Magnetic Fe3O4@SiO2–Pt and Fe3O4@SiO2–Pt@SiO2 Structures for HDN of Indole

The effect of a second porous SiO2 shell in the activity and selectivity of the Fe3O4@SiO2–Pt catalyst in the hydrodenitrogenation of indole is reported. The double Fe3O4@SiO2–Pt@SiO2 structure was prepared by coating Fe3O4 nanoparticles with tetraethyl orthosilicate (TEOS) with a further impregnation of 1.0 wt.% of Pt on the (3-aminopropyl)triethoxysilane functionalized Fe3O4@SiO2 structures. The second porous SiO2 shell, obtained by using a hexadecyltrimethylammonium bromide (CTAB) template, covered the Fe3O4@SiO2–Pt catalyst with a well-defined and narrow pore-sized distribution. The full characterization by TEM, inductively coupled plasma-optical emission spectroscopy (ICP-OES), XRD, and N2 adsorption isotherm at 77 K and vibrating sample magnetometry (VSM) of the catalysts indicates homogeneous core@shell structures with a controlled nano-size of metallic Pt. A significant effect of the double SiO2 shell in the catalytic performance was demonstrated by both a higher activity to eliminate the nitrogen atom of the indole molecule present in model liquid fuel and the improvement of the catalytic stability reaching four consecutive reaction cycles with only a slight conversion level decrease.


Introduction
The presence of polyaromatic and cyclic compounds in diesel negatively affects the environment and reduces the quality of diesel, as well as lowers the cetane number, which is indicative of the ease of ignition thereof. Hydrotreatment is one of the most efficient methods for the removal of N and S from refinery currents, as well as for the saturation of alkenes and aromatics [1][2][3][4]. It is important to highlight the multiplicity of reactions that come into play in hydrotreatment in this area of study. Taking into account the different elements that are considered contaminants in the refining processes of the petrochemical industry, within hydrotreatment we can distinguish hydrodesulfurization (HDS), hydrodenitrogenation (HDN), hydrodeoxygenation (HDO), hydrodemetalization (HDM), and hydrodearomatization (HDA). The HDS has been extensively studied because sulfur is a pollutant that is present in higher proportions in crude oils of lower quality and also as an important catalyst deactivator because of poisoning [5][6][7][8][9]. However, removal of nitrogen from organic compounds present in various fractions of crudes is harder than sulfur removal, which leads to the importance of the study of Pd have been highlighted as efficient and easily recyclable for the catalytic hydrogenation of various groups of olefins, alkynes, keto, and nitro groups [28,29]. One application of the combination of the properties of the core and the shell is the development of magnetically separable catalysts, a quality that represents a significant improvement when evaluating the reuse of a catalyst. Polshettiwar et al. [30] reported the synthesis of supported nanoparticles of Ni-ferrite with excellent magnetic recovery in hydrogenation and transfer hydrogenation reactions. Scähtz et al. [31] in their review article made an extensive analysis of various catalysts supported in ferrite nanoparticles, highlighting their properties for magnetic separation. The application of core@shell silica microspheres, with ultra-small encapsulated nanoparticles of Pt, in the catalytic hydrotreatment process was not found in the literature. The main objective of this work is to develop homogeneous core@shell structures with a controlled nano-size of metallic Pt and study the effect of the double SiO 2 shell in the catalytic performance in the HDN of indole.

Synthesis
All reagents were used without purification or treatment provided; these were acquired from Sigma ® (Darmstadt, Germany) and Merck®(Darmstadt, Germany) Company. The Fe 3 O 4 -NPs were synthesized by a solvothermal method following Long et al. [32], and FeCl 3 (Merck ® ) was dissolved in a polyvinyl pyrrolidone (PVP K30, Sigma ® ) solution in ethylene glycol (EG, Merck ® ) with sodium acetate as nucleating agent. The mixture was transferred to a Teflon autoclave and isothermally treated at 200 • C for 8 h. The solid obtained was separated by magnetization and washed several times with absolute ethanol. The Fe 3 O 4 -core NPs were coated with SiO 2 using the Stöber method [32]. Fe 3 O 4 -NPs were dispersed in a mixture of ethanol, water, and ammonia, and after that tetraethyl orthosilicate (TEOS, Merck ® ) was slowly added to the dispersion under stirring for 6 h. The Fe 3 O 4 @SiO 2 solid was separated by magnetization and washed several times with an ethanol-water mixture. In order to promote an active Pt immobilization and dispersion on the surface of the material, the surface of the Fe 3 O 4 @SiO 2 was functionalized with (3-aminopropyl)trimethoxysilane (AMPTS, Merck ® , 1 mL g −1 of solid), a coupling agent refluxing in toluene for 24 h under mechanical stirring. The solid was separated by magnetization, washed with a toluene-acetone mixture, and dried in an oven at 50 • C for 12 h. The functionalized material was dispersed in a K 2 PtCl 6 (Merck ® ) solution with an amount of precursor necessary to produce 1.0 wt.% systems with respect to Fe 3 O 4 @SiO 2 . The material was placed in contact with the solution for 3 h under mechanical stirring and then reduced with a fresh solution of NaBH 4 (Merck ® ). The second coating of SiO 2 was deposited on the Fe 3 O 4 @SiO 2 -Pt surface by a modified Stöber method reported previously [33] using TEOS as a precursor, triethanolamine (TEA) as a basic catalyst, and hexadecyltrimethylammonium bromide (CTAB, Sigma ® ) as a stabilized and soft template. Finally, the CTAB template was removed by ion exchange using ammonium nitrate in ethanol dissolution under reflux conditions, obtaining the Fe 3 O 4 @SiO 2 -Pt@mSiO 2 material.

Characterization
The morphology and microanalysis of the structures were studied by scanning transmission electron microscopy with energy dispersive X-ray spectroscopy (STEM-EDS) using an FEI Tecnai ST F20 microscope (FEI, Hillsboro, OR, USA) operating at 200 kV. Up to 300 individual metal particles were counted for each catalyst, and the surface area-weighted mean Pt diameter (d p ) was calculated using the software ImageJ 1.48 (Wayne Rasband, National Institute of Health, Bethesda, MD, USA). Adsorption isotherms were obtained at 77 K in a Micromeritics ASAP 2010 instrument (Norcross, GA, USA). X-ray powder diffraction profiles were obtained in a Rigaku Diffractometer with Cu K α radiation (λ = 1.5418 Å) and a nickel filter (Rigaku, Tokyo, Japan). The magnetic behavior was studied using a Lakeshore series 7400 vibrating sample magnetometer (VSM) in an applied field of 20 kOe at 27 • C (Lakeshore, New Orleans, LA, USA). Pt and Fe contents were determined by inductively coupled plasma-optical emission spectroscopy (ICP-OES) with a Perkin-Elmer Optima 2100 DV instrument (Perkin Elmer, Waltham, MA, USA). The catalyst contents were determined after digestion in a 1:3 mixture of HNO 3 :HCl. X-ray source spectrophotometry (XPS) was obtained on a STAIB Instruments brand RQ-300 X-ray Source spectrophotometer (XPS RQ300/2, StaibInstrumente GmbH, Langenbach, Germany). The monochromatic radiation used as an excitation source is that of Al Kα (hν = 11486.6 eV) operated at 75 W.

Catalytic Activity
The hydrodenitrogenation reactions of indole were carried out at 250 • C and 15 atm of H 2 and 500 rpm in a 600 mL stirred autoclave (Parr Pressure Reactor 4536, Parr Instrument Company, Moline, IL, USA). The typical procedure was as follows: 150 ppm of N as indole was dissolved in 50 mL of dodecane (0.01 mol L −1 ). The mixture was placed into the autoclave and the catalyst (250 mg) was transferred to the reactor. The reaction time was 8 h; samples were taken every hour. The products were analyzed with a HP 5890 Series II GC and HP-5 capillary column and identified by GC/MS. Figure 1 presents the high-resolution transmission electron microscopy (HR-TEM) for the image sequence obtained at each step of the catalyst synthesis; the Fe 3 O 4 -core image is shown in Figure 1a. Figure 1b reveals that the CSNs, where the Fe 3 O 4 core is covered by a layer of SiO 2 , were close to 45 nm thick (Table 1). In Figure 1c,d, a successful impregnation-reduction process obtaining Fe 3 O 4 @SiO 2 -Pt and Fe 3 O 4 @SiO 2 -Pt@mSiO 2 catalysts can be seen, respectively. Both catalysts present similar Pt size distribution and an average particle size of 3.5 nm (Table 1). The SiO 2 coating generated by the Stöber method did not modify the distribution or increase the average size of the Pt-NPs on the surface of the material. The micrographs show the formation of a uniform coating on the surface of both catalysts, with channels perpendicular to the surface that correspond to the porosity of the material formed during the removal of the organic CTAB template used as a porosity directing agent. The preferential alignment of the surfactant and the silica oligomer with the core@shell in the performed structures is noticeable.  The catalysts were characterized by ICP-OES to determine the Pt content, as summarized in Table 1. The Fe3O4@SiO2-Pt catalyst contains slightly lower Pt than the nominal value, while for the Fe3O4@SiO2-Pt@mSiO2 catalysts the Pt loading is largely lower than the nominal. The larger difference The catalysts were characterized by ICP-OES to determine the Pt content, as summarized in Table 1. The Fe 3 O 4 @SiO 2 -Pt catalyst contains slightly lower Pt than the nominal value, while for the Fe 3 O 4 @SiO 2 -Pt@mSiO 2 catalysts the Pt loading is largely lower than the nominal. The larger difference in the Pt content for Fe 3 O 4 @SiO 2 -Pt@mSiO 2 is attributed to the second mSiO 2 shell of core@shell particles and not due to the leaching effect. When the mesoporous SiO 2 shell is coated around the Fe 3 O 4 @SiO 2 -Pt structures, a uniform increase in the size of the second shell is observed, reaching 51 nm of thickness in mean diameter (see Table 1). The increase of the thickness provides a dilution of the Pt active phase compared to the Fe 3 O 4 @SiO 2 -Pt structures, in the same way as the increase of the SiO 2 shell thickness.  (2) 178 -

Characterization
(1) Second shell thickness SiO 2 mean diameter in brackets; (2) nominal value in brackets. Figure 2 illustrates the XRD patterns of the synthesized materials. All systems show the diffraction peaks characteristic of magnetite, Fe 3 O 4 (JCPDS 19-0629) [34]. The Fe 3 O 4 phase was not modified during the SiO 2 coating. This finding indicates that the crystalline phase of the material is restricted to the magnetite core. Only for the Fe 3 O 4 @SiO 2 -Pt is a diffraction peak at 2θ = 39 • detected, corresponding to the surface metallic Pt (JCPDS 04-0802). The low intensity of this signal is attributed to the average particle size of the Pt clusters below the detection limit of the XRD technique. The high dispersion of the Pt crystalline phase is in line with the homogenous distribution of Pt-NPs on the surface of the SiO 2 shell observed by HR-TEM, and the surface XPS technique confirms the presence of metallic Pt (see below). in the Pt content for Fe3O4@SiO2-Pt@mSiO2 is attributed to the second mSiO2 shell of core@shell particles and not due to the leaching effect. When the mesoporous SiO2 shell is coated around the Fe3O4@SiO2-Pt structures, a uniform increase in the size of the second shell is observed, reaching 51 nm of thickness in mean diameter (see Table 1). The increase of the thickness provides a dilution of the Pt active phase compared to the Fe3O4@SiO2-Pt structures, in the same way as the increase of the SiO2 shell thickness. (1) Second shell thickness SiO2 mean diameter in brackets; (2) nominal value in brackets. Figure 2 illustrates the XRD patterns of the synthesized materials. All systems show the diffraction peaks characteristic of magnetite, Fe3O4 (JCPDS 19-0629) [34]. The Fe3O4 phase was not modified during the SiO2 coating. This finding indicates that the crystalline phase of the material is restricted to the magnetite core. Only for the Fe3O4@SiO2-Pt is a diffraction peak at 2θ = 39° detected, corresponding to the surface metallic Pt (JCPDS 04-0802). The low intensity of this signal is attributed to the average particle size of the Pt clusters below the detection limit of the XRD technique. The high dispersion of the Pt crystalline phase is in line with the homogenous distribution of Pt-NPs on the surface of the SiO2 shell observed by HR-TEM, and the surface XPS technique confirms the presence of metallic Pt (see below).  Figure 3 shows the spectra of Pt 4f7/2 for the Fe3O4@SiO2-Pt catalyst, with a BE of 70.9 eV attributed to the metallic Pt species. Surface metallic Pt was not detected in the system with the second SiO2 coating; this was attributed to the coverage with the second shell.     Table 1 presents the results obtained for SBET calculated from the N2 adsorption-desorption isotherms at 77 K shown in Figure 4. The Fe3O4@SiO2-Pt shows an isotherm typical of type II materials and a SBET value of 11 m 2 g −1 . After the deposition of the second shell, the isotherm changes to mesoporous type IV materials with a large increase in the specific area. The catalyst Fe3O4@SiO2-Pt@mSiO2 displayed an isotherm type IV with a hysteresis loop type H3 pores typical for cylindrical pores according to the IUPAC (International Union of Pure and Applied Chemistry) classification [35]. The isotherm exhibits two distinguished inflections in the adsorbed amount. The first one at P/P 0 = 0.4-0.6 is attributed to the pores created after CTAB micelles removal [36,37], and the second one at P/P 0 close to 0.8 which can be ascribed rather to the effect of interparticle voids present between some coalesced aggregates of the material [38]. The pore size distribution shows a bimodal distribution with a narrow distribution centered at 3.8 nm and a wider distribution centered at 14 nm associated with the interstitial spaces of the aggregates of the particles. The narrow distribution of mesopores at 3.8 nm is attributed to the removed CTAB organic template. These results are in line with both the shape of isotherm's hysteresis loop and the TEM characterization for the observed core@shell aggregates.  Table 1 presents the results obtained for S BET calculated from the N 2 adsorption-desorption isotherms at 77 K shown in Figure 4. The Fe 3 O 4 @SiO 2 -Pt shows an isotherm typical of type II materials and a S BET value of 11 m 2 g −1 . After the deposition of the second shell, the isotherm changes to mesoporous type IV materials with a large increase in the specific area. The catalyst Fe 3 O 4 @SiO 2 -Pt@mSiO 2 displayed an isotherm type IV with a hysteresis loop type H3 pores typical for cylindrical pores according to the IUPAC (International Union of Pure and Applied Chemistry) classification [35]. The isotherm exhibits two distinguished inflections in the adsorbed amount. The first one at P/P 0 = 0.4-0.6 is attributed to the pores created after CTAB micelles removal [36,37], and the second one at P/P 0 close to 0.8 which can be ascribed rather to the effect of interparticle voids present between some coalesced aggregates of the material [38]. The pore size distribution shows a bimodal distribution with a narrow distribution centered at 3.8 nm and a wider distribution centered at 14 nm associated with the interstitial spaces of the aggregates of the particles. The narrow distribution of mesopores at 3.8 nm is attributed to the removed CTAB organic template. These results are in line with both the shape of isotherm's hysteresis loop and the TEM characterization for the observed core@shell aggregates.  Table 1 presents the results obtained for SBET calculated from the N2 adsorption-desorption isotherms at 77 K shown in Figure 4. The Fe3O4@SiO2-Pt shows an isotherm typical of type II materials and a SBET value of 11 m 2 g −1 . After the deposition of the second shell, the isotherm changes to mesoporous type IV materials with a large increase in the specific area. The catalyst Fe3O4@SiO2-Pt@mSiO2 displayed an isotherm type IV with a hysteresis loop type H3 pores typical for cylindrical pores according to the IUPAC (International Union of Pure and Applied Chemistry) classification [35]. The isotherm exhibits two distinguished inflections in the adsorbed amount. The first one at P/P 0 = 0.4-0.6 is attributed to the pores created after CTAB micelles removal [36,37], and the second one at P/P 0 close to 0.8 which can be ascribed rather to the effect of interparticle voids present between some coalesced aggregates of the material [38]. The pore size distribution shows a bimodal distribution with a narrow distribution centered at 3.8 nm and a wider distribution centered at 14 nm associated with the interstitial spaces of the aggregates of the particles. The narrow distribution of mesopores at 3.8 nm is attributed to the removed CTAB organic template. These results are in line with both the shape of isotherm's hysteresis loop and the TEM characterization for the observed core@shell aggregates.  Even though the relative lower range of value of M s 25-45 emug −1 was enough for the efficient removal below 1 min of the micro-sized particles, it reflected the ability of these catalysts to respond to an external magnetic field, which allows for quick separation from the liquid phase [39]. Fe3O4@SiO2-Pt and Fe3O4@SiO2-Pt@mSiO2 materials exhibit the ferromagnetic properties of the started Fe3O4-NPs, as can be seen in Figure 5. The magnetic saturation (Ms) value of the Fe3O4-NPs changes with subsequent coverage, due to the relative decrease of Fe3O4 content with respect to the presence of the SiO2 shell. The Fe3O4 core corresponds to magnetization curves with an absence of a hysteresis cycle, characteristic of superparamagnetic materials. The decrease of the Ms of 91 emug −1 for magnetite phase (Fe3O4) to 45 emu g −1 for Fe3O4@SiO2-Pt and 25 emu g −1 for Fe3O4@SiO2-Pt@mSiO2 is a result of the effect of the coating with SiO2. Even though the relative lower range of value of Ms 25-45 emug −1 was enough for the efficient removal below 1 min of the micro-sized particles, it reflected the ability of these catalysts to respond to an external magnetic field, which allows for quick separation from the liquid phase [39].

Catalytic Activity
The reaction network of the hydrodenitrogenation of indole proposed by Zhang and Ozkan is shown in Scheme 1 [40].

Catalytic Activity
The reaction network of the hydrodenitrogenation of indole proposed by Zhang and Ozkan is shown in Scheme 1 [40].  Figure 6 shows the conversion of indole as a function of time for the two synthesized catalysts. It is clearly observed that the catalyst with the mesoporous silica coating Fe3O4@SiO2-Pt@mSiO2 is much more active than the uncoated catalyst, reaching a complete conversion at 8 h of reaction, while Fe3O4@SiO2-Pt only reaches 50% conversion. Figure 7 shows the molar fraction of the reaction system for both catalysts. In the case of Fe3O4@SiO2-Pt (Figure 7a), we can observe that indoline is formed  Figure 6 shows the conversion of indole as a function of time for the two synthesized catalysts. It is clearly observed that the catalyst with the mesoporous silica coating Fe 3 O 4 @SiO 2 -Pt@mSiO 2 is much more active than the uncoated catalyst, reaching a complete conversion at 8 h of reaction, while Fe 3 O 4 @SiO 2 -Pt only reaches 50% conversion. Figure 7 shows the molar fraction of the reaction system for both catalysts. In the case of Fe 3 O 4 @SiO 2 -Pt (Figure 7a), we can observe that indoline is formed first, then OEA, and that the appearance of EB and ECH is very slow. Contrary for Fe 3 O 4 @SiO 2 -Pt@mSiO 2 (Figure 7b), the consumption of indole and the appearance of EB and especially ECH (the denitrogenated products) is very fast. The percentage of denitrogenated products (% HDN) was calculated and listed in Table 2. In the table, the results were compared with those of a typical HDN catalyst NiMo/Al 2 O 3 Criterion DN200 [11]. This catalyst was presulfided according to our previously reported results [11]. Catalyst Fe 3 O 4 @SiO 2 -Pt@mSiO 2 reached a higher % HDN value compared with the other two catalysts. A control reaction was carried out using Fe 3 O 4 @SiO 2 as a catalyst under the same conditions and no activity was obtained, confirming the non-catalytic activity of the support.  Figure 6 shows the conversion of indole as a function of time for the two synthesized catalysts. It is clearly observed that the catalyst with the mesoporous silica coating Fe3O4@SiO2-Pt@mSiO2 is much more active than the uncoated catalyst, reaching a complete conversion at 8 h of reaction, while Fe3O4@SiO2-Pt only reaches 50% conversion. Figure 7 shows the molar fraction of the reaction system for both catalysts. In the case of Fe3O4@SiO2-Pt (Figure 7a), we can observe that indoline is formed first, then OEA, and that the appearance of EB and ECH is very slow. Contrary for Fe3O4@SiO2-Pt@mSiO2 (Figure 7b), the consumption of indole and the appearance of EB and especially ECH (the denitrogenated products) is very fast. The percentage of denitrogenated products (% HDN) was calculated and listed in Table 2. In the table, the results were compared with those of a typical HDN catalyst NiMo/Al2O3 Criterion DN200 [11]. This catalyst was presulfided according to our previously reported results [11]. Catalyst Fe3O4@SiO2-Pt@mSiO2 reached a higher % HDN value compared with the other two catalysts. A control reaction was carried out using Fe3O4@SiO2 as a catalyst under the same conditions and no activity was obtained, confirming the non-catalytic activity of the support.  The better activity of Fe 3 O 4 @SiO 2 -Pt@mSiO 2 compared with the NiMo/Al 2 O 3 catalyst can be related to the high hydrogenating capacity of platinum in Fe 3 O 4 @SiO 2 -Pt@mSiO 2 . In the case of both core@shell structures, it is very clear that the difference in activity is due to the presence of the second porous shell and not to the dispersion of the platinum nanoparticles.
The large difference in activity has to be explained in relation to the presence of the mesoporous silica layer and in the greater surface area of Fe 3 O 4 @SiO 2 -Pt@mSiO 2 compared with Fe 3 O 4 @SiO 2 -Pt. Fe 3 O 4 @SiO 2 -Pt@mSiO 2 possesses cylindrical channels perpendicular to the surface that corresponds to the porosity of the material formed during the removal of the organic template. The synthesis of such coating was carried out using triethanolamine (TEA), which confers a basic property, leading to a negatively charged surface of the SiO 2 . Moreover, due to the removal of CTAB that was performed by ion exchange using ammonium nitrate in ethanol dissolution, this exchange by NH 4 + cations enables adherence as counterions. The presence of these cations generates a slight surface acidity in the porous SiO 2 of the second coating [41] which benefits HDN reactions, as we have reported previously [11]. This is also evidenced in the larger amount of ECH with respect to EB for the Fe 3 O 4 @SiO 2 -Pt@mSiO 2 catalyst, similar to other acid catalysts [11]. On the other hand, in line with previous reports for heterogeneous core-shell catalysts with porous coatings, the porosity of the second coating generates a confinement effect into the cylindrical pores that favor the adsorption of the reactants and the re-adsorption of the intermediates of the reaction, increasing the contact time between intermediates and the platinum active catalytic species [42].  The better activity of Fe3O4@SiO2-Pt@mSiO2 compared with the NiMo/Al2O3 catalyst can be related to the high hydrogenating capacity of platinum in Fe3O4@SiO2-Pt@mSiO2. In the case of both core@shell structures, it is very clear that the difference in activity is due to the presence of the second porous shell and not to the dispersion of the platinum nanoparticles.
The large difference in activity has to be explained in relation to the presence of the mesoporous silica layer and in the greater surface area of Fe3O4@SiO2-Pt@mSiO2 compared with Fe3O4@SiO2-Pt. Fe3O4@SiO2-Pt@mSiO2 possesses cylindrical channels perpendicular to the surface that corresponds to the porosity of the material formed during the removal of the organic template. The synthesis of such coating was carried out using triethanolamine (TEA), which confers a basic property, leading to

Reutilization Study
Catalyst reuse is important from an industrial point of view. In this case, the catalysts have been tested during four catalytic cycles. The samples were washed several times with a mixture of methanol and water previous to the reaction. Figure 8 shows that the activity slightly decreases after the third recycle in the case of Fe 3 O 4 @SiO 2 -Pt@mSiO 2 , but the loss in activity is higher in the case of Fe 3 O 4 @SiO 2 -Pt. The higher stability of Fe 3 O 4 @SiO 2 -Pt@mSiO 2 could be due to the presence of the second porous silica coating that inhibits the loss of the metal charge and prevents sintering.
The stabilization achieved after the second layer of silica allows the successive reuse of the catalyst. In addition, the magnetic characteristics of this catalyst facilitates the separation process, which is very important from the practical and economic perspective in any industrial process.
Catalyst reuse is important from an industrial point of view. In this case, the catalysts have been tested during four catalytic cycles. The samples were washed several times with a mixture of methanol and water previous to the reaction. Figure 8 shows that the activity slightly decreases after the third recycle in the case of Fe3O4@SiO2-Pt@mSiO2, but the loss in activity is higher in the case of Fe3O4@SiO2-Pt. The higher stability of Fe3O4@SiO2-Pt@mSiO2 could be due to the presence of the second porous silica coating that inhibits the loss of the metal charge and prevents sintering. The stabilization achieved after the second layer of silica allows the successive reuse of the catalyst. In addition, the magnetic characteristics of this catalyst facilitates the separation process, which is very important from the practical and economic perspective in any industrial process.

Conclusions
In this work, we studied the effect of a second porous SiO2 shell in the activity and selectivity of the Fe3O4@SiO2-Pt catalyst in the hydrodenitrogenation of indole. The core-shell structures were applied for the first time in the indole HDN process. We found that the presence of a mesoporous SiO2 coating dramatically increases the activity of the catalyst. The great difference in activity was explained in terms of a confinement effect of the intermediate's products of the reaction in the second SiO2 layer. This coating of slightly acidic character, and with parallel mesochannels, favored the readsorption and transformation into denitrogenated products, allowing the Fe3O4@SiO2-Pt@mSiO2 catalyst to remain active after several catalytic cycles, and its magnetic character allows its easy separation and recovery.

Conclusions
In this work, we studied the effect of a second porous SiO 2 shell in the activity and selectivity of the Fe 3 O 4 @SiO 2 -Pt catalyst in the hydrodenitrogenation of indole. The core-shell structures were applied for the first time in the indole HDN process. We found that the presence of a mesoporous SiO 2 coating dramatically increases the activity of the catalyst. The great difference in activity was explained in terms of a confinement effect of the intermediate's products of the reaction in the second SiO 2 layer. This coating of slightly acidic character, and with parallel mesochannels, favored the re-adsorption and transformation into denitrogenated products, allowing the Fe 3 O 4 @SiO 2 -Pt@mSiO 2 catalyst to remain active after several catalytic cycles, and its magnetic character allows its easy separation and recovery.