α-Perfluoroalkyl-β-alkynylation of alkenes via radical alkynyl migration† †Electronic supplementary information (ESI) available. See DOI: 10.1039/c7sc02175e

Transition metal-free radical α-perfluoroalkylation with concomitant β-alkynylation of unactivated alkenes is presented.


Introduction
Fluorinated compounds are highly abundant in agrochemical industry, medicinal chemistry and materials science. 1 The introduction of uorine atoms and of peruoroalkyl groups into organic compounds increases their solubility, lipophilicity, bioavailability and metabolic stability. 2,3 Therefore, it is important to develop novel synthetic methods for uorination or peruoroalkylation. Signicant progress has been made on alkene peruoroalkylation with concomitant b-functionalization in recent years. 4 In most cases, transition-metal catalysts such as Cu(I), 5 Ru(II), 6 Ir(III), 7 and Fe(II) 8 are required to run these reactions and peruoroalkylated products bearing alkyl, aryl, OR, Cl or CF 3 groups at the b-position can be obtained (Scheme 1). However, alkene peruoroalkylation with accompanying balkynylation is rare and only four papers were published along these lines. Guided by pioneering studies of the Fuchs group, 9 Li 10 and Yu 11 recently reported radical alkene tri-uoromethylation and subsequent intermolecular alkynylation using phenylethynyl p-tolyl sulfone and acetylenic trione as the alkynylating reagents. As drawbacks of these processes, alkynyl transfer is limited to phenylalkynylation and trimethylsilylalkynylation and the methods can only be used for tri-uoromethylation. During preparation of this manuscript, Zhu and co-workers reported an elegant photoredox catalyzed alkene a-triuoromethylation-b-alkynylation comprising an intramolecular alkynyl migration. 12 However, the scope of this process is limited to the triuoromethylation and the costly Umemoto reagent has to be applied along with a photoredox catalyst. Thus, novel and efficient general methods for alkene a-peruoroalkylation-b-alkynylation are highly demanded. Considering the economy, transition metal free processes are highly valuable.
Herein we disclose a simple and efficient novel method for the a-peruoroalkylation with concomitant b-alkynylation of unactivated alkenes proceeding via radical 1,4-or 1,5-alkynyl migration using peruoroalkyl iodides 13,14 as cheap and commercially available peruoroalkyl radical sources in combination with readily prepared tertiary propargylic alcohols as substrates (Scheme 1). Notably, the potential of radical alkene peruoroalkylation with concomitant b-functionalization via intramolecular formyl, 15a aryl, 15b and heteroaryl 4e migration has recently been documented. 15c,d

Results and discussion
Initial investigations were performed with 3-methyl-1phenylhept-6-en-1-yn-3-ol 1a and peruorobutyl iodide using Li-hexamethyldisilazide (LiHMDS) (1.2 equiv.) as a base in DME. The alcohol was rst deprotonated with LiHMDS at room temperature (0.5 h). Aer complete deprotonation, DABCO (1.5 equiv., to mediate chain initiation) 14 and peruorobutyl iodide 2a (1.8 equiv.) were added sequentially and the mixture was stirred under visible-light irradiation (using a Philips Master HPI-T Plus (400 W) bulb) at 50 C for 18 hours. To our delight, the target ketone 3a was obtained in 53% yield (Table 1, entry 1). The base heavily affects the reaction outcome and the yield dropped to 42% upon using LiOH (Table 1, entry 2) but a better result was achieved with NaOH (63%) ( Table 1, entry 3). Surprisingly, peruoroalkylation alkynyl migration failed using KOH as a base (Table 1, entry 4).
Having noted the importance of base, we started to screen mixtures of two bases. The addition of NaOH (0.42 equiv.) to LiHMDS did not affect the reaction outcome to a large extent ( Base was found to be important and the target 3a was formed in only 6% in the absence of any base (Table 1, entry 14). Thus, 1a (0.1 mmol), 2a (1.8 equiv.), LiHMDS (1.2 equiv.), LiOH (0.42 equiv.), and DABCO (1.5 equiv.) in 1.25 mL of DME with stirring under visible-light irradiation (using a Philips Master HPI-T Plus (400 W) bulb) at 50 C for 18 h were identied as the optimal reaction conditions for this sequence.
With the optimized reaction conditions in hand, we then turned to examine the scope with respect to the alkene component, keeping peruorobutyl iodide 2a as the C-radical precursor ( Table 2). The aryl group in these propargylic alcohols was rst varied by replacing the phenyl substituent in 1a with differently substituted aryl groups. Electronic effects at the para-position do not play an important role as good results were achieved for both electron-rich and also electron-poor systems (3b-f, Table 2, entries 1-5). A similar result was noted for the meta-methyl substituted alcohol 1g to provide 3g, and also its ortho-congener 1g gave the target ketone 3h in a good yield (Table 2, entries 6 and 7). Thus, steric effects at the aryl moiety are not important, a fact which is further supported by the successful transformations of ortho, ortho 0 -disubstituted aryl alkynes 1i and 1j (see 3i, j, Table 2, entries 8 and 9). Notably, the 1-naphthyl and 2-pyridyl groups are both tolerated as Rsubstituents, and the corresponding ketones 3k and 3l were obtained in good yields (Table 2, entries 10 and 11).
However, replacing the phenyl group in 1a by a 1-cyclohexenyl (1m) or n-butyl substituent (1n) led to signicantly reduced yields of the corresponding alkynylated ketones 3m and 3n (Table 2, entries 12 and 13). Likely the rate constant for the radical cyclization is smaller for these less activated alkynes. A good result was also achieved in the transformation of the silylated alkyne 1o to provide 3o in 82% isolated yield.  We continued the studies by replacing the methyl group in 1a with other alkyl and aryl substituents in combination with 2a as the C-radical precursor (Scheme 2). The 1-butyl and 1-iisopropyl propargylic alcohols 1p and 1q worked well and the dialkyl ketones 3p and 3q were isolated in 64% and 62% yield, respectively. An excellent yield was also obtained for the reaction with the tertiary benzylic alcohol 1r to give 3r (88%). Pleasingly, quaternary centres can be built up via the alkynyl migration as documented by the successful preparation of 3s (73%). The trisubstituted alkene 1t also acted as an acceptor and surprisingly an acceptable yield of the alkynyl migration product 3t derived from initial addition at the more hindered alkene position was obtained. The geminally dimethylated alcohol 1u provided the ketone 3u in 57% yield. Notably, the peruoroalkyl group in these product ketones is readily varied by switching the radical precursor (see 3v-y). ICF 2 CF 2 Cl reacted chemoselectively at the C-I bond to give 3z and the corresponding product derived from C-Cl cleavage was not identi-ed. We also tested other alkyl iodides such as cyclohexyl iodide and adamantyl iodide that are typical substrates in radical Iatom transfer reactions. As expected, alkylation/alkynyl migration did not work due to the fact that such non-uorinated alkyl iodides cannot be readily SET-reduced. In addition, we tested abromo acetophenone but could not identify the target product, and large amounts of starting material remained unreacted. 16 We are currently not sure whether the problem lies in the initiation step or in the intermolecular SET transfer from the ketyl radical anion to the bromo ketone (see mechanistic discussion below). Importantly, the novel sequence is not restricted to 1,4-alkynyl migration and reaction of alcohol 1v with peruorobutyl iodide gave the ketone 3va resulting from a 1,5-alkynyl migration in 66% yield.
To test whether alkynyl migration proceeds stereospecically, we repeated the reaction with enantioenriched propargylic alcohol 1r (25% ee) as the starting material in combination with peruorobutyl iodide and obtained the desired product 3r in 82% yield and 7% ee showing that only very low stereospecicity is achieved for this sequence (for details, see the ESI †).
The proposed mechanism is depicted in Scheme 3. Initiation occurs by irradiation of the peruoroalkyl iodide in the presence of DABCO 11 to give the corresponding peruoroalkyl radical. This C-radical then adds to the terminal position of the alkene in the deprotonated alcoholate 1-Li to give the alkyl radical A. 5-or 6-exo radical cyclization leads to vinyl radical B, which further reacts by regioselective b-C-C bond cleavage to generate the ketyl radical anion C. As recently shown by us, such ketyl radical anions are good SET reducing reagents. 17 Hence, electron transfer from C to the starting peruoroalkyl iodide gives the product ketone 3 along with the corresponding per-uoroalkyl radical thereby sustaining the chain, and the overall cascade therefore belongs to an electron-catalyzed process. 18 We also investigated the follow-up chemistry using 3a as the starting material (Scheme 4). Chemoselective reduction of the carbonyl group in 3a was achieved using LiAlH 4 at 80 C to provide alcohol 4 in 76% yield as a 1 : 1 diastereoisomeric mixture. The triple bond in 3a was chemoselectively hydrogenated with Pd/C and H 2 to form ketone 5 in 80% yield. TfOHcatalyzed cyclization of 3a afforded the a,b-unsaturated ketone 6 in 76% isolated yield. 19 The same product could also be obtained using Au-catalysis, albeit in a slightly lower yield. 20 Scheme 2 Variation of the radical acceptor and the perfluoroalkyl iodides. a 3.6 equivalents of CF 3 I were added in two parts and the reaction time was 24 hours.

Conclusions
In summary, we have developed a novel and efficient method for the transition metal-free a-peruoroalkylation b-alkynylation of unactivated alkenes via radical 1,4-or 1,5-alkynyl migration. Reaction conditions are mild and the substrate scope is broad. The starting propargylic alcohols are readily prepared and commercially available peruoroalkyl iodides are used as C-radical precursors.

Conflicts of interest
There are no conicts of interest to declare.