Catalytic Enantioselective Alkylation of Indoles with trans-4-Methylthio-β-Nitrostyrene

Abstract

The reaction of the rhodium aqua-complex (S_Rh,R_C)-[Cp*Rh{(R)-Prophos} (OH₂)][SbF₆]₂ [Cp* = C₅Me₅, Prophos = propane-1,2-diyl-bis(diphenylphosphane)] (1) with trans-4-methylthio-β-nitrostyrene (MTNS) gives two linkage isomers (S_Rh,R_C)-[Cp*Rh{(R)-Prophos}(κ¹O-MTNS)]²⁺ (3-O) and (S_Rh,R_C)-[Cp*Rh{(R)-Prophos}(κ¹S-MTNS)]²⁺ (3-S) in which the nitrostyrene binds the metal through one of the oxygen atoms of the nitro group or through the sulfur atom, respectively. Both isomers are in equilibrium in dichloromethane solution, the equilibrium constant being affected by the temperature in such a way that when the temperature increases, the relative concentration of the oxygen-bonded isomer 3-O increases. The homologue aqua-complex of iridium, (S_Ir,R_C)-[Cp*Ir{(R)-Prophos}(OH₂)][SbF₆]₂ (2), also reacts with MTNS; but only the sulfur-coordinated isomer (S_Ir,R_C)-[Cp*Ir{(R)-Prophos}(κ¹S-MTNS)]²⁺ (4-S) is detected in the solution by NMR spectroscopy. The crystal structures of 3-S and 4-S have been elucidated by X-ray diffractometric methods. Complexes 1 and 2 catalyze the Friedel–Crafts reaction of indole, N-methylindole, 2-methylindole, or N-methyl-2-methylindole with MTNS. Up to 93% ee has been achieved for N-methyl-2-methylindole. With this indole, the ee increases as conversion increases, ee at 263 K is lower than that obtained at 298 K, and the sign of the chirality of the major enantiomer changes at temperatures below 263 K. Detection and characterization of the catalytic intermediates metal-aci-nitro and the free aci-nitro compound as well as detection of the Friedel–Crafts (FC)-adduct complex involved in the catalysis allowed us to propose a plausible double cycle that accounts for the catalytic observations.

Introduction

Chiral catalysts based on transition metals are among the most efficient and versatile for the preparation of enantioenriched compounds.¹ Metallic catalysts activate organic substrates by coordination and chiral induction takes place when the reaction occurs within the asymmetric environment generated around the metal by enantiopure ligands.²

In particular, transition-metal complexes efficiently catalyze asymmetric Friedel–Crafts (FC) reactions,³ and several metallic systems have been successfully applied to the alkylation of indoles with nitroalkenes. Thus, following the pioneering work of Bandini et al.,⁴ homogeneous systems based on copper,⁵ zinc,⁶ and, to a lesser extent, nickel,⁷ platinum,⁸ or palladium⁹ catalyze this transformation. Additionally, metal-containing hydrogen bond donors also catalyze the FC reaction between nitroalkenes and indoles through second coordination sphere mechanisms,¹⁰ in some instances, enantioselectively.¹¹

However, theoretical or experimental studies about the operating catalytic cycles are very scarce and, therefore, no reliable data about the active metallic intermediates are available. Although it should be taken into account that intermediates detected under catalytic conditions may not be responsible for the catalysis, the study of the metallic intermediates involved in catalytic processes is one of the most powerful tools available for the chemists to get a deep insight into the reaction mechanisms.

In this context, we have recently reported on the application of the chiral half-sandwich rhodium and iridium complexes (S_M,R_C)-[Cp*M{(R)-Prophos}(OH₂)][SbF₆]₂ [Prophos = propane-1,2-diyl-bis(diphenylphosphano); M = Rh (1), Ir (2)] (Scheme 1)¹² as catalyst precursors for the Michael-type Friedel–Crafts reaction between indoles and trans-β-nitrostyrenes. High yield and excellent enantioselectivity were achieved. The true catalyst was shown to be the dicationic species of formula (S_M,R_C)-[Cp*M{(R)-Prophos}(nitrostyrene)]²⁺ in which the nitrostyrene coordinates with the metal through one of the oxygen atoms of the nitro group. aci-Nitro and FC-adduct rhodium complexes, involved in the process, were characterized by spectroscopic means. On the basis of these data, together with theoretical and kinetic studies, the catalytic stereochemical outcome was satisfactorily explained.¹³

Scheme 1. Chiral Complexes [Cp*M{(R)-Prophos}(H₂O)][SbF₆]₂.

Here, we disclose our findings about the behavior of trans-4-methylthio-β-nitrostyrene, instead of unfunctionalized β-nitrostyrene, as an electrophile in the abovementioned reaction.

Results and Discussion

Reaction of the Aqua-Complex (S_M,R_C)-[Cp*M{(R)-Prophos}(H₂O)][SbF₆]₂ with trans-4-Methylthio-β-nitrostyrene

When an excess of different trans-β-nitrostyrenes was added to solutions of (S_M,R_C)-[Cp*M{(R)-Prophos}(H₂O)][SbF₆]₂ (M = Rh (1), Ir (2)) in dichloromethane, an equilibrium between the starting aqua-complex and the corresponding nitrostyrene complex (S_M,R_C)-[Cp*M{(R)-Prophos}(nitrostyrene)]²⁺ was reached. Addition of 4 Å molecular sieves drives this equilibrium completely to the nitrostyrene-coordinated compound side. The reaction is stereospecific and takes place with retention of the configuration of the metal center. Spectroscopic and diffractometric data reveal that the nitrostyrene ligand coordinates with the metal through one of the oxygen atoms of its nitro group and unveil the strong electronic activation of the β-carbon atom of this ligand toward nucleophilic attack.¹³

However, when one equivalent of the functionalized nitrostyrene trans-4-methylthio-β-nitrostyrene (MTNS) was added to CD₂Cl₂ solutions of the rhodium complex (S_Rh,R_C)-[Cp*Rh{(R)-Prophos}(OH₂)][SbF₆]₂ (1), in the presence of 4 Å molecular sieves, at 223 K, the ³¹P{¹H} NMR spectra showed that the starting aqua-complex had completely reacted to give a mixture of two rhodium complexes in approximately 41/59 molar ratio. Thus, the two double doublets corresponding to complex 1 (Figure 1, trace A) disappeared and two pairs of double doublets emerged. We assign the new resonances to the two linkage isomers, 3-O and 3-S (eq 1), in which the nitrostyrene coordinates with the metal through one of the oxygen atoms of the nitro group or through the sulfur atom, respectively. Table 1 shows ¹H and ¹³C{¹H} NMR data for selected nuclei of the coordinated nitrostyrene ligand of these isomers. The variation of chemical shifts with respect to those measured for the corresponding nuclei of the free ligand (Δδ) is also tabulated. In particular, a comparison of this variation, for the H_β and SMe protons and for the C_β and SMe carbons of both isomers, allows us to distinguish between their coordination modes. Thus, ΔδH_β and ΔδC_β values of −0.98 and +8.57 ppm, respectively, were attributed to the O-coordinated isomer, whereas −0.04 and −0.14 ppm were ΔδH_β and ΔδC_β, respectively, for the S-coordinated isomer. On the other hand, −0.92 and +9.32 were Δδ measured for the protons and carbon of the SMe group, respectively, in the S-coordinated isomer, whereas in the O-coordinated compound, −0.01 and +0.06 ppm were Δδ recorded for these protons and carbon nuclei, respectively. These data strongly suggest that in the minor isomer, one of the NO₂ oxygen atoms completes the coordination sphere of the rhodium, and in the major isomer, the nitrostyrene coordinates with the metal through the sulfur atom of the SMe group. Regarding reactivity, it is convenient to be aware that the NMR data indicated that when the nitrostyrene oxygen binds to the metal, the β carbon is activated for nucleophilic attack (ΔδC_β = +8.57), but this activation is not apparent when coordination occurs through the sulfur atom (ΔδC_β = −0.14).

1

Notably, EXSY experiments revealed that the 3-O and 3-S linkage isomers exchange with each other and also with free trans-4-methylthio-β-nitrostyrene. Moreover, the equilibrium constant 3-O ⇆ 3-S is strongly affected by the temperature. Figure 2 shows the ³¹P{¹H} NMR spectra of 3-O ⇆ 3-S equilibrium mixtures at different temperatures. As the temperature increases, the concentration of 3-O increases at the expense of the concentration of 3-S and vice versa. Regarding the catalytic properties of this compound, it is interesting to notice that at temperatures below 233 K, complex 3-O accounts for less than 50% of the rhodium present in the solution. Additionally, below 213 K, the broadening of the doublets assigned to complex 3-S is apparent. Taking into account that in this isomer the sulfur atom is a stereogenic center, we tentatively propose that, whereas above 213 K epimerization at sulfur is rapid on the NMR time scale, below this temperature this process is significantly slowed down. On the other hand, above 263 K the four resonances slowly broaden, and the doublets assigned to 3-S coalesce at around 298 K. The thermal stability of the compound avoids reaching the temperature region of fast exchange between both isomers. In fact, at 298 K, a small amount of the decomposition cation¹⁴ [Cp*RhCl{(R)-Prophos}]⁺ was observed. Finally, from the linear dependence of ln K versus the reciprocal of the temperature (Figure 3), the thermodynamic parameters for the equilibrium 3-O ⇆ 3-S, ΔG° = 0.84 ± 0.28 kcal·mol^–1, ΔH° = −3.24 ± 0.13 kcal·mol^–1, and ΔS° = −13.93 ± 0.54 cal·mol^–1·K^–1, were derived.

Figure 1.

³¹P{¹H} NMR spectra of complex 1 (trace A) and a mixture of 3-O and 3-S (trace B) at 223 K.

Table 1. Selected ¹H and ¹³C{¹H} NMR Data (223 K) for the Complexes 3-O and 3-S, in CD₂Cl₂^a.

minor isomer (3-O)
1H NMR
δHβ	ΔδHβ	δSMe	ΔδSMe
6.95	–0.98	2.49	–0.01
13C {1H} NMR
δCβ	ΔδCβ	δSMe	ΔδSMe
147.63	+8.57	14.75	+0.06

major isomer (3-S)
1H NMR
δHβ	ΔδHβ	δSMe	ΔδSMe
7.89	–0.04	1.58	–0.92
13C {1H} NMR
δCβ	ΔδCβ	δSMe	ΔδSMe
138.95	–0.14	23.13	+9.32

^aChemical shifts in ppm; coupling constants in Hz.

Figure 2.

³¹P{¹H} NMR spectra of a mixture of complexes 3-O and 3-S at different temperatures. The asterisk denotes the decomposition species [Cp*RhCl{(R)-Prophos}]⁺.

Figure 3.

Plot of ln K versus 1/T for the equilibrium 3-O ⇆ 3-S.

The iridium analogue (S_Ir,R_C)-[Cp*Ir{(R)-Prophos}(OH₂)][SbF₆]₂ (2) also reacted with MTNS. However, only the S-coordinated isomer (S_Ir,R_C)-[Cp*Ir{(R)-Prophos}(κ¹S-MTNS)][SbF₆]₂ (4-S) was detected for this metal. The low solubility of the resulting compound in the usual organic solvents precludes its complete spectroscopic characterization, but single crystals for diffractometric studies were obtained from dichloromethane/pentane mixtures.

Molecular Structure of the Complex [Cp*M{(R)-Prophos}(MTNS)][SbF₆]₂ (M = Rh (3-S), Ir (4-S))

Single crystals of the complexes were grown by slow diffusion of pentane into dichloromethane solutions. Molecular representations of the cations are shown in Figure 4 and selected structural parameters are listed in Table 2.

Figure 4.

Molecular representation of the cations of the compounds 3-S and 4-S (one of the crystallographically independent molecules of each unit cell). For clarity, hydrogen atoms have been omitted.

Table 2. Selected Bond Lengths (Å) and Angles (deg) for 3-S and 4-S^a.

	3-S (M = Rh)		4-S (M = Ir)
	molecule 1	molecule 2	molecule 1	molecule 2
M-S(1)	2.379(3)	2.388(3)	2.373(5)	2.368(5)
M-P(1)	2.332(3)	2.330(3)	2.327(4)	2.323(5)
M-P(2)	2.397(2)	2.403(2)	2.366(4)	2.376(4)
M-Ga̲	1.899(6)	1.912(6)	1.910(9)	1.930(10)
S(1)-M-P(1)	85.11(9)	84.97(9)	84.77(17)	84.99(17)
S(1)-M-P(2)	89.67(9)	89.99(9)	89.45(16)	89.68(16)
S(1)-M-Ga̲	125.97(18)	125.68(18)	125.7(3)	125.8(3)
P(1)-M-P(2)	82.91(9)	82.91(9)	82.62(16)	82.90(16)
P(1)-M-Ga̲	125.60(18)	126.14(18)	125.9(3)	126.5(3)
P(2)-M-Ga̲	132.28(18)	131.95(17)	132.8(3)	131.7(3)

^aG represents the centroid of the Cp* ligand.

Both cations exhibit a “three-legged piano-stool” geometry. An η⁵-C₅Me₅ group occupies three fac positions and the two phosphorus atoms of the chelate diphosphano and the sulfur atom of the nitrostyrene ligand complete the coordination sphere of the metal. The absolute configuration of the complexes is S_M,R_C,S_S.¹⁵ Asymmetric units of both complexes comprise two crystallographically independent but chemically equivalent molecules.

At a molecular level, both complexes exhibit similar parameters. Furthermore, bond lengths and angles of a Cp*M{(R)-Prophos} fragment in complexes 3-S and 4-S agree with those found in parent aqua complexes.^12b Coordination of the sulfur atom is apparent in both cases, the two M–S bond distances being comparable, in the 2.368(5)–2.388(3) Å range. The 5-membered metallacycles M-P(1)-C(14)-C(23)-P(2) adopt a λ conformation. These puckered rings may be characterized by two parameters.¹⁶ On the one hand, their Cremer and Pople ring-puckering amplitude (Q ∼ 0.5 Å, see the Supporting Information) nicely agrees with those observed in other Cp*M{(R)-Prophos} complexes.^12b,17 On the other hand, their phase angle (ϕ) values between 62.84(3) and 65.33(1)°, characteristic of ³T₂ with a contribution of ³E conformations, have been found.

Catalytic Reactions

Complexes 1 and 2 catalyze the reaction between indoles and MTNS. Table 3 shows a selection of the obtained results together with the reaction conditions. The results shown are the average of at least two comparable reaction runs. The reactions were carried out in CH₂Cl₂, at 298 K, in the presence of 4 Å molecular sieves. A molar ratio cat*/MTNS/indole of 1/30/20 (5 mol % catalyst loading) was employed in all cases.

Table 3. Results of Selected Asymmetric FC Alkylation Reactions of Indoles with MTNS^a.

entry	cat*	R1	R2	t (min)	conv.b (%)	eec (%)
1	1	H	H	30	55	rac
2	2			30	22	–2
3	1	Me	H	30	90	+12
4	2			30	43	+15
5	1	H	Me	30	81	+3
6	2			30	48	+15
7	1	Me	Me	5	100	+93
8	2			30	100	–21

^aReagents and conditions: catalyst (0.03 mmol, 5.0 mol %), MTNS (0.90 mmol), indole (0.60 mmol), 4 Å molecular sieves (100 mg), and CH₂Cl₂ (4 mL).

^bBased on indole. Determined by ¹H NMR spectroscopy.

^cDetermined by high-performance liquid chromatography (HPLC). Changes in the sign indicate that the opposite enantiomer was preferentially obtained.

Catalytic runs were quenched after the reaction times shown in Table 3 by adding excess of a saturated solution of N(nBu₄)Br in methanol. The reactions are clean: only the addition product and the remaining unreacted reagents were detected in the NMR spectra of the crude reaction mixture. The rhodium complex is more active than the iridium analogue, but it is remarkable that the iridium compound catalyzes the reaction although, under the catalytic conditions, the presence of the O-coordinated isomer (the one in which the nitrostyrene is activated for nucleophilic attack) was not detected by NMR (see above). As the best results, in both reaction rate and enantioselectivity, were obtained for N-methyl-2-methylindole using the rhodium complex as catalyst, we chose this system as the model to continue our study.

Effect of the Temperature

The model reaction was performed at different temperatures in the range 298–233 K, the remaining conditions being maintained unchanged. At a fixed temperature, the ee changes as conversion increases and, in some cases, it changes even when indole (the limiting reagent) is no longer detected by ¹H NMR spectroscopy in the reaction medium. Table 4 shows the ee values measured when at least three identical ee values have been obtained at reaction times longer than that needed to achieve complete conversion. Changes in the sign of the enantioselectivity indicate that the opposite enantiomer was preferentially obtained. ee values up to 93% were obtained at temperatures higher than 263 K. Unexpectedly, by reducing the temperature from 298 to 263 K, a decrease from 93 to 55% in the enantioselectivity was observed. Notably, this decrease results in a switch in the sign of the major isomer at the lowest temperatures measured (see the Supporting Information).

Table 4. Effect of the Temperature^a.

entry	T (K)	ee (%)
1	298	+93
2	283	+73
3	273	+65
4	263	+55
5	243	–8 → −11b
6	233	–6 → −10b

^aFor reagents and conditions, see the footnote in Table 3.

^bRange of values measured after complete conversion was achieved.

To try to explain this behavior, we studied the catalytic process by NMR spectroscopy.

Reaction of Complexes 3 with N-Methyl-2-methylindole

Plausibly, the catalytic cycle starts with the interaction between the activated O-coordinated electrophile in (S_Rh,R_C)-[Cp*Rh{(R)-Prophos}(κ¹O-MTNS)]²⁺ and the indole. To confirm this proposal, at 193 K, the reaction of a 30/70 molar ratio mixture of 3-O/3-S (Figure 5, trace A) -formed in situ from complex 1 (1 equiv) and MTNS (10 equiv), in the presence of 4 Å molecular sieves- with N-methyl-2-methylindole (5 equiv) was monitored by NMR spectroscopy. After the addition of the indole, the ³¹P{¹H} NMR spectra showed the immediate disappearance of complexes 3 and the formation of two rhodium complexes, 5, in a molar ratio of around 80/20, characterized by two new pairs of double doublets (Figure 5, trace B). Both starting isomers, 3-O and 3-S, were transformed into 5; 3-O by direct reaction with the indole and 3-S, most probably, after conversion into 3-O via the equilibrium 3-O ⇆ 3-S (Scheme 2). The selected region of the ¹H NMR spectrum shown in Figure 5, trace B shows a pair of coupled signals centered at 6.12 (H_α) and at ca. 5.25 ppm (H_β, overlapped; major isomer) together with two broad doublets at 6.02 (H_α) and 4.76 ppm (H_β; minor). Signals at about 17 ppm, attributed to NO₂H groups, are also observed. Complexes 5 should be the aci-nitro complexes formulated in Scheme 2, resulting from the formation of the C–C bond between the C3 carbon of the indole and the C_β carbon of the coordinated nitroalkene in 3-O.¹⁸ Apart from the excess of indole (the C3 indole proton gives a singlet at about 6.2 ppm), trace B also shows the presence of two new organic compounds characterized by one singlet at around 14 ppm and a doublet centered at 5.45 ppm and a complex multiplet centered at about 5.2 ppm. The singlet and the doublet were attributed to the NO₂H group and the H_β proton, respectively, of the free aci-nitro ligand (6) displaced from the coordination sphere of the metal in 5 by a new molecule of MTNS (Scheme 2). The complex multiplet corresponds to the three aliphatic protons of the FC adduct 7 (Scheme 3), as determined by comparison with an authentic sample obtained in an independent catalytic experiment. Further treatment at 223 K gave rise to the gradual disappearance of the indole together with the progressive formation of the FC adduct 7 at the expense of the free aci-nitro compound 6 (traces C and D). The aci-nitro compound 6 rearranges to the FC adduct 7,¹⁹ most probably, through the prototropic tautomerism shown in Scheme 3.^13,18

Figure 5.

Trace A: selected regions of the ¹H and ³¹P{¹H} NMR spectra of a mixture of 3-0 and 3-S at 193 K. Trace B: measured at 193 K after the addition of N-methyl-2-methylindole (catalyst/nitroalkene/indole molar ratio, 1/10/5). Traces C, D: measured at 193 K after heating up to 223 K for 20 min. Trace E: measured at 193 K after total consumption of the indole. Trace F: selected regions of the ¹H and ³¹P{¹H} NMR spectra of a mixture of 1, racemic FC adduct 7, and 4 Å MS (1/7 molar ratio, 1/5) at 193 K. The asterisk denotes residual, not deuterated dichloromethane. A double asterisk denotes the decomposition species [Cp*RhCl{(R)-Prophos}]⁺ (see Figure 2 and the corresponding text).

Scheme 2. Reaction of Complexes 3 with N-Methyl-2-methylindole.

Scheme 3. Prototropic Tautomerism of the aci-Nitro Compound 6 to the FC Adduct 7.

The ³¹P{¹H} NMR spectrum of the resting state of the catalyst consisted of a pair of double doublets in about 20/80 molar ratio (8-S, trace E). To determine the nature of these complexes, excess of racemic FC adduct 7 was added to a dichloromethane solution of the rhodium aqua-complex (S_Rh,R_C)-[Cp*Rh{(R)-Prophos}(OH₂)][SbF₆]₂ (1). The ¹H and ³¹P{¹H} NMR spectra of the resulting solution are consistent with a mixture of the resting state complexes 8-S (Figure 5, trace E) and starting complex 1. The value of the chemical shift of the phosphorus nuclei indicates that the coordination of the FC adduct takes place through the sulfur (Scheme 4).

Scheme 4. Sulfur-Coordinated FC-Adduct Rhodium Complex 8-S, the Resting State of the Catalyst.

The solution’s spectroscopic data obtained from the reaction between MTNS and N-methyl-2-methylindole catalyzed by (S_Rh,R_C)-[Cp*Rh{(R)-Prophos}(OH₂)][SbF₆]₂ closely resemble those reported for the same reaction with parent trans-β-nitrostyrene.¹³ In both cases, metal-nitroalkene, metal-aci-nitro, free aci-nitro, and metal-FC-adduct intermediates have been spectroscopically detected and characterized. Moreover, also in both cases, at a fixed temperature, the ee changes with reaction time and, simply by changing the temperature, each of the two enantiomers of the FC adduct can be preferentially obtained.

The catalytic cycles A and B (Figure 6) were proposed for the parent nitrostyrene on the basis of theoretical and experimental studies reported in ref (13). From these studies it was also determined that the rate-limiting step was the prototropic tautomerism of the corresponding aci-nitro compound 5 or 6 (see Schemes 2 and 3). This prototropy could take place when the aci-nitro compound is not coordinated to the metal rendering the corresponding FC adduct (steps 6 → 8 → 7, Figure 6) or, alternatively, when it is coordinated to the metal (step 5 → 8-O). Additionally, the reported DFT studies for the unsubstituted nitrostyrene case¹³ revealed that a dynamic kinetic resolution (DKR) occurred on 5. At 298.15 K, the lowest barrier for the whole process 5 → 7 was for the S epimer at C_β through the free aci-nitro compound 6. However, at 233.15 K, the lowest barrier for the same process is for the R epimer at C_β through the FC-adduct complex 8-O. As the 5 → 3-O step is reversible, in both cases, the less-suited enantiomer goes back into the combination of the substrates, i.e., a nitrostyrene complex and indole.

Figure 6.

Catalytic cycle proposed for the reaction between N-methyl-2-methylindole and trans-β-nitrostyrene catalyzed by (S_Rh,R_C)-[Cp*Rh{(R)-Prophos}(OH₂)][SbF₆]₂.

From the behavior encountered for the FC reaction with trans-4-methylthio-β-nitrostyrene, identical to that found for the parent nitrostyrene, we assume that a similar mechanism would be operating in both cases. In fact, all of the experimental observations (no DFT calculations have been carried out for the reaction with trans-4-methylthio-β-nitrostyrene) made for the reaction with MTNS can be explained. Thus, ee increases when conversion increases because cycle A is reversible and a DKR occurs in the aci-nitro complex 5. As conversion increases, the turnover number of cycle A also increases in this way, improving the ee via DKR. At low temperatures, even when the amount of indole is negligible (not detected by NMR), the reversible cycle A works, improving the ee.

The decrease of the ee value as the temperature decreases can be accounted for by assuming that at higher temperatures the step 6 → 7 is kinetically favored for one of the enantiomers of the FC adduct 7 over the metal-assisted prototropy 5 → 8 but, at lower temperatures, the latter route, which preferentially gives the opposite enantiomer of 7, becomes competitive. Therefore, at low temperatures, the ee of the FC adduct is lower because a part of the product is obtained through the route 5 → 8 → 7. Consequently, at temperatures lower than 263 K, the ee further decreases, promoting a change in the sign of the chirality of the adduct. We have not carried out catalytic runs at temperatures below 233 K because the catalytic reaction becomes too slow.

Conclusions

One of the nitro group oxygens and the sulfur atom of MTNS can coordinate with the half-sandwich rhodium or iridium fragment “Cp*M{(R)-Prophos}” rendering κ¹O or κ¹S complexes. In solution, these linkage isomers are in equilibrium and in the iridium case, this equilibrium is strongly shifted to the complex that exhibits a κ¹S coordination mode. Only the κ¹O-coordinated nitroalkene was activated for the nucleophilic attack of indole, which is the first step of the catalytic cycle. However, the equilibrium between the κ¹S and κ¹O isomers facilitates the catalytic reaction. In fact, both metallic complexes catalyze the Michael-type FC reaction between the indoles and MTNS in quantitative yield and with up to 93% ee. Detection and characterization of the key intermediates metal-aci-nitro (5), metal-FC-adduct (8), and free aci-nitro compound (6) allow us to propose a mechanism that accounts for the effect of conversion and temperature over the measured ee.

Experimental Section

General Comments

All preparations have been carried out under argon. All solvents were treated in a PS-400-6 Innovative Technologies Solvent Purification System (SPS) and degassed prior to use. ¹H, ¹³C, and ³¹P NMR spectra were recorded on a Bruker AV-300 (300.13 MHz), a Bruker AV-400 (400.16 MHz), or a Bruker AV-500 (500.13 MHz) spectrometer. In both ¹H NMR and ¹³C NMR measurements, the chemical shifts are expressed in ppm downfield from SiMe₄. The ³¹P NMR chemical shifts are relative to 85% H₃PO₄. J values are given in Hz. COSY, NOESY, HSQC, HMQC, and HMBC ¹H–X (X = ¹H, ¹³C, ³¹P, ¹⁵N) correlation spectra were obtained using standard procedures. Analytical high-performance liquid chromatography (HPLC) was performed on an Alliance Waters 2695 (Waters 2996 PDA detector) instrument using a chiral column Chiralpak IB (0.46 × 25 cm²) and an IB guard (0.46 × 1 cm²). High-resolution mass spectra were obtained with a Micro Tof-Q Bruker Daltonics spectrometer.

Preparation of the Complex [Cp*M{(R)-Prophos}(MTNS)][SbF₆]₂ (M = Rh 3, Ir 4)

In an NMR tube, at 223 K, to a suspension of (S_M,R_C)-[Cp*M{(R)-Prophos}(OH₂)][SbF₆]₂ (M = Rh (1), Ir (2)) (0.024 mmol) in CD₂Cl₂ (0.45 mL), MTNS (4.7 mg, 0.024 mmol) and 4 Å MS (15 mg) were added. The resulting solution containing complexes 3 and 4 was analyzed by NMR without any further purification. Crystals suitable for X-ray analysis for both complexes were obtained by crystallization from CH₂Cl₂/n-pentane solutions. Yield: complex 3, 15.5 mg (49%); complex 4, 28.0 mg (83%).

(S_Rh,R_C)-[Cp*Rh{(R)-Prophos}(MTNS)][SbF₆]₂ (3)

3-O

¹H NMR (300.13 MHz, CD₂Cl₂, 223 K): δ = 7.98–6.90 (m, 24H, H_Ar), 7.24 (brd, 1H, H_α), 6.95 (brd, 1H, H_β), 3.56 (m, 1H, H₂₁), 3.09 (m, 1H, H₁₁), 2.49 (s, 3H, SMe), 2.41 (m, 1H, H₂₂), 1.39 (brs, 15H, C₅Me₅), 1.23 ppm (m, 3H, Me). ¹³C{¹H} NMR (75.48 MHz, CD₂Cl₂, 223 K): δ = 151.92–120.10 (C_Ar), 147.63 (C_β), 134.48 (C_α), 107.10 (brs, C₅Me₅), 30.04 (m, C₂), 31.54 (m, C₁), 15.82 (dd, J(P,C) = 17.5 and 4.6 Hz, Me), 14.75 (SMe), 10.23 ppm (C₅Me₅). ³¹P{¹H} NMR (121.42 MHz, CD₂Cl₂, 223 K): δ = 75.91 (dd, J(Rh,P) = 131.2 Hz, J(P,P) = 38.7 Hz, P¹), 51.40 ppm (dd, J(Rh,P) = 132.5 Hz, P²).

3-S

¹H NMR (300.13 MHz, CD₂Cl₂, 223 K): δ = 7.98–6.90 (m, 24H, H_Ar), 7.89 (brd, 1H, H_β), 7.50 (brd, 1H, H_α) 3.67 (m, 1H, H₁₁), 3.61 (m, 1H, H₂₁), 2.48 (m, 1H, H₂₂), 1.58 (s, 3H, SMe), 1.50 (brs, 15H, C₅Me₅), 1.31 ppm (m, 3H, Me). ¹³C{¹H} NMR (75.48 MHz, CD₂Cl₂, 223 K): δ = 141.58–120.97 (C_Ar), 138.95 (s, C_β), 135.90 (s, C_α), 108.57 (brs, C₅Me₅), 32.33 (m, C₁), 29.89 (m, C₂), 23.13 (s, SMe), 15.59 (dd, J(P,C) = 17.6 and 4.2 Hz, Me), 10.23 ppm (s, C₅Me₅). ³¹P{¹H} NMR (121.42 MHz, CD₂Cl₂, 223 K): δ = 73.34 (dd, J(Rh,P) = 123.6 Hz, J(P,P) = 38.9 Hz, P¹), 43.84 ppm (dd, J(Rh,P) = 129.6 Hz, P²).

NMR Studies for the Equilibrium 3-O ⇆ 3-S

In an NMR tube, at 193 K, to a suspension of (S_Rh,R_C)-[Cp*Rh{(R)-Prophos}(OH₂)][SbF₆]₂ (1) (20.4 mg, 0.018 mmol) in CD₂Cl₂ (0.50 mL), MTNS (3.5 mg, 0.018 mmol) and 4 Å MS (15 mg) were added. The resulting solution containing complexes 3-O and 3-S was analyzed by NMR at different temperatures. Thermodynamic parameters were obtained through Van’t Hoff analyses.

(S_Ir,R_C)-[Cp*Ir{(R)-Prophos}(κ¹S-MTNS)][SbF₆]₂ (4-S)

³¹P{¹H} (121.42 MHz, CD₂Cl₂, 223 K): δ = 38.26 (d, P¹), 8.71 ppm (d, P²).

Catalytic Procedure for the Friedel–Crafts Reaction

Under argon, to a Schlenk flask equipped with a magnetic stirrer, 0.03 mmol (5 mol %) of (S_M,R_C)-[Cp*M{(R)-Prophos}(OH₂)][SbF₆]₂ (M = Rh (1), Ir (2)), the solvent (4 mL), 0.90 mmol of MTNS, and about 100 mg of 4 Å MS were added. The resulting mixture was stirred for 15 min at the selected temperature and then the corresponding indole (0.60 mmol) was added. After the appropriate reaction time, the reaction was quenched by addition of a methanolic solution of [N(nBu)₄]Br. The resulting suspension was vacuum-concentrated until dryness. The residue was extracted with Et₂O (4 × 6 mL), and the resulting oil was analyzed by NMR and HPLC.

Procedure for the Synthesis of 5

At 193 K, in an NMR tube, to a suspension of (S_Rh,R_C)-[Cp*Rh{(R)-Prophos}(OH₂)][SbF₆]₂ (1) (17.0 mg, 0.014 mmol) in CD₂Cl₂ (0.3 mL), MTNS (27.3 mg, 0.140 mmol) and 4 Å MS (15 mg) were added. After 20 min, N-methyl-2-methylindole (10.2 mg, 0.070 mmol) in CD₂Cl₂ (0.15 mL) was added to the resulting solution at 193 K. The resulting solution was analyzed by NMR without any further purification and showed the presence of complexes 5 in 80/20 5a/5b molar ratio.

5a

¹H NMR (400.16 MHz, CD₂Cl₂, 193 K): δ = 17.6 (br, 1H, OH), 6.12 (d, J = 7.5 Hz, 1H, H_α), 5.25 (brd, 1H, H_β), 3.57 (brs, 3H, NMe), 2.64 (s, 3H, Me), 2.37 (s, 3H, SMe), 1.26 (brs, 15H, C₅Me₅), 0.92 ppm (br, 3H, Me). ¹³C{¹H} NMR (100.16 MHz, CD₂Cl₂, 193 K): δ = 118.43 (C_α), 38.35 (C_β), 34.86 (NMe), 14.75 (SMe), 14.27 (Me), 12.82 (Me_indole), 8.99 ppm (C₅Me₅). ³¹P{¹H} NMR (161.96 MHz, CD₂Cl₂, 193 K): δ = 75.96 (dd, J(Rh,P) = 129.9 Hz, J(P,P) = 42.4 Hz, P₁), 50.32 ppm (dd, J(Rh,P) = 132.4 Hz, P₂).

5b

¹H NMR (400.16 MHz, CD₂Cl₂, 193 K): δ = 17.5 (br, 1H, OH), 6.02 (brd, Hz, 1H, H_α), 4.76 (brd, 1H, H_β), 3.58 (brs, 3H, NMe), 2.64 (s, 3H, Me), 2.39 (s, 3H, SMe), 1.31 (brs, 15H, C₅Me₅), 1.11 ppm (br, 3H, Me). ¹³C{¹H} NMR (100.16 MHz, CD₂Cl₂, 193 K): δ = 113.70 (C_α), 40.62 (C_β), 29.65 (NMe), 14.70 (SMe), 14.57 (Me_indole), 14.34 (Me), 9.21 ppm (C₅Me₅). ³¹P{¹H} NMR (161.96 MHz, CD₂Cl₂, 193 K): δ = 76.83 (brdd, J(P,P) = 42.6 Hz, P₁), 51.92 ppm (brdd, J(Rh,P) = 135.3 Hz, P₂).

Preparation of the aci-Nitro Compound 6

At 193 K, in an NMR tube, to a suspension of (S_Rh,R_C)-[Cp*Rh{(R)-Prophos}(OH₂)][SbF₆]₂ (1) (12.4 mg, 0.010 mmol) in CD₂Cl₂ (0.3 mL), MTNS (58.6 mg, 0.300 mmol) and 4 Å MS (15 mg) were added. After 20 min, N-methyl-2-methylindole (29.0 mg, 0.200 mmol) in CD₂Cl₂ (0.15 mL) was added to the resulting solution. The gradual consumption of the indole with the concomitant appearance of the aci-nitro compound was monitored using NMR at 233 K. When the indole was not observed, compound 6 was fully characterized by NMR at 223 K in the presence of 7 (6/7 molar ratio: 85/15). 6. ¹H NMR (500.13 MHz, CD₂Cl₂, 223 K): δ = 13.41 (br, 1H, OH), 7.32–6.90 (m, 9H, H_Ar), 6.99 (d, J = 8.1 Hz, 1H, H_α), 5.56 (d, J = 8.1 Hz, 1H, H_β), 3.65 (s, 3H, NMe), 2.50 (s, 3H, SMe), 2.44 ppm (s, 3H, Me). ¹³C{¹H} NMR (125.77 MHz, CD₂Cl₂, 223 K): δ = 140-107 (C_Ar), 122.02 (C_α), 39.80 (C_β), 29.80 (NMe), 14.87 (SMe), 10.40 ppm (Me).

Characterization of the Friedel–Crafts Adduct 7

¹H NMR (400.16 MHz, CDCl₃, RT): δ = 7.30–7.17 (m, H_Ar), 7.06 (t, J = 12.5 Hz, 1H), 5.28-5.20 (m, 2H, H_α), 5.16-5.08 (m, 1H, H_β), 3.60 (s, 3H, NMe), 2.51 (s, 3H, SMe), 2.43 ppm (s, 3H, Me). ¹³C{¹H} NMR (100.62 MHz, CDCl₃, RT): δ = 143.67–106.60 (C_Ar), 77.37 (C_α), 38.86 (C_β), 28.11 (NMe), 13.25 (SMe), 9.00 ppm (Me). HPLC (Chiralpak IB with IB guard, n-hexane:iPrOH (70:30) 1 mL/min), t_R: 16.9 (minor), 18.4 min (major). HRMS (ESI): m/z = 341.4475 (calcd 341.4480 [C₁₉H₂₁N₂O₂S]⁺).

NMR Study of the Reversibility of Cycle A

At 193 K, in an NMR tube, MTNS (34.9 mg, 0.179 mmol) and 4 Å MS (15 mg) were added to a suspension of the complex (S_Rh,R_C)-[Cp*Rh{(R)-Prophos}(H₂O)][SbF₆]₂ (1; 6.8 mg, 0.006 mmol) in CD₂Cl₂ (0.3 mL). After 20 min, N-methyl-2-methylindole (17.3 mg, 0.119 mmol) in CD₂Cl₂ (0.20 mL) was added to the resulting solution at the same temperature. The gradual consumption of the indole with the concomitant appearance of the aci-nitro compound 6 was then monitored by NMR spectroscopy at 243 K. N-methyl-2-methylindole progressively reacted and when the solution consisted of a mixture of 6, 7, and indole in a molar ratio of 82:16:2, trans-β-nitrostyrene (26.7 mg, 0.179 mmol) was added at the same temperature. At 243 K, the ¹H NMR analysis showed the formation of the aci-nitro compound derived from trans-β-nitrostyrene and N-methyl-2-methylindole (6′). The measured 6/6′ molar ratios were 87:13 and 23:77 after 10 min and 1 h, respectively. Heating up to 298 K, 6 and 6′ evolved to the corresponding FC products 7 and 7′, respectively.

Procedure for the Preparation of [Cp*Rh{(R)-Prophos}(FC-Adduct)]²⁺ (8-S)

At 193 K, in an NMR tube, to a suspension of (S_Rh,R_C)-[Cp*Rh{(R)-Prophos}(OH₂)][SbF₆]₂ (1) (25.0 mg, 0.022 mmol) in CD₂Cl₂ (0.5 mL), 7 (37.3 mg, 0.110 mmol, racemic) and 4 Å MS (15 mg) were added. The resulting solution was analyzed by NMR without any further purification and consisted of a mixture of complexes 8-S and 1, in 75/25, 8-S/1, and 65/35, 8-Sa/8-Sb molar ratios.

8-Sa

³¹P{¹H} NMR (202.46 MHz, CD₂Cl₂, 263 K): δ = 72.78 (br, P¹), 42.80 ppm (dd, J(Rh,P) = 129.2 Hz, J(P¹,P²) = 34.3 Hz, P²).

8-Sb

³¹P{¹H} NMR (202.46 MHz, CD₂Cl₂, 263 K): δ = 72.78 (br, P¹), 43.21 ppm (dd, J(Rh,P) = 127.6 Hz, J(P¹,P²) = 33.1 Hz, P²).

Crystal Structure Determination for Complexes 3-S and 4-S

X-ray diffraction data were collected on an APEX DUO Bruker diffractometer, using graphite-monochromated Mo κα radiation (λ = 0.71073 Å). Single crystals were extracted from NMR tubes, manipulated, and selected at low temperature on an inert atmosphere (surrounded by dry ice). Compounds 3-S and 4-S crystallize on well-formed yellow and orange prisms, respectively. The prisms are stuck to each other. Several samples were tested before selecting the ones used in the data collections. Most of the tested samples exhibit doubled diffraction peaks and problems in reflection indexation, evidencing the existence of twinning.

Selected crystals were mounted on a fiber, coated with a protecting perfluoropolyether oil, and cooled to 100(2) K with an open-flow nitrogen gas. Data were collected using ω-scans with a narrow oscillation frame strategy (Δω = 0.3°). Selected crystals show isolated well-defined diffraction spots; more than 95% of the harvest reflections were indexed in an orthorhombic C222₁ space group. Consequent data integration (with SAINT program)²⁰ and structure solution (with direct methods with SHELXS)²¹ in C222₁ lead to a problematic structural model in which only the heaviest atoms (Rh, Ir, S, and P) can be anisotropically refined. Furthermore, the observed residual density peaks are anomalously large and the second parameter of the weighting scheme is too large.

A careful inspection of reflection pointed out the existence of a pseudomerohedric twin, with a monoclinic lattice. Therefore, data have been integrated in a monoclinic system, and the structures have been refined by full-matrix least-squares on F² with SHELXL program²² included in a Wingx program system²³ as 2-component twins. It is noteworthy that the same twin law has been observed in rhodium and iridium complexes 3-S and 4-S. Furthermore, several crystallization assays led to the same feature.

All of the nonhydrogen atoms have been anisotropically refined and several rigid-body restraints have been included in the model.

CCDC 1972185 and 1972186 contain the supplementary crystallographic data for this paper. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/structures.

Crystal Data for Complex 3-S

2(C₄₆H₅₀NF₁₂NO₂P₂RhSSb₂)·4(CH₂Cl₂); M_r = 2974.26; an orange plate, 0.101 × 0.167 × 0.198 mm³; monoclinic P2₁; a = 10.9772(6) Å, b = 46.096(3) Å, c = 10.9795(6) Å, β = 93.0010(10)°; V = 5548.0(5) ų, Z = 2, D_c = 1.780 g/cm³; μ = 1.628 cm^–1; min. and max. absorption correction factors: 0.7137 and 0.8251; 2θ_max= 56.454°; 93 923 reflections measured, 27199 unique; R_int = 0.0430; number of data/restraint/parameters 27199/37/1331; R₁= 0.0426 [25 806 reflections, I > 2σ(I)], wR(F²) = 0.1018 (all data); largest difference peak 1.551 e·Å^–3; Flack parameter: −0.005(7); BASF parameter: 0.505.

Structural Data for 4-S

2(C₄₆H₅₀IrNF₁₂NO₂P₂SSb₂)·3(CH₂Cl₂)·2(H₂O); M_r = 3103.94; a yellow prism, 0.053 × 0.098 × 0.104 mm³; monoclinic P2₁; a = 10.9760(18) Å, b = 46.000(8) Å, c = 10.9840(18) Å, β = 93.038(3)°; V = 5537.9(16) ų, Z = 2, D_c= 1.861 g/cm³; μ = 3.687 cm^–1; min and max absorption correction factors: 0.6212 and 0.7955; 2θ_max = 55.134°; 59 017 reflections measured, 25 261 unique; R_int = 0.0501; number of data/restraint/parameters 25261/115/1321; R₁= 0.0488[21281 reflections, I > 2σ(I)], wR(F²) = 0.1106 (all data); largest difference peak 1.574 e·Å^–3; Flack parameter: 0.012(4); BASF parameter: 0.352.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.0c03485.

• Ring-puckering parameters of 3-S and 4-S, thermodynamic parameters for the 3-O ⇆ 3-S equilibrium, HPLC chromatograms and UV spectra (PDF)

• Structural analyses of complexes 3-S and 4-S (CIF)

The authors declare no competing financial interest.