Chem. Commun., 2006, (Advance Article) DOI: 10.1039/b600568n	Feature Article
Discovery of indium complexes as water-tolerant Lewis acids

Teck-Peng Loh* and Guan-Leong Chua
Division of Chemistry and Biological Chemistry, School of Physical and Mathematical Sciences, Nanyang Technological University, No. 1 Nanyang Walk, Block 5 Level 3, Singapore. E-mail: teckpeng@ntu.edu.sg; Fax: +65 6791 1961; Tel: +65 6316 8899

The work on development of organic reactions that are tolerant to air and moisture, especially that of carbon–carbon bond forming reactions, had only started in the early 1990s. Our laboratory had approached this Course when it was still in its infancy and had developed methodologies for that end. In this article, our decade of work on indium complexes as Lewis acids is summarised.

Loh Teck-Peng

Teck-Peng Loh is a Professor of chemistry in Nanyang Technological University, School of Physical and Mathematical Sciences. He was born in Malaysia in 1962. He received his BSc (1987) and MSc (1989) from the Tokyo Institute of Technology. He had worked with Professor Takeshi Nakai and Professor Koichi Mikami for his research work there. Under the tutelage of Professor E. J. Corey, Professor Loh obtained his PhD (1994) from Harvard University. In 1994, he joined the National University of Singapore as a Lecturer, in 1999 as an Associate Professor and in 2004 as a full Professor. Professor Loh then moved to Nanyang Technological University in 2005 and is currently a Professor and the Head of Division (Chemistry and Biological Chemistry, School of Physical and Mathematical Sciences). His research interests are in the area of enantioselective reactions, development of new synthetic methodologies especially in the area of Green chemistry, and the total synthesis of natural products.

Guan-Leong Chua

Guan-Leong Chua was born in Singapore in 1971. He obtained his BSc (1992) and PhD (2001) from the National University of Singapore, working under the guidance of Professor Loh Teck-Peng. He had worked in industry for five years before joining Nanyang Technological University in 2005 as a Lecturer. His research interest is in synthesis methodologies.

Introduction

The traditional approach to catalysis using Lewis acids is changing rapidly from single-use, air- and moisture-sensitive metal complexes such as AlCl₃ and TiCl₄, to reusable and highly tolerant catalysts. Notably, salts from main group elements¹ are featured prominently in this progressive move towards mild and facile reactions, with sights set on the concept of green chemistry. Procedures that require organic solvents and toxic reagents are slowly being phased out. Above all, many laboratories are focussing on carbon–carbon bond forming reactions in the presence of air and water,^2,3 building the backbone of complex molecules with methodologies not conceivable in Victor Grignard's time. Very often, the total exclusion of organic solvents in every step of the synthesis protocol for most green methodologies is still not possible. However, the developments so far had established reaction procedures that are easy to execute, and such procedures often come with the added advantage of tolerance to many functional groups. This dispenses with the need for protection and deprotection steps along the synthetic pathway, thus simplifying procedures and conserving materials.

Our laboratory has been interested in developing water-tolerant reactions since its establishment. For this purpose, we initially chose indium salts for their low toxicities and stability in air and water, properties that matched our research goal. In this article, we will outline our decade of work on the development of indium salts⁴ in carbon–carbon bond forming reactions.

Allylations

In our early studies on catalysis using water-tolerant Lewis acids for development into aqueous-based or water-based reactions, we had investigated the potential of lanthanide triflates⁵ to promote an In-mediated allylation⁶ of a glucose-derived aldehyde in aqueous media.⁷ The use of Yb(OTf)₃ in our reaction system was found to effectively catalyse the reaction and imparted an appreciable increase in diastereoselectivity and yield (Table 1). The same catalyst was then applied to our work on the synthesis of enantiomerically pure -aminoalcohols. This was done through either an In- or Zn-mediated allylation on protected -aminoaldehydes.⁸ Although the rates of reactions were increased, we did not observe a significant contribution to yields and selectivities from added Yb(OTf)₃ and La(OTf)₃ (Table 2).

Entry	Conditions	Yield (%)	RS
1	DMF–H2O (64), 1 h, Yb(OTf)3 (1 eq.)	88	694
2	DMF–H2O (64), 2 h	82	1783
3	H2O–THF (41), 10 h	72	2476
4	H2O, 10 h	66	4159

Entry	Conditions	Yield (%)	antisyn
1	DMF–H2O (201), 0.5 h, In/La(OTf)3 (1 eq.)	88	928
2	DMF–H2O (201), In, 5 h	82	8713
3	H2O, In, 24 h	82	7327
4	DMF–H2O (201), 0.5 h, Zn/La(OTf)3 (1 eq.)	94	919
5	DMF–H2O (201), Zn, 5 h	88	9010
6	H2O, Zn, 24 h	Trace	—

With emerging reports on novel Lewis acids derived from main group elements at that time, we had started investigating the use of indium complexes for developing allylation reactions in water. Earlier works by Marshall and Hinkle⁹ and Baba and co-workers¹⁰ suggested InCl₃ as a suitable candidate.

Our first application of InCl₃ is in the Sn-mediated allylation of carbonyl compounds in water (Table 3).¹¹ Yields were moderate to high and high diastereoselectivities were observed for most of the substrates studied. Note that InCl₃ may only have moderate effects on the yields, but had significant influences on the final isomer ratios (entries 7 and 8, Table 3). The reactions were proposed to occur via transmetalation of the allyltin with the InCl₃, followed by preferential participation of the resulting trans intermediate in a cyclic complex to form the final product (Fig. 1). Such a transmetallation mechanism was also proposed by Marshall and Hinkle in their allylstannanes/InCl₃ system.¹² The anti products would have resulted from the most stable six-membered ring transition states whereby the bulky substituents on the aldehydes adopted the equatorial position. In the same piece of work, the convenient allylation of unprotected sugars¹³ was also studied and good to excellent yields were observed. Diastereoselectivies were generally excellent (Fig. 2). Note the strong syn preferences in these cases were attributed to a five-membered chelation transition state. The stereochemical outcome in this study drew a close parallel to the indium-mediated allylation system as observed by Paquette and Thomas.¹⁴

	Fig. 1 Proposed mechanism for the Sn/InCl3 system for allylation of aldehydes in water.

	Fig. 2 Allylation of unprotected sugars in water using the Sn/InCl3 system.

Entry	R1	R2	Conditions	Yield (%)	antisyn
1	Ph	Me	15 h	80	5050
2	Ph	Ph	15 h	45	991
3	Ph	CO2Et (E)	15 h	96	8515
4	CH2Br	CO2Et (E)	15 h	55	8020
5	3-C5H4N	CO2Et (E)	24 h	51	8020
6	cyclo-C6H11	CO2Et (Z)	15 h	65	6832
7	cyclo-C6H11	CO2Et (E)	18 h (no InCl3)	60	8515
8	cyclo-C6H11	CO2Et (E)	15 h	65	991

Our work in this new system of allylation converged with our interest in the synthesis of fluorinated compounds, especially compounds with CF₃ in place of CH₃. The CF₃ moiety, with its high C–F strength, electron-withdrawing properties, both hydro- and lipo-phobicities, often impart unique properties to the equivalent hydrocarbon counterpart.¹⁵ This is especially so for pharmacologically active compounds.

The Sn/InCl₃ allylation system was subsequently applied to the synthesis of trifluoromethylated homoallylic alcohols in water (Table 4).¹⁶ Results show that excellent yields and selectivities were accessible (entries 2–7), and that InCl₃ was indispensable (entry 1). This water-based methodology had allowed the convenient use of aqueous formaldehyde and hydrated glyoxylic acid, substrates which are not possible with classical C–C bond forming systems. Crucially, the predominantly anti products supported the six-membered ring transition state previously proposed (Fig. 3). The notable reversals of stereochemical outcomes from 2-pyridinecarbaldehyde and glyoxylic acid (Table 4, entries 6, 7, 12 and 13) are explained through the bulky groups adopting the axial position as a result of the five-membered ring chelation with the In center. Our earlier proposal for an initial transmetalation of the allyltin with InCl₃ was supported by the similar yields and selectivities obtained with the solely In-mediated reactions (Table 4, entries 8–13).

	Fig. 3 Allylation of unprotected sugars in water using the Sn/InCl3 system.

Entry	R	Condition	Yield (%)	antisyn
1	Ph	Sn	0	—
2	Ph	Sn/InCl3	87 (919 , 60E40Z)	955
3	H	Sn/InCl3	90 (only )	—
4	cyclo-C6H11	Sn/InCl3	92 (only )	>99<1
5	3-C5H4N	Sn/InCl3	95 (only )	928
6	2-C5H4N	Sn/InCl3	96 (only )	<1>99
7	CO2H	Sn/InCl3	83 (only )	<1>99
8	Ph	In	87 (only )	928
9	H	In	86 (only )	—
10	cyclo-C6H11	In	90 (only )	>99<1
11	3-C5H4N	In	95 (only )	>99<1
12	2-C5H4N	In	88 (only )	<1>99
13	CO2H	In	80 (only )	496

The Sn/InCl₃ or In-only systems of allylation in water were particularly useful in our synthesis of -trifluoromethylated alcohols from trifluoroacetaldehyde. Such an aldehyde is volatile and unstable, with its commercially available counterpart in the form of ethyl hemiacetal. Thus, a direct C–C bond formation using such a hemiacetal, not feasible with classical methods, would be a facile route to -trifluoromethylated alcohols. In this regard, our water-based methodolgy was found to achieve good to excellent yields with moderate diastereoselectivities even for trifluoroacetaldehyde hydrate (Table 5, entries 1–10). Note that unreactive allylic halides are unable to react even with the use of Yb(OTf)₃ (Table 5, entries 11–16). The absence of the ethoxy product could be due to the preference of the hard In center to coordinate with the hard OH base with OEt being the leaving group (Fig. 4). The comparable, if not moderately better, yields from In-only mediated reactions again reinforced our transmetalation proposal to account for our Sn/InCl₃ results.

	Fig. 4 Proposed mechanism of the allylation of trifluoroacetaldehyde ethyl hemiacetal.

Entry	R1	R2	R3	Condition	Yield (%) (diastereomeric ratio)
1	Et	H	H	Sn/InCl3 (0.1 eq.)	Trace
2	Et	H	H	Sn/InCl3 (1 eq.)	85
3	Et	H	CO2Me	Sn/InCl3 (1 eq.)	65
4	Et	Me	H	Sn/InCl3 (1 eq.)	72 (6733)
5	Et	H	H	In	95
6	Et	H	CO2Me	In	87
7	Et	Me	H	In	80 (6535)
8	H	H	H	In	81
9	H	H	CO2Me	In	82
10	H	Me	H	In	70 (6832)
11	H	CO2Et	H	In	0
12	H	CO2E	H	In, Yb(OTf)3	0
13	H	Me	Br	In	0
14	H	Me	Br	In, Yb(OTf)3	0
15	H	H	Me	In	0
16	H	H	Me	In, Yb(OTf)3	0

Recently, we applied the use of InCl₃ in catalyzing the In-mediated one-pot allylation of dihydropyrans and dihydrofurans in water (Table 6).¹⁷ The InCl₃ catalysed the formation of the lactol in situ and subsequent allylation gave the corresponding 1,4- and 1,5-diols in good yields.

Entry	Allyl bromide	n	Yield (%) (antisyn)
1		1	83
2		2	83
3		1	77
4		2	68
5		1	53
6		2	64
7		1	60
8		2	58
9		1	83 (4060)
10		2	78 (4060)
11		1	49 (2476)
12		2	52 (2377)

Current work on allylation is focused on the use of InCl₃ in asymmetric allylations in ionic liquids, with encouraging preliminary results.¹⁸

Aldol and related reactions

The aldol reaction is a powerful tool for the construction of carbon–carbon bonds that lends itself to acyclic stereocontrol.¹⁹ An important variant of the aldol reaction is the Mukaiyama aldol reaction²⁰ which is essentially a directed aldol reaction.

Our laboratory has strong interest in exploring the chemistries of aldol reactions, especially the Mukiayama aldol reaction. In particular, we envisioned protocols that would be tolerant of water and air, in contrast to the initially developed Mukaiyama aldol methodologies.

Kobayashi et al. pioneered the work in the adaptation of the Mukaiyama aldol reaction to work in aqueous media using lanthanide triflates as catalysts.^21,22 We were interested at that time to investigate the feasibility of conducting the reaction in water with catalysis from InCl₃.²³ However, our results (Table 7) were found to be inconsistent, with effects from sequence of addition of reactants playing a part in the final yield.²⁴ The heterogenous nature of the reaction would have also introduced inconsistencies in results from the varying efficiencies of mixing.

Entry	Aldehyde	Silyl enol ether	Sequence	Yield(s) (%)	antisyn
1			A	51	4852
2			B	74, 69	6139
3			C	79, 55	5248
4			C	82	4060
12	HO2CCHO·H2O		C	80	5644

In our unpublished work on the synthesis of the core structure of zaragozic acids, we had devised a diastereoselective one carbon extension of a glucose derived silyl enol ether. This method uses the commercially available formaldehyde with InCl₃ as catalyst, giving good yields and excellent selectivities (Table 8).²⁵ Note that both cis and trans isomer of the silyl enol ether gave the same major isomer. The formaldehyde nucleophile was proposed to approach only in one direction due to the steric restrictions imposed by the rigid silyl enol ether. Product stereochemistry would then depend on the preferred conformation of the molecule. An ab initio calculation was done and we found that it corresponded well with our experimental results (Fig. 5). The use of water in this case is restricted to the quantities introduced by the 37% formaldehyde solution. This tend to give more consistent results and subsequent studies on Mukaiyama aldol reactions were done with no solvent.

	Fig. 5 Proposed mechanism for to account for the product stereochemistry. Numbers in parentheses denote the relative energies in kJ mol–1.

Entry	EZ	Lewis acid (mol%)	Time	Yield (%)	RS
1	8020	InCl3 (40)	4–7 d	73	964
2	8020	In(OTf)3 (40)	30 min	0 (decomp.)	—
3	8020	In(OTf)3 (2)	0.5–1 d	38	937
4	8020	Yb(OTf)3 (40)	2–3 d	40	8812
5	0100	InCl3 (40)	4–7 d	68	8218
6	0100	In(OTf)3 (2)	0.5–1 d	37	7030
7	0100	Yb(OTf)3 (40)	2–3 d	35	5446

We proceeded to study the synthesis of 1,3-aminoalcohols through the Mukaiyama aldol reaction of a keto ester catalysed by InCl₃ under neat conditions.²⁶ Moderate to good yields were obtained (Table 9). Most importantly, high diastereoselectivities were observed that were explained through shielding of the carbonyl moiety by the phenyl group (Fig. 6). The proposal was supported by crystal structures of the keto ester. Note that -stacking does not play a role as replacement of the phenyl group with cyclohexyl did not appreciably affect the stereoselectivity.

	Fig. 6 Proposed mechanism to account for the stereochemical course of the Mukaiyama aldol reaction of the depicted keto ester.

Entry	Silyl enol ether	Lewis acid	RS (Yield, %)
1		BF3·H2O	No reaction
2		La(OTf)3	Trace
3		Yb(OTf)3	9010 (40)
4		InCl3, H2O	9010 (40)
5		InCl3	9010 (73)
6		InCl3	8614 (72)
7		InCl3	991 (56)
8		InCl3	8713 (58)
9		InCl3	991 (72)

The direct aldol reactions that circumvent the silyl enol ether intermediate were reported by several groups, and these procedures often require air- and moisture-sensitive Lewis acids.^27,28 Our laboratory had applied the use of InCl₃ as a catalyst in the direct aldol reaction of glyoxylic acid with ketones without the use of solvents. Sonication is required and excellent yields with moderate diastereoselectivities were obtained (Table 10).²⁹ An open-chain antiperiplanar transition state was put forward to account for the product stereochemistry. Our laboratory had also used Montmorillonite K10 with success to catalyse similar Mukaiyama-aldol reactions in water.³⁰ These synthetic methodologies are useful to furnish building blocks for complex natural products through the synthesis of -hydroxy acids.³¹

Entry	Silyl enol ether	InCl3 (mol%)	Time/d	Yield (%) (synanti)
1		No InCl3	3	10 (7327)
2		20	2	80 (6040)
3		50	3	95 (5941)
4		20	3	94
5		20	3d	91
6		20	2	80 (6238)
7		20	4	95 (7129)

Mannich and related reactions

The Mannich reaction is an important reaction for the preparation of -amino-ketones and -aldehydes.³² The products of such reactions are used in the synthesis of natural products and pharmaceuticals. An important variant of the Mannich reaction is the reaction between preformed imines from condensing aldehydes with amines, and silyl enol ethers generated from ketones. However, imines are frequently unstable, moisture sensitive and difficult to purify through distillation.

Encouraged by the expanding utility of InCl₃, we made use of its water-tolerance to devise a procedure to conduct the Mannich reaction in water. For such a methodology, we would need to generate the imine in situ and immediately react it with the silyl enol ether. We would thus need to investigate if the reaction with the imine would be faster than the competing aldol reaction.³³ Results show that this one-pot Mannich reaction produced moderate to excellent yields in most of the reactant combinations studied (Table 11). However, yields for enolizable imines, generated from aqueous formaldehyde in our study (entries 1, 4 and 7), were poor. Interestingly, there was a lack of deamination products as usually observed in conventional Mannich reactions. The expected competing aldol reactions were also not observed frequently. This suggests that the formation of Mannich products occur preferentially over the aldol products in our system. In fact, in the absence of InCl₃, the aldol products actually dominate, thus confirming the unique role that InCl₃ plays in this reaction. In the same paper, we had also reused the InCl₃ with no significant impact to the yield of product.

Entry	Aldehyde R1	Amine R2	Yield (%) (synanti)
					EZ = 41
1	H	Ph	30, 8a	91	58	46
2	Ph	Ph	54	75	80 (5743)	68 (5941)
3	2-Py	Ph	92	94	70, 10b (6931)	60 (5248)
4	H	4-ClPh	21	85	78	60
5	Ph	4-ClPh	23	60	26 (6040)	17 (7426)
6	2-Py	4-ClPh	90	91	55, 23b (5743)	35 (6436)
7	H	4-MeOPh	35, 17a	86	41, 5	20
8	Ph	4-MeOPh	30	40	45 (5149)	11 (7624)
9	2-Py	4-MeOPh	90	90	70, 11b (6634)	20 (7624)
10	CO2H	Ph	10	Decomp.	—	—
11	CO2H	4-ClPh	31	63	—	—
12	CO2H	4-MeOPh	10	Decomp.	—	—
a Yield of:   b Yield of aldol product.

The above procedure was not successful for an enantioselective one-pot Mannich reaction using chiral amines. In addition, the procedure was limited to nonenolizable aldehydes and aromatic amines. However, further work shows that the use of MeOH as a solvent addressed these issues, resulting in moderate to high yields with good to excellent diastereoselectivities in a variety of substrates (Fig. 7).³⁴ Diastereoselectivity was attributed to the chelation of InCl₃ by the carbonyl group of the ester and the nitrogen. The resulting rigid bidentate structure allows bulky groups to effectively block one face, leading to high facial selectivities (Fig. 8).

	Fig. 7 A typical one-pot Mannich reaction in MeOH catalysed by InCl3.

	Fig. 8 Proposed mechanism to account for the diastereoselectivities in the one-pot Mannich reaction in MeOH catalysed by InCl3.

An ionic liquid version of the one-pot Mannich procedure was also developed but shall not be discussed here.³⁵

Others

The Diels–Alder reaction is an important cycloaddition reaction for the construction of six-membered rings. Breslow had pioneered the study on carrying out the Diels–Alder reaction in water.³⁶ Our laboratory had studied the catalytic behaviour of InCl₃ for this pericyclic reaction and had demonstrated that significant improvements in yields and regioselectivities were observed in the presence of InCl₃ (Table 12).³⁷ InCl₃ was also demonstrated to improve the selectivity of the Diels–Alder reaction using Evan's chiral auxilliary³⁸ from low to moderate (Fig. 9).

	Fig. 9 Asymmetric Diels–Alder reaction catalysed by InCl3.

Entry	Reactants	Major product	Yield (%) (endoexo)
1			89 (919)
2a			60 (7426)
3			85 (9010)
4			99 (1000)
5			88 (9010)
6			95 (595)
7			88 (one regioisomer)
Conditions: water, room temperature, InCl3 (20 mol%), 4 h. a No InCl3.

The direct synthesis of -amino derivatives can be achieved through a conjugate addition of amines with ,-ethylenic compounds. We were able to carry out such a reaction under mild conditions in water with the yield of the reaction increased by the addition of InCl₃.³⁹ Moderate to good yields, albeit with poor diastereoselectivities, were obtainable with a wide variety of substrates (Table 13). Note that the reaction can occur without the use of InCl₃ but suffers a yield penalty.

Entry	Reactants	Conditions	Yield (%) (12 or de)
1		6 h	70 (3367)
2		No InCl3, 6 h	20 (3565)
3		16 h	82
4		No InCl3, 16 h	62
5		23 h	45 (14% de)
6		No InCl3, 23 h	40 (15% de)
7		3 d	70 (7% de)
8		No InCl3, 3 d	75
9		21 h	78 (60% de, 3R major)
10		No InCl3, 21 h	63 (60% de, 3R major)

Related to our work on aldol reactions, we found that under neat conditions, InCl₃ catalyses the 1,4-addition of ,-unsaturated carbonyl compounds with silyl enol ethers (Table 14).⁴⁰ Moderate to excellent yields were obtained in most of the reactant combinations. InCl₃ was proven to be an essential catalyst, without which no reaction can take place. InCl₃ can also be recovered and reused with no loss in catalytic activity.

Entry	,-Unsaturated carbonyl	Silyl enol ether	Conditions	Yield (%) (antisyn)
1		1	No InCl3, r.t. (1 h)	No reaction
2		1	r.t. (0.5 h)	67
3		2	0 °C to r.t.	82
4		3	r.t. (0.5 h)	86 (6535)
5		4	r.t. (0.5 h)	90 (6535)
6		5	0 °C (5 h), r.t. (2 h)	66
7		1	r.t. (2 min.)	68
8		2	0 °C (1 h)	12
9		3	r.t. (0.5 h)	51
10		4	–30 °C to r.t.	48
11		5	0 °C (0.5 h), r.t. (1 h)	18
12		1	r.t. (24 h)	64
13		2	r.t. (0.5 h)	90
14		4	r.t. (15 h)	14 (3070)

The versatility of the bifunctional cyanohydrins in the synthesis of complex molecules prompted us to develop synthetic methodologies that would realise these compounds with simple procedures. Evans et al. developed the use of TMSCN in place of traditional HCN as a safer source of CN for this addition reaction.⁴¹ Our intention was to study the scope and limitation of this reaction in water. We found that this reaction did not occur without Lewis acids but proceeded to give a good yield with InCl₃ (Table 15, entries 1–4).⁴² Interestingly, use of InF₃ gave good to excellent yields for a variety of substrates, including the hemiacetal form of the volatile trifluoroacetaldehyde (entries 5–10). Competitive studies also show that the reaction is highly chemoselective towards aldehydes, with TMSCN unable to react with ketones under our reaction condition.

Entry	Aldehyde	Catalyst (mol%)	Yield (%)
1		—	No reaction
2		KF (100)	No reaction
3		TBAF (100)	No reaction
4		InCl3 (100)	60 (conversion)
5		InF3 (30)	95
6		InF3 (30)	80
7		InF3 (30)	85
8		InF3 (30)	80
9		InF3 (30)	75
10		InF3 (30)	75

The formation of tetrahydropyran rings are often central to many syntheses of natural products. This Prins reaction is often performed under strong acidic conditions or, with water- and air-sensitive Lewis acids.⁴³ Prior to our work, Li and co-workers had achieved such a reaction using 100 mol% of InCl₃.^44,45 Our group had further modified the reaction by using allylchlorosilane as allylating agent and conducting the reaction in CH₂Cl₂. With our protocol, we require only 20 mol% of InCl₃ (Fig. 10).⁴⁶ The tolerance of such a procedure to water was demonstrated by the equivalent yields and selectivities obtained in the presence of 1% water.

	Fig. 10 Prins cyclisation reaction catalysed by InCl3.

Chiral indium complexes

One of the most efficient way to effect asymmetric synthesis of enantiomerically pure compounds from achiral substrates is the use of chiral catalysts. However, most often these catalysts are moisture- and air-sensitive thus limiting their utility. Our laboratory had been studying the feasibility of a water-tolerant chiral catalyst and recently, we had developed the first catalytic chiral indium complex that is water-tolerant.

Prior to several attempts on developing asymmetric catalytic systems,⁴⁷ our initial work on a chiral indium complex was developed through the study on the asymmetric allylation of carbonyl compounds.⁴⁸ With the optimization of an early promising result, we found that InCl₃, in the presence of 4Å molecular sieves in CH₂Cl₂, reacts with the chiral ligand (S)-BINOL to give an active chiral indium complex. This BINOL–In(III) complex catalyzes the allylation of benzaldehyde in a highly enantioselective manner with excellent yields (Fig. 11, Table 16).⁴⁹ The study shows that InCl₃ with 2 equivalents of allyl tributylstannane gave the best allylation result. The quantity of molecular sieves (MS) is also critical (entry 6) and 20 mol% of the catalyst is required to maintain the high yields and enantioselectivities (entry 7). A variety of aldehydes were also allylated using this procedure and good yields with high enantioselectivities were obtained (Fig. 12).

	Fig. 11 Catalytic enantioselective allylation with (S)-Tol-BINAP·AgNO3.

	Fig. 12 Enantioselective allylation of aldehydes using BINOL–In(III) complex.

The requirement for an excess allyltributyl stannane is actually surprising to us and our study shows that it is essential to form the final catalytic complex. Observations of several controlled experiments, including an NMR study, suggest its possible participation through transmetallation with InCl₃. The resulting indium species then form the catalytic complex by reaction with (S)-BINOL (Fig. 13). Mechanistic study is still on-going to confirm this hypothesis.

	Fig. 13 Postulated process for the formation of chiral indium complex.

An important aspect of the above protocol is the tolerance of the formed chiral indium complex to small amounts of water. We had attempted to develop this procedure towards an asymmetric allylation of aldehydes in water. However, we found that for a 30 mol% catalyst loading, 7.4 equivalents of water (relative to InCl₃) gave the best results in terms of yield without any penalty in enantioselectivities.⁵⁰ Further addition of water (above 22.2 equiv.) resulted in poor yields and enantioselectivities. These data, however, are only able to confirm the moisture tolerance of our methodolgy and efforts are still being directed towards an aqueous or water-only chiral catalytic system.

Extending an allylation methodology from aldehydes to ketones is usually not possible due to the lesser reactivities of ketones. Such allylations are normally carried out using the more reactive tetraallystannanes.⁵¹ However, we are able to adopt the chiral BINOL–In(III) complexes to the allylation of ketones to give chiral tertiary alcohols (Table 17).⁵² In this procedure, the use of the stronger Lewis acid, InBr₃, is necessary for good to excellent yields and enantioselectivities in a variety of ketone substrates.

In broadening the scope of the BINOL-In(III) complexes even further, we had exploited their use in the asymmetric Diels–Alder reaction.⁵³ Experiments show that the BINOL-In(III) complex is equally applicable for such reactions, with moderate to good yields and high enantioselectivities (Table 18).⁵⁴ A possible assembly of catalyst and dienophile to account for the stereochemistry of the products was proposed (Fig. 14).

	Fig. 14 Proposed assembly of the (S)-BINOL–In(III) complex with dienophile.

Entry	Reactants	Conditions (°C, h)	Major product	R1	Yield (%)	ee (%)
1	3 + 1	r.t. 20		Me	63	98
2	3 + 2	–20, 20		Br	74	98
3	3 + 2	r.t., 20 + H2O		Br	70	80
4	4 + 1	–20, 20		Me	71	98
5	4