Effect of TiO₂ morphology on the benzyl alcohol oxidation activity of Fe₂O₃–TiO₂ nanomaterials

Abstract

Three series of Fe loaded TiO₂ anatase (1, 3, 5 and 7 mol% Fe) nanomaterials with different morphologies: nanoparticles (NP), nanotubes (NT) and nanofibers (NF) were synthesized and calcined at 400 °C. The physico-chemical properties of the catalysts were studied by using elemental analysis, XRD, UV-vis, N₂-physisorption, SEM, TEM, XPS, pyridine adsorption using FTIR and H₂-TPR techniques. It was observed that iron oxide was highly dispersed on the TiO₂-NP support due to its strong interaction. The catalytic activity of the catalysts was tested in the oxidation of benzyl alcohol with hydrogen peroxide at mild reaction temperatures (70–100 °C) and atmospheric pressure. The highest activity was observed with 3 mol% Fe supported on TiO₂-NP at 100 °C. It seems that TiO₂-NP is unique in stabilizing small Fe₂O₃ nanoparticles. A greater number of surface hydroxyl groups in TiO₂-NT and NF tend to increase the density of adsorption sites and/or the affinity of the surfaces with the Fe₂O₃ precursor. This appears to favor agglomeration, which results in a higher density of larger iron oxide particles. Fe-TiO₂-NP catalysts display high activities due to a detrimental morphology effect, high surface area of the TiO₂ support, dispersion of Fe₂O₃, more Lewis acid sites and easy reducibility of total catalysts. It was also observed that the Fe-TiO₂ nanomaterials possessed two types of Fe³⁺ ions on the support surface; one dispersed in which Fe³⁺ ions interact with the TiO₂ surface and another in which Fe₂O₃ crystals are located on the surface of the catalyst. The catalyst which possessed the former species exhibited the best performance in the oxidation of benzyl alcohol. Metal oxide leaching studies prove the true heterogeneous nature of the reaction. The catalysts are found to be reusable and resistant to rapid deactivation.

---

1. Introduction

Selective oxidation of alcohols to more valuable aldehydes, ketones, and carboxylic acids is of great importance in both laboratory synthesis and industrial manufacturing.¹ Liquid phase oxidation of benzyl alcohol is a well-studied topic in modern organic synthesis. The design and development of a catalyst with high conversion and selectivity for partial oxidation must also be carried out with regard to the preservation of oil related resources.² For selective oxidation reactions, there is a tremendous challenge to prevent over oxidation of the products, which are often more sensitive to be oxidized than the reactants.³ The direct oxidation of benzyl alcohol to benzaldehyde is such a type of reaction. Benzaldehyde is an important raw material in synthesis of pharmaceuticals, plastic additives, perfumes, flavoring compounds and also in the preparation of certain aniline dyes.⁴ Several methods are available for alcohol oxidations using metal salts in the form of homogeneous catalysis⁵ or supported metal ions as heterogeneous catalysts.⁶ However the common methods of alcohol oxidation generally use toxic, corrosive, expensive oxidants such as Cr(VI), and setting up a severe condition, like high pressure or temperature, using strong mineral acids.

With the increase of environmental concerns, selective oxidation of alcohols to the corresponding aldehydes with environmentally benign oxidants has attracted great attention.⁷ Many studies have been reported on oxidation of benzyl alcohol to benzaldehyde with molecular oxygen where addition of supports and promoters, visible light irradiation, and process in combination with different oxidants have been used to improve the catalytic systems.⁸ A significant number of studies on catalytic oxidation using precious metal or metal-based compounds as catalysts, such as Au,⁹ Pd,¹⁰ Ru,¹¹ and Pt¹² have been reported for selective oxidation of benzyl alcohol. These catalysts suffer from high cost and limited availability and also structural stability, heterogeneity, deactivation rates and recyclability of the noble metal catalysts are still critical impending to the high catalytic performance in liquid phase reactions.¹³ Therefore, it is highly desirable to investigate and develop inexpensive catalysts consisting of ‘3d’ transition metals, because these metals offer environmentally benign and cost-effective alternative to noble metal catalysts.

Titanium dioxide (TiO₂) was utilized as a support for heterogeneous catalyst due to the effect of its high surface area stabilizing the catalysts in its mesoporous structure.¹⁴ TiO₂ supported metal catalysts have attracted interest due to its high activity for various reduction and oxidation reactions at low pressures and temperatures.¹⁵ Furthermore, TiO₂ was found to be a good support due to the strong metal–support interaction, chemical stability, and acid–base property.¹⁶ The aforementioned properties could make transition metal oxide supported TiO₂ catalysts very effective in selective oxidation of benzyl alcohol. Previously, researchers observed a strong relationship between the specific physical and chemical properties of nanostructured TiO₂ particles with its phase composition, specific surface area, pore size distribution, particle morphology and surface defects. Panagiotopoulou and Kondarides¹⁷ studied the effect of morphological characteristics of TiO₂-supported noble metal catalysts on their activity for the water–gas shift reaction. They observed that catalytic activity depends strongly on the structure and morphology of the support that TiO₂-supported noble metals with proper structural and morphological characteristics (high metal dispersion, small TiO₂ crystallite size) could be considered as promising candidates for the low temperature water–gas shift reaction.

It was also reported that well-defined anatase TiO₂ particles with exposed (001) crystal facets can enable a range of catalytic applications.¹⁸ Hence, it is necessary to develop synthetic strategies, in which size and morphology of materials can be precisely controlled with designed functionalities. A variety of methods were developed to prepare TiO₂ nanostructured materials, such as solvothermal,¹⁹ precipitation,²⁰ sol–gel²¹ and thermal decomposition of alkoxides.²² The properties of the TiO₂, such as crystal phase, crystallinity, surface area, porosity, and morphology were varied, when TiO₂ was synthesized using different methods.

The purpose of the work presented here is to investigate the influence of the morphology of nanostructured TiO₂ anatase sample. Three series of iron oxide loaded TiO₂ anatase (1, 3, 5 and 7 mol% Fe) nanomaterials with different morphology; nanoparticles (NP), nanotubes (NT) and nanofibers (NF) were synthesized, in the scope to obtain a representative panel of samples with various morphologies. The synthesized Fe-TiO₂ nanomaterials were characterized and evaluated for the benzyl alcohol oxidation activity. The synthesized materials have shown different behaviors depending on both their composition and morphology. The characteristics of the TiO₂ nanomaterials claimed to account for the observed different activities, with more or less pronounced synergetic effects, due to the use of different Fe loading and the specific morphology of the TiO₂ support.

2. Experimental

2.1. Materials

All reagents were analytical grade and used as received without any purification. Anatase TiO₂ [P25, Degussa AG] was purchased from Germany. Titanium isopropoxide [C₁₂H₂₈O₄Ti], ferric nitrate nonahydrate [(FeNO₃)₃·9H₂O], ethyl alcohol [C₂H₅OH], benzyl alcohol [C₆H₅CH₂OH], hydrogen peroxide [H₂O₂], sodium hydroxide [NaOH] solution and tetrapropyl ammonium hydroxide [(CH₃CH₂CH₂)₄N(OH)] solution were purchased from Aldrich, U.K.

2.2. Synthesis of TiO₂ nanomaterials

2.2.1 TiO₂ nanoparticles. TiO₂ nanoparticles have been synthesized by using modified sol–gel method. A calculated amount of titanium isopropoxide solution was added to 150 ml of ethanol under mechanical stirring and kept at ambient conditions for one hour. A colorless gel was obtained by drop wise addition of tetrapropyl ammonium hydroxide solution (pH = 10) under constant stirring. The colorless gel was transformed into a white colored precipitate after removing the excess solvent by heating at 80 °C, and then the white precipitate was washed three times with ethanol and dried in air at 100 °C for 12 h and finally calcined in air at 400 °C for 5 h. The obtained material was denoted as TiO₂-NP.

2.2.2 TiO₂ nanotubes. The preparation of TiO₂ nanotubes was based on the alkaline hydrothermal method proposed by Kasuga et al.;²³ typically, 6 g of Degussa TiO₂ powder was dispersed in 120 ml of 10 M NaOH solution. The resulting suspension was stirred for 30 min, and transferred into a Teflon-lined stainless steel autoclave, sealed, and maintained at 130 °C for 48 h. The resulting material was filtered, neutralized with 0.1 M HCl solution and washed with deionized water for five times to remove the NaCl formed. The material was dried in air at 100 °C for 12 h and finally calcined in air at 400 °C for 5 h. The obtained material was denoted as TiO₂-NT.

2.2.3 TiO₂ nanofibers. TiO₂ anatase nanofibers (NnF CERAM-TiO₂) were purchased from Pardam nanotechnology, Czech Republic. The purchased material was used as received without any further modification. The material was denoted as TiO₂-NF.

2.2.4 Fe₂O₃ supported TiO₂ nanomaterials. Iron oxide supported TiO₂ nanomaterials were prepared by soaking the calculated amount of calcined TiO₂ powder in solution of Fe(NO₃)₃·9H₂O that corresponded to 1.0 mol% to 7.0 mol% of iron oxide. The excess water was removed by slow drying and the dried materials were washed four times with distilled water. Portions of the synthesized materials were calcined at 400 °C for 4 h. The calcined materials were annotated using the following nomenclature; xFeTi-NP, xFeTi-NT, and xFeTi-NF (where x = 1, 3, 5 and 7 mol% of iron oxide) and NP is for nanoparticles, NT is for nanotubes and NF is for nanofibers samples respectively.

2.3. Characterization of synthesized materials

Chemical analysis of the materials was performed by using a Perkin-Elmer ICP-AES instrument. X-ray powder diffraction (XRD) studies were performed for all of the prepared solid samples using a Bruker diffractometer (Bruker D8 advance target). The patterns were run with copper Kα₁ and a monochromator (λ = 1.5405 Å) at 40 kV and 40 mA. The crystallite size of the ZrO₂ phase was calculated using Scherrer's eqn (1):

D = Bλ/β_1/2 cos θ (1)

where D is the average crystallite size of the phase under investigation, B is the Scherer constant (0.89), λ is wavelength of the X-ray beam used (1.54056 Å), β_1/2 is the full width at half maximum (FWHM) of the diffraction peak and θ is the diffraction angle. The identification of different crystalline phases in the samples was performed by comparing the data with the Joint Committee for Powder Diffraction Standards (JCPDS) files. SEM measurements were carried out using a JEOL JSM840A system. For SEM, each powder was attached to an aluminum block using double sided carbon tape. The samples were then coated in gold to make them conductive and compatible with the SEM technique. A Philips CM200FEG microscope, 200 kV, equipped with a field emission gun was used for TEM and HRTEM analyses. The coefficient of spherical aberration was Cs = 1.35 mm. The information limit was better than 0.18 nm. High-resolution images with a pixel size of 0.044 nm were taken with a CCD camera. The textural properties of the prepared samples were determined from nitrogen adsorption/desorption isotherm measurements at −196 °C using Autosorb ASiQ automated gas sorption system (Quantachrome, USA). Prior to measurement, each sample was degassed for 6 h at 150 °C. The specific surface area, S_BET, was calculated by applying the Brunauer–Emmett–Teller (BET) equation. The average pore radius was estimated from the relation 2V_p/S_BET, where V_p is the total pore volume (at P/P⁰ = 0.975). Pore size distribution over the mesopore range was generated by the Barrett–Joyner–Halenda (BJH) analysis of the desorption branches, and the values for the average pore size were calculated. The XPS measurements were carried out using Thermo Scientific Escalab 250 Xi XPS instrument both in survey and narrow scan modes using AlKα X-rays having a spot size of 650 μm. A flood gun was used in standard mode to compensate surface charge. Peak shift due to charge compensation was corrected using the binding energy of C1s peak. The samples were attached to a standard XPS sample carrier block using 10 mm diameter carbon stubs in the form of 0.5 mm thick and 8 mm diameter pellets prepared using handheld pelletizer. The data was acquired using pass energy of 100 eV, dwell time 200 ms with step size of 0.1 eV and 10–30 scans.

FTIR spectra of calcined catalysts were obtained at room temperature using a Perkin-Elmer Spectrum 100 FTIR spectrometer. Samples were then subjected to a pyridine adsorption analysis. The analysis was carried out over a catalyst disk which was treated under vacuum for 5 h. Later, the sample was treated with pyridine vapor and finally heated at 100 °C under vacuum for 30 min. FTIR spectra were collected at room temperature. The amount of Bronsted and Lewis acid sites was calculated via integration of the area of the absorption bands showing the maximum values of intensity at 1446 and 1536 cm⁻¹, respectively. Integrated absorbance of each band was obtained using the appropriate software by applying the corresponding extinction coefficient and normalized by the weight of the samples.

Temperature programmed reduction experiments were performed using a CHEMBET 3000 (Quantachrome USA) instrument. Calculated amount of catalyst was initially treated with 10% O₂–He at 300 °C for 30 min, before the sample tube was purged with helium gas. The sample temperature was then reduced to room temperature (25 °C) by passing the air into the furnace. Next, the flow of the gas changed to 5% H₂–He and the sample temperature increased to 600 °C at the rate of 5 °C min⁻¹. The H₂-TPR patterns were obtained by recording the TCD signal with respect to the time and temperature.

2.4. Liquid phase oxidation of benzyl alcohol (BzOH)

The oxidation of BzOH was carried out in a magnetically stirred round bottom flask (50 ml) by adding benzyl alcohol (5 ml; 50 mmol) to the flask containing dry catalyst (0.05 g; 0.04 mol%). Next, 30 wt% hydrogen peroxide (12 ml; 60 mmol) was added to the flask. The temperature of the flask was controlled using a thermocouple located in the oil bath. The samples were taken periodically (with a syringe) and centrifuged to separate the catalyst from the mixture. The products were analyzed using a gas chromatography, equipped with flame ionization detector (FID, SE-30 capillary column) and identified by comparison with known standards. External calibration method was used for quantitative analysis of the concentrations of reactants and products generated. The selectivity to the desired product benzaldehyde was expressed as the amount of particular product divided by amount of total products and multiplied by 100.

3. Results and discussion

Powder XRD measurements were performed to identify the crystalline phases presented in the synthesized samples. The XRD patterns of the pure TiO₂ and iron oxide supported TiO₂ (1, 3, 5 and 7 mol% Fe₂O₃) nanoparticles (NP), nanotubes (NT) and nanofibers (NF) are shown in Fig. 1(a)–(c) respectively.

Fig. 1 XRD patterns of pure and iron oxide supported TiO₂ nanomaterials (a) NP (b) NT and (c) NF.

All the TiO₂ nanomaterial samples exhibited reflections due to (101), (103), (004), (200), (105), (211), (204), (116), (220) and (215) planes for the anatase phase, revealing that the TiO₂ NPs, NTs and NFs are well-crystallized in the anatase phase (JPCDS: 21-1272).²⁴ The absence of major reflections at 2θ = 27° and 31° indicating that all the TiO₂ nanomaterial samples are free from rutile and brookite phases. It is clear that reflections due to iron oxide were not observed in the XRD patterns of the 1 mol% iron oxide loaded TiO₂ nanomaterial samples (NP, NT and NF); this is probably due to low concentration of iron oxide. Presence of two major reflections at 2θ = 33.3°, 36.0° corresponding to (104), (110) planes of hematite (α-Fe₂O₃) in the XRD patterns of samples in which the iron oxide content increased above 1 mol%, These results indicate that the formation of crystalline α-Fe₂O₃ particles occurred in 3, 5 and 7 mol% iron oxide loaded TiO₂-NP, NT and NF samples.

The reflections due to anatase phase are relatively strong and narrow in case of NT samples than that of NP and NF samples, indicating that the anatase crystallinity and average crystallite size is high in case of NT than that of NF and NP samples. The average crystallite sizes of all the TiO₂ nanomaterials were calculated from full-width at the half-maximum of (101) diffraction peak using the Scherrer's equation. TiO₂-NT sample possessed high crystallite size compared to TiO₂-NP and TiO₂-NF (Table 1), which may be because of different titanium sources, hydrolyzing agents and preparation conditions used to synthesize TiO₂-NP, NT and NF samples. The effect of iron oxide loading on the crystallite size of anatase can be observed. The calculated values indicating that the crystallite size of iron oxide supported TiO₂ is slightly larger than those of unsupported TiO₂ nanomaterials. Obviously, the impregnation of iron oxide resulted an increase of crystallite size. Wang et al.²⁵ also observed a very similar result that iron doping effected the crystallite sizes of TiO₂ support.

Table 1 Elemental composition and textural properties of the samples

Catalyst	Surface area (m^2 g^−1)	Pore volume (cm^3 g^−1)	Pore radius (Å)	Elemental bulk composition (ICP)				Elemental surface composition (XPS)
				Ti	O	Fe	Fe/Ti	Ti	O	Fe	Fe/Ti
TiO2-NP	70.0	0.077	37.0	57.0	43.0	—	—	57.0	43.0	—	—
1FeTi-NP	82.0	0.084	38.1	53.9	44.3	1.8	0.03	54.3	44.6	1.1	0.02
3FeTi-NP	125.0	0.119	39.2	52.3	43.2	4.5	0.08	52.6	43.4	4.0	0.07
5FeTi-NP	70.0	0.101	21.0	51.4	43.0	5.6	0.10	51.5	43.3	5.2	0.10
7FeTi-NP	65.7	0.089	21.0	50.5	40.5	9.0	0.17	51.0	41.0	8.0	0.15
TiO2-NT	53.3	0.376	195.2	56.6	43.4	—	—	56.6	43.4	—	—
1FeTi-NT	54.4	0.379	194.8	55.0	43.1	1.9	0.03	55.4	43.4	1.2	0.02
3FeTi-NT	57.5	0.407	195.5	52.2	43.0	4.8	0.09	52.4	43.4	4.2	0.08
5FeTi-NT	58.9	0.401	194.8	51.5	42.6	5.9	0.11	52.0	43.0	5.0	0.10
7FeTi-NT	53.4	0.384	194.5	49.5	40.7	9.8	0.19	50.0	40.8	9.2	0.18
TiO2-NF	45.2	0.062	17.0	57.2	42.8	—	—	57.2	42.8	—	—
1FeTi-NF	47.6	0.062	17.0	56.5	41.6	1.9	0.03	56.7	41.6	1.7	0.03
3FeTi-NF	47.4	0.100	17.0	54.5	41.0	4.5	0.08	54.5	41.3	4.2	0.08
5FeTi-NF	57.9	0.140	17.0	53.7	40.4	5.9	0.10	53.9	40.6	5.5	0.10
7FeTi-NF	37.4	0.120	17.0	51.0	39.7	9.5	0.18	51.4	39.7	9.1	0.18

---

Ballyy et al.²⁶ observed that 0.32 wt% Fe doping resulted the transformation of TiO₂ from anatase to rutile phase. In another study, Ranjit and Viswanathan²⁷ also noticed the formation of rutile and pseudobrookite TiO₂ phases upon doping of 0.11 to 1.76 wt% Fe in TiO₂. In another study, Zhu et al.²⁸ observed a shift in peak position of (101) plane of anatase phase upon doping of Fe into TiO₂. This was due to the distortion of TiO₂ crystal lattice by Fe dopant. Yue and Zhang²⁹ observed a change in the unit cell volume in case of 5 wt% Fe-doped TiO₂ compared with that of pure TiO₂ indicates that Fe³⁺ replaces Ti⁴⁺ in the TiO₂ lattice to form solid solution. In the present study, the phase structure and lattice parameters of Fe-TiO₂ nanomaterials were not changed even after impregnation of 7 mol% of iron oxide. The anatase structure is preferred over other polymorphs for many catalytic and photocatalytic applications because of its higher electron mobility, low dielectric constant and lower density.

FTIR spectroscopy is a sensitive technique which could be used to verify the main short range orders of the nanostructured TiO₂ materials. The FTIR spectra of the bare and iron oxide supported (1, 3, 5, and 7 mol%) TiO₂-NP, NT and NF samples are presented in Fig. S1 (ESI†). The FTIR spectra appears to be same for all the pure and iron oxide supported NP, NT and NF samples, however some changes could be observed in the spectra. The broad peaks observed between 3370 and 3435 cm⁻¹ in case of NT and NF samples can be assigned to stretching vibration of surface hydroxyl groups and adsorbed water.³⁰ This broad peak was very weak in case of TiO₂-NP samples. On other hand, presence of peaks at 1635 cm⁻¹ and 1410 cm⁻¹ which are due to the bending vibrations of Ti–OH and Ti–O–Ti functional groups could be observed in all TiO₂-NP, NT and NF samples. It is interesting to note that the intensity of these peaks are gradually decreased with increase of iron oxide loading over TiO_2-NP, NT and NF samples.

A broad envelope in the range of 400–800 cm⁻¹ was observed in TiO₂-NP, NT and NF samples. It was reported that the band at 467 cm⁻¹ can be assigned to bending vibration of Ti–O bond.³¹ Supporting a small amount of iron oxide (1 mol%) did not resulted any change in the IR spectra of bare TiO₂ nanomaterials. The bands were appeared due to presence of iron oxide in the range of 590–510 cm⁻¹; these observed bands could be assigned to the vibrations of FeO₆ structural units.³² The intensity of these bands increased with increase of iron oxide loading. The decrease of intensity of the peaks due to Ti–OH and Ti–O–Ti functional groups and increase of intensity of peaks due to FeO₆ functional groups indicating that there might be an interaction existed between Fe₂O₃ and TiO₂ nanomaterial (NP, NT and NF). However, the extent of interaction might be different for TiO₂ nanomaterial samples with different morphology.

The FE-SEM images for representative samples are showed in Fig. 2. It can be seen that pure TiO₂-NF sample consists of linear structured TiO₂ NFs with several micro meters length (≅3 μm) and diameters of 8–10 nm. However, the length of the fibers decreased from micro meters to 500 nm after iron oxide loading. Clearly, the TiO₂ NFs have changed into shorter NFs which were intertwined together showed macro size spaces. The impregnation of iron oxide precursor and calcination resulted the formation of iron oxide nanoparticles, some of were attached to the NFs; thus, it is reasonable to argue that the impregnation of iron oxide led to morphology change of TiO₂ NFs. The precursor (Degussa P25 TiO₂ anatase) used to synthesize TiO₂-NT consisted of spherical particles (Fig. S2, ESI†), were entirely transformed into tubes. The nanotubes have average diameter size 50 nm around and average length around 800 nm.

Fig. 2 FE-SEM images of pure and iron oxide supported TiO₂-NP, NT and NF samples.

It is known that during the treatment with NaOH solution, some Ti–O bonds of TiO₂ precursor powder breaks, leading to formation sodium titanate nanosheets and subsequent rolling into nanotubes during the hydrothermal treatment.³³ After impregnation of different iron oxide loadings (1 to 7 mol% of iron oxide), the xFeTi-NT samples showed very similar morphological characteristics. These observations indicating that supporting iron oxide had little effect on tube-like structure. The morphology of pure TiO₂-NP was found in irregular spherical particles with diameters ranging from 10–30 nm as seen in Fig. 2. The FE-SEM micrographs of iron oxide supported TiO₂-NP samples showed very similar morphology as bare TiO₂-NP powder. These samples consisted some portions of relatively larger (500 nm) agglomerates of powder particles. The primary particles of the 1 and 3 mol% iron oxide TiO₂-NP powders possessed relatively small size (40 nm) than those estimated for the 5 and 7 mol% iron oxide supported TiO₂-NP samples. The latter two samples consisted of 0.5–1 μm sized agglomerated particles of about 200 to 500 nm size. The FE-SEM micrographs of the samples also indicating that the degree of agglomeration was gradually increased with the increase of iron oxide loading over TiO₂-NP support.

It is known that TEM analysis can provide valuable information regarding the internal structure and accurate measurement of particle size and morphology. The TEM images of representative iron oxide supported TiO₂ nanomaterials were shown in Fig. 3. The TEM images of both 1FeTi-NP and 7FeTi-NP samples revealed an almost spherical shape with average particle size in the range of 10–30 nm. It is clear that the surface of the 7FeTi-NP sample has dense packed particles composed of the aggregation of nanoparticles of Fe₂O₃–TiO₂ composite. The TEM image results are consistent with the crystallite size calculated by the Scherrer's equation from the XRD measurements. The pure TiO₂ nanofibers with average diameter about 200 nm have a relatively smooth surface. The 7 mol% iron oxide impregnated TiO₂-NF sample showed nanofibers along with rod like nanostructures and have diameters of about 50 nm. From TEM image in Fig. 3, it is clear that the secondary Fe₂O₃ nanoparticles grow on the surface of TiO₂ NFs.

Fig. 3 TEM images of representative iron oxide supported TiO₂-NP, NT and NF samples.

TEM images of synthesized iron oxide supported TiO₂-NT clearly showed a typical morphology of the TiO₂-NTs and Fe₂O₃ nanoparticles (spherical black spots) after impregnation of iron oxide and calcination at 500 °C for 4 h. The anatase TiO₂-NTs are uniform, and the diameter size is around 5–8 nm with 200–400 nm in length. High or low iron oxide loading, there are no clear differences in the physical appearance of the TiO₂-NT and Fe₂O₃ nanoparticles, and only the compositions changed which was proved by the XRD analysis.

In the HRTEM image of the 3FeTi-NP catalyst, hetero-junction region was observed with two sets of lattice fringe spacing of 0.352 and 0.254 nm which are consistent with the (101) plane of the anatase crystal structure and the planes of the cubic hematite Fe₂O₃, respectively.³⁴ Furthermore, energy-dispersive X-ray spectroscopy (EDS) characterization also confirms that peaks include Fe, Ti, and O, indicating the presence of both Fe₂O₃ and TiO₂ (Fig. S3†). The EDS line scanning along the cross section of peaks (Fig. S4†) further shows that Fe is present only outside of the TiO₂-NPs, NTs, and NFs. As a result, the synthesized Fe₂O₃ supported TiO₂ nanomaterials can be considered as the Fe₂O₃–TiO₂ composite nanostructures (Fig. 4).

Fig. 4 HRTEM images of 3FeTi-NP catalyst.

Fig. 5 displays N₂ adsorption–desorption isotherms of pure and iron oxide supported TiO₂ (a) NP (b) NT and (c) NF samples. All the isotherms of TiO₂-NP samples reveal the stepwise adsorption and desorption branch of type IV curves, indicating the presence of mesoporous pores according to the IUPAC classification.³⁵ These isotherms can be categorized as type IV with small hysteresis loops observed at a relative pressure of 0.45–0.85 for all the samples. However, shape of the hysteresis loop changed after iron oxide impregnated over TiO₂-NP support and the areas of the hysteresis loops were decreased with increase of iron oxide loading. These observations indicating that the synthesized pure and iron oxide supported TiO₂-NP has mesoporous structure and iron oxide loading resulting the partial filling of mesopores of TiO₂-NPs.

Fig. 5 N₂ adsorption–desorption isotherms of pure and iron oxide supported TiO₂ (a) NP (b) NT and (c) NF samples.

All of the TiO₂-NT samples possessed type IV isotherms, which are representative of mesoporous materials according to the literature report.³⁶ The type IV isotherm has a hysteresis loop due to the capillary condensation of the N₂ gas in the pore. The bare TiO₂-NT sample displayed H1 hysteresis loop, indicating that the pore size distribution is broad in the mesoporous sample. However, after the impregnation of iron oxide, the hysteresis loops were changed to type H2, indicating that the samples have pores with narrow necks and wider bodies (ink-bottle pores). It is interesting to note that the areas of these hysteresis loops were almost same and the isotherms of the samples were not shifted downward, indicating that the BET surface have not changed significantly. A slight decrease of surface area and pore volume of iron oxide supported TiO₂-NT samples confirm that the iron oxide was deposited outside of the TiO₂-NTs.

The N₂ adsorption–desorption isotherms of TiO₂-NF samples [Fig. 5(c)] can be characterized as type III. The isotherms demonstrate a weak adsorption interaction between N₂ molecules and the TiO₂ NFs, suggesting that very small amount of micropores or mesopores existed in the TiO₂-NF sample. However, there is a sharp increase with adsorption volume in the relative pressure range of 0.89–0.99, indicating the presence of macropores. This may be due to the existence of the hollow structures or interparticulate spaces. To analyze the pore size and pore volume; pore size distribution patterns were plotted from desorption branch using BJH method (Fig. S5†).

The surface area, pore volume and average pore size determined from the N₂-physisorption were presented in Table 1. The pore volume of TiO₂-NP, 3FeTi-NP and 7FeTi-NP was found to be 0.077, 0.119 and 0.054 cm³ g⁻¹ respectively. The bare TiO₂-NT sample possessed the pore volume of 0.376 cm³ g⁻¹, with impregnation of Fe loading, a slight increase of pore volume was observed and the maximum pore volume in case of 3FeTi-NT sample (0.407 cm³ g⁻¹). It is clear that beyond the monolayer coverage, formation of crystalline particles of Fe₂O₃ was occurred and resulted the decrease of pore volume. A very similar behavior was observed in case of FeTi-NF samples.

The average pore radius of the FeTi-NP samples ranged from 21 Å to 39.2 Å. The bare TiO₂-NP sample showed broad pore size distribution (PSD) peak due to unimodal meso pores, while a sharp PSD peak at 21 Å was started to appear after impregnation of 1 mol% iron oxide (Fig. S5†). With increase of iron oxide loading to 5 and 7 mol%, the broad PSD peak was disappeared and only the sharp PSD peak was remained, which clearly indicating partial pore blocking due to deposition of iron oxide. The average pore radius of bare TiO₂-NT is 195 Å with broad pore size distribution and it has not changed considerably after iron oxide loading (for FeTi-NT samples). The pore size distribution for FeTi-NF samples exhibits a narrow range, with a peak pore radius of 17 Å. When the Fe content increased, the maximum pore sizes shift into higher mesoporous regions, indicating an increase in the inter-particle pore size.

The BET surface area for all the three series of samples was calculated from N₂ physisorption measurements. BET specific surface area of pure TiO₂-NP and 3 mol% Fe-impregnated TiO₂ was 70 and 125 m² g⁻¹, respectively. The BET results showed that impregnation of TiO₂ with Fe results in an increase of specific surface area. Therefore, it is possible that the changes in particle size and specific surface area of Fe-doped TiO₂ NPs could be due to the interaction between Fe₂O₃ and TiO₂. The TiO₂-NT and TiO₂-NF powders displayed a BET specific surface area of 53 and 45 m² g⁻¹ respectively. The BET surface area of the FeTi-NT and FeTi-NF catalysts was also increased with the increase of iron oxide loading. The highest surface area was observed for 5FeTi-NP and 5FeTi-NF samples. This is possibly due to the mesopore structure and good dispersion of iron oxide. The BET specific surface areas were decreased upon further increase of iron oxide loading, which may be caused by the aggregation of iron oxide NPs; this effect leads to the decrease in the specific surface area.

To determine the oxidation states of the metal ions, surface chemical composition and also understand the extent of interaction in the iron oxide supported TiO₂ nanostructures, we have carried out the XPS measurements. The core level Ti2p, Fe2p, O1s XPS spectra for FeTi-NP, FeTi-NT and FeTi-NF are shown in Fig. S6.† Two distinct peaks at 458.2 and 464.2 eV which can be assigned to Ti2p_3/2 and 2p_1/2 respectively³⁷ were observed in the Ti2p XPS core level spectra for three series of samples. The splitting energy between the Ti2p_3/2 and Ti2p_1/2 core levels is approximately 6.0 eV, which suggests the existence of a normal state of Ti⁴⁺. It can be seen that two peaks are highly symmetrical and no shoulders are observed, indicating that the presence of stoichiometric TiO₂. The binding energies of Ti2p_1/2 and Ti2p_3/2 peaks in FeTi-NP, NT, NF samples are shifted slightly towards the higher values, this slight shift in the binding energy of supported samples is due to the interaction of Fe ions with TiO₂.³⁸ The impregnation of Fe precursor on dry TiO₂ support, there is a possibility of interaction of Fe ions on the surface of the TiO₂ rather than incorporation into the TiO₂ crystal lattice. The Ti2p spectrum of 7FeTi-NP, NT, NF samples were fitted with the Gaussian function (Fig. 6) reveals a dominant Ti2p_3/2 peak with a binding energy at 457.4 eV, which is characteristic of a Ti⁴⁺ state in a TiO₂ lattice³⁹ along with two shoulders. The first shoulder at 456.4 eV corresponds to a Ti³⁺ state, due to an oxygen deficiency in TiO₂,⁴⁰ while the second shoulder at 459 eV arises from a Ti⁴⁺ state in the Ti–O–Fe species.³⁹ It is clear from the Ti2p spectra that the contribution due to oxygen deficient TiO₂ species is higher in case of TiO₂-NP than TiO₂-NT and TiO₂-NF samples.

Fig. 6 The deconvoluted XPS Ti2p, Fe2p and O1s spectra of 7FeTi-NP, NF and NT samples.

The core level Fe2p XPS spectra of FeTi-NP, NT and NF samples showed 2p_3/2 and 2p_1/2 peaks at 710.5 eV and 724.2 eV, which could be attributed to the Fe³⁺ species. The Fe2p_3/2 satellite peak at 718 eV is a characteristic peak of Fe³⁺, similar to α-Fe₂O₃.⁴¹ The deconvoluted Fe2p_3/2 peaks for 7FeTi-NP, NT, NF samples (Fig. 6) showed three sub peaks at 710.2–710.5 eV, 711.5–712.2 eV and 713.5–714.5 eV. The peak at 710.2–710.5 eV can be assigned to Fe³⁺ in interactive Ti–O–Fe species. Appearance of peak at 711.5–712.2 eV suggesting the presence of minor portion of Fe²⁺ ions in the samples.⁴² The peak in the region of 713.5–714.5 eV represents the Fe³⁺ presented in the isolated Fe₂O₃ clusters.

The O1s peak was also fitted into three peaks by applying the optimum combination of Gaussian parameters (Fig. 6). The peak at 528.4 eV corresponded to the lattice oxygen (O²⁻) in TiO₂ support and the other O1s peak at 529.3 eV due to the formation of Ti–O–Fe bonds on the surface of TiO₂. A small shoulder at 531.1 eV, which belong to O₂²⁻ or O⁻ belonging to defect-oxide or hydroxyl-like groups was also appeared for all samples. The quantitative results of the mole ratios of different atoms derived from XPS and peak-fitting results of different samples indicating that Ti–O–Fe and oxygen deficient species are higher for FeTi-NP catalysts. It was reported that interactive and oxygen deficient species are more reactive than lattice oxygen in oxidation reactions due to higher mobility.⁴³ Therefore, xFeTi-NP catalysts are beneficial to benzyl alcohol oxidation.

The bulk elemental composition of the catalysts was measured using the ICP analysis. The chemical composition observed in ICP analysis confirms that the targeted chemical composition (2–3% experimental error) was achieved in the samples (Table 1). The surface Fe/Ti ratios determined from XPS analysis was also presented in Table 1. The Fe/Ti surface ratios of FeTi-NT catalysts comparable to bulk ratio, which can be due to homogeneous distribution of iron oxide after calcination; the iron oxides are present in form of clusters. For FeTi-NF catalysts, differences were observed in bulk and surface Fe/Ti ratio, this may be attributed to the lower specific area of the catalysts. For FeTi-NP catalysts, the strong interaction of Fe₂O₃ and TiO₂ allowed the presence of isolated species on the TiO₂-NP surface. In these samples, the Fe/Ti surface ratio is higher than the bulk ratios, indicative of a surface enrichment in iron oxide.

The reducibility of the bare and iron oxide supported TiO₂ nanomaterials was evaluated by H₂-TPR technique. The H₂-TPR patterns of all samples are shown in Fig. 7. The bare TiO₂-NP, NT and NF samples showed very similar TPR pattern (a single reduction peak around 780 °C), probably due to the fact that all three samples possessed only TiO₂ anatase phase. Two prominent peaks can be identified in the TPR patterns of iron oxide impregnated TiO₂-NP, NT and NF catalysts. The low temperature reduction peak (peak I) can be assigned to the reduction of Fe₂O₃ (Fe³⁺) to FeO (Fe²⁺).⁴⁴

Fig. 7 H₂-TPR patterns of pure and iron oxide supported TiO₂ (a) NP (b) NT and (c) NF samples.

The broad high temperature peak (peak II) is related with the reduction of FeO (Fe²⁺) to metallic Fe (Fe⁰). It should be noted that the temperature maximum of ‘peak I’ and ‘peak II’ is depended on the morphology of TiO₂ nanomaterial and Fe₂O₃ loading. The ‘peak I’ appeared at 350 °C in case of iron oxide supported TiO₂-NP samples, it was appeared at 360 °C and 395 °C for iron oxide supported TiO₂-NT and TiO₂-NF samples respectively. This observation indicating that reduction of Fe³⁺ to Fe²⁺ is occurring relatively at low temperature in case of NP samples than that of NT and NF samples. And iron oxide loading did not affected the reduction temperature of the ‘peak I’ in case of NP, NT and NF samples. However, it is interesting to note that the ‘peak II’ which is assigned to reduction of Fe²⁺ to Fe⁰ appeared in the range of 480–590 °C for xFeTi-NP, 485–590 °C for xFeTi-NT and 525–590 °C for xFeTi-NF samples. This observation indicating that both morphology of TiO₂ and iron oxide have an influence on the reduction of Fe²⁺ species to metallic Fe. These results are suggesting that the Fe₂O₃ species deposited on the surface of in the TiO₂-NP are easier to reduce than the Fe₂O₃ species deposited with TiO₂-NTs and NFs. A very similar finding has been reported on the Fe particles confined in CNTs, which was attributed to the confinement effect of CNTs on the entrapped Fe.⁴⁵ The extent and easy reducibility of xFeTi-NP samples is an indication that these materials possessed higher mobility of oxygen species presented in these samples. In our previous publication, we observed that bulk Fe₂O₃ shows a primary reduction peak at 380 °C.⁴⁶ The reducibility of FeTi nanomaterials is actually higher than that of bulk Fe₂O₃. This indicates that higher mobility of lattice oxygen is obtained over FeTi nanomaterials samples due to the strong interaction between Fe and Ti species, which could be beneficial to enhance the oxidation ability.

Fig. 8 shows FTIR spectra of samples after pyridine adsorption in the spectral region between 1700 and 1400 cm⁻¹. After outgassing at room temperature for 30 min to remove any physisorbed pyridine, strong bands were observed at 1607, 1578, 1490 and 1447 cm⁻¹ for iron oxide loaded TiO₂ nanomaterials, whereas they show very low intensity on the bare TiO₂ nanomaterials. With increase of Fe loading, the intensity of the bands was increased. The observed bands can be ascribed to stretching modes of the pyridine ring coordinated to the Lewis acid sites, i.e. coordinated unsaturated Ti⁴⁺, Fe³⁺ species.⁴⁷.

Fig. 8 FTIR spectra of samples after pyridine adsorption.

It is clear from spectra that bands attributed to pyridinium species were not observed indicating that the samples possessed negligible Bronsted acid sites, as previously observed by Neri et al.⁴⁷ It is also interesting to note that the area of the bands corresponding to Lewis acid sites increased after deposition of iron oxide on TiO₂ supports, however the FeTi-NP samples predominantly exhibited high Lewis acidity than other samples.

3.1. Liquid phase catalytic oxidation of benzyl alcohol

In the present study liquid phase oxidation of benzyl alcohol has been carried out under a variety of reaction conditions over synthesized FeTi-NP, FeTi-NT and FeTi-NF catalysts. The reaction was carried out at different reaction temperatures and with benzyl alcohol to H₂O₂ mole ratios in order to optimize the reaction conditions. Furthermore the reaction was performed by varying the concentration of H₂O₂. The effect of time on the reaction was also studied by running the reaction for 12 h period. The major product identified was benzaldehyde for all the catalytic systems. A slight amount of benzoic acid was also observed. The catalytic activity was expressed as the percentage (%) conversion of benzyl alcohol. An attempt has been made to optimize the reaction variables such as temperature, catalyst amount, time, substrate to oxidant mole ratio etc. in order to study the effect of various parameters on the conversion of benzyl alcohol and selectivity to benzaldehyde. It also helps to know the condition for getting maximum conversion and product yields.

3.2. Effect of reaction temperature

The effect of reaction temperature on the oxidation of benzyl alcohol was studied in a temperature range of 70 to 100 °C, while all other reaction parameters were kept constant. The results are presented in Fig. 9. When the reaction temperature was increased from 70 to 90 °C, conversion rate increased gradually and a drastic increase was observed at 100 °C with benzaldehyde selectivity is around 98% in case of FeTi-NP catalysts. A very similar activity patterns were observed for FeTi-NT and FeTi-NF samples, however the highest conversion of benzyl alcohol was observed for 3FeTi-NP catalyst, but it is not the case for NT and NF samples. In case of NT and NF series samples, 5FeTi-NT and 7FeTi-NF catalysts offered the best catalytic activity in oxidation of benzyl alcohol oxidation. Further increase in reaction temperature may cause a decrease in conversion. It was reported that at higher temperatures (greater than 100 °C), the self-decomposition of H₂O₂ to molecular oxygen proceeds faster and it could not participate efficiently for the oxidation processes;²⁴ for this reason we have not conducted the reaction at higher temperature than 100 °C. For the present reaction the temperature selected was 100 °C in order to get high conversion. From Table 2 it is clear that a very small decrease in selectivity of benzaldehyde was observed by raising the temperature from 70 to 100 °C.

Fig. 9 Influence of temperature on benzyl alcohol oxidation over all FeTi-NP, NT and NF catalysts; reaction conditions: catalyst weight: 0.05 g, H₂O₂: 60 mmol, benzyl alcohol: 50 mmol, time: 6 h.

Table 2 Product distribution of benzyl alcohol oxidation over Fe₂O₃–TiO₂ nanomaterials at different reaction temperatures^a

Catalyst	Product distribution (%) at different temperatures
	70 °C		80 °C		90 °C
	Benzaldehyde	Benzoic acid	Benzaldehyde	Benzoic acid	Benzaldehyde	Benzoic acid
a Reaction conditions: catalyst weight: 0.05 g, H2O2: 60 mmol, benzyl alcohol: 50 mmol, time: 6 h.
TiO2-NP	96	4	95	5	94	6
1FeTi-NP	98	2	96	4	93	7
3FeTi-NP	98	2	96	4	92	8
5FeTi-NP	99	1	97	3	92	8
7FeTi-NP	98	2	97	3	92	8
TiO2-NT	95	5	94	6	94	6
1FeTi-NT	96	4	95	5	94	6
3FeTi-NT	96	4	95	5	94	6
5FeTi-NT	96	4	95	5	94	6
7FeTi-NT	96	4	95	5	94	6
TiO2-NF	95	5	94	6	95	5
1FeTi-NF	97	3	96	4	95	5
3FeTi-NF	97	3	96	4	95	5
5FeTi-NF	96	4	97	3	95	5
7FeTi-NF	96	4	97	3	95	5

---

3.3. Effect of reaction time

Effect of reaction time on the oxidation of benzyl alcohol is illustrated in Fig. 10. An appropriate reaction time is the main assurance for a perfect reaction. In the present study the reaction mixture was analyzed at various time intervals in order to study the effect of reaction time on the oxidation of benzyl alcohol. An increase in the conversion was observed as the time proceeds and reached to a plateau with further increase of reaction time. The selectivity towards benzaldehyde decreased slightly with increase of reaction of time. This is due to the consecutive oxidation of the product benzaldehyde, which was favored with increasing time.

Fig. 10 Influence of time on benzyl alcohol oxidation over all FeTi-NP, NT and NF catalysts; reaction conditions: catalyst weight – 0.05 g, benzyl alcohol – 50 mmol, H₂O₂ – 60 mmol, temp – 90 °C.

The maximum conversion of 91% is reached at 12 h of reaction time with selectivity of benzaldehyde 88%. For further studies of the present reaction a reaction time of 6 h was selected in order to achieve an appreciable conversion of benzyl alcohol and selectivity of benzaldehyde.

3.4. Effect of catalyst amount

The dependence of the amount of catalyst on the production of benzaldehyde by the oxidation of benzyl alcohol is presented in Fig. 11. The influence of catalyst amount was studied by taking different weight of the catalyst most active catalyst in each series of catalyst (3FeTi-NP, 5FeTi-NT and 5FeTi-NF), while keeping the other parameters constant. The conversion of benzyl alcohol was sharply increased from 43 to 55% as the amount of catalyst increased from 0.03 g to 0.05 g, however benzyl alcohol conversion remained steady with further increase of catalyst amount. Interestingly, there is a slight lowering in selectivity of benzaldehyde. This lowering in selectivity is attributed to the oxidation of benzaldehyde to benzoic acid. The optimized catalyst amount for further reaction was 0.05 g. The results of this study indicate that an appropriate conversion and selectivity was obtained even with the small amount of catalyst (0.05 g).

Fig. 11 Influence of catalyst amount on benzyl alcohol oxidation over all FeTi-NP, NT and NF catalysts; reaction conditions: temp – 90 °C, benzyl alcohol – 50 mmol, H₂O₂ – 60 mmol, time – 6 h.

3.5. Effect of H₂O₂ amount

Fig. 12 depicts the influence of H₂O₂ amount in the benzyl alcohol oxidation reaction. The reactions were carried out under previously optimized conditions and changing the volume of H₂O₂ from 0 to 90 mmol while keeping the benzyl alcohol concentration constant. Without H₂O₂ only 20% conversion was observed with 100% benzaldehyde selectivity. In the presence of H₂O₂ higher benzyl alcohol conversion was observed in case of FeTi-NP, FeTi-NT and FeTi-NF catalysts. In case of 3FeTi-NP catalyst, the conversion increased from 37 to 53% upon changing the H₂O₂ amount from 30 to 60 mmol. A small increase in conversion was noticed with further increase of H₂O₂ amount. The selectivity of benzaldehyde decreased further with 90 mmol of H₂O₂. The presence of excess oxidant favored further oxidation of the initially formed product benzaldehyde. A volume of H₂O₂, 60 mmol was found to be the optimum for further studies.

Fig. 12 Influence of H₂O₂ amount on benzyl alcohol oxidation over all FeTi-NP, NT and NF catalysts; reaction conditions: catalyst – 0.05 g, benzyl alcohol – 50 mmol, temp – 90 °C, time – 6 h.

Comparison of the benzyl alcohol oxidation activity results with literature data over non-precious metal-based catalysts reveals that high catalytic performance can be obtained over metal supported catalysts.⁴⁸ The results may be thought of as the proof that the performance of the FeTi-NP catalysts may arise from the existence of Fe₂O₃–TiO₂ interactive species in these catalysts. It was reported that TiO₂ with different morphologies had a significant impact on degradation and scavenging pollutants in water and air owing to their large surface area.⁴⁹ The spherical structures with a high surface area with large size pores found great potential in solar energy conversion.⁵⁰

It was also observed that the confinement effect could improve catalytic activity greatly and, prominently. In the previous reports, researchers fabricated confinement catalysts with metals such as Pt and Pd nanoparticles entrapped in TiO₂ nanotubes and proved that the confinement of Ti-NT could modulate the catalytic performance. In Scheme 1, we presented the schematic representation of interaction pattern between Fe₂O₃ and TiO₂-NPs, NTs, NFs. As it can be seen that the confinement effect (in other words the volume available for interaction) between Fe₂O₃ and TiO₂-NPs is higher than other two systems. It was reported that in the nanoparticle systems, the interactions are controlled by van der Waals forces and the adsorption of metal oxide nanoparticles on the support nanosystem (NP or NT or NF) obeys a simple quadratic dependence on the support nanosystem surface area. Changes in the geometric parameters of the support have pronounced effects on the affinity of metal oxide nanoparticles.⁵¹ It is clear that the TiO₂-NP support has an additional advantage on the catalytic oxidation activity due to the high specific surface area and the availability of charge carrier by modifying the internal electric field close to the surface (space charge effects). Moreover, a great anisotropy of nanoparticles induces distinct behaviors by giving more importance to some crystalline planes in contact with the reacting molecules.

Scheme 1 Schematic representation of coalescence of Fe₂O₃ with TiO₂ nanomaterials.

3.6. Reusability studies

One of the major objectives guiding the development of solid heterogeneous catalysts includes the easy separation of final products from the reaction mixture and efficient catalyst recovery. Reusability studies were carried out by removing the catalyst by filtration from the reaction mixture after completing the reaction, washed thoroughly with acetone and then dried and calcined at 400 °C for 12 h. The calcined catalyst was again used for carrying out the subsequent runs under similar reaction conditions. The reusability of the 3FeTi-NP catalyst was examined for the oxidation of benzyl alcohol under the same reaction conditions. Fig. S7† displays the conversion of benzyl alcohol and selectivity of the products obtained for benzyl alcohol oxidation reaction using regenerated catalyst 3FeTi-NP catalyst. The results suggests that the catalyst can be recycled at least six times without much loss in activity. Benzyl alcohol conversion showed a decrease of 1% at the end of 3rd cycle and a marked decrease of 4% was observed after 6th cycle of use. There is no significant variation in benzaldehyde selectivity during the recycling process. These suggest the resistance to rapid deactivation in this oxidation reaction of benzyl alcohol using 3FeTi-NP catalyst. A very similar results were obtained in case of 5FeTi-NT and 5FeTi-NF catalysts.

3.7. Leaching studies

In addition to thermal stability, chemical stability is also an essential requirement for heterogeneous catalysts. Leaching of the active metal ions can occur during a catalyzed reaction without an induction period and the nature of the reaction may gradually change from heterogeneous to homogeneous without any indication in the reaction profile.²⁹ Metal leaching studies give an idea about the nature of the reaction. For this the solid FeTi-NP, FeTi-NT and FeTi-NF catalysts refluxed for 6 h in hydrogen peroxide, after which the catalyst was removed. The leaching of soluble metal species in the hydrogen peroxide was investigated by assessing the activity of the residual catalyst in benzyl alcohol oxidation reaction.

The results obtained are presented in Table 3. The results obtained (Table 3) indicating that the conversion of benzyl alcohol at the time of filtration is as same as the fresh catalyst for the three types of catalysts. The filtrate was further subjected to qualitative analysis for testing the presence of leached metal ions. Qualitative analysis of the filtrate also confirmed the absence of any Fe ions in the filtrate (ICP analysis). The residual powder catalyst (7FeTi-NT) was characterized by XRD, TEM and BET surface area measurements (Fig. S8†) and the results revealed that the iron ions are not leached out from the catalyst during the reaction confirming their heterogeneous mode of action.

Table 3 Leaching studies of catalysts on benzyl alcohol oxidation^a

Catalyst	Reaction time	TOF (mmol h^−1 gcat^−1)		Selectivity to benzaldehyde
		Before reflux	After reflux	Before reflux	After reflux
a Reaction conditions: benzyl alcohol – 50 mmol, H2O2 – 60 mmol, catalyst weight – 0.05 g, temp – 80 °C.
3FeTi-NP	1 h	143	142	97	97
	6 h	100	100	97	97
5FeTi-NT	1 h	112	112	95	95
	6 h	58	57	95	95
5FeTi-NF	1 h	90	90	97	97
	6 h	42	42	97	97

---

4. Conclusions

In conclusion, Fe₂O₃ (1, 3, 5 and 7 mol% Fe) supported on TiO₂ anatase with different morphologies (nanoparticles, nanotubes and nanofibers) have been synthesized successfully. The catalytic activities of Fe₂O₃–TiO₂ nanomaterials for oxidation of benzyl alcohol were investigated. It has been observed that the Fe₂O₃–TiO₂ nanoparticles catalysts exhibited the highest benzyl alcohol conversion (3FeTi-NP-91.3%), whereas Fe₂O₃–TiO₂ nanotubes and nanofibers show relatively low catalytic activities. Characterization studies by different techniques revealed that iron oxide was highly dispersed on the TiO₂-NP support and more number of surface hydroxyl groups in TiO₂-NT and NF tends to increase the affinity of the surfaces with the Fe₂O₃ precursor. This appears to favor agglomeration, which resulted formation of larger iron oxide particles. It was also observed that the Fe₂O₃–TiO₂ nanomaterials possessed two types of Fe³⁺ ions on support surface; one highly dispersed in which Fe³⁺ ions are placed in the TiO₂ surface and another in which large Fe₂O₃ crystals are located on the surface of the catalyst. The catalyst which possessed former species exhibited the best performance in the oxidation of benzyl alcohol. The difference in catalytic activities in case of xFeTiO₂-NP catalysts could be attributed to the support morphology, high surface area of the TiO₂-NP, dispersion of Fe₂O₃, easy reducibility of total catalyst and more number of Lewis acid sites. The synthesized catalysts are found to be reusable and resistant to rapid deactivation.

Acknowledgements

The authors gratefully thank King Abdulaziz City for Science and Technology (KACST) for the financial support provided through grant no. AP-36-113. The team members at SABIC research and development center, Riyadh is also acknowledged for their technical support.

References

1. T. Mallat and A. Baiker, Chem. Rev., 2004, 104, 3037–3058.
2. M. M. Dell'Anna, M. Mali, P. Mastrorilli, P. Cotugno and A. Monopoli, J. Mol. Catal. A: Chem., 2014, 386, 114–119.
3. S. Xiao, C. Zhang, R. Chen and F. Chen, New J. Chem., 2015, 39, 4924–4932.
4. L. Tonucci, M. Nicastro, N. D'Alessandro, M. Bressan, P. D'Ambrosio and A. Morvillo, Green Chem., 2009, 11, 816–820.
5. R. A. Sheldon, I. W. C. E. Arends and A. Dijkman, Catal. Today, 2000, 57, 157–166.
6. R. S. Chandran and W. T. Ford, J. Chem. Soc., Chem. Commun., 1988, 46, 104–105.
7. U. R. Pillai and E. S. Demessie, Green Chem., 2004, 6, 161–165.
8. N. Hiroki and I. Akichika, Chem. Pharm. Bull., 2006, 54, 1620–1621.
9. V. R. Choudhary and D. K. Dumbre, Appl. Catal., A, 2010, 375, 252–257.
10. K. Liu, X. J. Yan, P. P. Zou, Y. Y. Wang and L. Y. Dai, Catal. Commun., 2015, 58, 132–136.
11. G. Csjernyik, A. H. Ell, L. Fadini, B. Pugin and J. E. Backvall, J. Org. Chem., 2002, 67, 1657–1662.
12. Y. Li, T. Bian, J. S. Du, Y. L. Xiong, F. W. Zhan, H. Zhang and D. R. Yang, CrystEngComm, 2014, 16, 8340–8343.
13. I. W. C. E. Arends and R. A. Sheldon, Appl. Catal., A, 2001, 212, 175–187.
14. Y. X. Tang, P. X. Wee, Y. K. Lai, X. P. Wang, D. G. Gong, P. D. Kanhere, T. T. Lim, Z. L. Dong and Z. Chen, J. Phys. Chem. C, 2012, 116, 2772–2780.
15. T. O. Eschemann, J. H. Bitter and K. P. de Jong, Catal. Today, 2014, 228, 89–95.
16. R. Nafria, P. Ramírez de la Piscina, N. Homs, J. R. Morante, A. Cabot, U. Diaz and A. Corma, J. Mater. Chem. A, 2013, 1, 14170–14176.
17. P. Panagiotopoulou and D. I. Kondarides, J. Catal., 2004, 225, 327–336.
18. Y. Q. Dai, C. M. Cobley, J. Zeng, Y. M. Sun and Y. N. Xia, Nano Lett., 2009, 9, 2455–2459.
19. B. Souvereyns, K. Elen, C. De Dobbelaere, A. Kelchtermans, N. Peys, J. D'Haen, M. Mertens, S. Mullens, H. V. Rul, V. Meynen, P. Cool, A. Hardy and M. K. V. Bael, Chem. Eng. J., 2013, 223, 135–144.
20. H. D. Nam, B. H. Lee, S. J. Kim, C. H. Jung, J. H. Lee and S. Park, Jpn. J. Appl. Phys., 1998, 37, 4603–4608.
21. (a) M. Hussain, R. Ceccarelli, D. L. Marchisio, D. Fino, N. Russo and F. Geobaldo, Chem. Eng. J., 2010, 157, 45–51; (b) E. I. Seck, J. M. Dica-Rodriguez, C. Fernndez-Rodriguez, O. M. Gonzalez-Diaz, J. Araia and J. Pérez-Peca, Chem. Eng. J., 2012, 203, 52–62.
22. H. Kominami, J. I. Kato, Y. Takada, Y. Doushi, B. Ohtani, S. I. Nishimoto, M. Inoue, T. Inui and Y. Kera, Catal. Lett., 1997, 46, 235–240.
23. T. Kasuga, M. Hiramatsu, A. Hoson, T. Sekino and K. Niihara, Langmuir, 1998, 14, 3160–3163.
24. A. K. Dalai, R. Sethuraman, S. P. R. Katikaneni and R. O. Idem, Ind. Eng. Chem. Res., 1998, 37, 3869–3878.
25. Q. Wang, S. Xu and F. Shen, Appl. Surf. Sci., 2011, 257, 7671–7677.
26. A. R. Ballyy, E. N. Korobeinikova, P. E. Schmid, F. Levy and F. Bussy, J. Phys. D: Appl. Phys., 1998, 31, 1149–1154.
27. K. T. Ranjit and B. Vishwanathan, J. Photochem. Photobiol., A, 1997, 108, 79–84.
28. J. Zhu, F. Chen, J. Zhang, H. Chen and M. Anpo, J. Photochem. Photobiol., A, 2006, 180, 196–204.
29. L. Yue and X.-M. Zhang, J. Alloys Compd., 2009, 475, 702–705.
30. Z. Y. Liu, D. D. Sun, P. Guo and J. O. Leckie, Chem.–Eur. J., 2007, 13, 1851–1855.
31. T. Lopez, J. A. Moreno, R. Gomez, X. Bokhimi, J. A. Wang, H. Yee-Madeira, G. Pecchi and P. Reyes, J. Mater. Chem., 2002, 12, 714–718.
32. A. M. Stoyanova, H. Y. Hitkova, N. K. Ivanova, A. D. Bachvarova-Nedelcheva, R. S. Iordanova and M. P. Sredkova, Bulg. Chem. Commun., 2013, 45, 497–504.
33. K. Narasimharao, M. A. Malik, M. M. Mokhtar, S. N. Basahel and S. A. Al-Thabaiti, Ceram. Int., 2014, 40, 4039–4053.
34. M. Sun, Y. Fang, Y. Wang, S. Sun, J. He and Z. Yan, J. Alloys Compd., 2015, 650, 520–527.
35. S. Brunauer, L. S. Deming, W. E. Deming and E. Teller, J. Am. Chem. Soc., 1940, 62, 1723–1732.
36. F. Lin, H. Song, S. Tian, X. Chen, J. Zhou and F. Wang, Electrochim. Acta, 2012, 83, 305–310.
37. C. Lin, H. B. Yang, H. Shuai, W. T. Jie, C. L. Su and J. Yong, Powder Technol., 2012, 227, 3–8.
38. M.-H. Pham, C.-T. Dinh, G.-T. Vuong, N.-D. Ta and T.-O. Do, Phys. Chem. Chem. Phys., 2014, 16, 5937–5941.
39. J. P. Espinos, A. Fernandez and A. R. Gonzalez-Elipe, Surf. Sci., 1993, 295, 402–410.
40. A. P. Grosvenor, B. A. Kobe, M. C. Biesinger and N. S. McIntyre, Surf. Interface Anal., 2004, 36, 1564–1574.
41. R. A. Jones, XPS evidence for Fe and As oxidation states and electronic states in loellingite (FeAs2), American Mineralogist, 2002, vol. 87, no. 11-12, pp. 1692–1698.
42. R. H. Gao, D. S. Zhang, X. G. Liu, L. Y. Shi, P. Maitarad and H. R. Li, Catal. Sci. Technol., 2013, 3, 191–199.
43. J. Y. Kim, G. Magesh, D. H. Youn, J. W. Jang, J. Kubota and K. Domen, Sci. Rep., 2013, 3, 2681–2688.
44. W. Chen, Z. L. Fan, X. L. Pan and X. H. Bao, J. Am. Chem. Soc., 2008, 130, 9414–9419.
45. T. T. Ali, K. Narasimharao, N. S. Ahmed, S. Basahel, S. Al-Thabaiti and M. Mokhtar, Appl. Catal., A, 2014, 486, 19–31.
46. J. M. Amores, V. S. Escribano, G. Ramis and G. Busca, Appl. Catal., B, 1997, 13, 45–48.
47. G. Neri, G. Rizzo, S. Galvagno, G. Loiacono, A. Donato, M. G. Musolino, R. Pietropaolo and E. Rombi, Appl. Catal., A, 2004, 274, 243–251.
48. Y. Du, Q. Wang, X. Liang and D. Q. Li, J. Catal., 2015, 331, 154–161.
49. M. M. Khin, A. S. Nair, V. J. Babu, R. Murugan and S. Ramakrishna, Energy Environ. Sci., 2012, 5, 8075–8109.
50. M. Pal, J. Garcia-Serrano, P. Santiago and U. Pal, J. Phys. Chem. C, 2007, 111, 96–102.
51. Y. Suzuki, M.-H. Berger, D. D'ella, P. Ilbizian, C. Beauger, A. Rigacci, J.-F. Hochepied and P. Achard, Nano, 2008, 3, 373–379.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra13958b

---