HETEROCYCLES

Review | Regular issue | Vol. 92, No. 8, 2016, pp. 1373-1395
Received, 18th March, 2016, Accepted, 19th April, 2016, Published online, 3rd June, 2016.
DOI: 10.3987/REV-16-842

■ Development of Solvent-Driven Iron-Catalyzed Reactions

Toshiyuki Itoh*

Department of Chemistry and Biotechnology, Graduate School of Engineering, Tottori University, 4-101 Koyama-minami, Tottori 680-8552, Japan

Abstract

The possibility of solvent-driven iron-catalyzed reactions has been investigated using ionic liquids and acetonitrile as key solvents. Three iron-catalyzed reactions, the intramolecular cyclization of cyclopropane dithioacetals, the [2+2]-cycloaddition of (trans)-anethole, and the [2+3]-type cycloaddition of styrene derivatives with 1,4-benzoquinone, were first developed using acetonitrile as solvent in the presence of 3~5 mol% of Fe(ClO4)3•Al2O3 or Fe(BF4)2•6H2O under air conditions. In particular, we found the rapid cycloaddition of iron-catalyzed [2+3]-type cycloaddition of styrene derivatives with 1,4-benzoquinone when the reaction was carried out in an ionic liquid which contained [PF6] or [Tf2N] anion. Homo-coupling reaction of aryl or alkynyl Grignard reagent was next discovered using 1 mol% of FeCl3 as catalyst and the reaction also proceeded very rapidly in an ionic liquid. We also found that 3~5 mol% of Fe(ClO4)2•Al2O3 or Fe(BF4)2•6H2O-successfully catalyzed Friedel-Crafts type alkylation of indoles or pyrroles with vinyl ketones or α-aryl-β-silylalcohol. In particular, the reaction of chiral α-aryl-β-silylalcohol with indole using Fe(ClO4)3•nH2O as catalyst proceeded with retention of the configuration of stereochemistry of the hydroxyl group. Then, we discovered the first example of iron-catalyzed enantioselective C-S bond formation via Michael addition of thiols to (E)-3-crotonoyloxazolidin-2-one using 10 mol% of Fe(BF4)2/Pybox. We further demonstrated that Fe(ClO4)3•Al2O3-catalyzed the Nazarov type cyclizations of thiophene, pyrrole, indole, benzofuran, and benzo[b]thiophene derivatives using ionic liquids as solvent. Solvents have been recognized as a by-player in chemical reactions in the process of “optimization of reaction conditions”. However, further investigation of the solvent-driven iron-catalyzed reaction will allow discovery of useful reactions for organic syntheses.

1. INTRODUCTION
Iron is one of the most abundant and environmentally friendly metals on the earth and there have been various types of iron metal-catalyzed organic transformations during the past decades.1 Solvents provide a favorable environment for the intermediates of the chemical reaction and enhances the reactivity of the catalysts. However, solvents have been recognized as a by-player in chemical reactions in the process of “optimization of reaction conditions”. We have been investigating the possibility of iron-catalyzed reactions by focusing on the solvent not as a supporting actor of the reaction but as an educator that brings out a masked ability of the iron cation: we have thus developed several of them, i.e., the intramolecular cyclization of cyclopropane dithioacetals,2 the [2+2]-cycloaddition of (trans)-anethole,3 the [2+3]-type cycloaddition of styrene derivatives with 1,4-benzoquinone,4 the Michael addition of β-ketoesters to vinyl ketones,5 the enantioselective Michael addition of thiols to unsaturated carbonyl compounds,6 Friedel–Crafts type alkylation of indoles or pyrroles with vinyl ketones7 or α-aryl-β-silylalcohol,8 homo-coupling reaction of aryl or alkynyl Grignard reagent,9 and the Nazarov type cyclization of thiophene,10,11 pyrrole,12,13 indole,14 benzofuran,14and benzo[b]thiophene derivatives.14 In this chronicle, I report the story of how we developed these iron salt-catalyzed reactions from the beginning to the most recent ones.

2. IRON-CATALYZED REACTIONS BEGAN FROM A SINGLE ELECTRON OXIDATION
Kratochvil and Long reported that the reduction potential of the iron(III)-(II) couple depended on the solvent system which was estimated to be 1.57 volts in acetonitrile (MeCN).15 I was strongly fascinated by the results because “1.57 volts” was a sufficient level for the single electron oxidation of organic molecules. On the contrary, the oxidation potential of a Fe3+ ion shown in textbooks of Inorganic Chemistry was 0.77 volt which was slightly insufficient to oxidize common organic molecules to generate cation radical species. Inspired by the results, I had decided to launch the present project of “development of a solvent driven iron-catalyzed reaction” in 1998. It was by sheer chance for me that I read the article of Kratochvil and Long: an abstract of the paper was in a neighboring column on a page of Chemical Abstracts when I looked for papers concerning the oxidation level of various metal cations as I prepared a lecture in graduate school. We had a strong interest in the iron(III)-catalyzed reactions emanating from a single electron oxidation in MeCN or an ionic liquid (IL) solvent system. ILs have now attracted great interest because of their characteristics of low-volatility, non-flammability, high ion conductivity, and a wide potential window;16 however, they were not popular in those days. In fact, there were neither commercially available ILs nor any report referring to the oxidation potential of Fe3+/Fe2+ when we began the project; the answer was reported by Yamagata and Katayama et al. to be 1.1 V in an ionic liquid, 1-butyl-3-methylimidazolium bis(trifluoromethylsulfonyl)amide ([bmim][Tf2N]) in 2005 (Figure 1).17 We also anticipated that the Lewis acidity of Fe3+ ion might be increased in ILs: because ILs consisted of non- or less coordinated anions and organic cationic molecules so that iron cation might exist as a “naked cation” in an IL solvent.

Since there were no commercially available ILs in those days, we prepared ILs for the solvents of iron-catalyzed reactions ourselves and also established a method of recycling them.18 This experience became a valuable asset for our group because we accumulated much important know-how on to use ILs during the study.
We had started the present project by the investigating a single electron oxidation followed by the ring-opening reaction of cyclopropanes using MeCN as a solvent, because it was well known that the cyclopropane ring opened easily to generate a carbon radical.19 Several examples of ring-opening reactions of the cyclopropane ring had been reported (Figure 2). Booker-Milburn and co-workers reported that iron(III)-mediated oxidative ring opening of the cyclopropane compound 1 took place and gave the cyclized product 2 through the intermediate 1#, although the reaction required 3 eq. of iron(III) salts vs. substrate 1 to complete the reaction (Figure 2, left reaction).20 Takemoto and co-workers reported the similar reaction using a catalytic amount of an oxidant (0.5 eq. of (p-BrPh)3NSbCl6), and succeeded in obtaining bicyclopentane 4 from p-tolylthio-substituted cyclopropane 3 through cation radical 3# as illustrated in the right reaction in Figure 2. 21

Based on the results, we hypothesized that ring-opening of cyclopropanonetrimethylenethioacetal 722 might occur by treatment of iron(III) salt in MeCN as a solvent, because it was reported the reduction potential of dithioacetal was estimated to be 0.88 volt.23 The desired reaction was indeed realized when 7 was treated with 10 mol% of iron(III) perchlorate (Fe(ClO4)3•nH2O) in MeCN under air, subsequent cyclization with the olefinic part and reduction by an additional 1,3-dithiane giving cis-fused oxabicyclo[3.3.0]octane-1,1-trimethylenethioacetals 8 (Figure 3).2 It was revealed that the key solvent was MeCN and no reaction took place in dichloromethane (CH2Cl2), hexane, toluene, THF, or H2O.2 The plausible mechanism was suggested by the PM3 calculation illustrated in Figure 3: the single-electron oxidation of the sulfur atom on 7 to afford the cation radical 7-1# which easily generated 7-2# by the ring opening reaction and subsequent cyclization to afford 7-3# that formed the final intermediate 7-4#, which was finally reduced by 1,3-dithiane to give 8.

We next found [2+2] cyclodimerization of (trans)-anethole (9) and synthesized cyclobutane 10 in 92% yield with perfect anti-selectivity in the presence of 3 mol% of iron(III) salts (Figure 4):3 iron(III) perchlorate (Fe(ClO4)3•Al2O3) catalyzed the reaction and anti-adduct 10 was produced as the sole product due to the presence of equilibrium of the final intermediate syn-10# and anti-10#.

We next found [2+2] cyclodimerization of (trans)-anethole (9) and synthesized cyclobutane 10 in 92% yield with perfect anti-selectivity in the presence of 3 mol% of iron(III) salts (Figure 4):3 iron(III) perchlorate (Fe(ClO4)3•Al2O3) catalyzed the reaction and anti-adduct 10 was produced as the sole product due to the presence of equilibrium of the final intermediate syn-10# and anti-10#.

We obtained very important results for further investigation in this study: the reaction rate drastically accelerated when the reaction was carried out in ILs as solvents (Scheme 1 and Table 1). To our surprise, the reaction was completed in less than 10 min in 1-butyl-3-methylimidazolium hexafluorophosphate ([bmim][PF6]), while it took 2 hours in MeCN with the same concentration (entries 1 and 2 in Table 1).

It was found that even iron(II) dichloride (FeCl2) worked as a good catalyst in the IL (entry 4), though a long period was required to complete the reaction. On the contrary, no product was obtained for iron(II) sulfate (FeSO4) (entry 5). We finally found that iron(II) tetrafluoroborate (Fe(BF4)2•6H2O) worked the best in [bmim][PF6] as a solvent (entry 6). Since the reaction proceeded smoothly under dry air conditions and it was well known that ferrous ion was easily oxidized by air to generate ferric ion, the real catalyst was speculated to be Fe3+ cation which caused the initial oxidation of (trans)-anethole (9). However, since Fe(BF4)2•6H2O provided no product in MeCN (entry 7), the reactivity of this salt was strongly dependent on the solvent system. We further found that the reaction was also influenced by the counter-anion of the ILs employed as solvent. Very rapid reaction was also recorded for [bmim][Tf2N] (entry 9), while the reaction rate was dropped for [bmim][BF4] (entry 8), and no reaction occurred when [bmim][OTf] or [bmim][n-C5H11SO4] was used as solvent (entries 10 and 11). Rapid reaction was again observed in N,N-diethyl-N-methyl-N-(2-methoxy)ethylammonium bis(trifluoromethylsulfonyl)amide ([N221ME][Tf2N]) (entry 13), and the reaction rate declined for [N221MEM][BF4] (entry 12). It was thus found that the reaction rate depended on the anions of the ILs in the following order, PF6≃Tf2N>>BF4. These results indicated that the counter-anions of both the iron cation and the ILs might play the key role in realizing this [2+3]-cycloaddition.
Engler et al. proposed that 2,3-dihydrobenzofuran 15 was generated by [2+2]-cyclization of anethole 9 with benzoquinone 13 and subsequent rearrangement of the resulting cation intermediate.24 The reaction was stopped by addition of 1 eq. of TEMPO, then the positive solvatochromism was observed when the reaction was conducted in [N221ME][Tf2N]; these results indicated that a radical cation or anion intermediate existed in the reaction course.4c ILs are highly polar liquids, hence the environment provided by the IL might be appropriate for generation of a radical cation intermediate but not stabilize it because [PF6] or [Tf2N] anion is a less coordinative anion. Therefore, marked acceleration was accomplished. Fuchigami and co-workers reported that a charge-separated intermediate was destabilized by an IL and accelerated the corresponding reaction.25 Another mechanism might also be possible: the “naked Fe3+ cation” may coordinate with 1,4-benzoquinone 13 and accelerate electron transfer reaction to (trans)-anethole (9). Although we were the first to find the activation of the iron salts mediated reaction by ILs, we did not pursue this further. Lee, Song and co-workers reported many examples of activation of several Lewis acids using ILs three years later after our publication; the activity of a Lewis acid depended on the counter anions of the solvent ILs following the order PF6>Tf2N>>BF4.26,27,28 This was almost the same as that we obtained in the reaction of iron-catalyzed [2+3]-cyclization of (trans)-anethole with benzoquinone in ILs.
Styrene derivative 16 reacted with benzoquinone 13 in the presence of Fe(ClO4)3•Al2O3 in MeCN gave the adduct 17 in 84% yield. Then, the acetyl group of 17 was removed by methanolysis to afford 18 which was converted to optically active acetate 19 with >99% ee via lipase-catalyzed transesterification as shown in Figure 6.4b The aim to catalyze the reaction of acetate 16, Fe(ClO4)3•Al2O3/ MeCN resulted better results as a catalyst than that of Fe(BF4)2•6H2O/[bmim][PF6].

This work revealed an interesting property of ILs in activating of electron transfer type reactions. Since electron transfer is the essential starting point of many chemical reactions, we next focused on the possibility of ILs-driven iron-catalyzed reactions.
Recently iron-catalyzed cross coupling reactions have gained strong interest among synthetic chemists.30 Among such examples, we turned our attention to the FeCl3-mediated homocoupling reaction of aryl Grignard reaction reported by Hayashi31 and Cahiez et al.32 The authors proposed that the reaction included electron transfer steps and the key step might be a reduction step of iron (III) cation by aryl Grignard reagent or a transmetallation step on the iron.31a However, ILs have been considered inappropriate for strong base-mediated reactions. Clyburn and co-workers reported the first example of Grignard reaction in an IL.33 Since then, several groups, including ours, reported successful examples of the designs ILs for strong base-mediated reactions such as the Grignard reaction.34-36 We succeeded in designing an IL by introduction of alkyl ether moiety on the side arm of phosphonium salt: methoxyethyl(tri-n-butyl)phosphonium bis(trifluoromethanesulfonyl)imide ([P444ME][NTf2]) was thus developed.36 We applied this IL as solvent of the homo-coupling reaction of aryl Grignard reagent in the presence of 1 mol% of FeCl3 using 1,2-dichloroethane or 1,2-diiodoethane as an oxidant (Figure 7).9
We found that the desired coupling reaction of 20 proceeded very rapidly; 21 was obtained in less than 5 min at 0 °C in excellent yields (86 ~ 100%) except for bulky 2,4,6-tirimethylphenylMgBr.9 The reaction using 2-thienylMgBr 22 was very successful and the desired bithiophene 23 was obtained in 91% yield after just 10 min reaction at 25 °C. The IL solvent system was also useful for the homo coupling reaction of alkynyl Grignard reagent.9 In this case, use of 1,2-diiodoethene was essential to afford the coupling product, and we obtained the desired 25 in 80% yield.9

3. IRON-CATALYZED FRIEDEL–CRAFTS TYPE ALKYLATION IN AN IL SOLVENT SYSTEM
We next investigated iron-catalyzed Friedel–Crafts type alkylation in ILs, focusing in particular, on the use of indole7a,7c and pyrrole7b derivatives as donor molecules, because the products of the reactions contain important structural motifs found in natural products and are essential intermediates in the preparation of pharmaceuticals and other chemicals.
The Lewis acid-mediated Friedel–Crafts type alkylation of indole with unsaturated carbonyl compounds has gained growing interest.37 However, due to the basicity of indole, protection of the nitrogen atom in the indole ring was essential to realize such alkylation.37 In the previous chapter, we found that Fe(ClO4)3•Al2O3 and Fe(BF4)2•6H2O worked as excellent catalysts of cycloaddition reaction. We anticipated that these iron salts might work as an effective catalyst for Friedel–Crafts type alkylation of the N-protecting group–free indole. We found that the reaction indeed proceeded smoothly in MeCN or ILs as solvent, while no reaction took place when the reaction was conducted in CH2Cl2, CHCl3, Et2O, THF, toluene, or benzene in the presence of 3 mol% of Fe(BF4)2•6H2O under air conditions at rt.7 We thus obtained 27a in 92% yield when indole 26 (R= H) was reacted with 5-phenylpent-1-en-3-one (Figure 8).7a The reaction of indole with cyclopentenone afforded 27b in 86% yield.7a Unfortunately, the reaction in the presence of chiral ligands, such as the (R,R)-Jacobsen38 or (S,S)-bis-(oxazolinyl)pyridine (pybox) ligand,39 gave the product with a racemic state under our reaction conditions, 7a although Xia and Huang et al. recently reported chiral Brφnsted acid directed FeCl3-catalyzed enantioselective alkylation of indole.40
Control of regioselectivity of electrophilic substitution reaction of indole is an important task to be solved. We expected that regioselective electrophilic alkylation might be obtained using weak Lewis acid Fe(BF4)2•6H2O or Fe(ClO4)3•Al2O3 as a catalyst. As expected, the reaction showed perfect regioselectivity: alkylation took place at 3-position of the indole 26 to give 3-alkylated product 27: we thus succeeded in obtaining the indole derivatives 27a and 27b were obtained in 92% and 86%, respectively (Figure 8, upper).7c We further established that the second alkylation took place at 2-position of 27. After optimization of the reaction conditions, interestingly, solvent free conditions gave the best results: 7-methylindole 26c was reacted with hept-1-en-3-one in the presence of 3 mol% of Fe(ClO4)3•nH2O to form 3-alkylated compound 27c, after completion the reaction, addition of non-1-en-3-one with additional 5 mol% of Fe(ClO4)3•nH2O afforded the desired 2,3-double alkylated indole 28 (Figure 8, down).7c In the case, it was supposed that substrate indole and vinyl ketone worked as good solvents for the iron-catalyzed reactions.

The pyrrole group is found in many natural products as an important structural motif. However, since pyrrole is very sensitive to strong Lewis or Brφnsted acid, no electrophilic alkylation method had been reported when we launched the project. Although we tested the reaction of pyrrole with 1-buten-3-one in the presence of 3 mol% of FeCl3, neither alkylated product nor starting compound and only polymerized compounds were formed.7b We found that 2,5-dialkylated pyrrole and 2,3,5-trialkylated pyrrole was obtained in 37% and 12% yield, respectively, when the reaction was carried out in the presence of 3 mol% of Fe(BF4)2•6H2O in MeCN.7b After optimization of the reaction conditions of 2-acetylpyrrole 29, we established double alkylation of 2-acetylpyrrole with vinyl ketones using 3 mol% of Fe(ClO4)3•Al2O3 and succeeded in preparing tri-substituted pyrrole derivative 30a, 30b, and 30c in an 80%, 52%, and 78% yield, respectively (Figure 9).7b Solvent free conditions again gave the best results, though the reaction itself proceeded very smoothly in an IL, [bmim][Tf2N], because extraction of the products from the IL layer was unsuccessful using conventional organic solvents such as Et2O, cyclohexane, ethyl acetate or CH2Cl2.

We next turned our interest to the Friedel–Crafts alkylation using α-aryl alcohols as alkylating agents,41 because several leading works had been reported.42 We found that the cationic iron salt Fe(BF4)2•6H2O was a very effective catalyst for the Friedel–Crafts alkylation of indole derivatives 31 with β-silyl- alcohols 32. The combination of indole derivatives with 2-furyl and 2-thienyl derivatives using Fe(BF4)2•6H2O was the best system and the corresponding products 33a to 33d were obtained in high yield; on the other hand, Fe(ClO4)3•nH2O gave better results for the reaction of phenyl derivative of β-silylalcohol, and 33h was obtained in good yield (Figure 10).8 However, no reaction took place for the reaction of 5-cyano- or nitro-1H-indole. In addition, reaction of indole with 1-(pyridin-2-yl)-2-(trimethylsilyl)ethanol also gave no desired product (Figure 10).

During the course of this study, we discovered that the Friedel–Crafts alkylation of chiral α-arylalcohols with a trimethylsilyl group was found to proceed with retention of the configuration of the hydroxy group as a leaving group (Figure 11).8 Although the memory of chirality is useful in carbanion chemistry,43 such reaction in carbocation chemistry is rare and there is no general means of achieving it.44 Ugi and co-workers first reported the retentive electrophilic substitution of the α-substituted alkylferrocenes.45 Another example is the electrochemical generation of the N-acyliminium ion and the in situ trap with a nucleophile reported by Matsumura and Onomura et al.46 To our surprise, chiral alkylation product (+)-(R)-33e (91% ee) was obtained when we subjected chiral β-silylalcohol (+)-(S)-32e (92% ee) to the reaction of indole 31a (Figure 11).8 The result indicated that the Friedel–Crafts alkylation of chiral β-silylalcohols proceeded with retention of the configuration. After optimization of the catalyst, we found that the same level of the memory chirality reaction was observed with other Lewis acids such as AlCl3, FeCl3, and Sc(OTf)3, but it was confirmed that our iron catalyst was most effective to realize the reaction.
The stereoselective reactions of carbocations have gained strong interest among synthetic chemists.47 Bach and co-workers reported syn-selective Friedel–Crafts alkylation at the benzylic position of 2,3,3-trimethyl-1-phenylbutan-1-ol using 1.2 molar equivalent of strong Lewis or Brφnsted acid through benzylic carbocation intermediate,48 and they also reported successful diastereoselective reactions via carbocation intermediates.48 But this reaction, however, is the diastereoselective reaction but the memory chirality reaction: we are now assuming that the origin of the reaction stems from the carbocation intermediate which is conformationally stable.8

4. IRON-CATALYZED CONJUGATED REACTION
Conjugate addition is one of the most important bond forming reactions in synthetic organic chemistry. The asymmetric conjugate addition of thiols to α,β-unsaturated substrates is also a very useful reaction because it produces optically active compounds such as pharmaceuticals.49 The first example was reported by Kanemasa et al. in 1999: Ni/DBFOX catalyzed the highly enantioselective reactions.50 However, there have been only limited examples of the transition metal catalyzed enantioselective addition of thiols to (E)-3-crotonoyloxazolidin-2-one. We found the chiral cobalt or iron catalyzed asymmetric conjugate addition of aromatic thiols (34) to (E)-3-crotonoyloxazolidin-2-one (35) (Figure 12).6 Several types of thiols were reacted with (E)-3-crotonoyloxazolidin-2-one to provide the desired chiral adducts with high optical purities (Figure 12).6 Aiming at obtaining high enantioselectivity, the reaction was conducted at a lower temperature (–20 ˚C) in the presence of molecular sieves 4Å and we thus succeeded in obtaining the addition products, 37a ~ 37e with high optical purity. The highest enantioselectivity (95% ee) was recorded for the reaction of the sterically hindered 2-methylbenzenethiol (37d), though it required a longer time to complete the reaction (Figure 12).6b

The 1,4-addition of appropriate donor molecules such as active methylene compounds with α,β-unsaturated carbonyl compounds is well recognized as an important C-C bond forming reaction in organic synthesis.51 Transition-metal catalysis of the 1,4-additions have recently been developed as a valuable alternative to the classic base catalyzed reaction.51 We found that the 1,4-addition reaction of β-ketoester 38 with vinyl ketones was accomplished by Fe(BF4)2•6H2O as catalyst to afford desired product 39a ~ 39c in modest (45%) to excellent (93%) yield, and the catalyst was used repeatedly more than 10 times in an ionic liquid solvent system (Figure 13).5

To gain the insight of the reaction mechanism in the present reaction, we tested the additive effect of 2,2,6,6-tetramethyl-1-piperidinyloxyl free radical (TEMPO). Interestingly, the inhibitory action of TEMPO was significantly dependent on the reaction medium. The reaction was completely inhibited by TEMPO with an equal amount of the catalyst (3 mol%) in MeCN. On the other hand, 1.0 eq. of TEMPO was needed to inhibit the reaction in the ionic liquid solvent. These results may indicate that the catalytic mode of Fe(BF4)2 depends on the solvent system. Although we attempted to isolate TEMPO-iron complex, we isolated neither TEMPO-substrate complex nor TEMPO-Fe salt complex from the reaction mixture.5

5. IRON-CATALYZED NAZAROV CYCLIZATION
Nazarov cyclization is promoted by Lewis or Brφnsted acids and is known as an important method for the preparation of five-membered carbocycles.52 Iron(III) chloride (FeCl3) is well recognized as the most typical strong Lewis acid and Denmark and Jones reported the first example of the FeCl3-catalyzed Nazarov cyclization;53 Kang and co-workers extended the catalytic reaction.54 Initially, we gave our attention to the use of FeCl3 for the Nazarov cyclization of the 3-substituted thiophene derivative 40, because thiophenes are respected as important components of many biologically active compounds and some optical materials. We established that the combination of FeCl3 (10 mol%) with toluene as a solvent was the best system to afford the desired product when the reaction was carried out at 60 °C (Figure 14).10 The reaction of substrate which has an electron donating group, such as the phenyl, p-tolyl, or p-methoxyphenyl group proceeded smoothly and gave the cyclized products 41a, 41b, and 41c in 90%, 82% and 83%, respectively. On the other hand, the reaction rates of substrates that have an electron withdrawing group or aliphatic group, such as p-cyanophenyl or cyclohexyl group, were significantly lower and 41d and 41e were obtained in 52% and 47% yield, respectively (Figure 14).10

We next found that the cyclized product 43a was obtained when substrate 42 (Ar= Ph) was heated at 60 °C in the presence of 20 mol% of Fe(OTf)2 with (S,S)-tb-pybox (L2) in a mixed solvent of CH2Cl2-hexane (2:1) with 83% ee.11 The enantioselectivity was increased when 42b (Ar= 4-CF3-C6H4) was subjected to the reaction using Fe(OTf)2/L to give 43b in 53% yield with 89% ee (Figure 15). The best enantioselectivity was recorded for 42c (Ar= 2-naphthyl) and 43c was obtained in 40% yield with 93% ee (Figure 15).11 This example was the first iron-catalyzed asymmetric Nazarov cyclization.

The pyrrole moiety is common in many biologically-active or functional molecules. Therefore, we next focused on the iron salts-catalyzed Nazarov cyclization of pyrrole derivatives. We anticipated that the key point to realize the desired Nazarov cyclization might be use of a weaker Lewis acid as mentioned earlier.
It was also assumed that the intermediate of Nazarov cyclization might be favorable in a highly polar IL solvent. Therefore, we tested the reaction using various combination of iron salts with solvents using pyrrole derivative 44 as a model substrate. The results were dependent on the combinations of iron salt with solvent but independent of the stereochemistry of the starting substrate: the desired cyclized product 45 was obtained in excellent yield from (E)-44 and (Z)-44 in 92% yield and 90% yield, respectively, when the reaction was conducted in the presence of 5 mol% of Fe(ClO4)3•Al2O3 in CH2Cl2 at 60 °C (Figure 16).12

An X-ray crystallographic analysis of the cyclized product 45 revealed the trans relationship between the α-carbomethoxy group and the β-phenyl group.12 The solvent choice was very important in the reaction and CH2Cl2 or the ionic liquids, 1-butyl-3-methylimidazolium bis(trifuluoromethyl)sulfonylamide ([bmim][NTf2]), tributyl(2-methoxyethyl)phosphonium bis(trifuluoromethyl)sulfonylamide ([P444ME][NTf2]), or N,N-diethyl-N-(2-methoxyethyl)-N-methylammonium bis(trifuluoromethyl)sulfonylamide ([N221ME][NTf2]) provided excellent results.12 Since avoiding the use of a haloalkane solvent is now a strong requirement from an environmental aspect, this feature is a very important element in our reaction.
As mentioned earlier, iron-catalyzed 1,4-addition reaction of β-ketoester with vinyl ketone was possible. In fact, we found that the compound 45 further reacted with but-3-en-2-one to afford the corresponding Michael product 46a with an excellent trans selectivity in 86~90% yield (Figure 17).12,13 Hence, we next attempted tandem type Nazarov cylization and 1,4-addition; the yield of desired 46a was only 32% due to formation of further alkylated compounds on the pyrrole moiety. Desired 46a (R1= Me) and 46b (R1= H) was obtained in 86% and 87% yield, respectively, when the reaction was conducted as a step-by-step method (Figure 17).

We then investigated the possibility of recycling the iron catalyst for the Nazarov cyclization/Michael reaction using an ionic liquid, [bmim][Tf2N] as solvent. We initially encountered an unexpected difficulty; isolation of the product 46a (R= Me) from the reaction mixture was unsuccessful because 46a was very soluble in [bmim][NTf2]. We finally resolved this by modification of the substrate and using the n-pentyl ester or benzyl ester instead of the methyl ester. Extraction of the n-pentyl ester 46c from the IL solvent was successful and we demonstrated repetition of the reaction using [bmim][NTf2] solution as follows: the reaction course was monitored by silica gel thin layer chromatography analysis and but-3-en-2-one was added when the spot of 44 had been consumed and the spot of 45 appeared, the reaction mixture was then stirred at 60 °C until the spot of 45 disappeared completely. The desired product 46c was isolated as the sole stereo isomer in 78% yield by a mixed solvent (hexane and Et2O= 2:1). The Nazarov/Michael product 46c was obtained in 66% yield after repetition of the reaction five times without any addition of the catalyst (Figure 17).13
The Nazarov cyclization of indole, benzofuran, and benzo[b]thiophene was also accomplished in the presence of catalytic amounts of the iron salts, though we had to choose an appropriate iron salt and solvent for the substrates. For the indole, the corresponding cyclization product 47 was obtained using Fe(ClO4)3•Al2O3 in [bmim][Tf2N] in 85% yield. On the other hand, 10 mol% of FeCl3 was essential to obtain 48 or 49 as catalyst (Figure 18).14 Since the iron catalyst remained in the IL layer, recycling the catalyst was, of course, possible for the Nazarov cyclization of indole derivatives. We thus succeeded in obtaining 47 in 79% yield after five repetitions of the reaction without any addition of the catalyst to the [bmim][NTf2] solvent system.14

6. CONCLUSION
The most interesting feature of iron is that it can act as various catalyst types for chemical reactions, the changing of counter anion and solvents. As described in this chronicle, we succeeded in developing many types of iron-catalyzed reactions (Figure 19). As mentioned in the introduction, the project took from when I recognized a paper by Kratochvil and Long15 of a reading room in the library of Okayama University. Since then we have been attempting to expand the scope of iron-catalyzed reaction using ionic liquids or acetonitrile as key solvents as summarized in Figure 19. Further investigation will allow us to develop further iron-catalyzed reactions by focusing on the solvent not as a supporting actor of the reaction but as an educator that brings out a masked ability of the iron cation.

ACKNOWLEDGEMENTS
The author is most grateful to Dr. Hiroyuki Ohara, Professors Eiichi Nakamura (The University of Tokyo) and Masaharu Nakamura (Kyoto University) for their great contribution to the early studies of the present project. He is also grateful to Professor Motoi Kawatsura of Nihon University who was an Associate professor of the author’s group and his great contribution to the Nazarov cyclization project, and also to Associate Professor Toshiki Nokami of Tottori University for his effort in completing the chirality retentive Friedel–Crafts alkylation. Finally, the author expresses his deep appreciation to all students (K. Kudo, H. Kiyokane, K. Kawai, H. Uehara, Y. Komatsu, M. Yamamoto, S. Nomura, K. Ogiso, J. Kobayashi, S. Matsui, K. Kude, M. Fujiwara, Y. Higuchi, K. Kajita, S. Hayashi, C. Ibara, M. Sakae, M. Natsuyama, Y. Yamane, and S. Oshitani) who pursued this project with him.

References

1. (a) C. Bolm, L. Legros, J. Le Paih, and L. Zani, L. Chem. Rev., 2004, 104, 6217; CrossRef (b) A. Correa, O. G. Mancheno, and C. Bolm, Chem. Soc. Rev., 2008, 37, 1108; CrossRef (c) E. B. Bauer, Curr. Org. Chem., 2008, 12, 1341; CrossRef (d) A. O. Sarhan, Abdelwareth, and C. Bolm, Chem. Soc. Rev., 2009, 38, 2730; CrossRef (e) K. Mori, J. Synth. Org. Chem. Jpn., 2010, 68, 75; CrossRef (f) D. Bezier, J-B. Sortais, and C. Darcel, Adv. Synth. Catal., 2013, 355, 19. CrossRef
2. H. Ohara, K. Kudo, T. Itoh, M. Nakamura, and E. Nakamura, Heterocycles, 2000, 52, 505. CrossRef
3. H. Ohara, T. Itoh, M. Nakamura, and E. Nakamura, Chem. Lett., 2001, 624. CrossRef
4. (a) H. Ohara, H. Kiyokane, and T. Itoh, Tetrahedron Lett., 2002, 43, 3041; CrossRef (b) T. Itoh, K. Kawai, S. Hayase, and H. Ohara, Tetrahedron Lett., 2003, 44, 4081; CrossRef (c) T. Itoh, H. Uehara, K. Kawai, S. Hayase, H. Ohara, and M. Oyama, ‘Analytical Mechanistic and Synthetic Organic Electrochemistry-6th International M. Baizer Award Symposium in Honor of Dennis H. Evans and Masao Tokuda’, ed. by J. Lessard, P. Hapiot, and I. Nishiguchi, The Electrochemical Society, Inc. USA, 2004, pp. 9-12.
5. H. Uehara, S. Nomura, S. Hayase, M. Kawatsura, and T. Itoh, Electrochemistry, 2006, 74, 635. CrossRef
6. , 　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　; CrossRef (b) M. Kawatsura, Y. Komatsu, M. Yamamoto, S. Hayase, and T. Itoh, Tetrahedron, 2008, 64, 3488. CrossRef
7. (a) T. Itoh, H. Uehara, K. Ogiso, S. Nomura, S. Hayase, and M. Kawatsura, Chem. Lett., 2007, 36, 50; CrossRef (b) M. Kawatsura, M. Fujiwara, H. Uehara, S. Nomura, S. Hayase, and T. Itoh, Chem. Lett., 2008, 37, 794; CrossRef (c) J. Kobayashi, S. Matsui, K. Ogiso, S. Hayase, M. Kawatsura, and T. Itoh, Tetrahedron, 2010, 66, 3917. CrossRef
8. T. Nokami, Y. Yamane, S. Oshitani, J. Kobayashi, S. Matsui, T. Nishihara, H. Uno, S. Hayase, and T. Itoh, Org. Lett., 2015, 17, 3182. CrossRef
9. K. Kude, S. Hayase, M. Kawatsura, and T. Itoh, Heteroat. Chem., 2011, 22, 397. CrossRef
10. M. Kawatsura, Y. Higuchi, S. Hayase, M. Nanjo, and T. Itoh, Synlett, 2008, 1009. CrossRef
11. M. Kawatsura, K. Kajita, S. Hayase, and T. Itoh, Synlett, 2010, 1243. CrossRef
12. M. Fujiwara, M. Kawatsura, S. Hayase, M. Nanjo, and T. Itoh, Adv. Synth. Catal., 2009, 351, 123. CrossRef
13. C. Ibara, M. Fujiwara, S. Hayase, M. Kawatsura, and T. Itoh, Sci. China Chem., 2012, 55, 1627. CrossRef
14. M. Sakae, S. Oshitani, C. Ibara, M. Natsuyama, T. Nokami, and T. Itoh, Heteroat. Chem., 2014, 25, 482. CrossRef
15. B. Kratochvil and R. Long, Anal. Chem., 1970, 42, 43. CrossRef
16. (a) N. V Plechkova and K. R. Seddon, Chem. Soc. Rev., 2008, 37, 123; CrossRef (b) J. P. Hallett and T. Welton, Chem. Rev., 2011, 111, 3508. CrossRef
17. (a) M. Yamagata, N. Tachikawa, Y. Katayama, and T. Miura, Electrochemistry, 2005, 73, 564; (b) M. Yamagata, N. Tachikawa, Y. Katayama, and T. Miura, Electrochim. Acta, 2007, 52, 3317. CrossRef
18. T. Itoh, J. Synth. Org. Chem. Jpn., 2009, 67, 143. CrossRef
19. M. Fedorynski, Chem. Rev., 2003, 103, 1099. CrossRef
20. (a) K. I. Booker-Milburn, Synlett, 1992, 809; CrossRef (b) K. I. Booker-Milburn, and D. F. Thompson, Synlett, 1993, 592; CrossRef (c) K. I. Booker-Milburn and D. F. Thompson, J. Chem. Soc., Perkin Trans. 1, 1995, 2315; CrossRef (d) K. I. Booker-Milburn, B. Cox, and T. E. Mansley, Chem. Commun., 1996, 2577; CrossRef (e) K. I. Booker-Milburn and R. F. Dainty, Tetrahedron Lett., 1998, 39, 5097; CrossRef (f) K. I. Booker-Milbum, A. Baker, and W. Braisford, Tetrahedron Lett., 1998, 39, 4373. CrossRef
21. Y. Takemoto, S-i. Furuse, H. Koike, T. Ohra, C. Iwata, and H. Ohishi, Tetrahedron Lett., 1995, 36, 4085. CrossRef
22. M. Nakamura, M. Toganoh, H. Ohara, and E. Nakamura, Org. Lett., 1999, 1, 7. CrossRef
23. Y. Deng, A. J. Illies, M. A. James, M. L. McKee, and M. Peschke, J. Am. Chem. Soc., 1995, 117, 420. CrossRef
24. T. A. Engler, M. A. Letavic, and J. P. Reddy, J. Am. Chem. Soc., 1991, 64, 2860; T. A. Engler, K. D. Combrink, M. A. Letavic, K. O. Lynch, Jr., and J. E. Ray, J. Org. Chem., 1994, 59, 6567. CrossRef
25. T. Sunaga, M. Atobe, S. Inagi, and T. Fuchigami, Chem. Commun., 2009, 956. CrossRef
26. B. Y. Park, K. Y. Ryu, J. H. Park, and S.-g. Lee, Green Chem., 2009, 11, 946. CrossRef
27. J. W. Lee, J. Y. Shin, Y. S. Chun, H. B. Jang, C. E. Song, and S.-g. Lee, Acc. Chem. Res., 2010, 43, 985. CrossRef
28. (a) C. E. Song, W. H. Shim, E. J. Roh, and S.-g. Lee, Angew. Chem. Int. Ed., 2004, 43, 6183; CrossRef (b) M. Y. Yoon, J. H. Kim, D. S. Choi, U. S. Shin, J. Y. Lee, and C. E. Song, Adv. Synth. Catal., 2007, 349, 1725; CrossRef (c) J. H. Kim, J. W. Lee, U. S. Shin, J. Y. Lee, S.-g. Lee, and C. E. Song, Chem. Commun., 2007, 4683; CrossRef (d) D. S. Choi, J. H. Kim, U. S. Shin, R. R. Deshmukh, and C. E. Song, Chem. Commun., 2007, 3482. CrossRef
29. C. -S. Chen, Y. Fujimoto, G. Girdauskas, and C. J. Sih, J. Am. Chem. Soc., 1982, 112, 7294. CrossRef
30. (a) G. Cahiez and H. Avedissian, Synthesis, 1998, 1199; CrossRef (b) A. Fürstner, A. Leitner, M. Méndez, and H. Krause, J. Am. Chem. Soc., 2002, 124, 13856; CrossRef (c) C. Duplais, F. Bures, T. Korn, I. Sapountzis, G. Cahiez, and P. Knochel, Angew. Chem. Int. Ed., 2004, 43, 2968; CrossRef (d) M. Nakamura, K. Matsuo, S. Ito, and E. Nakamura, J. Am. Chem. Soc., 2004, 126, 3686; CrossRef (e) R. B. Bedford, D. W. Bruce, R. Frost, and M. Hird, Chem. Commun., 2005, 4161; CrossRef (f) G. Cahiez, V. Habiak, C. Duplais, and A. Moyeux, Angew. Chem. Int. Ed., 2007, 46, 4364; CrossRef (g) G. Cahiez, C. Duplais, and A. Moyeux, Org. Lett., 2007, 9, 3253; CrossRef (h) T. Hatakeyama and M. Nakamura, J. Am. Chem. Soc., 2007, 129, 9844; CrossRef (i) T. Hatakeyama, S. Hashimoto, K. Ishizuka, and M. Nakamura, J. Am. Chem. Soc., 2009, 131, 11949; CrossRef (j) F. Vallée, J. J. Mousseau, and A. B. Charette, J. Am. Chem. Soc., 2010, 132, 1514; CrossRef (k) W. Liu, Angew. Chem. Int. Ed., 2010, 49, 1278; CrossRef (l) D. Castagnolo and M. Botta, Eur. J. Org. Chem., 2010, 3224; CrossRef (m) A. Nagaki, Y. Uesugi, Y. Tomida, and J-i. Yoshida, Beilstein J. Org. Chem., 2011, 7, 1064; CrossRef (n) O. M. Kuzmina, A. K. Steib, D. Flubacher, and P. Knochel, Org. Lett., 2012, 14, 4818; CrossRef (o) D. Toummini, F. Ouazzani, and M. Taillefer, Org. Lett., 2013, 15, 4690; CrossRef (p) S. Kawamura and M. Nakamura, Chem. Lett., 2013, 42, 183; CrossRef (q) O. M. Kuzmina, A. K. Steib, S. Fernandez, W. Boudot, J. T. Markiewicz, and P. Knochel, Chem. Eur. J., 2015, 21, 8242. CrossRef
31. (a) T. Nagano and T. Hayashi, Org. Lett., 2004, 6, 1297; CrossRef (b) T. Nagano and T. Hayashi, Org. Lett., 2005, 7, 491. CrossRef
32. (a) G. Cahiez, C. Chaboche, F. Mahuteau-Betzer, and M. Ahr, Org. Lett., 2005, 7, 1943; CrossRef (b) G. Cahiez, A. Moyeux, J. Buendia, and C. Duplais, J. Am. Chem. Soc., 2007, 129, 13788. CrossRef
33. T. Ramnial, D. D. Ino, and J. A. C. Clyburne, Chem. Commun., 2005, 325. CrossRef
34. (a) V. Jurcik and R. Wilhelm, Green Chem., 2005, 7, 844; CrossRef (b) S. T. Handy, J. Org. Chem., 2006, 71, 4659. CrossRef
35. M. C. Law, K.-Y. Wong, and T. H. Chan, Chem. Commun., 2006, 2457. CrossRef
36. T. Itoh, K. Kude, S. Hayase, and M. Kawatsura, Tetrahedron Lett., 2007, 48, 7774. CrossRef
37. (a) G. Bartoli, M. Bartolacci, M. Bosco, G. Foglia, A. Giuliani, E. Marcantoni, L. Sambri, and E. Torregiani, J. Org. Chem., 2003, 68, 4594; CrossRef (b) I. Komoto and S. Kobayashi, J. Org. Chem., 2004, 69, 680; CrossRef (c) H. Firouzabadi, N. Iranpoor, and F. Nowrouzi, Chem. Commun., 2005, 789; CrossRef (d) Z.-P. Zhan and K. Lang, Synlett, 2005, 1551; CrossRef (e) Z.-P. Zhan, R.-F. Yang, and K. Lang, Tetrahedron Lett., 2005, 46, 3859; CrossRef (f) W.-J. Li, X.-F. Lin, J. Wang, G.-L. Li, and Y.-G. Wang, Synlett, 2005, 2003; (g) Y.-H. Gao, L. Yang, W. Zhou, L.-W. Xu, and C.-G. Xia, Appl. Oraganometal. Chem., 2009, 23, 114; CrossRef (h) R. Jiang, X.-J. Wu, X. Zhu, X.-P. Xu, and S.-J. Ji, Eur. J. Org. Chem., 2010, 5946. CrossRef
38. W. Zhang, J. L. Loebach, S. R. Wilson, and E. N. Jacobsen, J. Am. Chem. Soc., 1990, 112, 2801. CrossRef
39. H. Nishiyama, M. Kondo, T. Nakamura, and K. Itoh, Organometallics, 1991, 10, 500. CrossRef
40. L. Yang, Q. Zhu, S. Guo, B. Qian, C. Xia, and H. Huang, Chem. Eur. J., 2010, 16, 1638. CrossRef
41. G. A. Olah, R. Krishnamurti, and G. K. S. Prakash, ‘Friedel–Crafts Alkylations’, Comprehensive Organic Synthesis, ed. by B. M. Trost, Pergamon, Oxford, 1991, Vol. 3, 229.
42. (a) M. Noji, T. Ohno, K. Fuji, N. Futaba, H. Tajima, and K. Ishii, J. Org. Chem. 2003, 68, 9340; CrossRef (b) M. Yasuda, T. Somyo, and A. Baba, Angew. Chem. Int. Ed., 2006, 45, 793; CrossRef (c) M. Niggemann and J. M. Meel, Angew. Chem. Int. Ed., 2010, 49, 3684. CrossRef
43. T. Kawabata and K. Fuji, J. Synth. Org. Chem. Jpn., 2000, 58, 1095. CrossRef
44. G. A. Olah and D. A. Klumpp, ‘Superelectrophiles and Their Chemistry’, John Wiley: Hoboken, New Jersey, 2008.
45. (a) D. Marquarding, H. Klisacek, G. Gokel, P. Hoffmann, and I. Ugi, J. Am. Chem. Soc., 1970, 92, 5389; CrossRef (b) W. G. Gokel, D. Marquarding, and I. Ugi, J. Org. Chem., 1972, 37, 3052. CrossRef
46. (a) Y. Matsumura, Y. Shirakawa, Y. Satoh, M. Umino, T. Tanaka, T. Maki, and O. Onomura, Org. Lett., 2000, 2, 1689; CrossRef (b) N. G. Wanyoike, O. Onomura, T. Maki, and Y. Matsumura, Org. Lett., 2002, 4, 1875. CrossRef
47. P. G. Cozzi and F. Benfatti, Angew. Chem. Int. Ed., 2010, 49, 256. CrossRef
48. (a) F. Mühlthau, O. Schuster, and T. Bach, J. Am. Chem. Soc., 2005, 127, 9348; CrossRef (b) F. Mühlthau, D. Stadler, A. Goppert, G. A. Plah, G. K. S. Pralash, and T. Bach, J. Am. Chem. Soc., 2006, 128, 9668; CrossRef (c) D. Stadler and T. Bach, Angew. Chem. Int. Ed., 2008, 47, 7557; CrossRef (d) D. Stadler and T. Bach, Chem. Asian J., 2008, 3, 272; CrossRef (e) D. Stadler, A. Goeppert, G. Rasul, G. A. Olah, G. K. S. Prakash, and T. Bach, J. Org. Chem., 2009, 74, 312; CrossRef (f) D. Nitsch, S. M. Huber, A. Poething, A. Narayanan, G. A. Olah, G. K. S. Prakash, and T. Bach, J. Am. Chem. Soc., 2014, 136, 2851. CrossRef
49. (a) C. Fraschetti, F. R. Novara, A. Filippi, N. A. Trout, W. Adcock, T. S. Sorensen, and M. Speranza, J. Org. Chem., 2007, 72, 4077; CrossRef (b) J. Y. L. Chung, D. Mancheno, P. G. Dormer, N. Variankaval, R. G. Ball, and N. N. Nsou, Org. Lett., 2008, 10, 3037; CrossRef (c) P -S. Lai, J. A. Dubland, M. G. Sarwar, M. G. Chudzinski, and M. S. Taylor, Tetrahedron, 2011, 67, 7586. CrossRef
50. S. Kanemasa, Y. Oderaotoshi, and E. Wada, J. Am. Chem. Soc., 1999, 121, 8675. CrossRef
51. (a) J. Christoffers and A. Baro, Angew. Chem. Int. Ed., 2003, 42, 1688; CrossRef (b) M. P. Sibi and S. Manyem, Tetrahedron, 2000, 56, 8033; CrossRef (c) J. Christoffers, Eur. J. Org. Chem., 1998, 1259; CrossRef (d) O. M. Berner, L. Tedeschi, and D. Enders, Eur. J. Org. Chem., 2002, 1877; (e) N. Krause and A. Hoffmann-Röder, Synthesis, 2001, 171; CrossRef (f) P. I. Dalko and L. Moisan, Angew. Chem. Int. Ed., 2004, 43, 5138. CrossRef
52. (a) M. A. Tius, Eur. J. Org. Chem., 2005, 2193; CrossRef (b) H. Pellissier, Tetrahedron, 2005, 61, 6479; CrossRef (c) W. T. Spencer III, T. Vaidya, and A. J. Frontier, Eur. J. Org. Chem., 2013, 3621; CrossRef (d) M. A. Tius, Chem. Soc. Rev., 2014, 43, 2979; CrossRef (e) T. Itoh, T. Nokami, and M. Kawatsura, Chem. Rec., 2016, 16, in press. CrossRef
53. S. E. Denmark and T. K. Jones, J. Am. Chem. Soc., 1982, 104, 2642. CrossRef
54. K-T. Kang, S. S. Kim, J. C. Lee, and J. S. U, Tetrahedron Lett., 1992, 33, 3495. CrossRef