Introduction

Hydrogen is a renewable energy carrier which can replace fossil fuels, reducing CO2 and other organic pollutant emissions [1]. Hydrogen can be produced by various methods such as methane reforming, biomass reforming, thermochemical and electrolysis methods [2,3,4,5,6,7]. Among all the carbon-free thermochemical cycles for hydrogen production, Iodine–Sulfur (I–S) process is reported to have the highest efficiency of about 51% at 1173 K [8,9,10,11,12]. This cycle operates in a closed loop with the three main reactions including sulfuric acid decomposition (oxygen evolution reaction), Bunsen (water splitting reaction), and HI decomposition (hydrogen generation reaction. Among these three reactions, the two-step sulfuric acid decomposition is the most endothermic, high-temperature and corrosive reaction, wherein, sulfuric acid is thermally decomposed to SO3 and H2O in the first step, and SO3 is further decomposed to SO2 and O2 in the presence of a potential catalyst in the second step [9, 13,14,15,16,17]. This SO3 decomposition exhibits a large kinetic barrier [18,19,20,21,22].

Sulfuric acid decomposition

$${\text{H}}_{{2}} {\text{SO}}_{{4}} \,\left( {\text{l}} \right)\, \to \,{\text{H}}_{{2}} {\text{O}}\,\left( {\text{g}} \right)\, + \,{\text{SO}}_{{3}} \,\left( {\text{g}} \right)\, \left( {{673}{-}{873}\,{\text{K}}} \right)$$
(1)

Sulfur trioxide decomposition

$${\text{SO}}_{{3}} \,\left( {\text{g}} \right)\, \to \,{\text{SO}}_{{2}} \,\left( {\text{g}} \right)\, + \,0.{\text{5O}}_{{2}} \,\left( {\text{g}} \right)\,\left( {{1}0{73}{-}{1273}\,{\text{K}}} \right)$$
(1A)

Bunsen reaction

$${\text{SO}}_{{2}} \, + \,{\text{2H}}_{{2}} {\text{O}} \, + \,{\text{I}}_{{2}} \, \to \,{\text{H}}_{{2}} {\text{SO}}_{{4}} \,\left( {{\text{aq}}.} \right)\, + \,{\text{2HI}}\,\left( {{\text{aq}}.} \right)\,\left( {{298}{-}{393}\,{\text{K}}} \right)$$
(2)

HI decomposition

$${\text{2HI}}\,\left( {\text{l}} \right)\, \to \,{\text{H}}_{{2}} \,\left( {\text{g}} \right)\, + \,{\text{I}}_{{2}} \,\left( {\text{g}} \right)\,\left( {{473}{-}{673}\,{\text{K}}} \right)$$
(3)

To date, several efforts have been made to enhance catalyst activity and stability for sulfuric acid decomposition into SO2 [23,24,25]. Many catalysts including platinum group metals, transition metal oxides, and complex metal oxides (perovskites and spinels) have been extensively researched in this reaction [26,27,28,29,30,31]. Among all reported catalysts, metal oxide-based catalysts seem to be the most promising substitute of platinum-based catalysts [18, 23]. The catalytic decomposition of the SO3 has been improved by immobilizing active metal over suitable support material such as SiO2 [18, 32, 33], TiO2 [32, 34], Al2O3 [27], SiC[35, 36]. In the recent past, Karagiannakis et al. [37], Giaconia et al. [38], Banerjee, and co-workers[23] reported Fe2O3 and chromium-doped Fe2O3 for sulfuric acid decomposition reaction. Ashish and co-workers prepared Fe2O3 supported over SiO2 and achieved high activity. It was expected that the high activity and stability of the catalyst were due to smaller size (∼2 nm) and Fe–O–Si linkages [18, 32]. They further synthesized Fe2O3 nanoparticles over commercial silica through polyol, solvothermal, and wet-impregnation methods. They obtained the α-Fe2O3 and ε-Fe2O3 phases in the particle size range of 10–20 nm. Rod-shaped, spherical, and elliptical-shaped nanoparticles were achieved through polyol, solvothermal and wet impregnation methods and achieved high activity and stability (79% at 1073 K). However, an earlier work of the same research group showed that a dramatic increase in the catalytic activity of the platinum nanoparticles dispersed over Al2O3 [24] was due to a smaller particle size ( ∼2 nm).

It has been observed that metal loss and sintering were more drastic at high temperatures, especially, in platinum-based catalysts. The catalytic deactivation was due to the increased mobility of nanoparticles at high temperatures. Henceforth, to avoid direct contact with acid vapors nanoparticles can be placed in the pores of mesoporous silica [39,40,41]. In this regard, Khan and co-workers placed platinum nanoparticles at the inner wall of hollow mesoporous silica spheres (Pt-HMSS) and the outer wall of hollow mesoporous silica spheres (HMSS-Pt) through the reduction and impregnation method [42]. The synthesized nanoparticles were in the range of 3.8–5.4 nm and prevented the metal loss when nanoparticles placed in inner wall hollow mesoporous silica spheres (Pt-HMSS) as compared to nanoparticles tethered to the outer wall of the mesoporous silica spheres (HMSS-Pt). However, catalyst activity reduced significantly after a 50-h run probably due to the cracking of the hollow spheres. They further extended the work and stabilized platinum nanoparticles on the porous wall of the hollow SiC (Pt/hSiC) [25]. However, Pt/hSiC exhibited poor stability after a 6 h run due to nanoparticles agglomeration and sintering.

Hence, extensive research has been made to prevent the active metal loss, support sulfation, and keep nanoparticles in their active state [32, 43,44,45,46,47]. Recently, Pt core silica shell structures were synthesized through sol–gel process to prevent the active platinum site loss due to sintering [48]. They achieved ~ 77% conversion at 1123 K and only 9% loss in active metal was observed after 100 h run. However, the surface area of the spent catalyst was shown to significantly deteriorate from 466 m2/g in pristine catalyst to 60.9 m2/g in the spent catalyst. Until now, most efforts have been dedicated to the preparation, and less has been devoted to studying them under extreme reaction conditions. It has been observed that the use of mesoporous silica-supported metal oxides has been exaggerated due to their size tunability, high surface area, and strong metal-support interaction. However, they are prone to several structural changes under reaction conditions. Therefore, in this work, we examine the core–shell structures that are prone to undergo such changes pronouncedly due to their high surface mobility.

Herein, iron oxide nanoparticles were immobilized through two different methods in the pores and core of the mesoporous silica for maintaining high catalytic activity and preventing metal loss. In the first method, the mesoporous silica support was synthesized, and iron oxide nanoparticles were immobilized in the pores of this mesoporous silica, while, in the second method, iron oxide nanoparticles were synthesized separately through the solvothermal method, and later, these were used as the core and a coating of mesoporous silica shell was accomplished through the sol–gel process. The prepared nanoparticles were characterized through XRD, FTIR, FESEM, HRTEM, and Raman. These nanoparticles were examined for their catalytic activity and stability for the highly endothermic oxygen-evolving reaction (both reactions (1) and (1.A) together i.e. sulfuric acid decomposition into SO2 and O2).

Experimental

Synthesis of the nanoparticles

Preparation of the catalyst support (SiO2)

For the synthesis of mesoporous silica, many routes have been reported, however, butanol assisted synthesis method was preferred due to its high order of mesoporous, large pore volume, and pore size [49, 50]. Catalyst support was prepared by mixing butanol and Pluronic P123 (EO20PO70EO20) in equal the ratio in 0.5 M HCl at 298.15–308.15 K. First of all, 12 g of P123 was dissolved in 434 g of distilled water and 23.6 g of concentrated HCL (35%) [49, 50]. After this, 12 g of butanol (Aldrich, 64%) was added while stirring at 308.15 K. After 1 h stirring, silica source TEOS (25.8 g, ACROS, 98%) was added at 308.15 K which made the solution mole ratio of TEOS:P123:HCL:H2O = 1:0.017:1.83:195:1.31. After this, the mixture was kept for stirring for 24 h at 308.15 K \(,\) followed by hydrothermal treatment at 373.15 K for 24 h in a closed polypropylene bottle. After hydrothermal treatment, solid the product was obtained which was filtered and dried at 373.15 K without washing. Finally, mesoporous silica (SiO2) was calcinated at 823 K for 5 h.

Impregnation of iron on support

10% iron was loaded on the support using the wet impregnation technique. In a typical synthesis batch, 725 mg of Fe(NO3)3·9H2O was dissolved in 10 ml of de-ionized water. This solution was added dropwise to 1 g of silica support material, till a paste was formed. This paste was kept for drying at 383 K for 2 h. Then it was taken out and the procedure was repeated till the entire precursor solution was expended. Then it was kept for drying overnight, followed by calcination at 823 K for 4 h. The prepared sample is named FSW. Using a similar approach, iron oxide precursor was dissolved in acetone and the resulting sample is named FSA. In another synthesis, the iron oxide precursor was dissolved in a solution of resins and acetone replacing de-ionized water and thus prepared sample is named FSR. FSW, FSA, and FSR were prepared through the wet impregnation method as described above.

Synthesis of iron oxide@mesoporous core–shell structures

Iron oxide embedded mesoporous core–shell structures were synthesized as per the method as previously described in the literature [51, 52]. The synthesis of iron oxide nanoparticles immobilized inside a mesoporous silica shell (FCSC) is a two-step process. The first is the synthesis of iron oxide nanoparticles through the solvothermal method. In the second step, these synthesized iron oxide nanoparticles in step 1 are used as the core and a coating of mesoporous silica shell is accomplished through the sol–gel process. In step 1, a certain amount of iron precursor, polyvinylpyrrolidone, and acetic acid was dissolved in 100 ml ethanol and stirred at room temperature for 1 h. Afterward, the solution was transferred to the autoclave and kept at 473 K for 24 h. Then the autoclave was cooled, and the resultant solution contains the iron oxide nanoparticles. In step 2, 100 ml of water and ethanol were added to this solution. Then, the solution was titrated with ammonia to achieve pH = 9. Afterward, 0.4 g cetyltrimethylammonium bromide (CTAB) and 5 ml tetraethoxysilane (TEOS) were added slowly. The solution was stirred for 24 h at ambient conditions. The prepared nanoparticles in the solution were filtered, washed, and dried at 373 K. The nanoparticles were calcined at 823 K for 4 h. The calcined sample is named FSCS throughout the manuscript.

Characterization of material

The catalyst was characterized using an X-ray diffraction technique using a Rigaku diffractometer, operated using a Siemens D500 diffractometer with Cu Kα radiation (λ = 0.15 nm). The data was collected from 4 to 90° (2θ) angle, taken at a step of 0.02° and a rate of 4°/min. The operating conditions were 40 kV and 15 mA. The crystallite mean size was calculated from the full width at half maximum (FWHM) using the Scherrer formula \(\tau =\frac{K\lambda }{Bcos\theta }\), where \(\tau\) is the mean size of the ordered (crystalline) domain, K is the shape factor (0.89), \(\lambda\) is the X-ray wavelength (0.15 nm), B is the line broadening at half the maximum intensity in radian, and \(\theta\) is the Bragg angle. Further, the nature of crystalline phases can be validated with Inorganic Crystal Structure Database (ICSD). The chemical composition can be analyzed through Fourier-transform infrared spectroscopy (FT-IR) measured in ambient conditions on a Nicolet Nexus 470 spectrometer in the range of 400–4000 cm−1 range in transmittance mode (pellets with KBr) or diffuse reflectance mode using Al mirror as a background. The morphology of the catalyst surface was analyzed using Field Emission Scanning Electron Microscopy (FESEM) images captured using ZEISS EVO Series EVO 50 at an accelerating voltage of 15 kV. The size and shape of the pristine and spent catalysts are estimated through FESEM analysis. Further, the size and shape of the nanoparticles including crystallite size are examined through low and high-resolution Transmission Electron Microscopy (TEM) images captured using Technai™ G2 20 at a voltage of 200 kV. Initially, powdered samples were suspended in ethanol under ultrasonic vibration. A drop of this suspension was deposited on a carbon-coated 300-mesh copper grid. The surface area of the catalysts was determined by Microtrac BELSORP MAX I. Pore size distributions were estimated from the desorption branch by the Barrett-Joyner-Halenda (BJH) method.

Experimental evaluation and analysis of the catalysts

The gas–solid reaction performance evaluation over the lab synthesized sinter resistant catalysts was conducted in a fixed bed continuous-flow quartz tube reactor (700 mm long with I.D. 8 mm). The typical experimental setup includes a syringe pump, quartz reactor placed in split tubular furnace, and a condenser set-up from where products and unreacted sulfuric acid are collected and analyzed. A schematic of the typical experimental reaction set-up for sulfuric acid decomposition is shown in Fig. S1, which indicates a tubular reactor inside a furnace. All the synthesized catalysts were loaded after being pelletized and crushed into particles between sizes of 0.5–0.6 mm. Ceramic beads were filled into the packed bed reactor on both sides of the catalyst packed bed for better heat distribution. The catalyst bed was sandwiched between two glass wool layers inside the reactor. Here, concentrated sulfuric acid (98 wt%) was pumped using a syringe pump into the quartz reactor maintained at a high temperature. All the products, and reactants except oxygen and inert nitrogen were neutralized in NaOH and moisture content in silica gel. The outlet flow rate was measured through a mass flow meter.

Catalytic activity measurements and reaction conditions

The activities of the prepared catalysts were evaluated in a continuous flow reactor as described in the previous section. Our previous studies showed the best reaction performance at a temperature ranging from 1000 to 1173 K and 8.1 h−1 WHSV (weight hour space velocity) [27, 36]. The nitrogen flow rate was fixed at 30 ml/min and the temperature was varied from 1023 to 1173 K at WHSV of 8.1 h−1. During all the experiments, 1 g catalysts were loaded into the reactor at the sulfuric acid flow of 4.4 ml/h. Sulfuric acid decomposition occurs in two steps: The first step is a thermal decomposition step, which usually takes place at ~ 100% conversion at the temperatures used in the experiment. The second step is the kinetically controlled step, resulting in the formation of SO2 and O2. Assuming that all the moisture and sulfur oxides were absorbed downstream of the condenser, only oxygen come out to be measured in the mass flow meter alongside nitrogen. The catalytic activity was measured in terms of percentage sulfuric acid conversion given by:

$${\text{Sulfuric acid conversion (}}X){\mkern 1mu} = {\mkern 1mu} \frac{{f_{{out}} - f_{{in}} }}{{Theoretical\, {\mkern 1mu} maximum\, {\mkern 1mu} oxygen\, {\mkern 1mu} flowrate}}{\mkern 1mu} \times {\mkern 1mu} 100$$
(4)

In experimental studies of transport in the heterogeneous catalytic reactor, the first step is to eliminate the effect of any external and internal mass transfer resistance region. Reaction kinetics and catalyst performance are strongly influenced by these external and internal mass transfer limitations. Henceforth, all the experiments were conducted in an internal and external diffusion-free region, ie., under the kinetic regime (Sections S1 and S2 in the Supplementary Information).

Results

Phases and lattice structure identification of the material

Iron oxide exists in many polymorphic phases (\(\alpha ,\beta ,\gamma\) and \(\varepsilon\)). Among these, \(\alpha\)-Fe2O3 also known as hematite is found to be the most stable iron oxide polymorphs. Phase identification of the dispersed iron oxide over support material is quite completed due to low metal loading and nanocrystalline nature. Phase identification in all synthesized sample was attempted by recording and analyzing the wide-angle powder XRD. The powder XRD pattern of all synthesized samples at 1273 K is shown in Fig. 1a. The broad hump at peak position 2 \(\theta \sim 22^\circ\) is attributed to the support SiO2 as evident from the literature [33]. In FACS, the characteristic peaks are clearer and more crystalline as compared to FSR, FSA, and FSW.

Fig. 1
figure 1

a XRD data for the calcined support and impregnated iron catalyst FSR, FSA, FSW, and FSCS. b Small-angle X-ray scattering profiles of all prepared catalysts. c FTIR spectra of SiO2, FSR and FSCS. d SAXS profile of FCSC with data fitting

Full size image

The XRD spectra of the iron oxide supported over SiO2 (FSR, FSA FSW, and FSCS) as shown in Fig. 1a are compared with the reference code (98–006-4599). The characteristic peaks at 2 \(\theta \sim 24.25^\circ , 33.24^\circ , 35.76^\circ , 49.58^\circ , 54.20^\circ , 62.5^\circ ,\) and \(64.14^\circ\) are attributed to Fe2O3 (hexagonal crystal system, space group R-\(\overline{3 }\) c (167), lattice: hexagonal, cell parameters: a = b = 5.0290 nm, c = 13.7360 nm, \(\alpha =\beta =90^\circ , \gamma =120^\circ\)) [012], [104], [110], [024], [116], [214], and [030]. In the case of FSCS, the calculated crystallite size using XRD patterns is found in the range of 24–34.6 nm. The peak at 2 \(\theta\) = 33.24 \(^\circ\) in the XRD peak corresponds to the low energy plane of Fe2O3 (III) (104) which works as an anchor site for obtaining an O from absorbed SO3. The observed XRD pattern is well matched with the reference data and no reflection of other impurities has been noticed in the pattern which reveals that the prepared nanoparticles are pure.

Particle characterization by small-angle X-ray scattering

In the recent past, the small angle X-ray scattering (SAXS) analysis has received much attention in nanoparticle research because it allows us to accurately determine the structure of particles and size distribution. Additionally, meso-structural ordering can also be confirmed using SAXS. The Small Angle X-rays Scattering (SAXS) spectra of all samples are shown in Fig. 1b. As illustrated in Fig. 1b, the SAXS pattern of all samples shows well-resolved scattering peaks in the \(2\theta\) range of 0.03–2 degree, suggesting an ordered hexagonal mesostructured which are also in accordance with the TEM observations (Fig. 2, 3a). The SAXS plot of FSCS shows a pothole which indicates a core–shell type structure while others do not represent the same. However, SiO2 and iron supported over silica (FSA, FSW, and FSR) exhibit the hexagonal structure with the occurrence of peaks due to (10), (11), (21), and (30) planes [49, 53,54,55,56,57]. The presence of these well-resolved peaks is associated with a highly ordered mesoporous silica structure. A smaller shift in these peaks is observed in the case of iron support over mesoporous silica which is in agreement with the reported literature [53, 54, 58]. Taking into account the pore size calculated from the TEM images or BET analysis, the pore wall thickness can be calculated and it is found to be 4.046 nm which is in agreement with the previous studies as shown in Figs. 2c–d [53, 54, 58].

Fig. 2
figure 2

a FESEM images of the FSCS b Particle size distribution of FSCS. c TEM images of mesoporous SiO2 and d fringes of SiO2

Full size image
Fig. 3
figure 3

a Iron core–shell structure. b diffraction images of the iron core–shell structure. c particle size distribution of the shell d particle size distribution of the iron core in FSCS

Full size image

By using scattering profiles, the structure of the nanoparticle can be calculated in terms of its radius of gyration, largest nanoparticle dimension (Dmax), and pair distance distribution function (PDDF) [59,60,61,62]. The form factor can be approximated by using a Gaussian curve at a small angle [62]. The average nanoparticle radii can be estimated in the low-q region by using Guinier's law as given in Eq. 5.

$$\ln \left[ {\Delta I\left( q \right)} \right]\, = \,\ln \,\left[ {a_{0} } \right]\, - \,\frac{{R_{G}^{2} }}{3}$$
(5)

Here, \(I\left(q\right)\), q, \({a}_{0}\), RG corresponds to scattered X-ray intensity, scattering vector, scattering intensity at q = 0 nm−1, and radius of gyration. The radius of gyration (RG) can be estimated from the slope in the linear fit of \(\mathrm{ln}\left[\Delta I\left(q\right)\right]\) vs \({q}^{2}\) and radius of the homogeneous sphere (Rh) using the following Eq. (6).

$$R_{h} \, = \,\sqrt{\frac{5}{3}} R_{G}$$
(6)

Figs. 1d, S2 and S3 show a natural logarithmic plot of intensity as a function of the square of the scattering vector plot of all samples and the calculated values of RG and Rh are shown in Table 1 which is in agreement with the results obtained from the TEM (Fig. 3). Additionally, the shape of the nanoparticles can be easily obtained from the Kratky plot (I(q)\(\times\) q2 vs. q) as shown in Figs. S2 and S3. This plot gives the information of the structure based on the high q region with q > 5/RG of the form factor equation. The bell-shaped curves indicate that the SiO2, FSA, FSR, and FSCS nanoparticles are globular in shape which is in agreement with the structure of the particle shown in TEM images (refer to Fig. 3).

Table 1 Sample details, composition, and crystallite size
Full size table

In order to gain knowledge of the structures of all synthesized catalysts, the PDDF, p(r) was estimated using SAS analysis software. This p(r) function can be estimated by an Indirect Fourier transformation of the form factor as given in Eq. (7) [61, 63].

$${\text{P}}\left( {\text{q}} \right){\mkern 1mu} = {\mkern 1mu} 4\pi {\mkern 1mu} \int\limits_{{ - \infty }}^{\infty } {\left( p \right)r\frac{{{\text{sin}}\left( {qr} \right)}}{r}dr}$$
(7)

Here P(q) corresponds to the form factor of the synthesized material.

P(r) is the plot of all the distance distribution of pairs of atoms within a system of particles with the same electron density. These PDDF plots can be used to estimate the shape of the particles. The maximum size of the particle (Dmax) can be evaluated by the last point of the P(r) curve intersecting the x-axis. Fig. S2 exhibits the P(r) plots of SiO2, FSR, and FSCS. The P(r) plots of the FSA, FSW, and FSR spent catalysts are shown in Fig. S3. The bell shape curve represents the globular-shaped nanoparticles. The maximum size of the FSCS is ~ 105.5 nm which is in agreement with the results obtained from TEM analysis. The average diameter of the shell calculated using Eq. 5 is 92.39 nm and the calculated diameter of the core is 35.1 nm which agrees with TEM results (Fig. 3).

Interface identification and analysis of the material components

Fourier transform infrared spectroscopy (FTIR) allow us to identify the material’s molecular composition by production absorption/transmittance spectrum. The FTIR spectra of the FSCS are shown in Fig. 1c. The transmittance band appearing at 1079 cm−1 is ascribed to asymmetrical stretching vibration. The band appearing at 464 cm−1 is attributed to the Si–O–Si bond vibrations while the peak near 800 cm−1 is attributed to the vibrations due to [SiO4] tetrahedron. The peak appearing in FSR and FSCS near 600 cm−1 is ascribed to the Fe–O stretching vibration in Fe2O3 lattice which is not the case in SiO2. These peaks indicate that the synthesized iron core and silica shell structure.

Morphological features of the developed interface and nanomaterial

The nanoparticles coated with PVP can be used as a core without further purification because purification led to the aggregation of nanoparticles. An important step in the preparation of iron oxide into silica mesospheres by a sol–gel process is the synthesis of iron oxide stock solution in ethanol. To investigate the morphological features of the synthesized samples, The FESEM and TEM images are captured. The FESEM images illustrate the uniform and monodisperse iron oxide silica core–shell structure (refer to Figs. 2a, b). TEM allows us to estimate the crystallite size of the synthesized nanoparticles. The crystallite size of d = 0.24 nm corresponds to the [110] planes of \(\alpha\)-Fe2O3. HRTEM images illustrate the presence of crystalline nanoparticles with an oval shape with an average size of 33.5 nm. Figure 3d shows the iron oxide distribution in the shell of the silica and the average particle size of Fe2O3 is found to be 33.5 nm. Similarly, Fig. 3c shows the size distribution of the silica shell, and the average particle size of the silica shell is found to be 103.5 nm. From Figs. 3a–d, it is evident that both silica shell and iron oxide core are well dispersed and don’t agglomerate and forms larger particles. Hence, it is expected that the well-dispersed nanoparticles can play a vital role in activity and stability.

The thickness of the silica layer can be adjusted by changing the amount of TEOS in the sol–gel process. The iron oxide nanoparticles are mainly confined to the center of each silica. Each silica sphere contains one monodispersed iron oxide nanoparticle. This clearly indicates the iron oxide is stabilized in the core of the silica shell. The average score and shell size obtained from HRTEM images (see Figs. 3a-d) are in agreement with the outcome obtained from SAXS analysis. The iron oxide core is highly crystalline as evident from Fig. 3b.

N2 Surface area characterization

The textural properties can be estimated through N2 adsorption and desorption analysis which was carried out at 77 K temperature maintained using liquid nitrogen. The surface area and pore size distribution of all synthesized samples are measured and analyzed through BET and BJH analysis. The nitrogen adsorption/desorption isotherms and pore size distribution of all prepared and core–shell nanoparticles are shown in Fig. S4. The nitrogen adsorption/desorption isotherms of samples can be represented as type IV isotherms, illustrating a uniform sized mesoporous structure whereas, sample FSCS can be classified as the combination of type III and IV isotherms.

BET surface area and pore size distribution are illustrated in Table 2. When the relative pressure is very low (p/p0 ≤ 0.3), the leap in the isotherm could be attributed to the mesopores, wherein, capillary condensation of nitrogen occurred. In the case of FSCS, the hysteresis loop between the adsorption and desorption is completely absent, indicating a monolayer of N2 on the wall of the pores. FSCS shows an adsorption leap at p/p0 = 0.3 and ~ 0.95 that is ascribed to the presence of the mesopores in the silica shell. BET surface area of the synthesized SiO2 is 631.1 m2/g with a pore size of 4.5 nm and pore volume of 1.01 cm3/g. After iron oxide loading, the surface area of FSR, FSA, and FSW significantly reduced indicating the stabilization of iron oxide in the pores, and pore volume is also reduced. This reduction in surface area and pore volume is minimum in the case of FSW as compared to the FSA and FSR suggesting that the iron oxide is stabilized over the superficial surface in FSW instead of pore filling. However, FSCS shows the highest surface area and pore volume as compared to FSR, FSA, and FSW. Based on the pore size, all the synthesized samples can be classified as mesoporous material owing to pore size between 2 and 50 nm. It is evident from Table 2 that all samples retained mesopore structure even after iron oxide deposition. The high pore volume and surface area of FSCS make it more accessible for reactant adsorption which is directly related to the performance of the catalyst.

Table 2 BET surface area, pore size, and pore volume
Full size table

Discussion

Reaction performance and sensitivity of reaction to temperature

The catalytic activities of all samples are tested in the temperature range of 1048–1173 K as shown in Fig. 4. It is clear from Fig. 4a that catalytic activity increases with an increase in temperature. The conversion has increased from 15% at 1048 K to 80% at 1173 K due to the improved kinetics. FSCS exhibits the highest sulfuric acid conversion at elevated temperature range and approaches thermodynamic conversion. FSCS attains a maximum of 92% at 1173 K which is equivalent to the thermodynamic conversion. It is expected that the high catalyst activity could be due to well-dispersed iron oxide nanoparticles without agglomeration, high surface area, and pore volume of the core–shell structure. The high catalytic activity of FSCS could be attributed to the large surface area and high pore volume as compared to the FSA, FSR, and FSW. After catalysts catalytic activity analysis, it is indispensable to investigate the long-run stability of the prepared nanoparticles. The calculated activation energies of the FSR and FSCS are found to be 143.53 kJ/mol and 127.39 kJ/mol, respectively (see Fig. 4c), which are in the range of iron oxide-based catalysts [2, 10]. The high values of activation energies infer that the experiments are performed in a kinetically controlled regime.

Fig. 4
figure 4

a The effect of temperature on catalytic activities of all iron supported catalysts (FSR, FSA, FSW and FSCS) in the temperature range of 1000 to 1173 K at 8.1 h−1 WHSV. b 100 h time-on-stream catalytic activity of FSR and FSCS at 1123 K and 8.1 h−1 WHSV. c Arrhenius plot of FSR and FSCS in the temperature range of 1000 to 1173 K at 8.1 h−1 WHSV. All the experiments were performed with a 30 ml/min nitrogen flow, 1 g of catalyst loading, and 4.4 ml/h sulfuric acid flowrate

Full size image

Long time-on-stream stability testing of catalysts

The FSR shows initial catalytic activity at ~ 72% at 1123 K which is equivalent to the Pt core silica shell structure–activity achieved at 1123 K. However, its activity starts decreasing after 50 h run. (Fig. 4b). The cause of the deactivation can be confirmed through detailed characterization of the spent catalyst. It is found that the nanoparticle agglomeration and decrease in surface area and pore volume due to support sulfation are the main reason for the deactivation. From the literature, it is evident that Pt loss was the main reason for deactivation in the case of platinum core silica shell structure [48]. On the other hand, FSCS catalyst shows the high catalytic activity even after 50 h and remains active more than 100 h and achieved ~ 82% conversion at 1123 K. There is no significant decrease in catalytic activity is observed during 100 h time-on-stream run.

Spent catalyst characterization after stability testing

Post 100 h time-on-stream catalytic performance, the spent catalysts are characterized through XRD and BET to understand the changes that occurred in structural properties due to 100 h prolonged run at 1123 K for sulfuric acid decomposition. The XRD spectra of spent catalysts are shown in Fig. S5. In the case of FSR spent catalyst, the high-intensity peak positioned at 2 \(\theta =\) 27.10 \(^\circ\) and other smaller peaks are attributed to the formation of bulk sulfate species Fe2(SO4)3(H2O)5 (ICDD code: 00–027-0245) [64]. The crystallite size of the peak at 2 \(\theta =\) 27.10 \(^\circ\) [110] calculated using Scherrer formula is found to be 21.6 nm which can be ascribed to the formation of metal sulfate species at high temperature (1123 K). On the other hand, in the case of FSCS spent, the crystallite size of the peak at 2 \(\theta =\) 35.6 \(^\circ\) [110] calculated crystallite size using Scherrer formula is found to be 27.8 nm. The change in the diffraction pattern of the FSCS spent catalyst is observed in comparison to the pristine FSCS. This change in pattern could be attributed to the formation of new phases of iron oxide at elevated temperatures (1123 K). The probable reason behind this could be high iron oxide loading which directly or explicitly leads to the formation of a thermally stable spinel form of iron oxide (Fe3O4, ICDD code: 01–088–0315) along with the substantial presence of parent hematite oxide at high temperature. The evidence of metal loss and broken shells in core–shell type structures are extensively recorded in the literature [42, 48]. Based on the TEM images, it has been seen that after 100 runs, SiO2 shells were broken when Pt nanoparticles were employed as a core in the SiO2 shell and metal loss up to 9% was observed [48]. To determine the metal content in the pristine and spent catalysts, the ICP-OES (Inductively coupled plasma-optical emission spectroscopy was employed. The metal loss in FSR and FSCS after a 100 h run was 16.8 and 3.8%. There is no support sulfation in FSCS, and no evidence of metal sulfate species is observed in the XRD pattern. It may be possible that sulfate species are present but, in less amount, which is beyond the limit of XRD or bonded weakly. It is expected that the high catalytic activity of FSCS could be due to the formation of a thermally stable spinel form of iron oxide (Fe3O4).

In order to gain the knowledge of the effect of 100 h time-on-stream run on textural properties of catalysts, N2 adsorption, and desorption analysis is carried out as shown in Fig. S6. The BET surface areas of the FSR and FSCS are decreased significantly after 100 h time on stream run as shown in Table 2 and Fig. S6. The reduction in surface area and pore volume could be attributed to the pore filling and pore blockage due to the bulk formation of metal sulfate species (Fe2(SO4)3) in the FSR spent catalyst.

To further understand the formation of metal sulfate species in both spent catalysts FSR and FSCS, FTIR spectra of spent catalysts are recorded as shown in Fig. 5a. In the case of FSR, a low-intensity peak appeared in the range of 1200–1000 cm−1 shouldering at 1070 cm−1 is attributed to the SO bond stretching in metal sulfate, suggesting the formation of metastable sulfate species (Fe2(SO4)3) at the surface of iron oxide. However, a similar peak is missing in FSCS, indicating the no formation of metal sulfate species which is in agreement with the results obtained from spent catalyst XRD spectra. The findings of this work provide new insights into the development of a highly active and stable catalyst. The confinement of iron oxide in silica shells can effectively avoid iron oxide aggregation and metal loss. However, the possible cause of the reduction in BET surface area could be due to the deactivation of the catalysts by sintering at high temperatures. The sintering initiation temperature can be expressed as

$$T\, = \,T_{m} \left( {1\, - \,\frac{2\omega \Delta \alpha }{{R\theta_{m} }}} \right)$$
(8)
Fig. 5
figure 5

a FTIR spectra of FSR and FSCS spent catalyst after 100 h time-on-stream run. b TEM image of FSCS spent catalyst

Full size image

Here T is the onset of agglomeration or sintering; Tm is the melting point of the bulk sample; ω is the volume per particle; Δα is the change in the surface energy across the interface; R is the radius of the particle; θm is the heat of melting. From the above equation, it can be deduced that if the melting initiate, it will fill the space between bigger particles solid–liquid agglomeration begins which is the possible case of reduction in surface area after 100 h run. In the case of SiO2, the onset of sintering temperature is ~ 1049 K. However, their structures are remained intact, and no evidence of shell broken, and collapse is found in TEM images of spent catalyst (Fig. 5b). Hence, silica-supported catalysts should be operated below this temperature and are more susceptible to sintering at elevated temperatures. It indicates that the nanoparticles embedded in core–shell structures can effectively avoid active metal loss. The spent catalyst characterization also confirms that the active phases are intact.

Conclusion

In this work, iron oxide nanoparticles supported over the porous structure of silica (FSW, FSA, and FSR) as well as nanoparticle encapsulated inside mesoporous silica (core–shell, FSCS) were successfully synthesized in the laboratory through wet impregnation and sol–gel methods to circumvent the problem of active phase agglomeration and active metal loss by sintering. SAXS and FTIR analysis confirmed that the synthesized iron oxide core and mesoporous silica shell in FSCS. FESEM and HRTEM analysis showed well-dispersed iron oxide nanoparticles encapsulated within silica shells. HRTEM further confirms the iron oxide nanoparticle average particle size of 33.5 nm and silica shell is of 103.5 nm. Among all the laboratory synthesized catalysts, this FSCS catalyst was found to be the most active and stable (during a 100 h long time-on-stream run) catalyst in this high-temperature, endothermic sulfuric acid decomposition reaction. All the spent catalysts were characterized using XRD, BET, FTIR to identify the changes related to their long-term behaviors. The FSR and FSCS catalyst activities were decreased significantly which is reflected in the reduction in their active surface area and pore volume due to pore blockage by bulk formation of metal sulfate species [Fe2(SO4)3] as seen as other smaller peaks in the spent catalyst XRD. This is further confirmed by the spent catalyst FTIR, wherein, a low-intensity peak appeared in the range of 1200–1000 cm−1 shouldering at 1070 cm−1 attributed to the SO bond stretching in metal sulfate, suggesting the formation of metastable sulfate species (Fe2(SO4)3) at the surface of iron oxide. However, in the case of iron oxide core and mesoporous silica shell, FSCS catalyst which showed high catalytic activity even after 50 h and remained active more than 100 h with ~ 82% conversion at 1123 K, XRD confirmed that the active phases are intact and there is no indication of any agglomeration and sintering in the spent catalyst. Further, this work emphasized the importance of core–shell structure based catalysts for high-temperature corrosive reactions. These iron oxide nanoparticles immobilized inside a mesoporous silica shell could be a substitute for Pt-based catalysts and can lead to revisiting various other high-temperature reactions where it is necessary to reduce cost and enhance overall reaction performance for industrial applications. These mesoporous silica core–shell catalysts have great potential to circumvent the problem of active phase agglomeration and metal loss and can be used at an industrial scale.