Iron-mediated oxidative C – H coupling of arenes and alkenes directed by sulfur : an expedient route to dihydrobenzofurans †

A novel route to medicinally-relevant dihydrobenzofurans utilises a sulfur-directed C–H ortho-coupling of arenes and unactivated terminal alkenes mediated by iron, and a palladium-catalysed deallylation/ heterocyclisation sequence. The iron-mediated coupling affords linear products of alkene chloroarylation in good yield and with complete regioselectivity. The coupling likely proceeds by redox-activation of the arene partner by iron(III) and alkene addition to the resultant radical cation.


Introduction
The search for more sustainable synthetic methods has led to increased interest in cross-couplings that unite two partners at the expense of a C-H bond in each, i.e. without the need for pre-functionalisation.Such processes would lead to lower costs, as fewer steps are required, and less waste results, when compared to standard metal-catalysed coupling reactions. 1he ubiquity of C-H bonds means that achieving regioselectivity in C-H functionalisation can often prove difficult, and directing groups are often necessary to facilitate functionalisation with directed ortho C-H bond activation on an aromatic or heteroaromatic ring being a common strategy. 2The oxidative ortho C-H coupling of arenes and heteroarenes with alkenes has been well investigated using various directing groups, reactive alkenes (e.g.acrylates and styrenes), and Pt group metals. 3In contrast, the use of a sulfur-based directing group to direct C-H coupling of an aromatic ring with alkenes is a recent advance and is largely unexplored (Scheme 1A). 4 The replacement of Pt group metals with inexpensive first row transition metals for synthetic procedures, including those involving C-H functionalisation, has clear benefits. 5The low cost of iron, its high abundance, and its low toxicity, make iron-based reagents ideal for new transformations.
Here we describe the Fe(III)-mediated oxidative C-H ortho coupling of aryl sulfides with simple, terminal alkenes to give linear β-chloroarenes. 6It is proposed that the two nucleophiles couple through oxidation of the arene partner by Fe(III), followed by addition of the alkene to quench the resulting radical cation. 7,8The sulfur-directed, iron-mediated coupling facilitates an expedient, novel route to decorated dihydrobenzofurans; a motif found in many natural products and drug molecules (Scheme 1B).The approach also features a Pd-catalysed deprotection/heterocyclisation and Ni-catalysed crosscouplings of sulfides (Scheme 1C).

Scope and limitations
Previously reported optimisation studies found that FeCl 3 was the most effective oxidant for the C-H coupling of diaryl sulfides and alkenes and that the reaction proceeded at ambient temperature under air in CH 2 Cl 2 . 6As the coupling is thought to involve reactive radical cation intermediates that are prone to decomposition, slow addition of the Fe(III) oxidant to a mixture of arene and alkene was employed to maintain a low concentration of radical cation.Slow addition of FeCl 3 was found to give improved mass balance and yield.Regardless of how the oxidant was added, crude 1 H NMR spectra showed clean product and mass balance was the only issue.
The scope with regard to the aryl sulfide partner was next assessed (Scheme 3).Although the coupling is sensitive to substitution in the aryl ring undergoing carbon-carbon bond formation (vide infra), the non-reacting, right-hand aryl ring can be varied: methyl (2k), bromide (2l), fluoride (2m and 2n), nitro (2o), trifluoromethyl (2p), and methoxy (2q and 2r) groups were compatible with the coupling and products were isolated in good yield.
Whilst 3,5-dioxygenation in the left-hand aryl ring of the diaryl sulfides has to date proved indispensible, allyl-and isopropoxyethers can also be used in the C-H coupling and products were isolated in moderate yield.The use of an allyl ether-containing substrate allowed interesting avenues for product manipulation to be explored (vide infra, Scheme 4).
Fig. 1 shows a selection of substrate combinations that did not result in successful coupling.Styrene was an ineffective partner potentially due to it being easily oxidised and thus reacting with FeCl 3 .Alkenes bearing oxygen-containing functionality, such as ketones and alcohols, also proved ineffective coupling partners.This is likely due to competitive coordination of Fe(III) to the Lewis basic sites in the alkene partner.Finally, internal alkenes, such as cyclohexene, did not undergo coupling.In terms of the arene partner, 3-methoxyphenyl phenyl sulfide did not undergo coupling with 1-octene (vide infra).Also, in contrast to substrates bearing nitro and trifluoromethyl groups in the right-hand aryl ring, the presence of an electron-donating substituent in the para-position of that ring adversely affected coupling.This may be due to formation of a thioquinone-type species.To date, alkyl aryl sulfides have also proved to be incompatible with the coupling reaction.Additional substrates that failed to undergo coupling can be found in the ESI.†

Mechanistic insights
Carbon-carbon bond formation at the terminal position of the alkene, and the formation of linear products, rather than coupling at the internal position of the alkene, and the formation of branched products, suggests that an S E Ar process is not operational. 9In addition, control experiments using nonredox-active Lewis acids did not result in product formation (Scheme 5).The preference for bond formation ortho to sulfur may indicate that chelation with the metal centre in D directs reaction towards this position. 12Substrates in which the arylsulfanyl substituent was replaced by an alkoxy or alkylamino substituent failed to undergo coupling, thus highlighting the important role played by sulfur (see the ESI †).
Our studies show that the oxidation potential 1a [+1.71V (vs.SHE) in MeCN] is compatible with the initial step of the proposed mechanism (Fig. 2).A carbocation is then formed by the oxidation of radical cation B with FeCl 3 that is then quenched by chloride to deliver the products 2. Indirect evidence for this mechanism comes from the observation that well-known oxidant CAN mediates a similar transformation: coupling of 1a with 1-octene gave the nitrate analogous to 2a in 56% yield (Scheme 7). 13,14Furthermore, the oxidation of diaryl sulfides to the corresponding sulfoxides using CAN is  proposed to proceed through radical cations analogous to A. 15 The terminal alkene partners are known to have higher oxidation potentials than the aryl sulfides. 16-hexenyl 'radical clock' cyclisation was used in an unsuccessful attempt to intercept radical B (Scheme 2: 2b was the only coupled product observed).Thus, under the experimental conditions, radical oxidation appears to be faster than cyclisation (i.e.>1 × 10 5 s −1 ) (Scheme 8). 17,18Interestingly, attempts to promote cyclative radical trapping using dienes bearing geminal substitution led to alternative coupling products 2y and 2z in low yield (Scheme 8).This may be due to the increased steric bulk impeding trapping of the carbocation intermediate with chloride.
Two factors appear crucial to the success of the cross-coupling process; (i) the ease of oxidation of the aryl sulfide, and (ii) the stability of the resultant radical cation A. For example, attempts to use 3-methoxyphenyl phenyl sulfide as a coupling partner with 1-octene led to only a trace of product even though the aryl sulfide starting material was consumed.We have measured the oxidation potential of 3-methoxyphenyl phenyl sulfide and found it to be similar to that of 1a [+1.72 V (vs.SHE) in MeCN] (Fig. 2), therefore, we propose that deleterious side reactions of the radical cation derived from 3-methoxyphenyl phenyl sulfide are likely responsible for inefficient carbon-carbon bond formation.Steric stabilisation of the radical cation A may therefore be the most important role played by substituents in the ring undergoing coupling in the aryl sulfide.
Switching the electronic properties of the para substituent in the non-coupling, right-hand aryl ring of substrates affected the yield of the process (Scheme 3).Coupling reactions of aryl sulfide substrates bearing nitro and trifluoromethyl electronwithdrawing groups, to give sulfides 2o and 2p, gave products in higher yield than processes involving comparable substrates lacking these substituents.As expected, the competition experiment shown in Scheme 9 using limiting FeCl 3 suggests that the rate of coupling is highest for electron-rich substrates: a mixture of 1-octene, 1a and 1f selectively delivered 2a with no trace of 2o.It appears likely that a decreased rate of aryl sulfide oxidation, thus avoiding large concentrations of radical cations that lead to decomposition, 19 results in the improved yields obtained for the coupling of aryl sulfides to give 2o and 2p (Scheme 3).

Product manipulation
The products of C-H arene/alkene coupling possess a range of functional handles for further manipulation (Fig. 3).For  example, directed ortho-metalation of 2a followed by reaction with various electrophiles (to form 3a-c), oxidation of sulfur (to give sulfone 3d), reduction and elimination of chloride (to form 3e and 3f, respectively), and S N 2 displacement of chloride (to give azide 3g) can be efficiently performed.
Notably, styrene 3f is formally the product of an oxidative Heck reaction involving aryl sulfide 1a and 1-octene. 20Coupling products can also undergo dehalogenation when treated with silver(I) salts to give the cyclic sulfonium salts (3h and 3i) as a 2 : 1 mixture of diastereoisomers.Carrying out the Fe(III)mediated C-H alkene/arene coupling in the presence of AgOTf allowed the sulfonium salt 3h to be prepared directly in moderate isolated yield (Scheme 10).
Pd-catalysed deprotection of the allyl ethers in the products of coupling 2s-w triggered cyclisation to give important benzofuran scaffolds.The reaction proceeds efficiently under mild conditions employing NaBH 4 and the allyl scavenger, morpholine.Interestingly, when NaH was used in place of NaBH 4 , only the mono-deprotected benzofuran product 4e was observed (Scheme 11).
Conversion of dihydrobenzofuran 4a to the corresponding triflate 5 (Tf 2 NPh, tBuONa, THF, 2 h; 87%) paved the way for further decoration of the heterocyclic scaffold.Pd-catalysed Suzuki-Miyaura cross-coupling 21 gave 6a-c in high isolated yield.Alternatively, copper-free, Pd-catalysed Sonogashira crosscoupling gave 6d in excellent isolated yield (Scheme 12). 22he arylsulfanyl directing group can be removed using RANEY® nickel in excellent yield (RANEY® Ni, EtOH, 1 h) (Fig. 4).Notably, reductive removal of the arylsulfanyl from the products of Fe(III)-mediated C-H coupling completes a concise approach to biologically important alkylresorcinols (e.g.7a). 23re attractively, the arylsulfanyl directing group can also act as a synthetic handle for the introduction of further diversity.Nickel-catalysed Kumada couplings with aromatic, heteroaromatic, and aliphatic Grignard reagents allow efficient  further decoration of the dihydrobenzofuran motifs (Scheme 13). 24Notably, nickel-catalysed coupling to give 8a and 8b, proceeded smoothly with no nickel insertion into thiophene C-S bonds. 25

Conclusions
A novel route to medicinally-relevant dihydrobenzofurans has been developed that utilises a sulfur-directed C-H ortho-coupling of arenes and unactivated terminal alkenes mediated by iron, and a palladium-catalysed deallylation/heterocyclisation.The iron-mediated coupling affords linear products of alkene chloroarylation with complete regioselectivity.Redox-activation of the arene partner by iron(III) and coupling of the resultant radical-cation with the alkene partner is proposed.The dihydrobenzofuran scaffolds can be readily decorated using Pd and Ni-catalysed cross-couplings.

Scheme 2
Scheme 2 Variation of alkene in the Fe(III)-mediated C-H coupling.

Scheme 4
Scheme 4 Variation of ether protection in the aryl sulfides undergoing Fe(III)-mediated C-H coupling.