Photocatalytic Reductive Radical‐Polar Crossover for a Base‐Free Corey–Seebach Reaction

Abstract A metal‐free generation of carbanion nucleophiles is of prime importance in organic synthesis. Herein we report a photocatalytic approach to the Corey–Seebach reaction. The presented method operates under mild redox‐neutral and base‐free conditions giving the desired product with high functional group tolerance. The reaction is enabled by the combination of photo‐ and hydrogen atom transfer (HAT) catalysis. This catalytic merger allows a C−H to carbanion activation by the abstraction of a hydrogen atom followed by radical reduction. The generated nucleophilic intermediate is then capable of adding to carbonyl electrophiles. The obtained dithiane can be easily converted to the valuable α‐hydroxy carbonyl in a subsequent step. The proposed reaction mechanism is supported by emission quenching, radical–radical homocoupling and deuterium labeling studies as well as by calculated redox‐potentials and bond strengths.

Despite the advances mentioned above,t he deprotonation of aliphatic dithianes stillr equires strong organometallicb ases (e.g. nBuLi), as the masked acyl anion exhibits ap Kav alue of approx.4 0(1,3-dithiane in DMSO). [13] This limitation prohibits the presenceo fb ase-and nucleophile-sensitive functional groups such as halides, esters and nitriles. Likewise, the stoichiometric use of an organometallic reagent generates undesired waste including metal salts.
In order to circumvent these issues, the dithiane would need to be activated by other means than direct deprotonation. With the CÀHb ond being in alpha position to two sulfur atoms it should be susceptible to af acile hydrogen atom transfer (HAT) opening an alluring alternative activation pathway. [14] This solution is especially attractive, as in recent years, the combination of photo-and HAT-catalysis enabled the development of several novel reactions and allowed the use of catalytic amounts of ah ydrogen abstracting reagent (HAT-catalyst) insteado fastoichiometricquantity. [15] Last year,w er eported that radicals generatedb yt he combination of photo-and HAT-catalysis can be reduced to render an anionic intermediate capable of reacting with electrophiles. [16] So far,t his methodw as limitedt os ubstrates stabilizing the radical and the anion intermediate by an aromatic group. We wonderedi ft his concepti sa pplicable to aliphatic dithianes forabase-free Corey-Seebach reaction, hence featuring an ovel andi mproved approachf or ac lassic and established strategy (Scheme 1D). Other photocatalytic methodst o access carbon nucleophilesf rom sp 3 CÀHb onds are rare, and requireacatalytic [17] or stoichiometric [18] amount of ac hromium source.T ot he best of our knowledge, the here presented methodi st he first example for ap hotocatalytic Corey-Seebach reaction via ac arbon nucleophile under metal-and basefree reaction conditions.
The investigation was started employing 2-methyl-1,3-dithaine (1a)a nd acetone (2a)a ssimple model substrates (Table 1). Gratifyingly,w ith 3DPA2FBN as photocatalyst and iPr 3 SiSH as HAT-catalyst the product was obtained in 32 %G Cyield (entry 1) using DMF as solventi nt he presence of 4 molecular sieve at 25 8C. Increasing the concentration (entry 2) improvedt he reactiono utcome slightly,w hile lowering the temperaturet o08Cp roved to be crucial (entry 3). The presence of bis(neopentyl glycolato)diboron (B 2 neop 2 )a sm ild Lewis acid was beneficial as well (entry 4). After completion of the reaction, degradation products of the HAT-catalyst were observed by GC-MS. Thus, the addition of as econd catalyst loading was tested as well, resultingi namodest yield gain (entry 5). Overall, an almostf ull conversion with ag ood GC-yield of 73 % translating into an isolatedy ield of 65 %w as achieved. Control experimentsr evealed the necessity of the combination of light, photo-and HAT-catalyst (entries 6-8). The full detailed optimization process is given in the Supporting Information.
With the optimizedc onditions, the substrate scope was established( Ta ble 2). The reactionw as successful with unsubstituted 1,3-dithiane (1b)a st he nucleophile, the alkyl chain could be prolonged (1c)a nd an additional heteroatom giving rise to other potentially cleavable CÀHb onds did not hamper the reaction( 1d). Indeed, the reactiont olerates base or nucle-ophile sensitive cyano-(1e)a nd ester functionalities (1g)a s well as halogen substituted aromatic moieties (1h-j)a nd an unsubstituted phenyl ring (1f).
The yield generally improves when increasing the equivalents of the electrophile. However,amaximum yield is reached at as pecific concentration, which cannot be increased by raising the electrophile concentration any further. Thus, evaluating the electrophile scope, the optimal amounto fr eactantf or every substrate was tested.P rolonging the alkyl chain was viable (2b), while an additional substituent in a-position decreasedt he yield due to an increased steric hindrance (2c). Cyclic Ketones (2d-f)a nd the presence of ah eteroatom (2f) were both well accepted. Ad ouble bond within the electrophile (2g)g ave the desired product despite ap otentialr adical addition as side reactiona nd ketones bearingaphenylr ing with variouss ubstituents (2h-k)w ere all tolerated. An exception is the presence of af ree -OH group (2l), as the anionic dithiane intermediate is likely to be protonated by the fairly acidic alcohol (pKa = 18.0 for phenol in DMSO) [19] rather than adding to its ketone moiety.A ccordingly,aprotected alcohol yielded the wanted product (2k). In terms of nitrogen containing functionalities, aB oc-protecteda mine (2m), at ertiary amine (2n)a nd an amide (2o)w ere viable electrophiles. AC F 3substituted cyclohexane (2p)g ave rise to ad iastereomeric product mixture separable by chromatography.AnX-ray crystal structure could verify the structure of the syn-isomer (syn-3ap). With as ubstrate containing both, an electrophilick etone and ester group (2q)t he nucleophilic addition proceeded exclusively at the more reactive ketone moiety,y ielding the corresponding lactone product (3aq)r esulting from an attack of the formed alcohol to the ester.A ldehydes were suitable electrophiles as well (2r-t). In this case, an excesso fd ithiane was  [20] leading to side reactions (tested lowand non-yielding substrates are listed in section 5o ft he Supporting Information). The described transformation is sensitive to steric hindrance, especially regarding the nucleophile (Scheme2). As terically demanding dithiane bearing ac yclohexyl ring (1k)c ould not be added to acetone giving the desired product 3ka.T he same waso bserved with ap henyl ring as substituent (1l). We reasoned that adding an electron donating methoxy substituent (1m)s hould increase the reactivity of the formed nucleophilic intermediate and indeed isolatable amountsofproduct 3 ma were formed in this case.
Consequently,w et ested aldehydes as sterically more accessible and reactive electrophiles for 1k-m,d ue to their ineffective addition to ketones.A se xpected, the desired products for all three nucleophile precursors (3kr-3 mr)c ould be isolated in moderate to good yields. With the reactive CÀHb ond of 1l and 1m being in ab enzylic positiona nd alpha to both sulfur atoms, it is highly susceptible to HATand the aldehyde could be added in slight excess. In contrast, an excesso fd ithiane 1k was required to arrive at ar easonable yield of 3kr,h owever, most of the surplus startingm aterial could be recovered after completion of the reaction(Scheme 2).
Scheme2.Sterically hindered dithianes as nucleophile precursors. Chem. Eur.J.2020, 26,12945 -12950 www.chemeurj.org 2020 The Authors. Published by Wiley-VCH GmbH al ower BDE. However,a na ldehyde is required as reaction partner in this case. No product could be isolated with acetone as electrophile.
Lastly,t he dithiane deprotection to valuable a-hydroxyk etones was investigatedf or selected examples. Among others, photocatalytic oxidations, [21] as well as an elegant Bi(NO 3 ) 3 ·5 H 2 Oc atalyzed method [22] have already been developed by other groups. After ap reliminary optimization (see Supporting Information), both strategies afforded the desired product in good yields, with the Bi(NO 3 ) 3 ·5 H 2 Om ethodb eing more efficient (Scheme 4). Ao ne-potp rocedure starting with the photocatalytic Corey-Seebachr eaction followed by the deprotection as the second step is possible, yetl ow yielding( see Supporting Information Scheme S3).
In the proposedm echanism, the oxidation of the HAT-catalyst by the excited photocatalyst is the first step. Thish ypothesis was supported by emissionq uenching studies indicating an interaction between the excited 3DPA2FBN and iPr 3 SiSH in the presence of 4 molecular sieve ( Figure 1A,l eft).
The same effect was not observedu sing 1a (Figure1A, right) or 2a ( Figure S5). The activated HATc atalyst( S-H BDE: 88.9 kcal mol À1 ) [23] should then abstract ah ydrogen atom from the dithiane substrate. The corresponding radical-radical homocoupling side product 5 could be detectedb yH RMS for 1m as as elected example after the reaction reached completion. Ah igher quantity could be detected by NMR when omitting the electrophile (Figure1B).
Based on the mechanistic studies and previous reports, followingm echanism is proposed (Scheme 5): the excited 3DPA2FBN photocatalyst oxidizes the iPr 3 SiSH HATc atalyst giving rise to the radicala nion of 3DPA2FBN and the iPr 3 SiSC radicala fter deprotonation. The thus activated HATcatalyst abstracts ah ydrogen atom from the most labile CÀHb ond, which is in alpha positiont ob oth sulfur atoms within the dithiane 1,r egeneratingt he HAT-catalyst and forming the correspondingr adical species 1C.T he photocatalytic cycle is then closed by reduction of 1C,y ieldinga nionic key intermediate 1 À À . This carbanion nucleophile is capable to attack non-activated Scheme3.Selectivef unctionalization of dithianes with two reactivesites.
In summary,w eh ave developedaphotocatalytic Corey-Seebach reaction operating under mild metal-a nd base-free conditions, employing solely catalytic additives. Base-andn ucleophile sensitive functional groups are tolerated and ketones as well as aldehydes are viable electrophilesg iving the desired product in moderate to good yields. The reaction is enabled by the combination of photo-and hydrogen atom transfer catalysis. This catalytic merger allows the activation of suitable CÀHb onds to carbanions capable of reacting with carbonyl electrophiles, the mechanism of which is supported experimentally as well as by calculated redoxp otentialsa nd bond strengths. The anionic intermediate seems to be less reactive than the classical lithiateds pecies, allowing regio-andc hemoselectivet ransformations. However,t he exact nature of the carbanion species remains as of now unknown and is under current investigation by time resolved spectroscopy and in situ NMR studies.