Rituparna
Das
,
Sourav
Ghosh
and
Milan Kanti
Naskar
*
Sol–gel Division, CSIR-Central Glass and Ceramic Research Institute, Kolkata 700032, India. E-mail: milan@cgcri.res.in
First published on 24th November 2017
Copper nanoparticles (CuNPs) confined in hollow silicalite-1 powders were synthesized by a selective “desilication–recrystallization” method using tetraethlyorthosilicate (TEOS), tetrapropylammonium hydroxide (TPAOH) and cupric chloride (CuCl2) as precursors. The synthesized product was characterized by XRD, Raman, XPS, FESEM, TEM and N2 adsorption–desorption studies. The presence of CuNPs of size 10–40 nm in the hollow silicalite-1 was confirmed by TEM. The BET surface area of the powders was 247 m2 g−1 composed of micropores and mesopores. The prepared CuNPs confined in hollow silicalite-1 showed excellent catalytic performance for the reduction of 4-nitrophenol (4-NP) to 4-aminophenol (4-AP) with apparent rate constant and activity parameter values of 5.6 × 10−3 s−1 and 44.09 s−1 g−1, respectively for 1 mg of catalyst.
Hollow zeolites can be the ideal matrix for the confinement of metal catalysts due to their high thermal and mechanical stability, unique shape selectivity, unique pore structure with large specific surface area etc.12 On the other hand because of their brilliant resistance under corrosive conditions and molecular sieving properties, crystalline zeolitic shells can also protect the catalyst under harsh reaction condition as a selective barrier against impurities, poisons, and undesirable reactions. By introducing mesoporosity in zeolite structures, the catalytic performance can be increased with decreasing diffusion path length. Therefore, hollow mesoporous zeolites with hierarchical porosity in particular, can be an interesting candidate as a catalytic support.
The preparation of hollow zeolites has been reported by the layer-by-layer assembly technique using polystyrene spheres as the soft template and nanozeolites as the building blocks.13,14 Desilication of framework Si in alkaline media is another efficient methodology to synthesize hollow zeolites.15–17 We have recently synthesized hollow mesoporous silicalite-1 confined with biogenic AgNPs of size 10–15 nm using green carambola extract.18 Li et al. reported noble metal nanoparticles located in hollow single crystals of silicalite-1 such as Au@silicalite-1 and Pt@silicalite-1.19,20 Recently Dai et al. synthesized hollow ZSM-5 single crystals with silicon-rich exterior surfaces encapsulating iron and carbon nanotubes.21
In this present study, we report the synthesis of CuNPs confined in hollow silicalite-1 via a desilication–recrystallization method. CuNPs confined in the hollow silicalite-1 exhibited enhanced catalytic reduction of p-nitrophenol, a water pollutant present in industrial effluents. The reduced form of p-nitrophenol, i.e., p-aminophenol, may be used as a drug, photographic developer, corrosion inhibitor etc. CuNPs confined in hollow silicalite-1 revealed enhanced catalytic efficiency with a rate constant of 5.6 × 10−3 s−1 and can be reused several times without significant loss of their original activity.
To investigate the composition and chemical state of Cu in the CuNPs confined in hollow silicalite-1 (CuHS-1R), X-ray photoelectron spectroscopy (XPS) was performed. Fig. 1b and c show the XPS spectra of the CuHS-1R sample representing the signals of Cu2p (Cu2p3/2 at 932.33 eV and Cu2p1/2 at 952.23 eV) and O1s (532.7 and 529.9 eV), respectively. It is difficult to identify the different oxidation states of Cu species from the Cu2p and O1s spectra as their peak positions are very close. However, the absence of the two main peaks at 954 eV (Cu2p1/2) and 934 eV (Cu2p3/2) along with shake-up satellite peaks centred at 943 eV confirmed the absence of CuO in the CuHS-1R sample. It is difficult to identify whether Cu2O is present in the CuHS-1R sample from the XPS study. From the O1s spectrum, the binding energy at 532.7 eV corroborated the presence of hydrophilic oxygen-containing groups, such as hydroxy/epoxy groups.23 More detailed results of the O1s spectrum show an oxygen band peak of Si–OH at 529.9 eV, which is related to the O1s spectra of the SiO2.24 Fig. S2 and S3, ESI† show the XPS spectra of CuS-1 and CuHS-1, respectively. Interestingly, in addition to Cu 2p signals (2p1/2 and 2p3/2), satellite peaks at higher binding energy at around 962 eV and 943 eV are observed in the Cu2p XPS spectra of the two samples, which could confirm the existence of CuO in the samples.25 It was reported that Cu(I) or Cu(0) species do not reveal any satellite peaks due to completely filled 3d shells.26 Interestingly, for the CuHS-1 particles the O1s peaks shift a little to lower energy (Fig. S3b, ESI†). The O1s spectra of SiO2 in silicalite-1 could contribute to the shifting of the peak.
To identify the different oxidation states of Cu in silicalite-1, Raman study was performed. The Raman spectra of the samples are shown in Fig. 2. Raman peaks appeared at around 283, 333, and 622 cm−1 for the CuS-1 and CuHS-1 samples. The peak at 283 cm−1 was assigned to the Ag mode, while the peaks at 333 and 622 cm−1 correspond to Bg modes of phonon vibrations of CuO.27 Some extra peaks at around 377 and 470 cm−1 also appear due to the presence of silicalite-1 zeolite.28 Interestingly, the characteristic peaks of CuO at around 282, 330, and 612 cm−1 are missing in the Raman spectra of S-1 and CuHS-1R particles. The absence of characteristic peaks of Cu2O at 218, 306, 435 and 625 cm−1 corresponding to the second-order Raman allowed mode (2 Γ12), second order overtones (2 Γ15(1)), fourth-order overtone (4 Γ12) and infrared allowed mode (Γ15(2)), respectively29,30 could confirm that the Cu species present in the CuNPs confined in hollow silicalite-1 (CuHS-1R) are in a zero valent state. These peaks are also absent in all the Raman spectra of the materials (Fig. 2).
Fig. 3a and b show the low and high magnification FESEM images of the CuNPs confined in silicalite-1 (CuHS-1R) representing the hollowness of the particles. The TEM images (Fig. 3c and d) clearly show the particles having a hollow interior surrounded by a thin shell. It also reveals that CuNPs of size 10–40 nm are present in hollow silicalite-1. The HRTEM image of CuHS-1R shows the lattice fringes of CuNPs with a d-spacing of 0.217 nm corresponding to the (111) plane of Cu (Fig. 3e), indicating the presence of CuNPs in the sample. The selected area electron diffraction (SAED) pattern corroborated the crystalline planes of CuNPs and silicalite-1 (Fig. 3f). The presence of Cu atoms of about 0.91 at% in the sample was determined by energy dispersive X-ray spectroscopy (EDS) (Fig. 3g). It is to be noted that CuNPs may be dispersed both in the exterior and interior surfaces of hollow silicalite-1 (Fig. S4, ESI†).
Fig. 3 (a and b) FESEM images, (c and d) TEM images, (e) HRTEM image, (f) SAED patterns and (g) EDS of CuNPs confined in hollow silicalite-1 (CuHS-1R). |
The line scan spectra of the elemental analysis confirmed the almost uniform distribution of Cu atom concentration (blue line) along a line representing both exterior and interior surfaces of hollow silicalite-1 (Fig. S5, ESI†). Before desilication of the sample CuS-1, the pseudo-hexagonal prismatic shaped particles of silicalite-1 were revealed by FESEM (Fig. S6a and b, ESI†) and TEM images (Fig. S6c and d ESI†). The characteristic SAED image of silicalite-1 is shown in Fig. S6e, ESI.† The EDS analysis indicated the presence of 1.49 at% Cu atoms in the sample (Fig. S7, ESI†). Interestingly, during the desilication process on the CuS-1 sample, the silicate oligomers are leached from the interior of the crystals and crystallize on the crystal surface, leading to regular hollow crystals with well-defined cavities and walls. Fig. S8a and b and c and d, ESI† show the FESEM and TEM images of CuHS-1, respectively after desilication of CuS-1. The SAED image of the CuHS-1 sample indicates the single crystalline nature of silicalite-1 (Fig. S8e, ESI†). The EDS of CuHS-1 shows the presence of 1.22 at% Cu atoms in the sample (Fig. S9, ESI†). It is to be pointed out that during desilication, some amount of CuO is dissolved in the alkaline medium rendering lower at% of Cu in CuHS-1.
The N2 adsorption–desorption isotherms and pore size distributions (PSDs) of the CuS-1 and CuHS-1 samples are revealed in Fig. S10 and S11 (ESI†), respectively. In the CuS-1 sample (Fig. S10a, ESI†), a steep rise in the isotherm at lower relative pressure, around p/p0 = 0.1, indicated the abundance of micropores in the sample. The mesoporosity in the sample was evidenced by the BJH PSD curve (Fig. S10b, ESI†).
The DFT pore size distribution curve reveals micropores generated at around 6 Å, which is the characteristic pore of silicalite-1 zeolite. The BET isotherms and PSDs of the CuHS-1 sample are shown in Fig. S11, ESI.† In this isotherm the hysteresis loop shifted towards higher p/p0 at around 0.5 indicating the abundance of mesopores (Fig. S11a, ESI†). It can be concluded that the mesoporosity is generated in the sample due to dissolution of silica with 0.3 M TPAOH treatment. The increase in mesoporosity in this sample was verified by the corresponding BJH pore size distributions (Fig. S11b, ESI†). The N2 adsorption–desorption isotherms of CuHS-1R are shown in Fig. 4a. The presence of mesoporosity and microporosity of silicalite-1 in the sample was confirmed by the pore size distribution (PSD) curves determined by the BJH (Fig. 4b) and DFT (Fig. 4c) methods, respectively. It is reflected from the isotherms that there is no significant change in the isotherms in both the CuHS-1 and CuHS-1R samples. From the isotherm, it is evident that at higher relative pressure H2 type hysteresis loops appeared indicating ink-bottle like pores. The large hysteresis loops at around p/p0 = 0.5 indicated the profusion of mesoporosity in the sample which was also evidenced by the BJH pore size distribution curve. The textural properties (BET surface area, total pore volume, and pore size) of the samples are shown in Table 1. The BET surface area, pore volume and pore size of CuHS-1R were found to be 247 m2 g−1, 0.387 cm3 g−1 and 6.2 nm, respectively. Table 1 indicates that the BET surface area and microporous surface area decreased in the order of CuS-1 > CuHS-1 > CuHS-1R. The total specific surface area is determined by the BET method, while the microporous contribution is examined by the difference between the BET surface area and the external surface area, i.e., the mesoporous surface area (derived from the slope of the t (statistical thickness)-plot).31 The t-plot graphs and linear fitted BET plots of CuS-1, CuHS-1 and CuHS-1R are shown in Fig. S12, S13 and S14, respectively, ESI.† For microporous materials like silicalite-1, the linear BET region occurs at p/p0 < 0.1, while the linear t-plot range is obtained at higher p/p0. The total surface area and the micropore surface area are gradually decreased; however, the ratio of the mesoporous surface area to the microporous surface area increased in the order of CuS-1 < CuHS-1 < CuHS-1R (Table 1).
Fig. 4 (a) N2 adsorption and desorption isotherms, and pore size distributions (PSD) by the (b) BJH and (c) DFT method of CuNPs confined in hollow silicalite-1 (CuHS-1R). |
Sample Id | S BET (m2 g−1) | S mic (m2 g−1) | S ext (m2 g−1) | S ext/Smicd | V p-total (cm3 g−1) | d P (nm) |
---|---|---|---|---|---|---|
a BET surface area. b Micropore surface area. c External surface area. d Ratio of external surface area to micropore surface area. e Total pore volume. f Average pore size. | ||||||
CuS-1 | 315 | 242 | 73 | 0.3016 | 0.205 | 2.6 |
CuHS-1 | 280 | 203 | 77 | 0.3793 | 0.389 | 5.5 |
CuHS-1R | 247 | 164 | 83 | 0.5060 | 0.387 | 6.2 |
This could be due to the structural distortion32 during desilication followed by the second and third calcination steps. Interestingly, the average pore size also significantly increased after desilication of silicalite-1. Furthermore, for the CuHS-1R sample, the increase in average pore diameter and reduction of surface area were attributed to the development of partial strain generated during the formation of CuNPs.33 The total pore volume of the CuHS-1 and CuHS-1R samples increased due to the presence of higher mesoporosity in the samples compared to that of the CuS-1 sample.
Fig. 5a shows that after addition of 1 mg catalyst, the absorbance at 400 nm of the p-nitrophenolate ion was reduced within 5 min accompanied by the simultaneous increase of the absorption peak at 300 nm which corresponds to the formation of 4-AP at room temperature. From the logarithm plot of the absorbance (−lnAt/A0) versus reaction time, the pseudo-first order rate constant (apparent rate constant) was calculated as 5.6 × 10−3 s−1 for 1 mg of catalyst at room temperature (Fig. 5b). The recyclability test of the catalyst was performed. Fig. 5c shows that the catalytic efficiency remained almost the same after reusing at least four times, which is reflected by nearly the same rate constant values.
Figs. S15 and S16, ESI† show (a) time-dependent UV-Vis spectra for the reduction of 4-NP, (b) the pseudo-first order plot of (−lnAt/A0) versus reaction time and (c) the apparent rate constant (k) for 4 consecutive cycles, for 2 mg and 3 mg of catalysts, respectively. The time of completion of the reaction, apparent rate constant, R2 values and activity parameter κ (rate constant per unit mass of Cu loading) of the reactions using different amounts of catalyst are shown in Table 2.
Serial no. | Amount of catalysts (mg) | Time (min) | k (s−1) | R 2 | κ (s−1 g−1) |
---|---|---|---|---|---|
1 | 1 | 5.0 | 5.60 × 10−3 | 0.996 | 44.09 |
2 | 2 | 2.5 | 12.61 × 10−3 | 0.995 | 49.60 |
3 | 3 | 4.0 | 7.76 × 10−3 | 0.970 | 20.36 |
The apparent rate constants are found to be 5.60 × 10−3, 12.61 × 10−3 and 7.76 × 10−3 s−1 for 1, 2 and 3 mg of catalyst, respectively. The apparent rate constant for pseudo-first order reaction can be changed with the amount of catalyst36,37 as well as the concentration of NaBH4. Here, the concentration of NaBH4 solution remains constant. Therefore, the apparent rate constant could vary with the catalyst concentration. The highest apparent rate constant i.e., 12.61 × 10−3 s−1 for 2 mg of catalyst is due to the increased mass transfer of reactants within the catalyst particles.38 For further increase in catalyst concentration, i.e. for 3 mg of catalyst, the mass transfer is restricted to some extent with the longer diffusional path through the tortuous pores causing a decrease in the apparent rate constant. The R2 values are closer to unity in each case which could be in good agreement with it being a pseudo-first order reaction. It is reported that the apparent rate constant (k) and the activity parameters (κ) are dependent on the amount of catalyst used.39,40 The activity parameter (κ) is calculated from the apparent rate constant (k) per g of Cu present in CuHS-1R (2.01 mmol Cu per 1 g of CuHS-1R, determined by ICP-MS analysis). The catalytic property of the CuNPs confined in hollow silicalite-1 was compared to the reported literature (Table S1, ESI†) based on the activity parameters (κ). The present value was found to be higher than previously reported values of Cu and other metal nanoparticles.
Furthermore, the catalytic activity of the samples S-1, CuS-1 and CuHS-1 was also studied. For pure silicalite-1 (S-1), the absorption peak of p-nitrophenolate ions at 400 nm remained almost unchanged even after 30 min of absorption suggesting no catalytic activity for the reduction of 4-NP (Fig. S17, ESI†). Fig. S18 and S19, ESI† illustrate (a) the time-dependent UV-Vis spectra for the reduction of 4-NP, and (b) pseudo-first order plot of (−lnAt/A0) versus reaction time for each of the samples CuS-1 and CuHS-1, respectively. The apparent rate constants (k) were calculated as 2.3 × 10−3 s−1 and 4.1 × 10−3 s−1 for 1 mg of CuS-1 and CuHS-1, respectively. However, k for 1 mg of CuHS-1R was found to be 5.6 × 10−3 s−1. This demonstrates that with increasing mesoporosity (mesoporous surface area with respect to microporous surface area) in silicalite-1 (Table 1), the diffusion path length is reduced rendering the catalytic activity in the order of CuS-1 < CuHS-1 < CuHS-1R. Here, CuNPs confined in hollow silicalite-1 (CuHS-1R) exhibit higher catalytic efficiency due to the presence of hollow architecture with increased mesoporous surface area with respect to microporous surface area. It is worth noting that hierarchical porosity has a significant contribution toward the catalytic efficiency of 4-NP. For comparison, the BET surface area and catalytic activity of CuNPs confined in a commercial 4A molecular sieve (Merck) was studied. The BET surface area of the CuNPs confined in a commercial 4A molecular sieve was found to be 30 m2 g−1 (isotherm shown in Fig. S20, ESI†) having only external pores, which is lower than that of CuNPs confined in hollow silicalite-1 (CuHS-1R) having hierarchical porosity. As a result, CuNPs confined in a commercial 4A molecular sieve give a lower apparent rate constant (2.3 × 10−3 s−1) for the catalytic reaction (Fig. S21, ESI†) than that of CuHS-1R. The lower apparent rate constant of CuNPs confined in a commercial 4A molecular sieve is due to the absence of hierarchical porosity with lower external surface area. During the catalytic reaction, in the presence of excess BH4− ions, the reduction of 4-NP takes place presumably on the catalyst surface in two steps. In the first step, both donor BH4− ions and accepter 4-NP molecules diffused on the surface of the catalyst through chemisorptions.41 Here, BH4− ions transfer hydrogen to the catalyst (CuHS-1R) surface for the reduction of 4-NP molecules in aqueous conditions.42 In this reaction the catalyst acts as a hydrogen shuttle and simultaneous reduction occurs to 4-AP via a desorption process. The presence of microporosity hinders the catalytic performance. CuO in CuHS-1 could show catalytic performance. However, due to the presence of higher microporosity in CuHS-1, its catalytic performance is lower than that of CuHS-1R. On the other hand, in catalytic reduction of 4-NP in the presence of CuO, the NaBH4 reduces Cu(II) to Cu(0). The BH4− ions and 4-NP are accumulated on the reduced Cu(0) surface, and the reduction of 4-NP to 4-AP takes place on the Cu(0) surface.43 However, for the same catalytic reaction by using Cu(0), the reaction becomes faster than that with CuO because Cu(0) is itself in a reduced state.
Footnote |
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c7nj04005a |
This journal is © The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2018 |