Nickel-catalyzed umpolung C–S radical reductive cross coupling of S-(trifluoromethyl)arylsulfonothioates with alkyl halides?

Yu-Zhong Yng,Gui-Fen Lv,Ming Hu,b,Yng Li,?,Jin-Heng Li,b,c,d,?

a Key Laboratory of Jiangxi Province for Persistent Pollutants Control and Resources Recycle,Nanchang Hangkong University,Nanchang 330063,China

b State Key Laboratory Base of Eco-Chemical Engineering,College of Chemical Engineering,Qingdao University of Science and Technology,Qingdao 266042,China

c State Key Laboratory of Applied Organic Chemistry,Lanzhou University,Lanzhou 730000,China

d School of Chemistry and Chemical Engineering,Henan Normal University,Xinxiang 453007,China

Keywords:Nickel Radical Reductive cross coupling S-(Trifluoromethyl)arylsulfonothioates Alkyl halides Alkyl aryl thioethers

ABSTRACT A new cooperative nickel reductive catalysis and N,N-dimethylformamide-mediated strategy for umpolung C–S radical reductive cross coupling of S-(trifluoromethyl)arylsulfonothioates with alkyl halides to produce alkyl aryl thioethers is described.This reaction features excellent selectivity,wide functionality tolerance,broad substrate scope,and facile late-stage modification of biologically relevant molecules.Mechanistic studies recognize initial generation of an amidyl radical anion via thermoinduced reduction of DMF with Sn,followed by umpolung reduction and single electron transfer of the nucleophilic sulfonyl moiety to form a sulphydryl radical and engage the Ni0/NiI/NiIII/NiI catalytic cycle.

Organosulfur compounds,including thioethers,are core moieties encountered in the structure of various drugs,natural products,herbicides,ligands and functional materials,as well as valuable synthetic building blocks and latent functional groups that can be modified to assemble complex target molecules in synthesis [1–16].As a result,significantly ongoing efforts have been devoted to the development and expansion of methods for catalytically forging highly valuable and functionality diverse thioether scaffolds in synthetic and medicinal chemistry [8–16].Conventional methodologies to straightforward access thioethers involve transitionmetal-catalyzed C–S cross coupling reaction [8–16],which is dominated by two different modes of reactivity,including a classicalpolarity method using the thiol functionality as a nucleophile(Scheme 1A-a) [17–32] and an umpolung approach employing the sulfur-based reactant as an electrophile (RSX,X=SR,SO2R,Cl,OR,NRR’ or CN;Scheme 1A-b) [33–42].While these polarity modes of catalytic C–S cross couplings of aryl halides or aryl organometallic reagents (such as arylboronic acids,arylmagnesiums and aryllithiums) with the thiolation reagents for producing aryl-tethered thioethers by incorporation of an aryl group onto a sulfur atom to form a C(sp2)-S bond have been well established and widely exploited [8–42],analogous versions to access alkyltethered thioethersviaintroduction of an alkyl group onto the sulfur atom to construct the C(sp3)-S bond have been less extensively studied [8–16,29],probably due to tendence to the facile side reactions (such asβ-hydrogen elimination) under strong alkaline and elevated temperature conditions.Furthermore,the vast majority of these reported protocols suffer from the use of highly toxic,air sensitive,odor disagreeable thiols and their oxidized derivatives,as well as only few commercially available alkyl thiols and alkyl disulfides,which significantly impede their widespread applications.Therefore,these challenges and the increasing importance of alkyl-tethered thioethers spur the synthetic chemists to develop mild,versatile strategies that (i) enable efficient incorporation of an alkyl group onto a sulfur atom to form the C(sp3)-S bond under base-free conditions;(ii) accommodate broad functionalized substrates,especially including diverse alkyl halides and readily accessible,bench-stable,odourless thiolation reagents;and(iii) are subject to facile late-stage modification of biologically relevant molecules.

Scheme 1.Synthesis of thioethers.

To circumvent these issues,transition-metal-catalyzed C–S reductive cross coupling reaction of organohalides with electrophilic thiolation reagents has recently been developed as a promising alternative to the conventional polarity types with preformed nucleophiles (Scheme 1B) [22,43–57].These approaches allow facile introduction of an electrophilic aryl or alkyl group onto the electrophilic sulfur atom to construct the sp2-and sp3-hybridized C–S bonds under mild and base-free conditions,and thus exclude side reactions,such asβ-hydrogen elimination.However,only few approaches have been reported to allow catalytic C–S reductive cross couplings of unactivated alkyl halides with electrophilic thiolation reagents (e.g.,disulfides and thiosulfonates) for producing alkyl-tethered thioethers.For example,the group of Wang/Ji has reported the first nickel-catalyzed C–S reductive cross coupling of unactivated alkyl bromides with thiosulfonates and Mn reductant [46–48],which is highlighted by the use of the simple,bench-stable and odorless thiosulfonates as the electrophilic thiolation reagents and by a plausible mechanism comprising an inner-sphere Ni0/II/III/I/IIcatalytic cycle directly engaged by the alkyl carbon-centered radicals from homolysis of alkyl halides.Later,this group developed a similar catalysis version to accomplish thiolation of alkyl bromides with arenesulfonyl cyanides as the electrophilic disulfide precursors for assembling alkyl aryl sulfides [50].Very recently,the group of Ackermann reported an electroreductive nickel-catalyzed radical thiolation by cross-electrophile coupling of alkyl bromides with functionalized thiosulfonates through Mg cathodic reduction to give alkyl-tethered thioethers [51].These methods rely on the generation of the alkyl carbon-centered radicalDfrom alkyl halides reacted with theinsituformed NiIintermediateC,which would sequentially execute single electron oxidation with the NiIIintermediateAto afford the NiIIIintermediateB(Pathway I;Scheme 1B) [45–50].On the basis,we hypothesized that initially generating the sulfur-centered radicalF,which are formed from homolysis of the electrophilic thiolation reagent components,would give rise to single electron oxidation to deliver the NiI-SR intermediateGfollowed by oxidative addition with alkyl halides to produce the NiIIIintermediateH(Pathway II;Scheme 1B),which would: (i) provide new radical reductive crosscoupling tactics comprising the engagement of the reaction with the sulfur-centered radicals thus resulting in access to otherwise poorly accessible or unobtainable molecular frameworks;(ii) expand the reactivity profile of Ni reductive catalysis;and (iii) innovate and advance radical chemistry.

Herein,we report the first nickel-catalyzed DMF-mediated umpolung C–S radical reductive cross coupling betweenS-(trifluoromethyl)arylsulfonothioates and alkyl halides involving a sulfur-centered radical formation (Scheme 1C).This reaction is initiated by DMF,Ni(ClO4)2?6H2O,4,4′-di–tert–butyl–2,2′-bipyridineL1and Sn,and enables the formation of the C(sp3)-S bonds through umpolung transformations ofS-(trifluoromethyl)arylsulfonothioates and sequential catalytic reductive cross coupling with alkyl halides.

To determine the role of arylesulfonothioates2as theS-based functional group sources,the umpolung C–S radical reductive cross coupling of 3-phenylpropyl bromide1awith PhSO2SCF32awas examined (Table 1).Screening various reaction parameters revealed that a combination of 5 mol% Ni(ClO4)2?6H2O,7.5 mol% 4,4′-di–tert–butyl–2,2′-bipyridineL1and 2 equiv.Sn in DMF (0.2 mol/L) at 60 °C for 12 h afforded the desired phenyl(3-phenylpropyl)sulfane3aain nearly quantitative yield with excellent chemoselectivity(entry 1).Unlike the previously reported results acted as the SCF3(often) or PhSO2source [58–68],PhSO2SCF32aserves as the PhS source.Both Ni catalysts and Sn are necessary to make the reaction successful as leaving out each led to no desired reaction (entries 2 and 16),and a lower loading of Ni(ClO4)2?6H2O (2 mol%)decreased the yield (entry 3).Other Ni catalysts,including NiCl2,NiBr2,NiCl2?DME,NiCl2(PPh3)2and NiCl2(Py)4,were highly active(entries 4–8),but all were less efficient than Ni(ClO4)2?6H2O.Opti-mization of the dinitrogen-based ligand effect indicated that these ligandsL1-L6served as promotors since omission of ligands the reaction could still run efficiently to tender3aain 86% yield (entries 9–15).Furthermore,ligandsL2,L4-L6could improve the reaction (entries 11 and 13–15),but 2,2′-bipyridineL3was detrimental to the reaction outcome attributing to strong coordination with the Ni catalyst lowering its catalytic activity (entry 12).Using the same equivalent amount of Ni(ClO4)2?6H2O andL1slightly diminished the yield (entry 10),suggesting that excessL1assists complete reduction of Ni(ClO4)2?6H2O to the active Ni0species avoiding consumption of Sn reductant.The yield raised from 84%to 95% with the increase of the Sn amount from 1.2 equiv.to 1.7 equiv.(entry 17).These observations indicate that the roles of Sn mainly include reduction of PhSO2SCF3and regeneration of the active Ni(0) species.Notably,the reaction is sensitive to the reducing reagents as the other common reductants,such as Mn,Mg,Zn and(EtO)3SiH,had no reactivity (entry 18).Surprisingly,the reaction was sensitive to solvents: Amides,such as DMF and MeCONMe2,were viable media (entries 1 and 19),but other solvents,such as MeCN,1,4-dixoane and ClCH2CH2Cl,were inert (entry 20).These results imply that amides may participate in the reaction besides as media.Decreasing temperatures led to diminishing yields (entry 21).The standard conditions were compatible with a scale up to 3 mmol1a,giving3aain excellent yield (entry 22).

Table 1Optimization of reaction conditions.a

After confirming the optimized conditions,we set out to study the generality of this umpolung C–S radical reductive cross coupling protocol (Scheme 2).Gratifyingly,a variety ofS-(trifluoromethyl)arylsulfonothioates2b-iefficiently underwent the reaction with bromide1a,Ni(ClO4)2?6H2O,L1and Sn,affording3ab-3aiin 85%–98% yields.Furthermore,several aryl functionalities,including 4-MeC6H4,4-tBuC6H4,4-MeOC6H4,4-ClC6H4,4-FC6H4,4-CF3C6H4,naphthalen-2-yl and thiophen-2-yl,were well tolerated.Whereas using 2 h reacted with NiBr2catalyst reduced the yield of3ahto 53%.

Scheme 2.Variation of the alkyl halides (1) and arylsulfonothioate (2).Reaction conditions: 1 (0.2 mmol),2 (0.22 mmol;1.1 equiv.),Ni(ClO4)2?6H2O (5 mol%),L1 (7.5 mol%),Sn (2 equiv.),DMF (0.2 mol/L;1 mL),argon,60 °C and 12 h.a NiBr2 (5 mol%)instead of Ni(ClO4)2?6H2O.

We next aimed to evaluate the scope of alkyl halides1(Scheme 2).Surprisingly,alkyl iodide,3-phenylpropyl iodide1b,was lower reactive for furnishing3aain 20% yield,attributing to readily decomposition of the C–I bonds.Using lower reactive 3-phenylpropyl chloride1cfailed to construct3aa.Strikingly,a wide range of functionalized alkyl bromides1d-aqaccommodated to this umpolung C–S radical reductive cross coupling (3da-aqa).For example,functionalized propyl bromides1d-gafforded3daga,respectively,in 70%–99% yields where a functionality,such as 4-ClC6H4,CN,CO2Et,and 4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl,at the positionγto the bromide atom was intact.This optimal conditions were compatible with (2-bromoethoxy)(tert–butyl)-dimethylsilane1),producing the high useful silyl-substituted product 3 ha in 80% yield.Using (3-bromoprop-1-en-1-yl)benzene1i,an alkene,furnished cinnamyl(phenyl)sulfane3iain good yield.The linear alkyl chains containing one to six carbon atoms were competent to the coupling,and several functional groups,including aryl,F,OH and Cl,were tolerated (3ja-sa).Alkyl bromides1t-wwith steric hindrance were suitable substates (3tawa).Broad secondary and tertiary alkyl bromides,including(1-bromoethyl)benzene1x,α-bromoketones (1y,1ad),four-to seven-membered cycloalkyl bromides (1z-ac),α–bromo ester (1ae)andα–bromo amide (1af),were subject to the coupling,furnishing the corresponding secondary and tertiary alkyl sulfanes3xaacain high to quantitative yields.Interestingly,dual umpolung C–S radical reductive cross couplings of alkyl dibromides1ag-akexecuted successfully to access disulfanes3aga-aka,which highlights the applicability of our protocol in organic and material synthesis.A number of natural product-or bioactive molecule-based alkyl bromides1al-aq,such as L-alaninate derivative [69],telmisartan derivative [70],cholesterol derivative [71],4-androstene-3,17–dione derivative [72],estrone derivative [73] and estradiol pentanoate derivative [74],exposed to the optimized conditions resulted in selective transformation of the C(sp3)–Br bonds to the C(sp3)–S bonds to produce highly valuable complex products3alaana,3aob,3apa-aqa,thus providing a powerful route to selective late stage modification of complex bioactive substrates with multiple potential sites of reaction.Unfortunately,aryl halides,such as bromobenzene and iodobenezen,had no reactivity for the reaction.

In contrast to alkene-containing bromides1i,1ai,and1anao(3ia,3aia,3ana,3aob,Scheme 2),6-bromohex-1-ene1arwas converted to a mixture of the desired product3araand the intramolecular alkene difunctionalization product4arain 71% total yield with 5.1:1 chemoselectivity (Eq.1,Scheme 3) [45].Moreover,increasing concentrations of the Ni(ClO4)2?6H2O/L1catalytic system shifted the chemoselectivity toward the coupling,and using 10 mol% Ni(ClO4)2?6H2O led to occurrence of the coupling exclusively.These radical clock experiments support that the reaction proceedsviaa radical chain process [45,75–78].Gratifyingly,3-bromoprop-1-yne3aswas a suitable substrate,efficiently affording3asdin 98% yield (Eq.2).It was noted that no conversion of bromide1awas observed in the absence of PhSO2SCF3(Eq.3),supporting initiation of this coupling not from the alkyl bromide component.Using PhSO2S(4-ClC6H4)2jresulted in the selectivity toward direct C–S reductive cross coupling with the S(4-ClC6H4) moiety [45–49],not the PhSO2moiety,to afford3ajin 19% yield along with 4-chlorobenzenethiol5jin 10% yield,but PhSO2SCH2(CH2)2Ph2khad no reactivity (Eq.4).The different chemoselectivity support that this current protocol performs a different mechanism from the previously reported reductive C–S cross coupling transformations[22,43–57],probably attributing to both the electron effect of the SCF3group and reduction behavior of Sn.In the presence of Sn or Mn reductant,PhSSCF32lcould react with bromide1ato access3aain 10%–15% yields (Eq.5),but leaving out each led to no detectable product3aa.These results suggest that the PhSSCF3is not the key intermediate during this current process,and the reductant can simultaneously assist both the S–S bond cleavage and the C–S bond formation.Subsequently,a series of the sulfur sources,including PhSSPh (2m),PhSO2CN (2n),PhSO2CF3(2o),PhSO2F (2p),PhSO2Na (2q),PhSO3Na (2r),PhSO2N3(2s) and PhSO2NCO (2t),were examined,but all had no reactivity under the optimized conditions (Eq.6).It is noteworthy that both electrophilic PhSSPh (2m)and PhSO2CN (2n),the reported highly reactive thiolation reagents[43,44,48],are inert,thus ruling out the generation of PhSSPh as the key intermediate.

Scheme 3.Variations of the other reaction components.

To further understand the mechanism,control reduction experiments with (4-ClC6H4)SO2SCF32ewere conducted (Eqs.7 and 8).In the presence of Ni(ClO4)2?6H2O,L1and Sn,substrate2ewas reductively decomposed to 4-ClC6H4SH5jin 74% yield and 4-ClC6H4SS(4-ClC6H4)6ein 13% yield (Eq.7).Reduction of (4-ClC6H4)SO2SCF32ewith Sn run smoothly,affording 4-ClC6H4SH5jexclusively in 95% yield;however,the SnBr2additive (10 mol%)is detrimental and decreased the yield of5jslightly to 92% yield(Eq.8).The reason may be that the SnBr2salt can promote the formation of disulfide [48],which would suppress the current coupling.It is noted that the reduction reaction is also sensitive to reductants: other reductants,such as Mn,Mg,Zn or (EtO)3SiH,are inert,and no reduction of the (4-ClC6H4)SO2SCF32ewas observed without reductants (Eq.8).These findings are consistent with the results observed in Table 1 (entries 1,16 and 18),and support that the generation of the active benzenethiol-type intermediate,not the reported active PhSH and/or PhSSPh intermediates [22,43–57],is the key step.Under the optimized conditions,4-ClC6H4SH5jwas less reactive than (4-ClC6H4)SO2SCF32eas using 4-ClC6H4SH5jdirectly reacted with alkyl bromide1adelivered a lower yield of3ae(85% yield) (Eq.9) than that of (4-ClC6H4)SO2SCF32e(96%yield,Scheme 2).It is because among the current coupling processes thermoinduced reduction of 4-ClC6H4)SO2SCF32eoccurs to generate the higher reactive 4-ClC6H4S-based intermediate,not 4-ClC6H4SH5j,to directly react with the active Ni species,thus avoiding further umpolung step of 4-ClC6H4SH to form the reactive 4-ClC6H4S-based intermediates (such as disulfides) and side reactions.

As shown in Scheme 4,the reaction of bromide1awith PhSO2SCF32awas inhibited by radical scavengers,such as TEMPO,BHT and hydroquninone (Eq.10).In the presence of TEMPO,the methylated products8and8′was detected by GC–MS analysis and no phenylpropyl-substituted product7from bromide1awas observed (Eq.9).Identical results were obtained from the reaction of1aalone in the presence/absence of Ni(ClO4)2?6H2O andL1(Eq.11).These observations speculate that the methyl radical is generated from DMF,and DMF may really engage the umpolung C-S radical reductive cross coupling reaction.To verify thespeculations,control transformations of DMF with TEMPO were examined (Eq.12).No reaction of DMF with TEMPO occurred when performing at 60 °C for 12 h.Using 2 equiv.Sn resulted in the formation of8in 3.7% GC yield.The optimized conditions that comprise a combination of 5 mol% Ni(ClO4)2?6H2O,7.5 mol%L1and 2 equiv.Sn were further confirmed,thus giving8in the highest 4.5% GC yield.Increasing loading of Ni(ClO4)2?6H2O andL1led to diminishing GC yields of8.The reason may be that the optimal loadings of the Ni(ClO4)2/L1system efficiently initiate the generation of the radicals and effectively improve their reactivity,whereas the higher loadings of the Ni(ClO4)2/L1system over activate the radicals to cause some unwanted side-reactions.The high reduction potentials of tin (Sn;-0.45 Vvs.SCE) and DMF (-1.95 Vvs.SCE) are proven to be useful reductants [79–83].These findings indicate that thermoinduced reduction of DMF with Sn occurs to generate an amidyl radical anion intermediate [83–88],and DMF as an organic catalyst mediated the umpolung C-S radical reductive cross coupling reaction.

Scheme 4.Control experiments and synthetic utilizations.

In the presence of 2 equiv.Sn,a stoichiometric amount of the Ni(II) complex9(Eq.13) exhibits identical catalytic activity to the NiBr2/L1catalytic system (3ah,Scheme 2).However,neglecting Sn led to no detectable C–S cross coupling (Eq.13).These observations prove the importance of the reduction process and the Ni0species,not the NiIIsalts,is the real active catalyst,which are further verified by the results using a stoichiometric amount of Ni(ClO4)2?6H2O (Eq.14).The C–S radical reductive cross coupling of1awith2aand 1 equiv.Ni(ClO4)2?6H2O in the presence/absence of Sn (Eq.14): Neglecting Sn caused no desired reaction after 12 h,but supplementing Sn to the same pot resulted in the formation of3aain 45% yield for 12 h.

Synthetic utilizations of phenyl(3-phenylpropyl)sulfane3aawere conducted under oxidative conditions (Eq.14)[22,43–57]: Sulfane3aawas converted to highly valuable ((3-phenylpropyl)sulfinyl)-benzene10aaand ((3-phenylpropyl)-sulfonyl)benzene11aa,respectively,in quantitative yields (Eq.15).However,both substrates10aaand11aacould not be transformed to3aaunder the optimized conditions,excluding the possibility of the umpolung C–S radical reductive cross couplingviathe10aaand/or11aaformation process.

Based on the current results and precedent literatures [22,43–57,75–88],the plausible mechanism for the Ni-catalyzed umpolung C–S radical reductive cross coupling reaction was proposed(Scheme 5).Initially,thermoinduced reduction of DMF with Sn affords an amidyl radical anion intermediateJ[79–88].Meanwhile,coordination of the NiIIspecies with the dinitrogen-based ligandLforms the active Ni0species.Subsequently,the reaction of the active Ni0species with the sulphydryl sulfur-centered radical (PhS?)intermediateF,which is formed from the umpolung reduction and single electron transfer (SET) of PhSO2SCF32awith Sn and the intermediateJ,occurs to produce the LnNiISPh intermediateG.Oxidation addition of the intermediateGwith 3-phenylpropyl bromide1aaffords the Ph(CH2)2CH2(Ln)NiIIIBr(SPh) intermediateH,followed by reductive elimination of the intermediateHto give the LnNiIBr intermediateIand the desired product3aa.Finally,reduction of the intermediateIby Sn regenerates the active Ni0species to start a new catalytic cycle.

Scheme 5.Possible reaction mechanism.

In summary,we have disclosed a novel catalytic radical reductive strategy for umpolung transformation ofS-(trifluoromethyl)arylsulfonothioatesviacooperative DMF and nickel reductive catalysis.This strategy was developed in a umpolung C–S radical reductive cross coupling ofS-(trifluoromethyl)arylsulfonothioates with unactivated alkyl halides to assemble alkyl aryl thioethers.The reaction involves the formation of a sulfur-centered radical through thermoinduced umpolung reduction ofS-(trifluoromethyl)arylsulfonothioates with DMF and Sn,as well as features excellent selectivity and wide functional group tolerance,which can be of great synthetic value for organic synthesis,such as applications in late-stage derivatization of pharmaceuticals and naturally occurring molecules,and creation of new reactions to access value-added derivatives of feedstocks.Mechanistic experiment evidence suggests that thermoinduced reduction of DMF by Sn readily occurs to generate the amidyl radical anion followed by umpolung reduction and SET ofS-(trifluoromethyl)arylsulfonothioates with the amidyl radical anion and Sn to produce a sulfur-centered radical that engages a process of single electron oxidation of the active Ni0species,unlike the previously explored alkyl carbon-centered radical counterparts.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

We thank the National Natural Science Foundation of China (No.22271245),the Jiangxi Province Science and Technology Project(Nos.20212AEI91002 and 20202ACBL213002) and the Open Research Fund of School of Chemistry and Chemical Engineering,Henan Normal University (No.2021ZD01) for financial support.