1 Introduction

Noble metals are capable of catalyzing most of the organic transformations [1,2,3], wherein, palladium is one of the key transition metals, generally engaged as an efficient catalyst for the cross-coupling reactions of Mizoroki–Heck and Suzuki–Miyaura reactions, which are the most promising synthetic routes in pharmaceutical industries. Most of these industries rely on homogeneous Pd catalysts, but the involvement of difficulties in catalyst separation and its reusability are unavoidable. Palladium contamination in the product is also one of the notable issues. To overcome these difficulties the research directed towards the development of heterogeneous catalysts. Unambiguously, the problem of catalyst separation and recyclability was solved with heterogeneous catalysts, but the leaching and agglomeration of palladium [4,5,6,7] are the most unwanted issues still to be resolved. Phosphine ligands are being used to stabilize Pd species, prevention of Pd agglomeration followed by inactive palladium black formation [8, 9]. Product contamination with ligand residues is one of the notable issues in pharmaceutical industries. The permissible limit of contamination is generally below 10 ppm [10], hence developments of alternative eco-friendly heterogeneous catalysts are highly desirable. To circumvent the problem of palladium leaching, different porous materials having peculiar porous network were used as supports, which are capable of preventing the leaching by accommodate palladium particle in the pore texture. But, still complete prevention of palladium leaching is not yet been achieved. In this context, immobilization of palladium with certain functional groups that are anchored to supports like silica, carbon, polymer, Pd@CDNS–g–C3N4, ordered mesoporous SBA-15/PrSO3 and metal–organic framework [11,12,13,14,15,16,17,18] is one of the important options.

Supports having high surface area, suitable porosity, held with organic functional groups capable of firm anchoring of palladium are highly desirable to prevent the leaching. Hydrogen titanate nanotubes are intensively studied materials, attracting much attention in the past decade due to unique textural and physicochemical properties including versatile applications, particularly in catalysis and photo catalysis. Titanate nanotubes that are synthesized from TiO2 nanoparticles (particle size < 25 nm) by hydrothermal method possess high specific surface area (≈ 200–400 m2/g), which are widely used because of their flexibility towards desired modifications. Albeit, the preparation of titanate nanotubes from TiO2 nanoparticles well established, preparing from commercial TiO2 powder are limited.

Amine and imine functional groups are the most common chelating agents which are anchored to the surface of the catalyst supports. Polydopamine (PDA) deposition on the silicone nanofilament is one of the best known example [19].The suppression ofPd leaching by nitrogen and oxygen containing polymer is exploited in nitrogen rich triazine functionalized mesoporous covalent organic polymer [20].

According to earlier reports TiO2 and SiO2 were modified with urea and termed as N-doping or N-modification, but the actual N-species has not been identified. The subsequent studies revealed that the N-doped species is C3N4 on the surface of TiO2 powder using the surface –OH of bulk TiO2. It can be ascertain that it is not only a simple nitrogen insertion but also surface modification with triazine ring derivatives (malon) [21]. Pyrolysis of urea beyond 400°C gives graphitic carbon nitride [22], similarly, pyrolysis of urea in presence of titania and silica at 350–400°C produces a unique materials of (C3N4 triazine) derivatives covalently attached to the semiconductor which is also called as covalently coupled inorganic–organic semiconductor [23].

The commonly used solvents for Miziroki–Heck and Suzuki–Miyaura C–C coupling are dimethylformamide, dimethylsulfoxide, N-methylpyrrolidin-2-one and N,N-dimethylacetamide. Of late, attention is diverted towards non-toxic bio-mass derived renewable solvent like γ-Valerolactone [24]. But less attention is being paid to use water as a solvent even though it is cheap, abundant and clean.

The present work deals with the design and synthesis of novel heterogeneous palladium catalyst (Pd–C3N4@TNT) for effective utilization in Mizoroki–Heck and Suzuki–Miyaura carbon–carbon bond forming reactions under eco-friendly water medium. To the best of our knowledge synthesis of melon (C3N4) anchored titanium nanotubes (TNT), immobilization of palladium in melon network to attain atomic level dispersion of palladium, maintain the leaching-free, agglomeration-free environment, application as catalyst for Mizoroki–Heck and Suzuki–Miyaura carbon–carbon bond forming reactions under eco-friendly water medium is the first report. Herein, details of catalyst preparation, characterization and catalytic applications are delineated.

2 Experimental Section

2.1 Synthesis of Pd–C3N4@TNT Catalyst and Catalytic Activity

Synthesis of catalyst involves (i) the transformation of bulk TiO2 powder into titanate nanotubes (ii) initial formation of sodium form of titanate nanotubes, (iii) transformation of sodium form of titanate nanotubes (Na-TNT) into H form of titanate nanotubes (H-TNT), (iv) anchoring of C3N4 (melon) to H-TNT, and (v) immobilization of palladium in the C3N4 network. The detailed catalyst synthesis is as follows.

2.2 Synthesis of H-TNT

Typically, 2.5 g of commercially available TiO2 (Sigma–Aldrich) powder was dispersed in 10 M NaOH (200 ml) aqueous solution and subjected to hydrothermal treatment at 130 °C in a Teflon-lined autoclave (250 ml capacity) for 48 h under autogeneous pressure in static condition. After cooling to room temperature, precipitate was separated by filtration and washed thoroughly with distilled water. The resultant white precipitate is the sodium form of titanate nanotubes (Na-TNT), which was treated with 0.1 M HCl (1000 ml) at room temperature under constant stirring for 12 h, filtered, washed, dried at 100 °C for 12 h, which results the formation of H form titanate nanotubes, hereafter this material is used as H-TNT [25].

2.3 Anchoring of C3N4 to H-TNT

Anchoring of C3N4 to H-TNT was achieved by pyrolysis of urea. In a typical procedure, 1 g of HTNT was dispersed in 50 ml of aqueous urea solution (5 g urea in 50 ml of water) stirred at RT for 1 h and evaporated the excess water on a water bath, dried at 60°C for 6 h followed by hot air oven at 100 °C for 12 h. The resultant material was taken in a glass reactor in between the two layers of glass wool. On the top layer of glass wool 5 g of urea was placed and increased the temperature to 400 °C in the flow of N2 gas and calcined at 400°C for 2 h in N2 gas environment (inert atmosphere). The obtained yellow materials was washed with water (200 ml) and ethanol (50 ml) and dried at 100°C in a hot air oven for 12 h, yielded the graphitic-C3N4 anchored to H-TNT, which is hereafter presented as C3N4@TNT.

The proposed mechanism of C3N4 formation and its anchoring on TNT was displayed in Scheme 1. As per previously reported in literature [21, 23] the thermal pyrolysis of urea to melamine which give to polymerized to C3N4 lead in four steps (1) urea decomposition to isocyanic acid (2) the reacts with isocyanic acid forms Ti–NH2 and CO2 (3) some of isocyanic acid convert to its tautomer cyanic acid (4) cyanic acid reacts with Ti–NH2 gives cyanamide. It is necessary to presence of OH groups containing a catalyst to convert cyanamide, which can trimerize to melamine, in the absence of the heterogeneous catalyst, melamine is formed only under high-pressure conditions. Calcination environment is a crucial step to get C3N4@TNT. Unless N2 atmosphere is maintained the resultant product is N-doped TiO2 nanotubes (N-TNT) which is confirmed by 13C NMR.

Scheme 1
scheme 1

Schematic representation of C3N4@TNT synthesis

Full size image

2.4 Immobilization of Palladium

0.975 g of C3N4@TNT was dispersed in 30 ml of DCM and sonicated for 30 min. 0.053 g of palladium acetate dissolved in 50 ml of DCM was slowly added to the sonicated C3N4@TNT under content stirring at room temperature for 3 h, centrifuged and the obtained solid material was kept in oven for drying at 100°C for 12 h, followed by annealing at 250 °C for 1 h under N2 flow. The resultant palladium immobilized C3N4@TNT is coded as Pd–C3N4@TNT, where the palladium loading is around 2 wt% determined by SEM–EDX analysis Fig. 1.

Fig. 1
figure 1

SEM-EDX of Pd–C3N4@TNT

Full size image

3 Results and Discussion

The elemental mapping of Pd–C3N4@TNT was made and the images are presented in Fig. 2, which shows that the spacial arrangement of C, N, Pd, Ti, and O respectively and are homogeneously distributed.

Fig. 2
figure 2

a SEM image of Pd–C3N4@TNT the resultant color-coded Single-element distribution maps of Pd–C3N4@TNT containing, b C, c N, d Pd, e Ti, f O

Full size image

Surface area and pore volume of H-TNT, C3N4@TNT and Pd–C3N4@TNT are listed in Table 1, gradual decrease in surface area and pore volume from H-TNT to C3N4@TNT confirming the anchoring of C3N4 onto H-TNT. Since surface area and pore volume of C3N4@TNT and Pd–C3N4@TNT are more or less equal, immobilization of palladium within C3N4 networked titanate nanotubes can be ascertained. The amount of Pd in Pd–C3N4@TNT was measured by ICP-OES and it is 1.7 wt%.

Table 1 BET surface area of TNT catalysts
Full size table

TEM images were shown in Fig. 3, convey the titanate nanotubes produced from titanium powder retained their tubular structure with internal diameter 4.834 nm and interlayer distance of 0.35 nm even after the anchoring of C3N4 and the palladium immobilization From Fig. 4, the C3N4 formation has been confirmed using 13C NMR spectroscopic analysis. Solid state 13C NMR chemical shifts values for Pd–TNT@C3N4 obtained in the region of 163.86 ppm and 155.54 ppm. The 13C NMR spectral signals at 165 ppm and 155.1 ppm represent the poly(tri-s-triazine) structure which is the characteristic of carbon nitride [26,27,28,29].Close observation of 13C spectral signals at 163.86 ppm and was assigned to C atoms that were not associated with sp3 N, whereas a second peak at 155.54 ppm was suggested to represent C atoms close to the bridging –N groups rather than at 165 ppm and 155.1 revealing the interaction of C3N4 with titanate nanotubes as well as immobilized palladium species. It is one of the evidences for the formation of malon structure of graphitic carbon nitride (C3N4) and its interaction with both titanate nanotube and immobilized palladium. Similar observation was made by Gao et al. [30] when potassium was interacted with C3N4.

Fig. 3
figure 3

TEM images a titanate nanotubes, b Pd–C3N4@TNT

Full size image
Fig. 4
figure 4

Solid state 13C NMR of C3N4@ TNT and N-TNT

Full size image

The XRD pattern of commercially purchased TiO2 powder shown in Fig. 5 is in the anatase form in accordance with the diffraction peak positions at 2θ = 25.4, 37.9, 48.0, 54.0, 55.1,62.7, 68.8,70.4 and 75.1° and the corresponding (101), (004), (200), (105), (211), (204), (116), (200) and (215) crystal planes (JCPDS: 01-07-2486C). The XRD pattern of H-TNT is poorly resolved confirming the transformation of highly crystalline anatase TiO2 into nanoscopic material (H-TNT). Albeit, the XRD pattern of H-TNT is poorly crystalline, four diffraction peaks were observed at 2θ = 9.33, 24.1, 28.3, and 48.4°, which are typical of monoclinic layered hydrogen titanate (JCPDS-ICDD Card No. 47-0561). The diffractograms of C3N4@TNT and Pd–C3N4@TNT are similar to that of the diffractogram of H-TNT, except the appearance of XRD reflection at 2θ = 12.4° implying the distance between fragments of tri-s-triazine (C3N7) in (100) plane of graphitic–C3N4, (C6N7)–N–(C6N7), the strongest peak at 27.6° (002) plane is a characteristic interplanar stacking peak of 2 aromatic systems.

Fig. 5
figure 5

XRD patterns of commercial TiO2 powder, H-TNT, C3N4@TNT and Pd–C3N4@TNT

Full size image

According to X-ray photoelectron spectra (Fig. 6) the binding energies of carbon at 288.24 eV and N at 398.78 eV are responsible for sp2 C=N bond in the g-C3N4 triazine ring and the peak for N at 400.99 assigned to (C)3–N or (C)2–N–H. The peaks at 288.24 eV and 284.6 eV in the C zone are attributed to electrons originating from a sp2 carbon (C) atom attached to an NH2 group and to an aromatic carbon atom [31, 32]. Whereas, the weak peak at 286.4 eV is ascribed to C–O from adsorbed CO2 and absence of a peak at 290.7 eV indicating the complete consumption of urea [33]. The binding energies observed for both Pd–C3N4@TNT annealed catalyst (at 250°C) and used catalyst are at 343.0 eV, 337.8 eV and 341.0 eV, 335.8 eV respectively corresponding to Pd 3d3/2and Pd 3d5/2 indicating that metallic Pd and Pd2+ in both the catalysts Fig. 6d and e, confirming that both the catalysts reduced some extent during annealing and reaction, this phenomenon has been reported elsewhere. In Fig. 6c and f the shift in XPS spectra of titania 2p for H-TNT and C3N4@TNT were found at 459.29 and 464.6 eV of TiO2 correspond to the Ti 2p3/2 and 2p1/2 respectively [34, 35].

Fig. 6
figure 6

XPS spectra of Pd–C3N4@TNT a N region, b C region, c Ti in H-TNT, d Pd in fresh catalyst, e Pd in used catalyst, f Ti in C3N4@TNT

Full size image

In FT-IR spectra (Fig. 7) several bands are observed in the range of 1200–1650 cm−1 for C3N4@TNT, is characteristic of the usual stretching modes of C–N heterocycles. Another band found at 808 cm−1 is contributing from the ring sextant out-of-plane bending vibrations of triazine rings. The absence of a band at 2500 cm−1 indicating the absence of urea moiety, the broad bands at appears 3330 cm−1 and 3160 cm−1 are indicative of stretching vibration modes for OH and NH. These results further confirming that g-C3N4 structure in TNT prepared by the supramolecular aggregation from urea.

Fig. 7
figure 7

FT-IR spectra of H-TNT, C3N4@TNT and Pd–C3N4@TNT

Full size image

3.1 Mizoroki–Heck Reaction

The catalytic activity of Pd–C3N4@TNT was determined in the liquid phase. In a typical procedure, 1 mmol acrylate 1 mmol iodobenzenes, 5 mol percent of TBAI, 10 mg of the Pd–C3N4@TNT catalyst and 3 ml of water were added to a RB flask and stirred magnetically at 75–80 °C in atmospheric conditions 45 min, the products were extracted with ethyl acetate and DCM mixture 3:1 the catalyst was separated by centrifugation. The catalytic activity of Pd–C3N4@TNT catalyst for the Mizoraki–Heck reaction was determined and data obtained at optimized conditions has been displayed in Table 2 and the scope of the catalyst performance has also been determined and the data displayed in Table 3. The TON and TOF of the Pd–C3N4@TNT catalyst for this reaction is 625 and 833 h−1 respectively.

Table 2 Optimization of reaction parameters over Pd–C3N4@TNT catalyst for Mizoroki–Heck coupling
figure b
Full size table
Table 3 Scope of Mizoroki–Heck coupling over Pd–C3N4@TNT catalyst
Full size table

3.2 Suzuki–Miyaura Coupling Reaction

The reaction was performed by taking 1.2 mmol of phenyl boronic acid, 1 mmol of iodobenzene, 1.5 mmol of Na2CO3 and 3 ml of water–ethanol (2:1 v/v) solvent at 80 °C for 1 h. The product mixture was extracted using ethyl acetate-dichloromethane (3:1, v/v) and analyzed by GC, prior to regular analysis the products were identified by GC–MS (Table 4).

Table 4 Scope of Suzuki–Miyaura coupling over Pd–C3N4@TNT catalyst
figure c
Full size table

3.3 Catalyst Reusability and Leaching Test

At the end of each cycle the catalyst was separated by centrifugation, washed with ethyl acetate-DCM (3:1, v/v) mixture and reused. Within 5 repeated cycles no significant loss in activity of Pd–C3N4@TNT catalyst was observed, revealing the robustness against agglomeration of active component, i.e., palladium species. Palladium leaching test was conducted by stopping the reaction at about half of the reaction time (after 30 min) and separated the catalyst from product mixture by centrifugation and continued the reaction for 3 h, but no change either in the conversion or in the product yield, confirming the absence of leached Pd in the product mixture. The product mixture was also analyzed by ICP-AES, no Pd species were observed. The spent Pd–C3N4@TNT catalyst was also characterized by powder XRD and found that there is no significant structural change was observed. These results reveal that the Pd–C3N4@TNT is heterogeneous in nature. The high stability (neither leach nor agglomerate) of Pd–C3N4@TNT catalyst is due to reasons (i) Pd immobilization in the C3N4 network (ii) firm anchoring of C3N4 to TNT through elimination of water molecule from Ti–OH and –NH2 group of C3N4 via Ti–NH bonding. As it is a solid catalyst ease of separation. High efficiency of Pd–C3N4@TNT is mainly due to isolated Pd species (Table 5, Fig. 8).

Table 5 Comparison of catalytic performance of Pd–C3N4@TNT catalyst with reported catalytic systems
Full size table
Fig. 8
figure 8

XRD patterns of spent Pd–C3N4@TNT

Full size image

Activity point of view the present catalyst (Pd–C3N4@TNT) is the best catalyst, where as in the case of reference 37, 100% yield was claimed. In this particular work the amount of reactants taken are much lower and also carried out at high temperatures using organic solvent (DMA).

4 Conclusion

A novel palladium immobilized titanate nanotube (Pd–C3N4@TNT) catalyst was synthesized. The performance of this heterogeneous catalyst towards Mizoraki–Heck and Suzuki–Miyaura carbon–carbon bond forming reaction is on par with the reported superior catalysts and in addition this catalyst can be operated in water medium. The high performance of this catalyst may be due to atomic level Pd immobilization in the titanate nanotubes. Undeniably, this catalyst is robust against leaching and agglomeration, contributing the green chemistry.