Chromium-catalyzed cyclopropanation of alkenes with bromoform in the presence of 2,3,5,6-tetramethyl-1,4-bis(trimethylsilyl)-1,4-dihydropyrazine

Chromium-catalyzed cyclopropanation of alkenes with bromoform was achieved to produce the corresponding bromocyclopropanes. In this catalytic cyclopropanation, an organosilicon reductant, 2,3,5,6-tetramethyl-1,4-bis(trimethylsilyl)-1,4-dihydropyrazine (1a), was indispensable for reducing CrCl3(thf)3 to CrCl2(thf)3, as well as for in situ generation of (bromomethylidene)chromium(iii) species from (dibromomethyl)chromium(iii) species. The (bromomethylidene)chromium(iii) species are proposed to react spontaneously with alkenes to give the corresponding bromocyclopropanes. This catalytic cyclopropanation was applied to various olefinic substrates, such as allyl ethers, allyl esters, terminal alkenes, and cyclic alkenes.


Introduction
Cyclopropane is a strained three-membered carbocycle, and a common structural motif in pharmaceutical and biologically active compounds. 1 The synthesis of cyclopropanes from easily available starting materials is in high demand, and several stoichiometric synthetic protocols for the C3 ring have been developed: (1) classical reductive cyclization of 1,3-dihalopropanes or b-haloalkenes using metal-based reductants such as lithium and magnesium, 2 (2) cyclopropanation of alkenes using haloform (CHX 3 ) and a strong base in phase-transfer conditions to afford geminal dihalocyclopropanes, 3 and (3) cyclopropanation of alkenes using nitrogen-, phosphonium-, and sulfur-ylides, 4 in situ-generated zinc carbenoid from Zn reagents and CH 2 I 2 (Simmons-Smith reaction), 5 and in situ-generated chromium carbene species from excess amounts of CrCl 2 , diamine ligands, and RCHI 2 . 6 In contrast to these stoichiometric reactions, metal-catalyzed cyclopropanation of alkenes using diazomethane and its derivatives is an alternative effective protocol, despite the use of explosive diazomethane derivatives. 7 To avoid the use of explosive compounds, the development of metal-catalyzed cyclopropanation reactions using non-explosive reagents was recently explored. 8 Uyeda et al. reported that some nickel and cobalt complexes serve as catalysts for Simmons-Smith type reactions of alkenes with less reactive CH 2 Cl 2 and CH 2 Br 2 in the presence of excess zinc powder (Fig. 1a). 8f-8i Takai et al. reported that chromiumcatalyzed cyclopropanation of alkenes with Me 3 SiCHI 2 proceeds in the presence of catalytic amounts of chromium complex and excess Mn powder as a reducing reagent, from which gem-dichromiomethane complexes (Cr 2 -SiMe 3 ) are isolated (Fig. 1b), 9a and, similarly, Anwander et al. isolated an iodomethyl-bridged dichromium complex by treating CrCl 2 Fig. 1 Metal-assisted cyclopropanation of alkenes with di-and trihalomethanes; (a) cyclopropanation with excess zinc powder, (b) cyclopropanation with excess or catalytic amounts of chromium, and (c) bromocyclopropanation with catalytic amounts of chromium and organosilicon reductant 1a (This Work).

Results and discussion
We then screened conditions by tuning reductants, additives, and supporting ligands to optimize the chromium-catalyzed cyclopropanation of allyl benzyl ether (2a) with bromoform as a model reaction, and the results are summarized in Table 1. When we used a 1 : 1 mixture of CrCl 3 (thf) 3 (5 mol%) and TMEDA (5 mol%) in the presence of 1a (2 equiv.) in 1,2-dimethoxyethane (DME) at 50 C for 24 h, bromocyclopropane 3a was obtained in 98% yield with high trans (89%) selectivity (entry 1). Cyclopropanation at 25 C resulted in a slightly lower yield (81%) of 3a with almost the same trans selectivity (entry 1 vs. 2). No cyclopropanation product was obtained when organosilicon compounds 1b-d were used as the reducing reagents (entries 3-5), although 1b-d reduced CrCl 3 (thf) 3 to CrCl 2 , probably due to coordination of the reduction byproducts, 2,5-dimethylpyrazine (from 1b), pyrazine (from 1c), and 4,4 0 -bipyridyl (from 1d), to the chromium center, as conrmed by the inhibition of the catalytic reaction when pyrazine was added under the standard conditions. Screening of several multidentate nitrogen-based ligands revealed that TMEDA was the best ligand for this catalytic reaction (entry 1 vs. 12-17; amines, phosphines, and other ligands in ESI †). Notably, no reaction was observed when using typical organic and inorganic reductants, such as tetrakis(dimethylamino)ethylene (TDAE), Zn, and Mn powder (entries 6-8). The coordination of TMEDA to the chromium center was essentially required to produce the catalytic activity: the addition of either ZnCl 2 (2 equiv.) or MnCl 2 (2 equiv.) to the standard reaction conditions resulted in no reaction (entry 9) or lowered the yield of 3a (entry 10), respectively, due to the removal of TMEDA from the chromium center, 9a while under ligand-free conditions, the yield of 3a decreased signicantly (entry 11). When isolated CrCl 3 (tmeda) (5 mol%) was used as the catalyst, the yield of 3a was comparable with that of the in situ CrCl 3 (thf) 3 and TMEDA system (entry 18).
With the optimized reaction conditions in hand, we analyzed the substrate scope of the alkenes (Table 2). Allyl phenyl ether (2b) was converted to the corresponding bromocyclopropane 3b in 92% yield with high trans selectivity. Other allyl aryl ethers 2cg with electron-withdrawing and -donating substituents on the phenyl ring were transformed to the corresponding cyclopropanes 3c-g in moderate to high yields, with a cyano group or halogen atoms at the para-position of the aryl ring remaining intact during the cyclopropanation reaction. Reaction of CHBr 3 with allyl butyl ether (2h) afforded 3h in 81% yield with a trans-: cis ratio of 87 : 13. The carbonyl group also tolerated the reductive conditions to produce cyclopropanes; benzoylsubstituted alkene 2i was converted to 3i in 75% yield, while allyl carbonate 2j, which is typically used for allylic substitution of nucleophiles, afforded 3j in 60% yield without any decomposition of 2j. Allylamine 2k was also applicable and the corresponding cyclopropylmethylamine 3k was obtained in 64% yield. Simple a-olens, such as allylbenzene 2l, 5-hexenyl acetate 2m, 1octene 2n, and vinylcyclohexane 2o, gave the corresponding cyclopropanes 3l-o in good yield. When we applied substrates possessing two olenic moieties, a terminal and monosubstituted olenic part was selectively cyclopropanated to give 3p and 3q in moderate yield. Internal alkenes with cis-conguration were also applicable to our catalytic system: cis-1,4diacetoxy-2-butene (2r) showed a moderate reactivity to give the corresponding cyclopropane 3r in 47% yield, while some cyclic alkenes such as cycloheptene (2s), cyclooctene (2t) and acenaphthylene (2u) were applicable to afford polycyclic compounds 3s, 3t, and 3u in moderate to high yields, though debromination of initially formed bromocyclopropane might be involved for the formation of 3u. Other olens such as styrene, 1,1-disubstituted alkenes, acyclic internal alkenes, and dienes were not applicable in this cyclopropanation reaction (see ESI † for the limitations of this cyclopropanation).
To elucidate the reaction mechanism, we carried out a kinetic study for the formation of 3a, and the resulting data were analyzed by variable time normalization analysis (see ESI †). 15 The overall reaction rate did not change with various concentrations of chromium catalyst (0.004-0.01 M) and alkene 2a (0.08-0.12 M), giving a rate dependence of [Cr] 0 [2a] 0 , which is in sharp contrast to the report of Takai et al. who found that chromium-catalyzed cyclopropanation with Me 3 SiCHI 2 obeys rst-order dependence on the concentrations of both a chromium carbene complex and 2a, giving a rate dependence of [Cr] 1 [2a] 1 . 9a Such a difference was further observed in the reaction prole; no induction period was observed under the various reaction conditions. 16 Next, to understand how 1a functioned to generate a catalytically active species, we performed some control experiments. Direct activation of CHBr 3 by 1a was excluded because no significant rate acceleration was observed when a mixture of CHBr 3 and 1a was pre-treated by stirring at 50 C for 1 hour before adding the chromium catalyst (see ESI †). Although we tried repeatedly to isolate the dichromium species having a bridging bromomethyl group, the target complex could not be isolated and characterized, probably due to the instability of the bromomethyl-bridged dichromium species (see ESI †). In previous reports, however, gem-dichromiomethane complexes (Cr 2 -X) was isolated as key intermediates prior to generating reactive mononuclear carbene species via disproportionation (Fig. 2). Takai et al. reported the rst example of an isolated gem-dichromiomethane complex (Cr 2 -SiMe 3 ) by introducing a bulky trimethylsilyl-substituent on a carbon atom of diiodomethane, from which silylcyclopropanes were obtained upon treatment with alkenes. The related germanium derivative, Cr 2 -GeMe 3 , was also isolated and used for cyclopropanation. Anwander et al. independently observed the formation of a gem-dichromiomethane complex (Cr 2 -I) from the reaction of CrCl 2 and CHI 3 at low temperature.
We next conducted a stoichiometric cyclopropanation reaction of alkene 2a with bromoform in the presence of excess CrCl 2 (Scheme 2). The desired cyclopropane 3a was not obtained even at 80 C (Scheme 2a), although formation of the corresponding cyclopropanes was observed when iodoform and diiodomethane derivatives were used. Moreover, under the catalytic conditions using 1a, the yield of 3a gradually decreased as the catalyst loading was increased from 5 to 100 mol% (Scheme 2b). The lower product yield caused by increasing the amount of the chromium salt suggested that involvement of gem-dichromiomethane species was less likely in our metal-salt free system with 1a compared with other chromium-catalyzed cyclopropanation developed by Takai et al.
On the basis of these ndings, we propose the reaction mechanism shown in Scheme 3. The initial step is the activation of bromoform by chromium(II) species A to form (dibromomethyl)chromium(III) species B accompanied by the formation of an equimolar amount of chromium(III) trihalide C, which can be reduced by 1a or in situ-generated chromium(I) halide F (vide infra). Species B is dehalogenated by 1a to afford (bromomethylidene)chromium(III) D along with the elimination of Me 4pyrazine and 2 equiv. of Me 3 SiX (X ¼ Cl, Br), whose reactivity is assumed due to the reductive dehalogenation of vicinal  This journal is © The Royal Society of Chemistry 2020 Chem. Sci., 2020, 11, 3604-3609 | 3607 dihaloalkanes by the organosilicon-based reductant 1d leading to the formation of alkenes. 12a In addition, the generation of metal carbene species by the dehalogenation of metal carbenoids with zinc powder was proposed for nickel-or cobaltcatalyzed cyclopropanation of alkenes with dibromomethane or dichloromethane (Scheme 3). 8b,8h Finally, the reaction of D with alkenes gives 4-membered metallacycle E, whose reductive elimination affords the desired bromocyclopropane together with a low valent nascent chromium(I) species F. The resulting F reacts with chromium(III) trihalide C to regenerate chromium(II) species A through comproportionation. Accordingly, 1a has dual functions to reduce not only a catalyst precursor, CrCl 3 (tmeda), at the initial step, but also mainly the chromium(III) species B for generating mononuclear chromium carbene species D as a key intermediate.

Conclusions
In summary, we developed chromium-catalyzed bromocyclopropanation of alkenes with bromoform using an organosiliconbased reductant 1a. The desired bromocyclopropanes were obtained in moderate to high yields with good trans selectivity, and the reaction was applicable to allyl ether derivatives, allyl carbonate, allylamine, and simple a-olens. Control experiments suggested that 1a played an important role in reducing the (dibromomethyl)chromium(III) species to generate mononuclear (bromomethylidene)chromium(III) as a key intermediate. Further exploration to discover the unique metal salt-free reductive transformation of organic compounds is ongoing in our laboratory.

Conflicts of interest
The author declares no conict of interest.