2. Results and Discussion
Friedel–Crafts alkylation (F–C alkylation) of indoles is an important reaction for the formation of new carbon-carbon bonds and construction of versatile indole-containing scaffolds [
35,
36]. Organocatalysis in this area has received tremendous success under nearly all the major catalytic activation modes. Particular interest was focused on H-bonding catalytic F–C alkylation due to the mildness of reaction conditions and the broad tolerance of functional groups [
1,
36,
37,
38]. A valuable unprotected indole N–H moiety made the F–C adducts readily accessible to bioactive compounds [
37]. Importantly, a free N–H readily accepts H-bond donor from a nucleophilic catalyst.
In the context of the ordinary H-bond [
36,
37,
39] and the metal-enhanced H-bond catalysis,[
40] we expanded exploration of the novel H-bond donor implemented by the vicinal positive on the cyclopropenium in F–C alkylation of indole
1 with trans-β-nitroalkene
2. Tri(phenylamino)cyclopropenium chloride (TPAC·Cl) was selected as a quintessential HBD catalyst (
Table 1). The TPAC·Cl was able to promote the F–C alkylation of
1a with
2a at 25 °C in dichloromethane efficiently (
Table 1, entry 2); in contrast, the noncatalyzed background reaction was negligible (entry 1). The TPAC·F ion pair, replacing the TPAC·Cl, performed no activity in the same reaction by changing the chloride to fluoride anion (entry 3). Poor catalytic performance implied that the chloride anion may be necessary for F–C alkylation. A possible reason was that the fluoride anion was easily coordinated with the HBD of the cationic TPAC, similar to the normal fluoride anion receptors with HBD structures [
9,
41,
42]. Poor catalytic performance suggested that TPAC·F formed a tight ion pair [
43] and the cationic TPAC preferably paired with the fluoride anion rather than to activate a substrate.
Dialkylamino-substituted cyclopropenium was tested on the F–C alkylation to verify the assumption that N–H moieties of the TPAC as an HBD were essential in the H-bonding catalysis (entries 4 and 5). The tris(dimethylamino)cyclopropenium chloride (TDAC·Cl) displayed no catalytic performance on the F–C alkylation (entry 4). The possible reason was the lack of N–H moiety on the cationic core TDAC and loss of the ability to work as an H-bond donor. Although fluoride anion, possessing strong nucleophilic, was supposed an excellent H-bond acceptor, the TDAC·F as catalysts was not workable on F–C alkylation (entry 5). These experimental results supported that the N–H moieties of TPAC as an HBD were essential.
The tetrabutylammonium chloride (TBA·Cl) along with tetrabutylammonium fluoride (TBA·F), counterparts to TPAC·Cl and TPAC·F, respectively, performed inactive in the benchmark F–C alkylations (entries 6 and 7). The discrepancy between TPAC·Cl and TBA·Cl suggested the necessity of the cationic structure of tris(monoalkylamino)cyclopropenium. We suggested that the H-bond donor realized by the vicinal cyclopropenium show excellent catalytic performance.
With primary success, we expanded substrates by using various indoles along with nitroalkenes (entries 8 to 15). Indoles
1a–
e bearing different substituents on both the benzene ring and the pyrrole ring were conductive in reactions with nitroalkene
2a. Whereas the reactions of bare indole
1a and indoles with electron-donating groups (
1b and
1c) afforded the corresponding products
3aa,
3ba, and
3ca in good yields (
Table 1, entries 8–9), an electron-withdrawing chlorine on the 5-position caused
1d proceeded in a moderate yield of
3da (
Table 1, entry 10). Steric hindrance at 7-position retarded the reaction (
Table 1, entry 11), which could be accounted for by the unfavorable interference to chloride anion. The applicability of the catalyst was further supported by the variation of the nitroalkene partners. Nitroalkenes with substitutions on the benzene ring (
2b–
d), both electron-donating and -withdrawing, decreased the yields (
Table 1, entries 12–14) as compared with a non-substituted one. Nevertheless, thienly nitroalkene
2d reacted stably with indole
1a to obtain the corresponding products
3ad in good yield (
Table 1, entry 14).
The single crystal of TPAC·Cl was prepared to show the visual view of the construction in the solid, including the interionic distances. A cubic system was confirmed by an X-ray diffraction analysis of the catalyst of TPAC·Cl. The chloride anion was closer to the benzene ring than the formally cationic (C
3N
3)
+ core (
Figure 1, a). The chloride anion is coplanar with cyclopropenium core (
Figure 1, b), but each phenyl group is slightly skewed out of the plane. Three TPAC coordinate to one Cl (
Figure 1, c), while each TPAC aligns in one of the three orthogonal planes of x, y, and z (
Figure 1, d). The structure of TPAC·Cl is C
3v symmetry. The distances, between the chloride anion and the positive core (C
3)
+ of the three carbons, showed distinctly larger distances than the normal ones, viz. 4.1086 Å, 4.8124 Å and 5.3485 Å, respectively [
44]. These data described the 3D architecture of TPAC·Cl in the solid state.
Two plausible mechanisms were proposed (
Scheme 2) based on the HBD implemented by cyclopropenium. One possible competitive interaction was the H-bonding attraction between the cationic TPAC and the counter anion, which decreased the activation of cationic HBD to the substrate, especially the fluoride anion (
Scheme 2, A). Thus, the changes of the anion from fluoride to chloride anion promoting the catalytic performance may be the reason for the weak attraction between the TPAC and the chloride anion. The high-lying closed-shell HOMO of aminocyclopropenium cation contending against the closed-shell HOMO of chloride anion will counteract the ionic electrostatic attractions [
45]. The phenomenon is called “ion pair strain” [
46,
47]. The cationic TPAC and the anionic Cl
–, in this case, will keep away from each other due to the resistance, and reach obviously larger interionic distances under the readjusted dynamic equilibrium. The counter chloride anion is potential HBA to activate the substrate nucleophilically. The TPAC as an H-bond donor, implemented by vicinal positive charged on the cyclopropenium core, coordinated to the two oxygen of the triangle planar nitro group (
Scheme 2 B). The counter chloride anion Cl
–, cooperatively, possibly coordinating to hydrogen of the N–H on the indole ring, played the role of an H-bond acceptor (HBA). Different from the prevailing viewpoint of HBD and Lewis base as co-catalysis, we preferred to suggest tris(monoalkylamino)cyclopropenium as HBD and counter chloride anion as potential HBA in cooperative catalysis.
NMR titration experiments were performed to verify the H-bonding interaction between the N–Hs of TPAC·Cl and the substrate of nitroalkene
2a (
Figure 2). The chemical shifts of the methine of
2a exhibited downfield shifts from 8.247 to 8.267 ppm by increasing the ratio of [TPAC·Cl]/[
2a]
0 from 0 to 2 (
Figure 2). The two different methines were due to the geometric isomerism of the
2a by C=C. These shifts were important evidence that the catalyst cation of TPAC·Cl as HBD could activate the nitro compounds of
2a by H-bonding. To validate the chloride anion as a potential H-bond acceptor (HBA) with N–H of indole
1a, NMR titration experiments were performed (
Figure 3). The chemical shifts of the H-bonding N–H of indole exhibited downfield shifts from 11.103 to 11.307 ppm by increasing the ratio of [TPAC·Cl]/[
1a]
0 from 0 to 2 (
Figure 3). These were important evidence that the counter anion of catalyst TPAC·Cl as HBA could activate the indole
1a by H-bonding.
3. Materials and Methods
The organic solution was concentrated using Buchi rotary evaporator or IKA rotary evaporator. The machine of nuclear magnetic resonance was a type of Bruker-AV-400 (400 MHz). The detecting temperature was 25 °C and the protic solvent was CHCl3 or DMSO. The substrates of indoles and nitroalkenes were purchased from Sigma Aldrich without additional purification. All experiments were executed by standard Schlenk reaction techniques under an argon atmosphere. Dichloromethane was stirred with CaH2 for 10 h and distilled under an argon atmosphere. The purified dichloromethane was stored in 3 Å molecular sieve pellets. Toluene, sodium and diphenyl ketone were heated and stirred until dark purple color came flooding out. The purified toluene was deposited in 3 Å molecular sieve pellets.
Preparation of N-trimethylsilylaniline [
31,
32,
33]
Argon airflow was employed to protect all operations progress under standard Schlenk techniques. Freshly distilled 3.0 mL aniline mixed with 1.7 mL chlorotrimethylsilane in 20.0 mL dry benzene at reflux for 1 h. Aniline hydrochloride was separated out in this system and removed by filtration. The filtrate was dried by rotary evaporator to obtain N-trimethylsilylaniline as a yellow oil: 2.46 g, 76 % yield. 1H NMR (400 MHz, CDCl3) δ 7.17 (dd, J = 8.5, 7.5 Hz, 2H), 6.73 (t, J = 7.5 Hz, 1H), 6.68 (d, J = 8.5 Hz, 2H), 3.45 (brs, 1H), 0.30 (s, 9H).
Preparation of the catalyst TPAC·Cl [
34]
Argon airflow was employed to protect all operations progress under standard Schlenk techniques. Freshly prepared 2.10 g N-trimethylsilylaniline was placed to 0.5 mL tetrachlorocyclopropene in 50.0 mL dry dichloromethane and stirred for 6 h. The white precipitate was precipitated gradually. The dichloromethane was used to clean up the white solid. Finally, the white solid was recrystallized from methanol: 0.78 g, 52 % yield; m.p, 207.3 °C (decomp); 1H NMR (400 MHz, DMSO-d6) δ 11.11 (s,3H), 7.41 (t, J = 7.8 Hz, 6H), 7.36 (d, J = 7.6 Hz, 6H), 7.16 (t, J = 7.2 Hz, 3H); 13C NMR (100 MHz, DMSO) δ 138.78, 129.68, 123.94, 118.02, 112.83; HRMS (ESI-TOF) m/z: [M + H]+ calcd for C21H18N3 312.1495; Found 312.1463.
Argon airflow was employed to protect all operations progress under standard Schlenk techniques. Thin-layer chromatography (TLC), combined with UV light, was used to monitor the reaction process. Purification was performed by flash column chromatography with silica gel 60 N (Kanto Chemical Co., Inc) or Isolera one with SNAP Ultra Column. In 10 mL reaction tube, 1.0 mmol nitroalkenes 2a–d along with the 0.0347 g TPAC·Cl were weighted in 10.0 mL dichloromethane, the 1.5 mmol indoles 1a–e were placed. The reaction tube was then placed at room temperature for 24 h, the product 3 was obtained by column chromatography (n-hexane/EtOAc mixtures).