyl (-CF 2 CF 2 R) and even difluoromethyl groups (-CF 2 H) on the nitrogen atom all underwent the reaction and provided products 2 in high to excellent yields.Despite the robustness of the method, certain limitations were observed.Triazole 1n underwent rearrangement; however, the product was not stable (see the ESI † for details).The presence of a 5-halo substituent (iodo in particular) in triazoles 1 improved isolated yields for difficult substrates (4-alkyl, 4-cycloalkyl) compared to unsubstituted examples.This effect is strikingly strong in the case of N-difluoromethyl substrates, where the iodo product formed in high yield (2ak) and the unsubstituted one did not form at all (2am).N-Alkyl triazole (1an) was not a competent substrate in this reaction.Generally, ketenimines 2 did not require any purification; solvent removal was sufficient to obtain suitably pure samples.Their purification using silica gel column chromatography is also possible.Ketenimines 2 are mostly liquids (2i and 2v are solids; see the crystal structure in Table 1) with limited air and moisture stability as pure substances; in solution (DCE, pentane) they were stable for weeks at room temperature in an inert atmosphere.
Having demonstrated the broad scope of the reaction, we investigated the mechanism of this unprecedented transformation.First, we synthesized isotopically labelled triazoles 3-5 from 13 C-phenylacetylene. 35All three triazoles underwent ring opening and rearrangement to form ketenimines 6-8, which contained the label exclusively in the central sp-hybridized carbon atom, proving that the reaction proceeds via a [1,2]shift of the R group, similarly to the Wolff rearrangement (Scheme 3).
As the Wolff rearrangement can proceed via a concerted or step-wise mechanism involving carbenes, 36 we performed ab initio calculations.We optimized the structures with the multi-reference CASSCF(4,4)/6-31+g* method, as it was able to localize open-shell systems such as carbene and nitrene (unlike, e.g., density functional approaches).Single point energies were calculated using the coupled cluster CCSD(T) method with the aug-cc-pVDZ basis set.The differences in Scheme 2 Synthesis of ketenimine 2a by the thermal decomposition of triazole 1a.thermal corrections were found to be on the order of 10 −4 eV (at DFT/BMK/6-31+g* level), so entropic effects do not play any role here.The effect of the solvent was also found to be negligible.The lowest energy pathway involved the expected diazo intermediate and final ketenimine product (Fig. 1A, blue box).This result is consistent with a concerted Wolff rearrangement, which also requires an s-cis configuration of the diazo and heteroatom groups.
Carbene and its hypothetical rearrangement products such as 1H-azirine, and nitrene species were found to be very unstable (see the ESI † for more details), too high in energy and are not involved in the reaction mechanism (Fig. 1B).
Next, the synthetic utility of the prepared ketenimines in [2 + 2] cycloadditions and in additions of nucleophiles was explored.Thermal cycloaddition of phenylacetylene proceeded with expected regioselectivity and produced cyclobutenimines 9 in good yields as E/Z mixtures with moderate to high stereoselectivities towards E-isomers (Scheme 4).Ketenimines with halogen substituents (R = Cl, Br, I) were unreactive, but the alkyl group (R = n-Pr) was tolerated well.Internal alkyne diphenylacetylene proved to be a competent substrate in cycloaddition under microwave heating affording cyclobutenimines 10 in good yields and exclusively as E-isomers.The crystal structure of cyclobutenimine (E)-10b is presented in Scheme 4. On the other hand, the electron acceptor terminal or internal alkynes (ethyl propiolate, diethyl acetylenedicarboxylate) exhibited no reactivity even under heating.
Scheme 3 Synthesis of 13 C-labelled triazoles 3-5 and ketenimines 6-8.stilbene), were unlike internal alkynes unreactive, even upon heating.The E/Z selectivity is governed by the sterical factors.Cyclobutane and cyclobutene imines are rare compounds and their N-fluoroalkyl derivatives were never synthesized before.
Finally, the reactivity of ketenimines 2 with N-, O-and S-centred nucleophiles was investigated (Scheme 5).Fast reactions were observed with an excess of ethanol and ethanethiol at room temperature to afford iminoesters 12 and thioimidates 13.In the cases of 12b and 13b, dehydrofluorinated sideproduct 12b′ and 13b′ formed in a small amount, and instead of thioimidate 13c, imidoyl fluoride 13c′ was isolated.All imino products formed in E-configuration.With primary amines, the addition took place, but the products decomposed and a complex reaction mixture was formed; however, quantitative formation of amidines 14 was observed using secondary amines as nucleophiles.Stable products of hydrolysis N-trifluoroacetylamidines 15 were obtained from N-pentafluoroethyl derivatives, using column chromatography.The crystal structure of amidine 15f was determined (Scheme 5).Very recently, related N-CF 3 amidines, imidates and thioimidates were prepared by the reaction of PhI(CF 3 )Cl with nitriles, followed by the addition of nucleophiles to the intermediate N-CF 3 nitriliums. 33

Conclusions
Structurally diverse N-( per)fluoroalkyl-1,2,3-triazoles rearranged under microwave heating to novel and surprisingly stable N-(per)fluoroalkyl ketenimines, which were isolated and fully characterized by spectroscopic methods and in one case also by X-ray diffraction.The mechanism, supported by control experiments, an isotopic labelling study and ab initio calculations, proceeds via triazole ring opening, nitrogen elimination and group rearrangement (1,2-shift).Ketenimines reacted with alkynes or alkenes in [2 + 2] cycloaddition reactions under mild conditions to yield novel cyclobutenimines or cyclobutanimines, respectively.The addition of alcohols, thiols, or secondary amine nucleophiles to ketenimines afforded imidates, thioimidates, or amidines, respectively.Thus, N-fluoroalkyl-1,2,3-triazoles have been demonstrated to be useful precursors of N-fluoroalkylated ketenimines, which served as valuable building blocks in the expansion of the chemical space of accessible N-fluoroalkyl compounds.Further transformations of these new N-fluoroalkyl building blocks are currently being investigated in our laboratory.