In order to analyze parent nanocomposite and its further functionalization, FT-IR spectroscopy was employed. A band was observed at 585 cm⁻¹ in the IR-spectrum of MNPs depicting the Fe-O stretching absorption (Figure 1A). (Zhu et al., 2011) The intensity of this band reduced on silica-coating with the appearance of three new sharp bands in the region of 806, 957 and 1,099 cm⁻¹, corresponding to the symmetric Si-O-Si, symmetric Si-O(H) and asymmetric Si-O-Si stretching vibrations respectively (Figure 1B) (Kooti and Afshari, 2012) Further functionalization of SMNPs with APTES resulted in absorption at 2,924 cm⁻¹ and 1,644 cm⁻¹, which corresponds to CH₂ and NH₂ from aminopropyl moiety of APTES (Figure 1C) (Yamaura et al., 2004) The immobilization of ligand DF onto ASMNPs was confirmed by the band at 1,662 cm⁻¹ accredited to C=N stretching frequency (Figure 1D) and to this, metal was immobilized using CuI which shifted the prominent band at 1,662 cm⁻¹ to a lower wavenumber indicating strong metal-ligand interaction (Figure 1E). (Masteri-Farahani and Tayyebi, 2011; Esmaeilpour et al., 2012).

To assess the crystalline nature of synthesized MNPs and SMNPs, powder X-Ray diffraction measurements were carried out. For MNPs (Figure 2A), six characteristic peaks were observed at 2θ: 30.366^o, 35.663^o, 43.024^o, 53.6^o, 57.299^o, and 62.865^o corresponding to the (220), (311), (400), (422), (511) and (440) crystallographic faces of magnetite (Abu-Reziq and Alper, 2012). These peaks were in accordance with the standard XRD data provided by the Joint Committee on Powder Diffraction Standards (JCPDS) card number 19–0,629 and is ascribed to inverse cubic spinel Fe₃O₄ crystal (Abu-Reziq et al., 2006). The average crystallite size of the MNPs was calculated by the Scherrer equation {D_hkl = Kλ/(/(β_hklcosθ)}, where D_hkl represents the size of the axis parallel to the (hkl) plane, k is a constant with a common value of 0.89 for spherical particles, λ is the wavelength of radiation, β _hkl is the full-width at half-maximum (FWHM) in radians, and θ is the diffraction angle. The mean crystallite size was found to be ∼10.6 nm for the (311) reflection. Besides these six diffraction peaks, a weak broad hump at 2θ = 20–24^o is observed in the XRD pattern of SMNPs showed in Figure 2B, which is attributed to amorphous silica (Zhang et al., 2012).

To investigate the topography of the synthesized nanoparticles, scanning electron microscopic (SEM) analysis was performed and it was found that the smooth surface of MNP (Figure 3A) turns spongy on silica coating (Figure 3B). The spherical morphology of the final Cu-DF@ASMNPs catalyst was seen with slight agglomeration and appears the same as that of SMNP (Figure 3C). This suggested that the surface modification methods did not alter the morphology of the nanocatalyst. Besides this, the SEM image of the recovered catalyst (Figure 3D) also indicates that the reaction did not affect the morphology of the catalyst.

TEM studies were performed to study the morphological changes of the synthesized nanomaterials, Figure 4A depicts that MNPs are polydisperse in nature and display slight agglomeration (Wang et al., 2010; Wang et al., 2013). An array of bright diffraction rings was observed in the selected area electron diffraction pattern (SAED) (Figure 4B) which confirmed the crystalline nature of these nanocomposites and also in accordance with the resultant XRD pattern. The average interplanar distance of the MNPs was measured from a high-resolution transmission electron microscopy (HR-TEM) image and was found to be ∼0.20 nm, which correlates with the (311) plane of inverse spinel Fe₃O₄ structure (Figure 4C). A dark core-shell of MNP, with an almost uniform silica coat of 4–5 nm thickness, was observed in the TEM image of SMNP (Figure 4D). TEM images of final catalyst and recovered catalyst are shown in Figure 4E and Figure 4F respectively, which further confirm that the structural morphology remain unchanged after the coupling reaction. In order to find the average particle size of MNPs, 52 colloidal aggregates were analyzed and it was found to be in the range of 10–11 nm (Supplementary Figure S1) which is in well accordance with the XRD results.

Energy dispersive X-ray analysis was performed to detect the composition of the synthesized nanocomposites, and the EDS spectrum displayed well-defined peaks of copper, silicon and iron (Figure 5) that substantiate the effective grafting of copper on the Cu-DF@ASMNPs. Moreover, to determine the amount of copper present in the final catalyst, ICP analysis was conducted and the metal loading was found to be 0.3217 mmolg⁻¹.

The field-dependent magnetization measurement of synthesized nanocomposites is provided in Figure 6. The superparamagnetic behaviour of these nanoparticles was confirmed by their magnetization curves which display no hysteresis at room temperature. This was further corroborated by the inset in Figure 6 where both coercivity and remanence were negligible in the absence of an externally applied magnetic field. The saturation magnetization value of MNPs, SMNPs, ASMNPs, and Cu-DF@ASMNPs were found to be 59 emu g⁻¹, 40 emu g⁻¹, 29 emu g⁻¹, and 19 emu g⁻¹ respectively. This decrease in the M_s value is due to the non-magnetic nature of the functionalizing groups. (Hu et al., 2005; Digigow et al., 2014). Despite of lower value of magnetization, the net magnetism of Cu-DF@ASMNPs was high enough for its effortless removal via an external magnet.

In order to test the efficacy of the prepared nano-catalyst, Cu-DF@ASMNPs and to discover the optimal reaction conditions, phenylpropiolic acid (0.6 mmol) and iodobenzene (0.5 mmol) were selected as the coupling partners. For the optimization of decarboxylative cross-coupling reaction, reaction parameters like solvent, base and catalytic amount were assessed. Figure 7 represent several combinations of base and solvent, and out of them highest yield was obtained when the base was Cs₂CO₃ (1 mmol), and toluene (2 ml) was the solvent. For the determination of the optimal catalytic amount, a blank test was carried out, where no significant yield was obtained. Although, the reaction gave product with 10 mg of catalyst and on increasing the amount of catalyst, significant increase in yield of the product was noticed. However, no noticeable increase in the product yield was found when 30 mg of catalyst was used and the best yield was achieved with 25 mg of catalyst (Supplementary Table S1). Also, the reaction was performed under diverse range of temperatures while keeping other parameters constant and 100 C was found to be the optimum temperature to carry out the coupling with 25 mg of synthesized catalyst (Supplementary Figure S2). Therefore, all the reactions were performed using toluene as the reaction solvent, Cs₂CO₃ as the base for 12 h at 100 C in the presence of 25 mg of Cu-DF@ASMNPs under N₂ atmosphere.

To demonstrate the efficiency of this method, various haloarenes including iodo and bromo derivatives were coupled with phenylpropiolic acid using the optimized reaction conditions (Table 1). To check the scope of this reaction, we initially examined a variety of iodoarenes possessing both activating and deactivating groups including methoxy, methyl, naphthyl, nitro, and chloro. It was observed that the reaction went smoothly for both electron donating and withdrawing groups on the iodoarene and excellent yields were obtained for various internal alkynes. For p-iodotoluene and p-iodoanisole the yield was similar, 90 and 88% respectively (entries 2 and 3). However, slight increase in yield was obtained when haloarene bearing electron withdrawing substituent was employed (entry 5).

To gauge the efficacy of reaction, more practical coupling partner aryl bromides were used. To our delight they also worked very well for this coupling reaction and both electron withdrawing as well as electron donating aryl bromides were efficiently converted into corresponding internal alkynes in excellent yields. To further assess the potential of the catalyst, an aliphatic alkynoic acid, 2-butynoic acid (entry 11) was tested for this reaction but this afforded slightly lower yield. Isolated yields were obtained for few selected reactions consisting of haloarene bearing electron donating group (entry 3), neutral group (entry 4), a halogen substituent (entry 6) to check for any kind of interference, and electron withdrawing group (entry 10).

To test the reusability of catalyst, after each experiment (conducted under optimized conditions), the catalyst was magnetically separated, washed with ethyl acetate and ethanol and dried under vacuum. This was then used to perform subsequent reactions. It was observed that the catalyst was active up to seven runs without any notable decrease in its performance (Figure 8). SEM and TEM images further confirmed the unaltered structure and morphology of the recovered catalyst (Figure 3D and Figure 4F).

In order to estimate the leaching rate and heterogeneous nature of the catalyst, two sets of corresponding experiments were conducted. A standard reaction was conducted for the first set where the catalyst was magnetically removed after completion of reaction, and filtrate was analyzed under ICP, which showed no leaching. A split test was performed with the second set, in which the standard reaction was conducted with catalyst for 2 hours, which roughly corresponds to 20% conversion by GC-MS. Afterwards, the nanocatalyst was magentically separated from the reaction mixture and the reaction was further pursued. No coupling product in the reaction mixture was observed up to 10 hours under the same reaction conditions, which authenticate the truly heterogeneous nature of the nanocatalyst.

Figure 9 depicts the proposed mechanism that has been derived from earlier reports. (Okuro et al., 1993; Ray et al., 2008; Gonda et al., 2010; Lauterbach et al., 2010). The reaction between Cu-DF@ASMNPs A and alkynoic acid produces intermediate B, which undergoes decarboxylation to yield C, an alkynyl copper intermediate. Further addition of haloarene results in the formation of another intermediate D, which then undergoes reductive elimination, to give the product while regenerating the catalyst A.

Finally, in order to show the superiority of the synthesized catalyst, we compared our obtained results with the previously reported work (Table 2) and found that our catalyst was far more efficient in terms of reaction conditions, reaction time and catalytic recovery. Also, it is the first copper based heterogeneous system for synthesis of internal alkynes.

In summary, a highly effective palladium-free Cu-DF@ASMNPs nanocatalyst was fabricated successfully and applied towards the synthesis of internal alkynes via decarboxylative cross-coupling reaction. These nanocomposites endowed low metal loading, high stability, and good functional group tolerance with excellent yields and high turn-over numbers. It is noteworthy that this catalytic system is the first report of copper based magnetic nanocatalyst that represents a practical and low-cost route to prepare internal alkynes. In addition, the effortless magnetic recovery and reusability of the catalyst for at least seven runs without any marked loss in its performance makes it an efficient protocol to produce a wide variety of unsymmetrical alkynes.