The indole nucleus, recognized as a biologically significant moieties^[1-4], has been the target of numerous synthetic methods^[5-9]. Often, synthetic strategies introduced copper compounds to catalyze the indole derivatives formation reaction in excellent yields^[10-12], and proceeded via a presumed mechanism (Scheme 1). Unfortunately, many of these transformations applied to the indole ring system need a high dosage of copper catalyst, give merely moderate yields (34%~53%)^[10-11] and are not favorable for the sufficient se-paration of catalysts and/or products. Overall, it would not be most welcome from atom-economy and environment perspective.

Graphitic carbon nitride (g-C₃N₄) as a sort of polymeric super molecular materials exhibits excellent stability as well as good environmental acceptance. The surface of g-C₃N₄ contains a small amount of hydrogen present as primary and/or secondary amine groups on the terminating edges, which can form a quantity of surface defects along with its unique dimensional layered structure. It is favorable for hybridizing with other components, due to this structure fragment and the electronic properties (a larger size of the π system). Accordingly, g-C₃N₄ features almost the precursor of transition metal heterogeneous catalyst^[13-16]. Herein, we would like to synthesize the Cu₂O/g-C₃N₄ on the pioneering work of predecessor, and utilize it to catalyze the formation of indole-2-carboxylic esters in order to avoid catalyst's separation problems, the low yields and the high loading of catalyst, starting from 2-bromobenzaldehyde and ethyl acetamidoacetate.

1 Experimental section 1.1 Preparation of the g-C₃N₄ polymeric materials

Firstly, melamine (2.5 g) powder was put into a semi-closed alumina crucible with a cover under N₂ gas flow to 550 ℃ for 4 h at a heating rate of 5 ℃/min, followed by naturally cooling to room temperature. After cooling, the yellow product was milled and collected.

1.2 Preparation of the Cu₂O/g-C₃N₄ composite

0.2 g of g-C₃N₄ and 0.075 g of CuCl₂·2H₂O were ultrasonically dispersed in 100 mL mixture solution of H₂O/EtOH (the volume rate is 1:1) for 1.5 h, and 100 mL of aqueous solution containing 0.165 8 g of NaBH₄ was added dropwise under stirring at 25 ℃. The yellow-brown solution changed gradually to deep blue within 0.5 h. Then the deep blue solution was stirring for 16 h. The product were centrifuged and washed with distilled water and EtOH for several times, and dried in a vacuum oven at 60 ℃. The brown product was obtained in 0.23 g.

1.3 Characterization

The X-ray diffraction patterns of the Cu₂O/g-C₃N₄ and the g-C₃N₄ were recorded by diffractometer (D/MAX Ultima Ⅳ, Rigaku, Japan). The X-ray photoelectron spectroscopy spectra were examined on the X-ray energy spectromer (Kratos Analytical Axis Ultra, United Kingdom). Fourier transform infrared spectroscopy spectra were scanned on the Nicolet-6700 spectrometer. SEM measurements were carried out by the JSM-6510LV scanning electron microscopy (Rigaku, Japan).

1.4 General procedure for indole-2-carboxylic esters

Caesium carbonate (652 mg, 2 mmol) and catalyst (30 mg) were added to a dried reaction tube. The tube was filled with nitrogen before DMSO (2 mL) was added. Then 2-bromobenzaldehyde (185 mg, 1 mmol) and ethyl acetamidoacetate (156.2 mg, 1.2 mmol) were added prior to heating the slurry to 100 ℃ under a nitrogen atmosphere for 12 h with good agitation. After cooled down, the mixture was added DI water (5 mL) and extracted with ethyl acetate (5 mL) for 3 times. The combined organic fraction was washed with DI water (5 mL) and brine (5 mL). After dried over sodium sulfate and filtered, the organic phase was concentrated to give crude product. The crude product was further analyzed by GC Agilent 7890A gas chromatography using HP-5 column (30 m × 0.32 mm × 0.25 μm) and FID detector

2 Results and Discussion

Schematic illustration of the synthesis process of the Cu₂O/g-C₃N₄ composite is indicated in Scheme 2. The first step requires the calcination of melamine at high temperature to obtain the entitled product (g-C₃N₄). The second step is the adsorbing of Cu²⁺ on the surface defects of g-C₃N₄ with the addition of CuCl₂·2H₂O. The third step is the reduction reaction of Cu²⁺ under the presence of reducing agent.

The actual images of the Cu₂O/g-C₃N₄ and g-C₃N₄ are exposed in Fig. 1. The g-C₃N₄ is off-white powder. Compared to g-C₃N₄, the Cu₂O/g-C₃N₄ is a small fulvous block-liker, which suggests that Cu₂O is successfully adsorbed on the surface defects of g-C₃N₄ from the facade of the prepared product.

As shown in Fig. 2, the FTIR spectrum of the Cu₂O/g-C₃N₄ is similar with that of the g-C₃N₄, because of no characteristic peak of Cu₂O. The peak at 810 cm^-1 is attributed to out of bending modes C-N heterocycles, while the peak between 1 250 and 1 650 cm^-1 are assigned to aromatic C=N and C-N stretching vibration. The measurement result is correspond to the reported literature^[17].

XRD and XPS analysis were performed to further confirm the formation of the Cu₂O/g-C₃N₄. The full range XPS spectrum of Cu₂O/g-C₃N₄ (Fig. 3) reveals that the composite contains C, N, O, and Cu element. It was clear that the peak appearing at 530.6 eV implied the presence of O^2- and the peak at 287.9 eV was attributed to the carbon of C—C, while the characteristic peak at 932.6, 952.5 eV were assigned to Cu^+[18-20].

The diffraction peaks of Cu₂O/g-C₃N₄, Cu₂O and pure Cu can be observed simultaneously, according to the information provided by the schematic diagram (Fig. 4). The peak's intensity (at 42.2°, 61.6°, 73.6°) of pure Cu is much weaker than that (at 37.7°) of Cu₂O, which reveals the presence of a light amount of pure Cu. The high intensity peak located at 27.6° is corresponded to the pure g-C₃N₄, which is in accordance with previous report^[21].

Additionally, SEM measurement was performed to investigate the microstructure of the Cu₂O/g-C₃N₄ and g-C₃N₄. From the SEM image (Fig. 5A) of g-C₃N₄, it can be seen that the g-C₃N₄ appears the sheet spatial structure and bending aggregation. It is visible from Fig. 5B that the appearance characteristics of g-C₃N₄ are modified, which is caused by the adsorption of numerous Cu₂O. All of these measurement results can confirm that the prepared sample is the Cu₂O/g-C₃N₄ composite though there is a trace amount of pure copper.

During the course of work, we set out to research the catalytic potentiality of Cu₂O/g-C₃N₄ for the synthesis of indole-2-carboxylic esters, using 2-bromobenzaldehyde and ethyl acetamidoacetate as reaction substrates in DMSO. To begin with, what is investigated that the catalytic efficiency of this transformation is affected by the embedded quantity of Cu in the case of the Cu₂O/g-C₃N₄ particles (see Table 1).

A low yield of 7.70% and 13.5% was obtained when the reaction was proceeded with the pure g-C₃N₄ and Cu₂O respectively (Table 1, entry 5 and 6). Meanwhile, the mixture of Cu₂O and Cu gave the yield of 8.20% (Table 1, entry 7). As anticipated, the catalytic activity of this reaction was improving as the loading of Cu increasing until 12.3% in the case of the Cu₂O/g-C₃N₄ (Table 2, entries 1-4). So, the Cu loading (12.3%) of the Cu₂O/g-C₃N₄ was selected as the heterogeneous catalyst in the following investigation. After examining of the effects of various embedded amount of Cu, various bases and reaction media were explored. Following the bases and solvents screen (Table 2), we found no reaction took place at all using 1-decanol as solvent (Table 2, entry 7). Meanwhile, a trace amount of the desired product was observed in the DMF which was proven to be preferable solvent for this reaction^[10]. Other solvents such as NMP, Toluene, 1, 4-dioxane failed to provide more than a trace amount of the desired products (Table 2, entries 2-4). To our delight, the yield still resulted in 35.9% and 31.5% respectively when DMSO and Ethyl acetate were employed as the solvent (Table 2, entries 1 and 5). Subsequently, the influence of the base was evaluated in DMSO. A series of base (K₂CO₃, K₃PO₄, CH₃COONa, LiOH, KOH, NaOH, KHCO₃) were investigated in the reaction, and we found that the screened base (Cs₂CO₃), giving good yield (35.9%), was preferred for this transformation (Table 2, entry 1 vs entries 8-14). A slight decrease in the amount of Cs₂CO₃ resulted in an obvious decrease in the yield of the desired product (Table 2, entry 15). Considering the moderate basicity, we chose Cs₂CO₃ (2 equiv) as base in the formation of indole-2-carboxylic esters.

Furthermore, we attempted to investigate the influence of reaction temperature and the catalyst loading (the Cu₂O/g-C₃N₄) in the transformation. As seen in Table 3, the yields were found to increase with a raising in the temperature (Table 3, entries 1-3) although the condition of high temperature (120 ℃) retarded the reaction (Table 3, entry 4). In the case of catalyst (the Cu₂O/g-C₃N₄) dosage, the experiments performed similarly with that of the temperature: increasing the dosage of the Cu₂O/g-C₃N₄ composite improved the yield of the reaction (Table 3, entries 3, 5-6). Surprisingly, the high dosage (40 mg) of catalyst led to a slight decrease in the yield (Table 3, entry 7). From what has been discussed above, the optimal condition of the one-pot three-step formation of indole-2-carboxylic esters is refined as follow: the Cu₂O/g-C₃N₄ (30 mg, 5.8% Cu), Cs₂CO₃ (2.0 equiv), DMSO (2 mL) and 100 ℃.

With our reaction conditions refined in hand, we further probed the scope of this novel methodology. Thus, a series of starting substrates were subjected to the reaction in DMSO (Table 4). All of benzaldehydes containing electron-rich groups reacted to give the expected indole product in the range of 30.6%~34.4% yield (Table 4, entries 5 and 6), while electron-poor substrates did not react well with this system (Table 4, enties 2-4).

3 Conclusion

In summary, the Cu₂O/g-C₃N₄ composite was synthesized by the calcination of melamine, ultrasonic exfoliation, and reduction of Cu²⁺ adsorbed on the defects of g-C₃N₄, and characterized by FTIR, XRD, XPS and SEM. Meanwhile, the Cu₂O/g-C₃N₄ was utilized to catalyze the formation of indole-2-carboxylic esters and gave desired yield although the low dosage (30 mg, 5.8% Cu) of catalyst was used. To the best of our knowledge, this procedure has not been reported to date.