Electrochemical Vicinal Difluorination of Alkenes: Scalable and Amenable to Electron‐Rich Substrates†

Abstract Fluorinated alkyl groups are important motifs in bioactive compounds, positively influencing pharmacokinetics, potency and conformation. The oxidative difluorination of alkenes represents an important strategy for their preparation, yet current methods are limited in their alkene‐types and tolerance of electron‐rich, readily oxidized functionalities, as well as in their safety and scalability. Herein, we report a method for the difluorination of a number of unactivated alkene‐types that is tolerant of electron‐rich functionality, giving products that are otherwise unattainable. Key to success is the electrochemical generation of a hypervalent iodine mediator using an “ex‐cell” approach, which avoids oxidative substrate decomposition. The more sustainable conditions give good to excellent yields in up to decagram scales.

The inclusion of fluorine in bioactive compounds is becoming more important: [1] since 2006, prevalence has increased from 6% to 31 %i nt he top 100 best-selling small-molecule drugs. [2] Fluorinating alkyl groups can increase potency, lipophilicity and metabolic stability, [3] while reducing the basicity of neighbouring amines, [4] all of which can improve bioactivity and pharmacokinetics.T he vicinal difluoroethane unit has attracted recent attention, particularly as abioisostere of ethyl or trifluoromethyl groups [5] (Figure 1A)a nd for its unique propensity to adopt ag auche conformation in solution. [6] Exploiting this stereo-electronic effect is an emerging strategy for molecular design, [7] and has found application in, for example,o rganocatalysis [8] and peptide mimics. [9] Simple vicinal difluorides have been prepared from alkenes with the use of ambiphilic fluoride reagents,F 2 and XeF 2 . [10] However,their high reactivity,toxicity and high cost render their use impractical. While the required oxidizing equivalents are self-contained in these reagents,s ubsequent methods have relied on the combination of HF-salts and oxidants ( Figure 1B). Oxidation with the use of electrochemistry [11] or Selectfluor [12] provided the earliest confirmations of this strategy.H owever,i nb oth cases the product selectivity is poor and the alkene-types amenable to the reaction are limited, with success demonstrated on only very simple substrates.Y oneda employed p-tolyl difluoro l 3iodane (1a)a st he oxidant, [13,14] which improved product selectivity,h owever 1a is light-and temperature-sensitive, highly hygroscopic and expensive. [15] Thus,t he in situ formation of 1a using aryl-iodide,H F-salts and Selectfluor or mCPBA, was the subject of elegant work by both Gilmour [16] and Jacobsen. [17] Theuse of these stoichiometric oxidants then permits the use of sub-stoichiometric quantities of the aryliodide catalyst.
While these examples represent great advances in accessing vicinal difluorides,t here is no general method for accessing these motifs in compounds containing electron- rich moieties;d ue to competitive substrate oxidation, resulting in either decomposition or unselective fluorination. Moreover,owing to the potential of fluorinated alkyl groups in high-value bioactive compounds,amethod that is readily scaled, and therefore is safe,i nexpensive and does not produce much waste,iss till required.
To address these shortfalls,w es ought to access the l 3iodane 1 mediators [18] using electrochemical [19] oxidation. The unique spatial control of redox events,along with the control of potential and rate,s hould facilitate the expansion of substrate classes.Aswell as the inherent safety and scalability of electrochemistry, [20] the addition of ac hemical oxidant is not required, as protons can ultimately accept electrons at the cathode to form H 2 as the sole by-product, thereby rendering the process more atom-economical. [21] Thee lectrochemical generation of many hypervalent l 3iodane species from aryl-iodides is known. [22] Difluoro l 3iodanes (1)h as been comparatively less explored, [23] with Waldvogel recently reporting the only example in the presence of alkenes, [23e] which are liable to preferentially oxidise.Anumber of additional problems are reported to occur when 1 is generated at an electrode, [24] which include dimerization, benzylic fluorination and the formation of "many other complicated products". [23d] Thea pplication of high potentials with HF-salts also causes anode passivation, where an on-conducting polymer coating forms on the electrode surface that suppresses faradaic current and can attenuate reaction. [25] Thus,o ur primary objective was to address these issues in our optimisations.
We started by examining different aryl-iodide mediators for the electrochemical difluorination of allylbenzene (2a) (Figure 2A). No conversion to difluorinated alkane 3a was observed with the use of 4-iodo-anisole (R = 4-OMe), the most readily oxidized derivative we tested. However,w e observed the formation of 3a using aryl-iodides with higher oxidation potentials.F urther increases in potential beyond R = Me (tolyl) led to as ubsequent decline in yield, as direct substrate oxidation out-competes aryl-iodide oxidation (ca. E ox = 1.9 V). Although iodotoluene gave the highest yield of 3a,substantial quantities of benzylic fluorination (to 4)were observed (< 18 %) ( Figure 2B). We confirmed that 4 itself is av ery poor mediator,a sonly 7% of 3a was returned when using 4 in place of 1a. [26] As this side reaction requires deprotonation, we reasoned that it should be attenuated by reducing the availability of basic fluoride by increasing the proportion of HF to amine (in mixes of commercially available 3HF·NEt 3 and 9HF·py (amine = py or NEt 3 ). Indeed, by increasing this ratio from 3t o5 .6, benzylic fluorination decreased with an accompanying increase in product 3a ( Figure 2B). This trend also mirrors the enhanced activation of 1awith acid expected from the presence of more HF.F urther increases beyond 5.6 maintained the lack of 4 formation, but led to ad ecline in product 3a,p ossibly reflecting ad ecrease in fluoride activity.T his sweet-spot demonstrates the delicate balance between activation of 1a and deactivation of fluoride.
Thei nfluence of solvent was then examined. MeCN performed worse than CH 2 Cl 2 ,w hich was unsurprising considering anode passivation is particularly predominant under these conditions. [25] This insulating effect was reflected by al arge rise in cell potential during reaction compared to other solvents tested, [26] none of which improved the yield beyond that of CH 2 Cl 2 .H owever,w hen hexafluoro-isopropanol (HFIP) was added as ac o-solvent, higher yields of 3a were observed, as also noted in other halogenation [27] and electrochemical reactions. [28] 30 %HFIP in CH 2 Cl 2 led to the greatest enhancement. [26] Control reactions in the absence of alkene revealed that the concentration of HFIP determined the amount of 1a formation ( Figure 2C)a fter the same amount of charge was passed. Thea ddition of allylpentafluorobenzene (2e)t ot hese mixtures confirmed its more efficient transformation to 3e in the presence of more 1a. [26] To rationalise the downward trend of 1a ( Figure 2C), we observed ar educed solubility of iodotoluene with increased HFIP.C Vs tudies revealed that the oxidation potential of iodotoluene under the reaction conditions decreases with more HFIP present, [26,29] resulting in am ilder oxidizing environment, which should contribute to improved functional group tolerance.F inally, 19 FNMR analysis of ag enuine sample of 1a,s upported our assumption that 1a was formed under the optimized conditions. [26] Thei mportance of the iodotoluene mediator to product selectivity was confirmed by control reactions in its absence ( Figure 2D). Without iodotoluene,direct oxidation of 2as led to benzylic,r ather than alkene,f luorination with low mass balance.T he use of sub-stoichiometric quantities of the mediator led to ad ecline in yield of 3a, [26] reflecting am ismatch of reaction-rates in the electrochemical and chemical steps (EC mechanism). We could drop the loading of iodotoluene to 20 mol %byapplying 5cycles of 0.7 Fmol À1 with 6hstirring in between each cycle. [26] However, the vastly increased reaction time was deemed to be an inferior adjustment to the conditions than using an equivalent of iodotoluene and running the reaction in one go.Moreover,we are able to recover pure iodotoluene in greater than 70 % yields,a nd thus recycle it for use in subsequent reactions.
With this "in-cell" optimised method now in hand, we proceeded to explore the substrate scope ( Figure 3). Previously published protocols do not report substrate classes that contain electron-rich moieties,therefore,weinitially avoided these substrates.Indeed, electron-poor allyl arenes were well tolerated (2a-e)a nd gave difluorinated products in good to excellent yields.Long-chain alkenes, 2f-h,returned excellent yields,d emonstrating that ap roximal aromatic ring is not necessary for reactivity.Allylic ethers and amines (2i-t)were both tolerated if containing electron poor (hetero)aromatics. Ester (2h), alcohol (2f), sulfonate (2g)a nd halide (Br, Cl, e.g., 2l and 2q)f unctional groups were all untouched, providing useful products for further derivatization. While acid sensitive functionality,s uch as boc groups,w as not tolerated, [26] ac yclopropyl ring (2t)a voided competing oxidative ring opening. [30] Thes calability of the method was demonstrated by yielding gram-scale quantities of products 3e, 3i and 3s and adeca-gram-scale quantity of 3s.Ineach of these cases,70+ %o fp ure iodotoluene was recovered. Thec ommercially available ElectraSyn 2.0 set-up was also tested-in combination with aP TFE vial-and was found to give product 3d in acomparable yield to our set-up,validating the robustness of the conditions. These "in-cell" conditions performed worse with the more electron-rich allyltoluene (2u) ( Figure 4), returning only am oderate yield of difluorinated product 3u.T his reactivity is consistent with previous methods that also struggle with readily oxidized substrate classes. [16] Therefore,anew approach was sought to specifically gain access to products containing electron-rich moieties.A nalysis of ar ange of electron-rich substrates by cyclic voltammetry revealed their preferential oxidation to iodotoluene, [26] thereby eluding the vital formation of l 3 -iodane 1a.T oavoid this problem, an "excell" method was devised that spatially and temporally separates the electrochemical oxidation and fluorination steps,t hereby avoiding competitive direct oxidation of the substrates.T hus,c onditions were re-optimized for the initial formation of 1a in adivided cell, followed by the subsequent addition of substrate 2u. [26] With this approach the yield of 3u was raised from 45 %to73 % ( Figure 4).
Thes cope of electron-rich or easily oxidised substrates was now tested with this "ex-cell" method ( Figure 4). Electron-rich allyl arenes (2u-x)were now tolerated, returning moderate to very good yields of product. Success was achieved with the very electron-rich dimethylaniline 2w, however, the anisole derivative 2y was less well tolerated, which may be due to the known CÀHa ctivation pathway of these arenes-types to generate diaryliodonium species. [31] A pharmaceutically relevant morpholine amide (2z)w as also tolerated.
Anilines are asubstrate class that have also not previously been demonstrated, as they are very readily oxidized. Aniline 2aa posed problems using the "in-cell" method, cf.only 25 % 3aawas isolated. However, by adopting the "ex-cell" method, this was increased to 70 %. Thegreater electron withdrawing effect of nosyl (3ab)v s. tosyl (3aa)t ranslated into greater yields.Good to excellent yields of other difluorinated anilines containing electron rich and poor rings were generated (3acag), including mesityl. Other terminal alkenes containing electron-rich moieties were tolerated, including an allyl amine (2am)a nd ether (2as). Styrenes were tolerated if they were electron poor, such as 2ai,o therwise geminal fluorination occurred (2ah), which is ap rocess that has previously been observed with this substrate class. [32] Substituted styrenes (2aj-al)were also well tolerated.
Substituted alkenes that are unactivated are problematic substrates for other methods,a nd so we were pleased to discover that our "ex-cell" conditions readily translate electron-rich di-and tri-substituted terminal and internal alkenes (2ag, 2an-ar,2at-au). Good to excellent yields were produced in each case.Both cis and trans isomers of internal alkenes were viable substrates,a nd their stereochemical information was translated to their products with adiastereoselectivity between 5-9:1. Readily oxidised trialkylamines (2aq-ar)w ere again tolerated functionality.T oa scertain the effect of steric bulk or hydrophobic shielding on internal alkenes,4regio-isomers of trans-octene were tested in the reaction (2av-ay). It was found that yields improved as the bulk either side of the alkene decreased.
Thed ifluorination of several electron-rich substrates using our electrochemical method was compared to methods in the literature,including those that employ Selectfluor and   (Figure 4), thereby validating the importance of the "ex-cell" approach. Thes ustainability of the electrochemical method is reflected in the low E-factor [33] (ratio of total waste to product) calculated for the reaction, [34] which was consistently lower than the Selectfluor or mCPBAm ethods.T he main improvement to waste reduction originates from the lack of astoichiometric oxidant, which will also contribute to enhanced safety [35] and lower cost [26] on-scale.
In summary,anelectrochemical vicinal 1,2-difluorination of alkenes has been described, using as imple and userfriendly 2-electrode setup with nucleophilic fluoride and iodotoluene as am ediator.M oderate to excellent yields of fluorinated products are demonstrated in aw ide substrate scope.T he "ex-cell" method allows access to new substrate classes that have otherwise remained unattainable,i ncluding electron rich moieties,a nilines and substituted internal alkenes.T he method is sustainable (lower E-factor), safe, and high-yielding gram and decagram scale reactions demonstrate the practicality of the process.W et herefore expect this method to facilitate access to this important motif in aw ider variety of compounds and contexts.