The complete procedure for the preparation of the titled bimetallic catalyst is detailed in Scheme 1. The freshly prepared Fe3O4 nanoparticles, obtained from the reaction of FeCl2⋅4H2O and FeCl3⋅6H2O, were treated with Si(OEt)4 to afford Fe3O4@SiO2 core-shell. Next (3-glycidyloxypropyl)trimethoxysilane was used to introduce the epoxy group (I). Then, the synthesized 1- (2-propynyl)-1H-imidazole (II) and chlorodiphenylphosphine were allowed to react with epoxy functionalized magnetic nanoparticles(I) to achieve phosphorous and IL modified MNPs. Then, alkyne group was subjected to CuAAC reaction to introduce diol group. Finally, modified MNPs were treated with Au and Cu salts followed with reaction with NaBH4 to give AuCu supported on modified MNPs. This material refers to as Fe3O4@Phos-IL-AuCu NPs throughout of this article.
The contents of Cu and Au in the Fe3O4@Phos-IL-AuCu were quantified by ICP to be 0.03 and 0.07 mmol/g, respectively. Every stage of the preparation of Fe3O4@Phos-IL-AuCu is analyzed by FT-IR spectroscopy at (Fig. 1). Peaks located at 461 cm-1, 570 cm-1, and 638 cm-1 corresponded to the Fe-O vibrations. Signals positioned at 800 cm-1, 954 cm-1, and 1119 cm-1 derived from the Si-O, Si-OH, and Si-O-Si vibrations, respectively.23,24 Broad bands centered at 2867 cm-1 and 2927 cm-1 were caused by C-H stretching vibrations.25 The spectra of Fe3O4@SiO2-OPPh2-propargylimidazole and Fe3O4@Phos-IL compounds reveal that the peak at 1650 cm-1 is related to the C = N and C = C bonds of imidazole.26–28 However, this peak overlaps with the bending vibration of absorbed water molecules. Additionally, the present peak at 1071 cm-1 in the spectrum of these compounds is due to the vibrations of the P-O bond, which is overlapping with the peak corresponding to the stretching vibration of the Si-O-Si bond.29 In addition, the peak at 2120 cm-1 in the spectrum of Fe3O4@SiO2-OPPh2-propargylimidazole is ascribed to the existence of the terminal alkyne group.30 The disappearance of this peak in the final stage of the compound (Fe3O4@Phos-IL) signifies that the successful completion of the CuAAC 1,3-dipolar reaction occurred. In the Fe3O4@Phos-IL spectrum, the peak at 3138 cm-1 is caused by the stretching vibration of the C-H bond of the triazole moiety, while the peak at 1636 cm-1 is the result of overlapping bending vibrations of water, as well as C = C and C = N vibrations of imidazolium.31, 32 In all cases the wide band centered at 3400 cm-1 is associated to O-H stretching vibration.33 (Fig. 1).
The ionic liquid on increased the hydrophilic nature of the catalyst. This effect was studied by the calculation of the contact angle of the Fe3O4@C2H4OPPh2@1-(2-propynyl)-1H-imidazole and Fe3O4@Phos-IL aggregates. Results showed that contact angle of Fe3O4@Phos-IL is lower than the contact angle observed for Fe3O4@C2H4OPPh2@1-(2-propynyl)-1H-imidazole confirming the impressive effect of ionic liquid in increasing hydrophilicity (Fig. 2).
The results of transmission electron microscopy (TEM) analysis of Fe3O4@Phos-IL-AuCu NPs samples revealed a distinct core-shell structure, with evidence of the presence of AuCu nanoparticles on the surface of the substrate with the determined average size of 1.1 nm for AuCu NPs (see Fig. S1 and Fig. 3).
The scanning electron microscopy (SEM) images revealed the formation of Fe3O4@Phos-IL-AuCu NPs nanoparticles with a uniform distribution, good dispersion, and near spherical morphology (Fig. 4).
X-ray photoelectron spectroscopy (XPS) of Fe3O4@Phos-IL-AuCu in Fe, Si, C, N, Cl, P, Au, and Cu areas were analyzed (Fig. 5). The XPS plot of the Fe 2p region showed peaks at 709, 710, 712.5, and 716 eV, which correspond to Fe 2p3/2. Also, in this spectrum, the peaks in the regions of 723.9, 725.5 and 728.5 eV corresponded to Fe 2p1/2 (Fig. 5a).34–36 In the Si 2p spectroscopy, the peaks at 101.6, 102.8, and 103.5 eV corresponded to Si-C, Si-O-C, and O-Si-O, respectively (Fig. 5b).37, 38 XPS analysis of C 1s plot showed peaks located at 284.8, 286, and 288.5 eV which were ascribed to C-C, C-OH, and N-C = N, individually (Fig. 5c).39,40 In the N 1s spectroscopy, the peak at 399.5 eV was associated to the C-N bond, the 400 eV peak was attributed to the C = N bond, and the peak at 401.6 eV in this region was associated to the NR3+ species in imidazolium (Fig. 5d).41–43 In the Cl 2p spectral analysis, the peaks at 197.5, 199.4, and 201.7 eV were attributed to the Cl 2p3/2, while the peak at 200.4 eV was associated with the Cl 2p5/2 (Fig. 5e).44–47 XPS of phosphorus revealed two peaks in the 2p region, located at 131.2 and 133.4 eV, corresponding to the bonding states of P-C and P-O, respectively (Fig. 5f).48, 49 Additionally, the XPS plot of the gold 4f area revealed two peaks located at 83.5 and 87.5 eV, both of them conforming to the Au(0) (Fig. 5g).50 Finally, the XPS analysis was performed on the Cu 2p area, revealing two signals at 931.5 and 932 eV, corresponding to Cu2O. The peak observed at 933.5 eV was attributed to Cu 2p3/2, while the peaks at 953.4 and 955 eV were assigned to Cu 2p1/2 in CuO. Additionally, smaller peaks in the regions of 935, 940, and 943 eV were attributed to metallic copper(II), copper(II) oxide and copper(II) hydroxide (Fig. 5h).51–53
The X-ray diffraction analysis of the Fe3O4@Phos-IL-AuCu NPs catalyst demonstrated the existence of magnetic Fe3O4 nanoparticles, evidenced by the Bragg reflections at 2θ = 30.1°, 35.5°, 43.1°, 57.2°, and 62.8°.54,55 The reflections at 2θ = 43.85° and 37.58° corresponded to the gold nanoparticles in the bimetallic gold-copper structure, while the reflections at 2θ = 50.0° and 73.8° corresponded to the copper nanoparticles.56–59 Additionally, a wide band around 2θ = 22.0 was related to SiO2 (Fig. 6).
The superparamagnetic behavior of Fe3O4@Phos-IL-AuCu NPs was studied by vibrating-sample magnetometer (VSM) in which result indicated zero-coercivity together with remanence on the magnetization-loop proving superparamagnetic nature of the material and possibility of easy separation by external magnetic field (Fig. 7).
After characterization of Fe3O4@Phos-IL-AuCu NPs, its catalytic activity was initially assessed in the reduction of nitro aromatic compounds. The reduction of the starting compound 4-chloronitrobenzene was selected as the reference reaction and various factors like type of reductant, solvent, catalyst loading (mol%), and reaction time were analyzed (Table 1). Initial results indicated that a good yield was obtained by using catalyst with 0.04 mol% Au and 0.09 mol% Cu by using NaBH4 (2 mmol) in H2O during 2 h at room temperature (Table 1, entry 1). Results showed that enhancement of catalyst amount to Au (0.06 mol%) and Cu (0.13 mol%), a 65% yield of the desired amine was obtained, whilst with higher amounts of Au (0.09 mol%) and Cu (0.19mol%), quantitative yield was achieved (Table 1, entries 2 and 3). Also, the yield of the reaction involving 4-chloroaniline decreased using lower amounts of NaBH4 (1 mmol) (Table 1, entry 4) and shorter reaction times (Table 1, entries 5, and 6). Other reluctant agents such as hydrazine, formic acid and glycerol were investigated giving lower yields of the expected amine (Table 1, entries 7–9). By using of other solvents such as CH3CN, DMF and toluene obtained lower yields (Table 1, entries 10–12). Therefore, conditions including catalyst with 0.09 mol% of Au, 0.19 mol% of Cu, NaBH4 as reducing agent, in neat H2O, at room temperature, was selected as the optimal reaction conditions. This reference reaction, in the absence of catalyst, did not do any progress (Table 1, entry 13). To highlight the advantage of using bimetallic Fe3O4@Phos-IL-AuCu, we have examined the activity of single-metal catalysts, Fe3O4@Phos-IL-Au and Fe3O4@Phos-IL-Cu, in this reaction and using the optimized conditions. The experimental results indicated that both single-metal catalysts offered poorer yields than the reaction performed with bimetallic catalyst (Table 1, entries 14, and 15).
After determination of optimum reaction conditions, the reduction of different nitroarenes compounds were assessed (Table 2). The results showed that high to excellent yields were obtained in the reduction of nitro aromatic compounds with both groups of electron-donating groups such as -Me, - CH2OH, NH2, and –OH and electron‐withdrawing groups such as -F, -Cl, -Br, –CN, and ―COMe. In general, the results show that the reduction of electron-rich nitroarenes was slower than the substrates-containing electron deficient arrangements (Table 2, entries 2–15). Also, in the example run with 4-nitrobenzaldehyde, the aldehyde group was also reduced along with nitro group, (Table 2, entry 13). However, during the reduction of acetylnitroarenes only the nitro group was reduced (Table 2, entries 14, 15). In addition, nitro aromatic compounds with two nitro groups were completely reduced, using more amount of NaBH4 in longer times (Table 2, entries 16–19). Also, compound with both electron‐donating and electron‐withdrawing substitutions had good yields (Table 2, entry 20). Reduction of 4- nitrobenzylchloride was performed very good and corresponded amine obtained in excellent yield (Table 2, entry 21). Finally, the reduction of heterocyclic nitro compounds were carried out efficiently and the expected amines were isolated in 83–89% yields (Table 2, entries 22–24). Besides, the carbonyl group of the amide remained intact (Table 2, entries 22 and 23).
The catalytic performance of Fe3O4@Phos-IL-AuCu in the reduction of organic dyes, including methyl red (MR), methyl orange (MO), and rhodamine B (RhB) was evaluated. Progress of reduction was recorded by using UV-Vis spectrophotometer. The absorption of all MR (λmax = 435), MO (λmax = 467 nm), and RhB (λmax = 556 nm) dyes decreased in time when using Fe3O4@Phos-IL-AuCu as catalyst together with NaBH4 as reducing agent. Generally, MO (λmax = 467 nm), MR (λmax = 435), and RhB (λmax = 556 nm) dyes completely declined in 3 min, 2 min and 1 min, respectively and during desired times all dyes became colorless (Fig. 8a-c).60,61 Diagrams of ln (At/A0) vs reaction time for the dyes reduction were drawn and the rate constant (k) of MO, MR, and RhB were 1.5, 2.1, and 2.6 min− 1, respectively, which indicates the good performance of the titled synthesized catalyst (Fig. 8d).
In order to compare the activity of the prepared bimetallic catalyst with the activity of the single metal catalysts such as Fe3O4@Phos-IL-Au and Fe3O4@Phos-IL-Cu, the reduction of dyes MO, MR, and RhB was investigated (Fig. S2). The results showed that 31.5%, 27%, 11.5% yield were obtained in the reduction of MO, MR, and RB dyes when using Fe3O4@Phos-IL-Au catalyst. Also, 5%, 3%, and 9% yields were obtained in the reduction of MO, MR, and RB dyes with Fe3O4@Phos-IL-Cu catalyst. These results confirmed the higher catalytic activity of Fe3O4@Phos-IL-AuCu due to synergetic effect between Cu and Au species.
Finally, the catalytic activity of the as-prepared Fe3O4@Phos-IL-AuCu catalyst has been assessed in the tetracycline (TC) degradation by using the ammonium peroxodisulfate at diverse pH (3, 5, 7, 9, and 11) (Fig. 9a-e). The reaction process was recorded by UV-Vis spectrophotometer. Results showed that the absorbance of tetracycline gradually decreased when increasing the reaction time, and, similarly to the experience in previous reports, the best result was obtained under alkaline media.
Degradation of tetracycline with Fe3O4@Phos-IL-Au (Fig. S3), and Fe3O4@Phos-IL-Cu (Fig. S4) were also investigated. Here, the higher efficiency of bimetallic catalyst Fe3O4@Phos-IL-AuCu is demonstrated once more versus the reactions run in the presence of monometallic catalytic species (Table S1).
Since the reduction of 4-nitrophenol was proceeded very efficiently (Table 2, entry 8), we studied synthesis of acetaminophen via one-pot reduction/amidation of 4-nitrophenol. Results show that reaction proceed very effectively and desire product was obtained in 84% isolated yield (scheme 2).
Efficiently recovering and reusing magnetic heterogeneous catalysts is crucial for both environmental and economic perspectives. In this regard, the recovery and the recycling facilities of Fe3O4@Phos-IL-AuCu were studied in the reduction of 4-nitrophenol under optimized reaction conditions (Fig. 10). To achieve this objective, the catalyst was recovered after 30 min using an external magnet, washed with ethyl acetate and after drying, it was utilized in another new batch of the reaction. According to these findings, the catalyst demonstrated the ability to be reused for 16 consecutive runs with only a slight decrease in activity. However, during the 17th run, the yield dropped to 81%.
To determine whether the Fe3O4@Phos-IL-AuCu is heterogeneous or homogeneous in nature, the filtration test was used for the reduction of 4-chloronitrobenzene under optimal conditions. During the course of the experiment, the reaction proceeded to a level of 39% completion after an initial reaction time of one hour. At this stage, the reaction mixture was subjected to filtration, following which the filtrate was allowed to continue the reaction for an additional hour. Subsequent analysis using gas chromatography (GC) revealed that the progress of the reaction was slowed down and the product was formed at a yield of only 49% (Fig. 11). This result indicated negligible leaching of active catalysts to reaction mixture and possible heterogeneous nature of the catalyst.
In order to investigate the stability of the catalyst, the structure of the reused catalyst was analyzed by VSM (Fig. 7), XPS (Fig. 12), XRD (Fig. 13), and TEM (Fig. 14). VSM analysis of reused catalyst showed preservation of magnetic property of the catalyst during recycling. The information revealed by XPS spectrum in Fe, Si, C, N, Cl, P, Au, and Cu regions indicated that similar patterns of elements to the fresh catalyst existed, confirming the high stability of the catalyst during the reaction.
Furthermore, the XRD spectra obtained from the reused catalyst showed peaks corresponding to Fe3O4, Si, Au and Cu, which are closely to the analogous ones detected in the newly prepared catalyst, providing evidence of the catalyst's durability (Fig. 13).
TEM images of reused catalysts was also showed very similar pattern to fresh catalyst with small increase in average size of nanoparticles confirming stability of the catalyst (Fig. 14 and Fig. S5).