Synthesis of ZnO@γ-Fe₂O₃ core–shell nanocomposites by a facile thermal decomposition approach and their application in photocatalytic degradation of congo red

Abstract

ZnO@γ-Fe₂O₃ core–shell nanocomposites were synthesized by a facile thermal decomposition approach. ZnO nanorods were first synthesized by calcination of zinc acetate at 300 °C, in air. γ-Fe₂O₃ nanoparticles were then deposited on the surface of ZnO nanorods by the thermal decomposition of iron acetylacetonate at 200 °C in diphenyl ether. The structure, composition, optical and magnetic properties of the nanocomposites were studied using an array of techniques. XRD results suggest the presence of γ-Fe₂O₃ nanoparticles and ZnO, and FE-SEM images indicate formation of shell of iron oxide on the ZnO nanorods. Transmission electron microscopy studies clearly show that ZnO possesses rod morphology (length = 1.1 ± 0.1 μm, diameter = 40.1 ± 7 nm) and TEM images of the ZnO@γ-Fe₂O₃ nanocomposites show uniform shell of γ-Fe₂O₃ coated on the ZnO nanorods and thickness of the γ-Fe₂O₃ shell varies from 10 to 20 nm. Diffuse reflectance spectra of ZnO@γ-Fe₂O₃ nanocomposites reveal extended optical absorption in the visible range (400–600 nm) and photoluminescence spectra indicate that the ZnO@γ-Fe₂O₃ nanocomposites exhibit enhanced defect emission. The ZnO@γ-Fe₂O₃ core–shell nanocomposites show superparamagnetic behaviour at room temperature. The core–shell nanocomposites exhibit enhanced visible-light driven photocatalytic degradation of congo red in an aqueous solution as compared to pure ZnO nanorods and γ-Fe₂O₃ nanoparticles. The enhanced photocatalytic activity is attributed to good visible-light absorption and effective charge separation at the interface of ZnO@γ-Fe₂O₃ core–shell nanocomposites.

Introduction

Nowadays, research efforts are being focused on the use of semiconductors as photocatalysts for the degradation of toxic organic compounds discharged from various industries (Chen et al. 2010a). Among various metal oxide semiconductor photocatalysts used for this purpose, ZnO has been of tremendous interest due to its high photosensitivity and wide band gap (3.3 eV). ZnO possesses many advantages that include low cost, high quantum efficiency and high photocatalytic activity. There are two drawbacks associated with ZnO as the photocatalyst, namely, high charge recombination rate and low efficiency when utilized under solar light (Kumar and Rao 2015). Various methods have been developed to reduce the recombination rate of photogenerated electrons and holes in ZnO and thus increase the utilization of visible light. Hybrid semiconductor nanostructures have been shown to be promising for the removal of toxic organic pollutants from waste water (Wang et al. 2013). The combination of semiconductors with small band gap with wide band gap semiconductors increases absorption of radiation in the visible range. γ-Fe₂O₃, a narrow band gap (1.9–2.2 eV) n-type semiconductor with absorption in the visible region, is a potential candidate that can act as a sensitizer under sunlight irradiation. The enhanced photocatalytic efficiency has been explained by a fast transfer of photogenerated electrons and holes from γ-Fe₂O₃ to ZnO (Liu et al. 2012).

ZnO–Fe₂O₃ nanocomposites have been intensively investigated for a wide variety of applications such as waste water treatment, water splitting, sensors, magnetic materials, lithium ion batteries and solar cells (Guskos et al. 2010; Hsu et al. 2015; Liu et al. 2012; Qin et al. 2012; Reda 2010; Zhang et al. 2011). ZnO–Fe₂O₃ nanocomposites have been prepared by different reported methods. Liu et al. have synthesized ZnO–Fe₂O₃ core–shell nanocomposites using microwave irradiation followed by thermal decomposition (Liu et al. 2012). Hsu et al. have synthesized ZnO@Fe₂O₃ core–shell nanocomposites by a two-step (hydrothermal and wet chemical) method (Hsu et al. 2015). Guskos et al. have synthesized ZnO–Fe₂O₃ nanocomposites by a wet chemical method (Guskos et al. 2010). Reda, Maya-Trevino et al. and Hernandez et al. have synthesized ZnO–Fe₂O₃ nanocomposites by sol–gel method (Hernández et al. 2007; Maya-Treviño et al. 2014; Reda 2010). Yin et al. have synthesized α-Fe₂O₃ decorated ZnO nanorods by impregnation deposition method (Yin et al. 2014). Abdullah Mirzaie et al. have synthesized ZnO–Fe₂O₃ nanocomposites by solid state reaction (Abdullah Mirzaie et al. 2012). Liu et al. have synthesized ZnO–Fe₂O₃ nanotube composites by photochemical deposition under UV light (Liu et al. 2015a, b). Fu et al. have synthesized ZnO–γ-Fe₂O₃ nanocomposites by solution route (Fu et al. 2008). Si et al. have prepared ZnO–Fe₂O₃ core–shell nanorods by solution phase hydrolysis method (Si et al. 2006). Balti et al. have synthesized Fe₂O₃ decorated ZnO nanorods by polyol method (Balti et al. 2014). Zhou et al. have synthesized ZnO–Fe₂O₃ core–shell heterostructures by hydrothermal method (Zhou et al. 2015). Chu et al. have synthesized ZnO microrods coated with iron oxide nanoparticles by low temperature hydrothermal method (Chu et al. 2008). Wu et al. have synthesized ZnO-γ-Fe₂O₃ hybrid nanomaterials by layer-by layer assembly technique (Wu et al. 2010). Mo et al. have synthesized ZnO-FeO _x core–shell nanorods by thermal decomposition of Fe(acac)₃ at 265 °C under nitrogen atmosphere (Mo et al. 2009). Most of the methods, reported so far, are complex, involve multi-steps, require surface modification of ZnO nanorods by polyelectrolytes (e.g. PAH/PSS/PAH, PAH = poly(allylamine hydrochloride), and PSS = poly(sodium 4-styrenesulphonate)) with reaction time of 30 min to 4 h and also require often special conditions such as an inert atmosphere.

In the present study, ZnO@γ-Fe₂O₃ core–shell nanocomposites have been prepared by a novel thermal decomposition method. ZnO nanorods with length 1.1 ± 0.1 μm and diameter 40.1 ± 7 nm were prepared by heating zinc acetate dihydrate at 300 °C, in air. Uniform layer of γ-Fe₂O₃ was then deposited by the thermal decomposition of Fe(acac)₃ in diphenyl ether at 200 °C in the presence of ZnO nanorods. After thorough characterization, the nanocomposites were explored as catalyst for photocatalytic degradation of congo red in an aqueous solution.

Experimental details

Chemicals and materials

The chemicals used were zinc acetate dihydrate (Rankem), iron acetylacetonate (Fe(acac)₃) (ACROS), diphenyl ether (Aldrich), congo red (SRL), ammonium oxalate (Sarabhai Chemicals), benzoquinone (Alfa Aesar), tertiary butyl alcohol (Rankem), methanol (Rankem), and Millipore^® water. All the chemicals were used as received.

Synthesis

The ZnO@γ-Fe₂O₃ nanocomposites were prepared via a thermal decomposition method.

Synthesis of ZnO nanorods

The first step was the preparation of ZnO nanorods by solid-state thermal decomposition of zinc acetate dihydrate (Chandraiahgari et al. 2015). 2.0 g of zinc acetate dihydrate was crushed with the help of a mortar and pestle and was heated in a muffle furnace at 300 °C for 3 h at a heating rate of 2 °C/min, in air.

Synthesis of ZnO@γ-Fe₂O₃ core–shell nanocomposites

The ZnO@γ-Fe₂O₃ core–shell nanocomposites were synthesized by the thermal decomposition of Fe(acac)₃ in diphenyl ether at 200 °C in the presence of ZnO nanorods. Four different core–shell nanocomposites (ZF1, ZF2, ZF3 and ZF4) were prepared by varying the concentration of Fe(acac)₃ (Table 1). The reagents were added to 10 mL of diphenyl ether (boiling point = 257 °C) in a round bottom flask and refluxed in air at about 200 °C for 35 min. The precipitates obtained were centrifuged, washed with methanol and dried overnight under vacuum to obtain the ZnO@γ-Fe₂O₃ core–shell nanocomposites. For comparison, pure γ-Fe₂O₃ nanoparticles were also synthesized using the same method.

Table 1 Designation of various ZnO@γ-Fe₂O₃ nanocomposites prepared using different molar ratios of Fe(acac)₃ and ZnO

Photocatalytic activity

The photocatalytic experiments using ZnO@γ-Fe₂O₃ core–shell nanocomposites (ZF1, ZF2, ZF3 and ZF4) as the catalysts for the photodegradation of congo red in aqueous solutions were conducted as follows: 50 mg powder of each catalyst was suspended in 50 mL of congo red aqueous solution (50 mg L⁻¹) and the contents were stirred in dark for 60 min to reach adsorption–desorption equilibrium. Afterwards, the suspension was exposed to sunlight for various time intervals (0–90 min). The photocatalytic experiments under sunlight were carried out at Indian Institute of Technology Roorkee, Roorkee, India in the month of April 2015 between 12.00 noon and 2.00 pm. The intensity of solar radiation in Roorkee in the month of April was 280 W/m² (Ramachandra et al. 2011). At regular time intervals, 3 mL of the suspension was taken and centrifuged at 3500 rpm for 5 min to completely remove the catalyst. The concentration of congo red in the supernatant solution was analysed with the help of UV–Visible spectroscopy (Shimadzu UV-2450) by measuring the absorbance at 495 nm. The photocatalytic activity of pure ZnO nanorods and γ-Fe₂O₃ nanoparticles towards the photodegradation of congo red was also studied for comparison with that of the ZnO@γ-Fe₂O₃ core–shell nanocomposites. To verify the role of different oxidation species (h⁺, OH^· and O ^·−₂ ) in the photocatalytic degradation of congo red, scavenger tests were carried out using different scavengers. The scavengers used were ammonium oxalate (h⁺), benzoquinone (O ^·−₂ ), and tertiary butyl alcohol (OH^·).

Characterization

Powder XRD patterns of the ZnO@γ-Fe₂O₃ nanocomposites were recorded using a Bruker AXS D8-Advance X-ray diffractometer with Cu-Kα radiation (λ = 1.5406 Å) with a scan speed, 2°/min. The acceleration voltage for the XRD analysis was 40 kV and 30 mA with a 2θ interval of 0.018°. A Thermo Nicolet Nexus FT-IR spectrometer was used for recording IR spectra of the nanocomposites in the range 400–4000 cm⁻¹ using KBr pellets. Differential thermal analysis (DTA) of the nanocomposites was recorded on an EXSTAR TG/DTA 6300 instrument under air (flow rate = 200 mL/min) from 25 to 800 °C with a heating rate of 10 °C min⁻¹. The morphology and elemental analyses of the core–shell nanocomposites were studied with the help of a field emission scanning electron microscope (Carl Zeiss (ULTRA plus)) operating at an accelerating voltage of 15 kV coupled with an energy dispersive X-ray analysis (EDXA) facility. The transmission electron microscope images, selected area electron diffraction patterns (SAED) and high-resolution transmission electron microscope images (HRTEM) were acquired using a FEI TECNAI G² electron microscope operating at an accelerating voltage of 200 kV. For the TEM measurements, the nanocomposite powders were dispersed in ethanol with the help of a sonicator and a drop each from the dispersions was allowed to dry on carbon-coated copper grids in air. Diffuse reflectance spectra of the nanocomposites were recorded using a Varian Cary 5000 UV–VIS–NIR spectrophotometer attached with a diffuse reflectance accessory. BaSO₄ was used as the reference and the spectra were recorded in the range of 300–800 nm. Photoluminescence spectra were recorded with a Shimadzu (RF-5301PC) spectrofluorophotometer with λ _excitation as 325 nm. The emission spectra of ZnO@γ-Fe₂O₃ core–shell nanocomposites were recorded by dispersing about 5 mg of each nanocomposite powder in 5 mL water followed by sonication for 10 min. Surface area of the nanocomposites was measured using a NOVA 2200e (Quantachrome) instrument by BET method using nitrogen physisorption at 77 K. Prior to the analysis, the samples were degassed at 150 °C for 2 h. Magnetic measurements (M–H curves at 300 and 5 K) of pure γ-Fe₂O₃ nanoparticles and the ZnO@γ-Fe₂O₃ core–shell nanocomposites were carried out on a 3 Tesla cryogen-free mini Vibrating Sample Magnetometer with a pulse cryocooler (CRYOGENICS Ltd, UK). Thermal variation of the magnetization between 5 and 300 K was also studied using the same instrument in the zero-field-cooled/field-cooled (ZFC/FC) mode under an applied magnetic field of 0.05 T.

Results and discussion

The XRD patterns of pure iron oxide nanoparticles, ZnO nanorods and ZnO@γ-Fe₂O₃ core–shell nanocomposites (ZF1, ZF2, ZF3 and ZF4) are shown in Fig. 1a. Pure iron oxide nanoparticles do not show any peaks. The XRD patterns of pure ZnO and all the nanocomposites show peaks at 31.7°, 34.2°, 36.1°, 47.6°, 56.5°, 62.7°, 66.3°, 67.7°, 72.5°, 76.9° and 81.5° corresponding to (100), (002), (101), (102), (110), (103), (200), (112), (004), (202), (104) planes of wurtzite ZnO (JCPDS file no. 80-0075). In XRD patterns of the nanocomposites, reflections due to iron oxide are not observed. This is attributed to the presence of very small nanocrystals of γ-Fe₂O₃ that are not detectable by X-ray diffraction. The proof for the presence of γ-Fe₂O₃ comes from XRD, FT-IR, DTA, SAED, and magnetic measurements (discussed latter). The mean crystallite size of pure ZnO and ZnO in the core–shell nanocomposites was estimated using Scherrer equation (Si et al. 2006). The crystallite size of pure ZnO is 25.7 nm and that of ZnO in the nanocomposites varies from 20.6 to 27.7 nm.

Fig. 1

a XRD patterns of ZnO, γ-Fe₂O₃ and ZnO@γ-Fe₂O₃ core–shell nanocomposites (ZF1–ZF4), and b XRD patterns of as-prepared iron oxide (prepared by the present thermal decomposition method) and after calcination at 400 °C. The XRD patterns of as prepared Fe₃O₄ nanoparticles and after calcination at 400 °C are also shown

To identify the phase of iron oxide nanoparticles prepared by the present thermal decomposition method, Fe₃O₄ nanoparticles were prepared for comparison using a reported method (Ayyappan et al. 2008). Fe²⁺ and Fe³⁺ salts (ferric nitrate and ferrous sulphate) were mixed in distilled water in appropriate proportions and co-precipitated with the help of aqueous ammonia solution at 70 °C. The XRD patterns of as prepared iron oxide prepared by the thermal decomposition method and that of Fe₃O₄ nanoparticles are shown in Fig. 1b. The XRD pattern of as-prepared iron oxide prepared by the thermal decomposition method shows no discernible peaks. The XRD pattern of the same sample after calcination at 400 °C for 3 h shows peaks at 33.1°, 35.4°, 40.7°, 43.3°, 49.5°, 54.1°, 62.4° and 64.1° corresponding to (121), (110), (120), (020), (220), (132), (130) and (211) planes of α-Fe₂O₃ (JCPDS file No. 85-0987). The XRD pattern of Fe₃O₄ nanoparticles, prepared by the reported method, shows weak reflections at 35.5° and 63.1° corresponding to (311) and (440) planes of Fe₃O₄ (JCPDS file no. 85-1436). The XRD pattern of Fe₃O₄ nanoparticles calcined at 400 °C shows weak reflections at 35.6° and 63.8° corresponding to (311) and (440) planes of γ-Fe₂O₃. In general, it is difficult to distinguish γ-Fe₂O₃ and Fe₃O₄ from the XRD patterns. When γ-Fe₂O₃ is heated at 400 °C in air, it transforms to α-Fe₂O₃ and when Fe₃O₄ is heated at 400 °C in air, it transforms to γ-Fe₂O₃ (Ayyappan et al. 2008; Jayanthi et al. 2015; Teja and Koh 2009). Based on the XRD results of samples after calcination, the as-prepared iron oxide nanoparticles, prepared by the present thermal decomposition method, are proposed to be γ-Fe₂O₃ which on calcination transforms to α-Fe₂O₃.

The IR spectra of pure ZnO nanorods and the ZnO@γ-Fe₂O₃ core–shell nanocomposites were recorded (Fig. S1a). Two broad bands at 3434 and 1600 cm⁻¹ are attributed to O–H stretching and bending, respectively, due to adsorbed water molecules. The IR bands near 2920 and 2840 cm⁻¹ are due to C–H stretching and are attributed to the presence of organic molecules adsorbed on the surface of ZnO@γ-Fe₂O₃ core–shell nanocomposites (Hernández et al. 2007). The band at 1432 cm⁻¹ is attributed to stretching vibration of COO⁻. The band at 1096 cm⁻¹ is due to C–O–C bond (Yang et al. 2015). These observed bands suggest the presence of organic content adsorbed on the surface of the nanocomposites (Yang et al. 2015). The bands at 553 and 453 cm⁻¹ are attributed to characteristic absorption of γ-Fe₂O₃ and the band at 444 cm⁻¹ is attributed to ZnO (Maiti et al. 2015; Morales et al. 1999). Figure S1b shows the FT-IR spectra for as-prepared iron oxide nanoparticles prepared by the present thermal decomposition method, after calcination at 400 °C and the as-prepared Fe₃O₄ nanoparticles. The as-prepared Fe₂O₃ nanoparticles show IR bands at 453 and 553 cm⁻¹ attributed to Fe–O bond of γ-Fe₂O₃ (Maity and Agrawal 2007; Morales et al. 1999). The iron oxide nanoparticles (prepared by the thermal decomposition method) calcined at 400 °C show IR bands at 480 and 549 cm⁻¹ attributed to Fe–O bond of α-Fe₂O₃ (Apte et al. 2007; Pandey et al. 2014; Suresh et al. 2014). Fe₃O₄ nanoparticles show characteristic IR bands at 426 and 580 cm⁻¹ (Saffari et al. 2015; Song et al. 2015). The as-prepared iron oxide nanoparticles prepared by the thermal decomposition method and after calcination at 400 °C show IR bands at 1390 and 1031 cm⁻¹ attributed to the stretching vibration of COO⁻ and C–O–C, respectively (Yang et al. 2015). These bands are attributed to the presence of organic molecules adsorbed on the surface of iron oxide nanoparticles.

Figure 2 shows the DTA curves for pure ZnO, as-prepared γ-Fe₂O₃ nanoparticles and the ZnO@γ-Fe₂O₃ core–shell nanocomposites. The DTA curve for ZnO does not show any peak. The DTA curve for as-prepared pure γ-Fe₂O₃ nanoparticles shows an exothermic peak at 212 °C which is attributed to the removal of organic content adsorbed on the surface of iron oxide nanoparticles (Hernández et al. 2007; Lee et al. 2008). To verify whether the sharp peak at 212 °C is due to crystallization, the as-prepared iron oxide was calcined at 225 °C for 3 h and XRD pattern was recorded for the calcined sample. The XRD pattern did not show any discernible peak indicating that the sharp peak in the DTA pattern at 212 °C is not due to crystallization. The ZnO@γ-Fe₂O₃ core–shell nanocomposites show broad exothermic hump from 200 to 250 °C which is attributed to the removal of organic content adsorbed on the surface of ZnO@γ-Fe₂O₃ core–shell nanocomposites. The DTA curves for as prepared iron oxide nanoparticles and all the ZnO@γ-Fe₂O₃ core–shell nanocomposites show a broad exothermic hump from 250 to 360 °C attributed to the phase transition of γ-Fe₂O₃ to α-Fe₂O₃ (Phu et al. 2011; Schimanke and Martin 2000; Yu et al. 2013). Phu et al. have reported DTA results for the phase transition of γ-Fe₂O₃ to α-Fe₂O₃ in the temperature range of 310–360 °C (Phu et al. 2011). Schimanke and Martin have reported phase transition of γ-Fe₂O₃ to α-Fe₂O₃, based on DTA results in the range from 290 to 400 °C (Schimanke and Martin 2000). Yu et al. have reported TG/DSC results, suggesting phase transition of γ-Fe₂O₃ to α-Fe₂O₃ in the range from 200 to 500 °C (Yu et al. 2013). From the DTA results, it is evident that the as-prepared iron oxide is γ-Fe₂O₃ and, when it is heated in air, it transforms to α-Fe₂O₃ as reported (Azizi and Heydari 2014; Chen and Xu 1998; Phu et al. 2011; Schimanke and Martin 2000; Yu et al. 2013). The DTA results are in accordance with the XRD and FT-IR results.

Fig. 2

DTA curves for pure γ-Fe₂O₃ nanoparticles, ZnO and the ZnO@γ-Fe₂O₃ core–shell nanocomposites (ZF1, ZF2, ZF3 and ZF4)

FE-SEM and TEM measurements were performed to investigate morphology and size of ZnO, γ-Fe₂O₃ and the ZnO@γ-Fe₂O₃ core–shell nanocomposites. Figure 3 shows the FE-SEM images of pure ZnO, γ-Fe₂O₃ nanoparticles and the ZnO@γ-Fe₂O₃ core–shell nanocomposites (ZF1–ZF4). Pure ZnO exhibits nanorods and γ-Fe₂O₃ shows small agglomerated nanoparticles. After the deposition of γ-Fe₂O₃, the shape of ZnO is retained but the surface of ZnO nanorods is not smooth. Nanocomposites ZF1 and ZF2 show uniform shell formation of γ-Fe₂O₃ over the ZnO nanorods. Nanocomposites ZF3 and ZF4 show formation of γ-Fe₂O₃ shell over the ZnO nanorods with aggregation. The aggregation of γ-Fe₂O₃ nanoparticles increases as one increases the amount of Fe(acac)₃ used during the synthesis of the core–shell nanocomposites (ZF1–ZF4). Table 2 shows EDX analysis results on the ZnO@γ-Fe₂O₃ core–shell nanocomposites. The nanocomposites ZF1, ZF2 and ZF4 show uniform elemental distribution as compared to nanocomposite ZF3.

Fig. 3

FE-SEM images of ZnO nanorods, γ-Fe₂O₃ nanoparticles and ZnO@ γ-Fe₂O₃ core–shell nanocomposites (ZF1, ZF2, ZF3 and ZF4)

Table 2 EDX analysis data for the ZnO@γ-Fe₂O₃ core–shell nanocomposites (ZF1–ZF4). The analysis was done at three different spots

Figure 4 shows the TEM images and high resolution TEM (HRTEM) images of ZnO and γ-Fe₂O₃ nanoparticles. ZnO nanorods (Fig. 4a) are 1.1 ± 0.1 μm in length and 40.1 ± 7 nm in width. The SAED pattern of ZnO (inset of Fig. 4a) shows spots attributed to (101), (103), (202), (004) and (002) planes of wurtzite ZnO (JCPDS file no. 80-0075). Figure 4b shows small γ-Fe₂O₃ nanoparticles (2.1 ± 0.3 nm) and the SAED pattern (inset of Fig. 4b) shows rings attributed to (311), (421) and (440) planes of γ-Fe₂O₃ (JCPDS file no. 39-1346). The HRTEM images of pure ZnO nanorod (Fig. 4c) and γ-Fe₂O₃ nanoparticles (Fig. 4d) exhibit interplanar spacing of 0.24 and 0.48 nm which are attributed to (101) plane of ZnO and (111) plane of γ-Fe₂O₃, respectively.

Fig. 4

TEM images of a ZnO nanorods, b γ-Fe₂O₃ nanoparticles, c HRTEM image of ZnO nanorods and d HRTEM image of γ-Fe₂O₃ nanoparticles

The TEM and HRTEM images of ZnO@γ-Fe₂O₃ core–shell nanocomposites ZF1 and ZF2 are shown in Fig. 5. The ZnO@γ-Fe₂O₃ core–shell nanocomposites ZF1 (Fig. 5a) and ZF2 (Fig. 5c) show formation of uniform γ-Fe₂O₃ shell on the surface of ZnO nanorods. The inset of Fig. 5a shows the TEM image of a single nanorod of core–shell nanocomposite ZF1 coated uniformly with a shell of γ-Fe₂O₃ (~15 nm). The HRTEM image of nanocomposite ZF1 (Fig. 5b) shows interplanar spacing values of 0.24 and 0.48 nm for the core and shell regions corresponding to ZnO (101) and γ-Fe₂O₃ (111) planes, respectively. This indicates that the γ-Fe₂O₃ shell with (111) plane is deposited on crystalline ZnO nanorod core with (101) plane. The inset of Fig. 5c shows the TEM image of a single nanorod of core–shell nanocomposite ZF2 coated uniformly with γ-Fe₂O₃ shell (~20 nm). The HRTEM image of core–shell nanocomposite ZF2 (Fig. 5d) exhibits 0.24 nm as interplanar spacing attributed to (101) plane of ZnO.

Fig. 5

TEM and HRTEM images of ZnO@γ-Fe₂O₃ core–shell nanocomposites; ZF1 (a, b), and ZF2 (c, d)

Figure 6 shows the TEM and HRTEM images of ZnO@γ-Fe₂O₃ core–shell nanocomposites ZF3 and ZF4. The core–shell nanocomposites ZF3 (Fig. 6a) and ZF4 (Fig. 6c) show formation of γ-Fe₂O₃ shell on the ZnO nanorods. The insets of Fig. 6a, c show the TEM images of single nanorods of the core–shell nanocomposites coated with γ-Fe₂O₃ shell of thickness 10 nm (ZF3) and 15 nm (ZF4). The HRTEM image of core–shell nanocomposite ZF3 (Fig. 6b) exhibits 0.24 nm as interplanar spacing attributed to (101) plane of ZnO and the HRTEM image of core–shell nanocomposite ZF4 (Fig. 6d) exhibits interplanar spacing values of 0.24 and 0.48 nm for the core and shell regions corresponding to ZnO (101) and γ-Fe₂O₃ (111) planes, respectively. This again indicates that γ-Fe₂O₃ shell with (111) plane is deposited on crystalline ZnO nanorods (core) with (101) plane.

Fig. 6

TEM and HRTEM images of ZnO@γ-Fe₂O₃ core–shell nanocomposites; ZF3 (a, b), and ZF4 (c, d)

The SAED patterns of core–shell nanocomposites ZF1, ZF2, ZF3 and ZF4 are shown in Fig. S2. The SAED patterns of nanocomposites ZF1–ZF3 (Fig. S2a–c) show spots which are attributed to (101), (102), (103), (002) and (004) planes of hexagonal ZnO and rings due to (311), (440) and (310) planes of γ-Fe₂O₃. The SAED pattern of nanocomposite ZF4 (Fig. S2d) shows spots which are attributed to (101), (110), (103), (201) and (202) planes of hexagonal ZnO and rings due to (311), (440), (321) planes of γ-Fe₂O₃.

Diffuse reflectance spectra of pure ZnO nanorods, γ-Fe₂O₃ nanoparticles (before calcination) and the ZnO@γ-Fe₂O₃ core–shell nanocomposites (ZF1, ZF2, ZF3 and ZF4) are shown in Fig. 7. ZnO shows absorption in the UV region at about 385 nm. Pure γ-Fe₂O₃ nanoparticles show absorption at about 653 nm. The ZnO@γ-Fe₂O₃ core–shell nanocomposites show an extended optical absorption in the visible region. One can observe two absorption bands in the diffuse reflectance spectra of the ZnO@γ-Fe₂O₃ core–shell nanocomposites. The first one is observed at low energy in the visible region (between 550 and 655 nm) which is due to the absorption of γ-Fe₂O₃. The second absorption (at about 380 nm) is attributed to the absorption of ZnO.

Fig. 7

Diffuse reflectance spectra of ZnO nanorods, as-prepared γ-Fe₂O₃ nanoparticles (before calcination) and the ZnO@γ-Fe₂O₃ core–shell nanocomposites (ZF1–ZF4)

The absorption of ZnO@γ-Fe₂O₃ core–shell nanocomposites in the visible region increases with an increase in the concentration of precursor for γ-Fe₂O₃ (i.e. Fe(acac)₃) used during the synthesis of the nanocomposites. The band gap of ZnO nanorods, γ-Fe₂O₃ nanoparticles and ZnO@γ-Fe₂O₃ core–shell nanocomposites was estimated by plotting $\left( {\alpha h\nu } \right)^{2}$ versus $h\nu$ (Tauc plots) using the relation Chen et al. (2010b):

$$\alpha h\nu = A\left( {h\nu - E_{g} } \right)^{1/2}$$

where α is the absorption coefficient, $h\nu$ is the photon energy, E _g is the optical band gap energy and A is a constant. From the Tauc plots (Fig. S3), E _g was estimated by extrapolating a straight line from the linear portion of the absorption curve to the X-axis. The band gap values of pure ZnO nanorods and pure γ-Fe₂O₃ nanoparticles are 3.2 and 1.9 eV, respectively. Band gap values of 2.24, 2.2, 2.18 and 1.9 eV due to γ-Fe₂O₃ are observed for the ZnO@γ-Fe₂O₃ core–shell nanocomposites ZF1, ZF2, ZF3 and ZF4, respectively. The insets of Tauc plots (Fig. S3) of ZnO@γ-Fe₂O₃ core–shell nanocomposites (ZF1, ZF2, ZF3 and ZF4) show band gap of 3.3 eV due to ZnO. The band gap of γ-Fe₂O₃ in the core–shell nanocomposites shows red shift from ZF1 to ZF4. This red shift is attributed to an increase in particle size of γ-Fe₂O₃ with an increase in the precursor concentration of γ-Fe₂O₃ (i.e. Fe(acac)₃) used during the synthesis of the ZnO@γ-Fe₂O₃ nanocomposites (Hernández et al. 2007; Yin et al. 2014).

The UV–Visible absorption spectra of pure ZnO nanorods, iron oxide nanoparticles (as prepared by the thermal decomposition method, calcined at 400 °C and Fe₃O₄) and the ZnO@γ-Fe₂O₃ core–shell nanocomposite ZF1 are shown in Fig. S4. The spectra were recorded using dispersions in ethanol. The absorption spectrum of ZnO shows band edge absorption at about 376 nm (3.3 eV) which is blue shifted from that of bulk ZnO (3.2 eV) due to quantum confinement (Hernández et al. 2007). The as-prepared iron oxide nanoparticles (prepared by the thermal decomposition method) show absorption in the visible region (400–800 nm) with a broad band at about 440 nm which corresponds to that of γ-Fe₂O₃ nanoparticles (Wu and Wang 2014). The absorption spectrum of γ-Fe₂O₃ nanoparticles calcined at 400 °C shows absorption at about 575 nm characteristic of α-Fe₂O₃ (Pandey et al. 2014; Suresh et al. 2014). Fe₃O₄ nanoparticles show band edge absorption at about 503 nm (Wu and Wang 2014). The absorption spectrum of ZnO@γ-Fe₂O₃ core–shell nanocomposite ZF1 shows a small shoulder at about 565 nm and a band edge absorption at about 376 nm (3.3 eV) attributed to γ-Fe₂O₃ and ZnO, respectively (Hernández et al. 2007; Reda 2010; Zhang et al. 1997). The UV–Vis spectral studies also suggest that the phase of as-prepared iron oxide nanoparticles, prepared by the thermal decomposition method, is γ-Fe₂O₃.

Figure 8 shows the photoluminescence spectra of pure ZnO nanorods, γ-Fe₂O₃ nanoparticles prepared by the thermal decomposition method (before calcination) and the ZnO@γ-Fe₂O₃ core–shell nanocomposites (ZF1, ZF2, ZF3 and ZF4). The PL spectra of pure ZnO and ZnO@γ-Fe₂O₃ core–shell nanocomposites show emission bands around 382 and 469 nm. The emission band at about 382 nm is due to band-edge emission of ZnO. The broad emission centred at 469 nm, observed in ZnO nanorods, γ-Fe₂O₃ nanoparticles and all the core–shell nanocomposites (ZF1, ZF2, ZF3 and ZF4), is attributed to the presence of oxygen vacancies and defects (Liu et al. 2012). In the ZnO@γ-Fe₂O₃ core–shell nanocomposites, the defect emission at 469 nm is enhanced as compared to pure ZnO nanorods and γ-Fe₂O₃ nanoparticles. A possible reason is reduction in the number of hydroxyls groups on the surface of ZnO nanorods due to the formation of γ-Fe₂O₃ shell. Depletion of hydroxyls groups is known to increase oxygen vacancies and surface defects on the surface of ZnO nanorods (Chu et al. 2008). The oxygen vacancies/defects are good electron scavengers and thus expected to improve the photocatalytic activity (Abdullah Mirzaie et al. 2012). The defect emission decreases in the ZnO@γ-Fe₂O₃ core–shell nanocomposites from ZF1 to ZF4 with an increase in the concentration of Fe(acac)₃, used during the preparation of the core–shell nanocomposites. The presence of more and more γ-Fe₂O₃ nanoparticles on the surface of ZnO nanorods due to increase in the concentration of Fe(acac)₃, used during the synthesis, decreases surface defects and oxygen vacancies on the surface of ZnO nanorods (Wu et al. 2010).

Fig. 8

Photoluminescence spectra of ZnO nanorods, γ-Fe₂O₃ nanoparticles (prepared by the thermal decomposition method) and the ZnO@γ-Fe₂O₃ core–shell nanocomposites (ZF1, ZF2, ZF3 and ZF4)

Magnetic measurements were carried out for pure γ-Fe₂O₃ nanoparticles, prepared by the present thermal decomposition method, and the ZnO@γ-Fe₂O₃ core–shell nanocomposites. Magnetic measurements were done only for those nanocomposites which exhibited uniform elemental distribution (ZF1, ZF2 and ZF4). Nanocomposite ZF3 did not exhibit uniform elemental distribution (see Table 2 on EDXA results). The magnetization versus magnetic field (M–H) plots for pure γ-Fe₂O₃ nanoparticles and the core–shell nanocomposites (ZF1, ZF2 and ZF4) are shown in Fig. 9. All the samples exhibit superparamagnetic behaviour at 300 K. At 5 K, however, the samples exhibit weak ferromagnetic behaviour with small coercivity (H _c) and remanent magnetization (M _r). In general, for a superparamagnetic system, the coercivity and remanent magnetization increases below the superparamagnetic–ferromagnetic transition (Wilson et al. 2004; Zhang et al. 1997). The observed magnetic behaviour and magnetic parameters for pure γ-Fe₂O₃ nanoparticles and ZnO@γ-Fe₂O₃ core–shell nanocomposites are summarized in Table 3. The coercivity of pure γ-Fe₂O₃ nanoparticles (0.05 T), and the core–shell nanocomposites ZF2 (0.055 T) and ZF1 (0.05 T) are almost the same. The coercivity of core–shell nanocomposite ZF4 (0.045 T) is less than that of pure γ-Fe₂O₃, ZF1 and ZF2. This is attributed to an increase in the particle size of iron oxide nanoparticles in the nanocomposite ZF4 (Lin et al. 2006). Remanent magnetization values for the nanocomposites ZF1, ZF2 and ZF4 are 0.38, 0.88 and 0.66 emu/g, respectively. Remanent magnetization is lower in all the core–shell nanocomposites as compared to that of pure γ-Fe₂O₃ nanoparticles (2.7 emu/g). This is attributed to the presence of diamagnetic ZnO nanorods, which decreases the effective mass of iron oxide (Zhang et al. 2013). The unexpectedly high remanent magnetization in case of nanocomposite ZF2 is attributed to the formation of thicker shell of iron oxide (~20 nm) over ZnO nanorods as compared to the other nanocomposites (ZF1 and ZF4) (Ma et al. 2015). At 300 K, the saturation magnetization (M _s) values of the core–shell nanocomposites are less than that of γ-Fe₂O₃ nanoparticles and bulk maghemite (76 emu/g). The decrease of M _s in the core–shell nanocomposites is attributed to the presence of diamagnetic ZnO, surface spin disorder and canting (Wilson et al. 2004; Zhang et al. 2013). The saturation magnetization increases with an increase in the iron content of the nanocomposites ZF1–ZF4 (Table 3). In addition, there is an increase in the particle size of γ-Fe₂O₃ nanoparticles in the core–shell nanocomposites with an increase in the concentration of precursor for iron oxide (Fe(acac)₃) used during the synthesis (Kishore and Jeevanadam 2011). Figure 9 reveals that the magnetization at 5 K does not saturate for the as-prepared pure γ-Fe₂O₃ nanoparticles and also for the core–shell nanocomposites at 3 T. This is attributed to frozen spins on the surface as has been reported for maghemite nanoparticles (Balti et al. 2014; Nadeem et al. 2011).

Fig. 9

M–H curves of pure γ-Fe₂O₃ nanoparticles, prepared by thermal decomposition method, and the core–shell nanocomposites (ZF1, ZF2 and ZF4)

Table 3 Magnetic parameters for pure iron oxide nanoparticles and ZnO@γ-Fe₂O₃ core–shell nanocomposites

The zero-field-cooled (ZFC) and field-cooled (FC) magnetization curves for the as-prepared pure γ-Fe₂O₃ nanoparticles and for the ZnO@γ-Fe₂O₃ core–shell nanocomposites (ZF1, ZF2 and ZF4), under an applied field of 0.05 T, are shown in Fig. 10. Pure γ-Fe₂O₃ nanoparticles and the core–shell nanocomposites (ZF1, ZF2 and ZF4) show broad maxima in the ZFC curves (Table 3). The temperature corresponding to the broad maxima is called as the blocking temperature (T _B), and it corresponds to the superparamagnetic transition of iron oxide nanoparticles.

Fig. 10

ZFC and FC curves of pure γ-Fe₂O₃ nanoparticles, prepared by thermal decomposition method, and ZnO@γ-Fe₂O₃ core–shell nanocomposites (ZF1, ZF2 and ZF4)

At temperatures higher than T _B, the magnetization decreases and follows a Curie–Weiss law (Liu and Ma 2015). The observed blocking temperature for pure γ-Fe₂O₃ nanoparticles is 43 K and for the core–shell nanocomposites ZF1, ZF2 and ZF4, the T _B values are 30, 45 and 55 K, respectively. The increase in the blocking temperature in the nanocomposites ZF2 and ZF4 compared to ZF1 is attributed to an increase in dipolar interaction due to increase in the particle volume (Mikhaylova et al. 2004). The broad maxima in the ZFC curves of γ-Fe₂O₃ nanoparticles and the core–shell nanocomposites (ZF1, ZF2 and ZF4) indicate broad particle size distribution of iron oxide nanoparticles (Kishore and Jeevanadam 2011). The particle volume (V) can be calculated from T _B using the following equation (Laurent et al. 2008).

$$T_{B} = \frac{{K_{\text{eff}} V}}{{25k_{B} }}$$

where K _eff is the effective anisotropy constant (for bulk γ-Fe₂O₃, K _eff = 4.7 × 10⁴ erg cm⁻³) and k _B is the Boltzmann constant. The average particle sizes of γ-Fe₂O₃ calculated from the T _B of pure γ-Fe₂O₃ nanoparticles and the core–shell nanocomposites ZF1, ZF2, ZF4 are 18.2, 16.2, 18.5 and 19.8 nm, respectively. The ZFC curves of pure γ-Fe₂O₃ and nanocomposites ZF1, ZF2 and ZF4 also show Morin transition at about 266, 265, 266 and 260 K, respectively. The observation of Morin transition is attributed to the presence of a small amount of α-Fe₂O₃ nanoparticles due to the structural transformation of γ-Fe₂O₃ into α-Fe₂O₃ nanoparticles (Mashlan et al. 2004). In α-Fe₂O₃, Morin transition is a phase transition in which a weakly ferromagnetic state converts into an antiferromagnetic state (Mukherjee et al. 2008).

Figure S5 shows M–H curves for as prepared iron oxide nanoparticles prepared by the thermal decomposition method followed by calcination at 400 °C (i.e. α-Fe₂O₃), ZnO@γ-Fe₂O₃ core–shell nanocomposite ZF1 calcined at 400 °C, and Fe₃O₄ nanoparticles prepared by the reported method (before calcination). The observed magnetic parameters such as coercivity (H _c), remanent magnetization (M _r) and saturation magnetization (M _s) have been summarized in Table 4. Pure γ-Fe₂O₃ nanoparticles calcined at 400 °C show weak ferromagnetic behaviour at both 300 and 5 K with H _c values 0.016 and 0.65 T, respectively. The magnetization does not saturate at both the temperatures at a maximum field of 3 Tesla. The magnetic behaviour of γ-Fe₂O₃ nanoparticles calcined at 400 °C is similar to that reported for α-Fe₂O₃ nanoparticles (Bødker et al. 2000; Jayanthi et al. 2015; Pandey et al. 2014). The core–shell nanocomposite ZF1 (after calcination at 400 °C) shows superparamagnetic behaviour at room temperature. At 5 K, it shows weak ferromagnetic behaviour with a H _c value of 0.1 T. The M–H curves of Fe₃O₄ nanoparticles show superparamagnetic behaviour at room temperature and weak ferromagnetic behaviour at 5 K with H _c and M _r values of 0.48 T and 21.4 emu/g, respectively. The saturation magnetization values for Fe₃O₄ nanoparticles at 300 and 5 K are 73 and 89 emu/g, respectively.

Table 4 Magnetic parameters of as prepared iron oxide nanoparticles prepared by thermal decomposition method, followed by calcination at 400 °C, Fe₃O₄ nanoparticles and core–shell nanocomposite ZF1 (calcined at 400 °C)

The ZFC/FC curves of γ-Fe₂O₃ nanoparticles, prepared by the thermal decomposition method, followed by calcination at 400 °C, ZnO@γ-Fe₂O₃ core–shell nanocomposite ZF1 calcined at 400 °C, and Fe₃O₄ nanoparticles are shown in Fig. S6. The ZFC curve for γ-Fe₂O₃ nanoparticles calcined at 400 °C (Fig. S6a) shows a broad maximum around 200 K, corresponding to the blocking temperature. The FC curve departs from the ZFC curve at T _irr (blocking temperature of the largest particle) and the magnetization remains constant with decrease in temperature and this is due to strong inter-particle interaction (Sreeja and Joy 2007). The ZFC curve for the ZnO@γ-Fe₂O₃ core–shell nanocomposite ZF1 calcined at 400 °C (Fig. S6b) shows peak at about 52.5 K indicating the superparamagnetic blocking transition. Also, the ZFC curve shows a small hump at 262 K due to Morin transition indicating the formation of α-Fe₂O₃. Pure γ-Fe₂O₃ nanoparticles prepared by the thermal decomposition method and calcined at 400 °C do not show Morin transition. According to Zysler et al. (Zysler et al. 2001), when iron oxide is prepared in the form of small particles, Morin transition is a complicated phenomenon due to the presence of strain and defects. The Morin transition temperature reduces as the particle size decreases and vanishes for iron oxide particles smaller than 8-20 nm (Amin and Arajs 1987). The particle size values of iron oxide nanoparticles, prepared by the thermal decomposition method followed by calcination at 400 °C, and the ZnO@γ-Fe₂O₃ core–shell nanocomposite ZF1 calcined at 400 °C, estimated from their blocking temperatures, are 19.5 nm and 30.4 nm, respectively.

The ZFC curve for Fe₃O₄ nanoparticles (Fig. S6c) shows a broad maximum around 39 K due to superparamagnetic blocking temperature. The particle size of Fe₃O₄ nanoparticles, calculated from the blocking temperature, is 17.6 nm. In general, the saturation magnetization at room temperature reported for γ-Fe₂O₃ nanoparticles is between 1.3 and 69.6 emu/g (Balti et al. 2014; Chu et al. 2008; Morales et al. 1999; Wu et al. 2010; Zhang et al. 2013) and for Fe₃O₄ nanoparticles, the M _s value is between 20 and 84 emu/g (Chen et al. 2010b; Chu et al. 2008; Si et al. 2005). The reported coercivity at 5 K for γ-Fe₂O₃ nanoparticles is between 0.005 and 0.413 T (Tsuzuki et al. 2011; Zysler et al. 2001) and for Fe₃O₄ nanoparticles, the value is between 0.01 and 0.058 T (Saffari et al. 2015; Wu et al. 2015). The reported remanent magnetization values at 5 K for γ-Fe₂O₃ nanoparticles are between 0.62 emu/g and 16.7 emu/g (Woo et al. 2004) and for Fe₃O₄ nanoparticles, the values are between 13 and 18.8 emu/g (Jayanthi et al. 2015; Woo et al. 2004). From the above results, one can conclude that there is difference in the magnetic properties of γ-Fe₂O₃ and Fe₃O₄ nanoparticles. In spite of having nearly the same particle size for the as-prepared iron oxide nanoparticles prepared by the thermal decomposition method (18.2 nm) and Fe₃O₄ nanoparticles (17.6 nm), there is difference in their magnetic behaviour. For the as-prepared iron oxide nanoparticles, the coercivity (H _c  = 0.05 T) and remanent value (M _r = 2.7 emu/g) at 5 K and saturation magnetization (M _s = 7.8 emu/g) at room temperature are less than that of Fe₃O₄ nanoparticles (H _c = 0.48 T, M _r = 21.4 emu/g and M _s = 73 emu/g). This suggests that the phase of as-prepared iron oxide nanoparticles, prepared by the present thermal decomposition method, is γ-Fe₂O₃.

ZnO nanorods possess hydroxyl groups on their surface (Chandraiahgari et al. 2015). During the formation of ZnO@γ-Fe₂O₃ core–shell nanocomposites, iron (III) acetylacetonate interacts with the surface of ZnO via hydrogen bonding between the π-system of acetyl acetone and protons of the hydroxyl groups present on the surface of ZnO (Scheme 1). Iron (III) acetylacetonate decomposes in air, at about 200 °C, to produce iron oxide nanoparticles (maghemite) with by-products (Kishore and Jeevanandam 2012).

$$4\left( {{\text{C}}_{5} {\text{H}}_{7} {\text{O}}_{2} } \right)_{3} {\text{Fe}} + 17{\text{O}}_{2} \,\xrightarrow{{\sim200\,{^\circ }{\text{C}}}}\,2\gamma {\text{-Fe}}_{2} {\text{O}}_{3} + 30{\text{CO}} + 10{\text{CO}}_{2} + 20{\text{CH}}_{4} + 2{\text{H}}_{2} {\text{O}}$$

Since the thermal decomposition is carried out at 200 °C in diphenyl ether in the presence of ZnO nanorods, ZnO@γ-Fe₂O₃ core–shell nanocomposites are produced.

Scheme 1

Attachment of iron (III) acetylacetonate (Fe(acac)₃) on the surface of ZnO nanorods and its further decomposition to iron oxide nanoparticles

In order to study the photocatalytic activity of ZnO nanorods, γ-Fe₂O₃ nanoparticles and the ZnO@γ-Fe₂O₃ core–shell nanocomposites, photodegradation of congo red in an aqueous solution under sunlight was carried out. Congo red (C₃₂H₂₂N₆O₆S₂Na₂) is an anionic azo dye used widely in printing, textile, and photographic industries (Liu et al. 2015a, b). The residual congo red in water can cause serious environmental hazard and health problems (Abdullah Mirzaie et al. 2012). Hence, it is necessary to remove congo red from waste water by an efficient and low-cost process. Due to its complex aromatic structure, congo red is stable and it is hard to remove it by traditional methods of waste water treatment. The UV–Visible absorption spectrum of congo red is characterized by a band at 495 nm due to the azo chromophore which leads to its dark red colour. The other two bands at 237 and 343 nm are attributed to benzene and naphthalene rings, respectively (Wang et al. 2008). Figure 11a shows the UV–Visible spectral results on the photodegradation of congo red using pure ZnO nanorods, γ-Fe₂O₃ nanoparticles (before calcination) and all the nanocomposites (ZF1–ZF4) as the catalyst. It can be seen that ZnO@γ-Fe₂O₃ nanocomposite ZF1 shows maximum photodegradation compared to pure ZnO nanorods, γ-Fe₂O₃ nanoparticles and all the other three nanocomposites (ZF2, ZF3 and ZF4). Figure 11b shows the effect of varying the amount of nanocomposite ZF1 on the photocatalytic degradation of congo red. The amount of catalyst was varied from 20 to 100 mg in 50 mL of congo red solution (50 mg L⁻¹). 50 and 100 mg of the catalyst show almost the same extent of decrease in the concentration of congo red and hence, 50 mg of catalyst (ZF1) in 50 mL of congo red solution was optimized as the best condition to carry out further photocatalytic degradation experiments.

Fig. 11

a UV–Visible spectral results on the photocatalytic degradation of congo red in an aqueous solution by ZnO nanorods, γ-Fe₂O₃ nanoparticles and ZnO@γ- Fe₂O₃ core–shell nanocomposites (ZF1, ZF2, ZF3 and ZF4) and b UV–Visible spectral results on the photocatalytic degradation of congo red using different amounts of ZnO@γ-Fe₂O₃ core–shell nanocomposite ZF1 as the catalyst

Figure 12 shows the kinetics of photocatalytic degradation of congo red by ZnO, γ-Fe₂O₃ nanoparticles and the ZnO@γ-Fe₂O₃ core–shell nanocomposites (ZF1, ZF2, ZF3 and ZF4). In the case of ZnO@γ-Fe₂O₃ core–shell nanocomposite ZF1, the intensity of absorption band of congo red at 495 nm decreases with time and almost becomes zero in about 90 min. On the other hand, pure ZnO, γ-Fe₂O₃ nanoparticles and the other ZnO@γ-Fe₂O₃ nanocomposites (ZF2, ZF3 and ZF4) are unable to complete the degradation during this period. Figure 13a shows comparison of kinetic results on photodegradation of congo red using ZnO, γ-Fe₂O₃ nanoparticles and the ZnO@γ-Fe₂O₃ core–shell nanocomposites as the catalyst where C _o is the initial concentration of congo red and C is its concentration at different times. Nanocomposite ZF1 shows better activity than that of pure ZnO nanorods, γ-Fe₂O₃ nanoparticles and all the other (ZF2, ZF3 and ZF4) nanocomposites. This is because it exhibits effective charge separation, more surface area (The BET surface area values for ZF1, ZF2, ZF3 and ZF4, ZnO and γ-Fe₂O₃ are 62.1, 53.3, 50.3, 38.2, 24.4, and 51.1 m²/g, respectively) and more number of oxygen vacancies are present on its surface. As elucidated by PL studies in Fig. 8, nanocomposite ZF1 has enhanced defect emission compared to pure ZnO nanorods, γ-Fe₂O₃ nanoparticles and all the other ZnO@γ-Fe₂O₃ nanocomposites (ZF2, ZF3 and ZF4). The photocatalytic activity of metal oxides increases with an increase in the amount oxygen vacancies. The defects and vacancies act as electron acceptors and trap the photogenerated electrons to reduce the surface recombination of electrons and holes (Wang et al. 2009).

Fig. 12

Kinetics of photodegradation of congo red in aqueous solutions as indicated by UV–Visible spectroscopy using pure ZnO nanorods, γ-Fe₂O₃ nanoparticles and ZnO@γ-Fe₂O₃ core–shell nanocomposites ZF1, ZF2, ZF3 and ZF4 as the catalyst

Fig. 13

a Comparison of kinetics of photocatalytic degradation of congo red using ZnO nanorods, γ-Fe₂O₃ nanoparticles and ZnO@γ-Fe₂O₃ core–shell nanocomposites (ZF1–ZF4) as the catalyst (C _o is the initial concentration and C is the concentration of dye at different times), and b pseudo-first-order kinetics plots for congo red photodegradation. Symbols are experimental values and solid lines are the fits obtained using pseudo first-order kinetic model

The photocatalytic activity of ZnO@γ-Fe₂O₃ core–shell nanocomposites decreases with an increase in the concentration of Fe(acac)₃ used during the synthesis of the nanocomposites. The ZnO@γ-Fe₂O₃ core–shell nanocomposite with less amount of γ-Fe₂O₃ nanoparticles (ZF1) shows enhanced photocatalytic activity. It is proposed that presence of higher γ-Fe₂O₃ content (e.g. ZF2, ZF3, and ZF4) causes reduction in the distance between photoelectrons and holes, which results in decrease in the photocatalytic activity due to enhancement in the recombination (Abdullah Mirzaie et al. 2012; Hernández et al. 2007). ZnO@γ-Fe₂O₃ core–shell nanocomposites with lower concentration of Fe(acac)₃ (0.15 and 0.1 M) were prepared and their photocatalytic activity was tested and compared with that of nanocomposite ZF1 (prepared using 0.25 M of Fe(acac)₃). It was found that nanocomposite ZF1 possesses better activity than the nanocomposites prepared using lower concentrations of Fe(acac)₃ (0.15 M and 0.1 M). It is proposed that when the amount of γ-Fe₂O₃ in the nanocomposites is too low, an efficient interfacial electron transfer from γ-Fe₂O₃ to ZnO does not occur. Nanocomposite ZF1 contains an optimum amount of γ-Fe₂O₃ and exhibits maximum degradation efficiency due to the sensitization of ZnO with γ-Fe₂O₃ with effective visible-light absorption and charge separation at the interface.

In general, the photodegradation of congo red catalyzed by the semiconductors follows pseudo-first-order rate law (Wu et al. 2012).

$$\ln \left[ {\frac{{{\text{C}}_{\text{o}} }}{C}} \right] = k_{1} t$$

where C _o and C represent the concentration of congo red at t = 0 and at any time t and k ₁ is the apparent first-order rate constant. Figure 13b shows the fitting of kinetics data using pseudo-first-order rate law for the photodegradation of congo red using ZnO nanorods, γ-Fe₂O₃ nanoparticles and ZnO@γ-Fe₂O₃ core–shell nanocomposites (ZF1–ZF4) as the catalysts. The kinetic parameters obtained using the pseudo-first-order kinetic model are summarized in Table 5. The pseudo-first-order rate constant of congo red degradation using ZnO@γ-Fe₂O₃ core–shell nanocomposite ZF1 is 0.022 min⁻¹ which is higher than the values for pure ZnO (0.005 min⁻¹), pure γ-Fe₂O₃ (0.006 min⁻¹) and other ZnO@γ-Fe₂O₃ core–shell nanocomposites (ZF2 (0.019 min⁻¹), ZF3 (0.014 min⁻¹) and ZF4 (0.012 min⁻¹). The enhanced activity of ZnO@γ-Fe₂O₃ core–shell nanocomposites compared to ZnO and γ-Fe₂O₃ is attributed to faster charge separation at the interface of ZnO@γ-Fe₂O₃ core–shell nanocomposites.

Table 5 Estimated pseudo-first-order kinetic parameters for the photocatalytic degradation of congo red using pure ZnO, γ-Fe₂O₃ nanoparticles and ZnO@γ-Fe₂O₃ core–shell nanocomposites as the catalysts

The ability of reusing the ZnO@γ-Fe₂O₃ nanocomposite ZF1 was evaluated by performing repeated photocatalytic degradation experiments using congo red solution (50 mg L⁻¹). The suspension containing the catalyst was first irradiated in sunlight for 90 min. The photocatalyst was recovered from the suspension after every cycle by centrifugation and subsequent washing with double distilled water for three times. The recovered catalyst was dried in vacuum for overnight and was reused for the photodegradation. The degradation efficiency of the ZnO@γ-Fe₂O₃ core–shell nanocomposite ZF1 decreases from about 97 to 85 % after four cycles. Magnetic parameters are important for the nanocomposites since they will lead to development of nanocatalysts which are magnetic too. In liquid phase reactions, magnetic separation is an useful approach for removing and recycling catalysts by applying an external magnetic field (Xuan et al. 2009; Ye et al. 2010). A photograph showing magnetic separation of nanocomposite ZF1 (which is superparamagnetic) by applying an external magnetic field from an aqueous suspension, is shown in Fig. S7. The mechanism of photodegradation of congo red by ZnO@γ-Fe₂O₃ nanocomposites is discussed below.

ZnO has absorption in the UV region due to its wide band gap (3.28 eV) and γ-Fe₂O₃ has absorption in the visible region due to its narrow band gap (1.9–2.2 eV) (Yin et al. 2014). The absorption spectra of ZnO@γ-Fe₂O₃ core–shell nanocomposites have contribution from both the components. All the core–shell nanocomposites (ZF1–ZF4) show better photocatalytic efficiency compared to pure ZnO nanorods and γ-Fe₂O₃ nanoparticles. The detailed positions of energy bands for ZnO and γ-Fe₂O₃ are shown in Scheme 2 (Hsu et al. 2015; Kaneti et al. 2014; Pradhan et al. 2012; Wu et al. 2012; Zhang et al. 2015). The presence of γ-Fe₂O₃ nanoparticles on the surface of ZnO nanorods leads to formation of a heterojunction barrier at the interface. This is due to the difference in their work functions (5.0 eV for ZnO and 5.88 eV for γ-Fe₂O₃) (Kaneti et al. 2014). Since the work function of γ-Fe₂O₃ is higher than that of ZnO, the Fermi level of γ-Fe₂O₃ is lower than that of ZnO. Transfer of electrons occurs from the conduction band of γ-Fe₂O₃ to that of ZnO to equalize the Fermi energy levels which would make the conduction band of ZnO lower than that of γ-Fe₂O₃ (Hsu et al. 2015; Zhang et al. 2011, 2015). The conduction band and valence band potentials of ZnO shift to 0.32 and 3.62 eV, respectively. On sunlight irradiation, electrons in the valence band (VB) of γ-Fe₂O₃ and ZnO are excited to their conduction bands (CB) and at the same time, equal amount of holes are left in the VB. The photogenerated electrons in the CB of γ-Fe₂O₃ are transferred to that of ZnO, while the photogenerated holes in VB of ZnO are transferred to that of γ-Fe₂O₃. As a result, separation of the photogenerated electrons and holes are achieved at the ZnO@γ-Fe₂O₃ interface. This would reduce recombination of electrons and holes and enable them to move to the surface of γ-Fe₂O₃ and ZnO, respectively. The electrons on the surface of ZnO react with dissolved O₂ to give super oxide radical anions (O ^·−₂ ). The superoxide radicals hydrolyze to form hydroperoxyl radicals (${\text{HO}}_{2}^{ \cdot }$) which undergo subsequent hydrolysis to form OH^· radicals (Wu et al. 2012). The photogenerated holes in the VB of γ-Fe₂O₃ cannot produce OH^· radicals by oxidizing H₂O since the VB potential of γ-Fe₂O₃ (+2.48 eV vs. NHE) is higher than E(OH^·/H₂O) (+2.68 eV vs. NHE) (Kumar et al. 2013). The photogenerated holes directly oxidize the congo red (Chen et al. 2010b; Konstantinou and Albanis 2004). The photocatalytic mechanism is summarized as follows.

$${\text{ZnO}}@\gamma{\text{-}} {\text{Fe}}_{2} {\text{O}}_{3} + {\text{ h}}\nu \to \left( {\text{ZnO}} \right)e^{- } + \, (\gamma - {\text{Fe}}_{ 2} {\text{O}}_{3} ){\text{h}}^{ + }$$

$$\left( {\text{ZnO}} \right){\text{e}}^{ - } + {\text{O}}_{2} \to {\text{O}}_{2}^{ \cdot - }$$

$${\text{O}}_{2}^{ \cdot - } + {\text{H}}_{2} {\text{O}} \to {\text{HO}}_{2}^{ \cdot } + {\text{HO}}^{ - }$$

$${\text{HO}}_{2}^{ \cdot } + {\text{H}}_{2} {\text{O}} \to {\text{H}}_{2} {\text{O}}_{2} + {\text{OH}}^{ \cdot }$$

$${\text{H}}_{2} {\text{O}}_{2} \to 2{\text{OH}}^{ \cdot }$$

$${\text{Congo red }} + {\text{h}}^{ + } , {\text{O}}_{2}^{ \cdot - } , {\text{OH}}^{ \cdot } \to {\text{degradation products}}$$

Scheme 2

Schematic diagram of band configuration and electron–hole separation at the interface between ZnO nanorods and γ-Fe₂O₃ nanoparticles under sunlight irradiation

To verify the role of various oxidizing species (h⁺, OH^· and O ^·−₂ ) in the photocatalytic degradation of congo red, different scavengers were added to the congo red solution before addition of the catalyst (nanocomposite ZF1). The scavengers used were ammonium oxalate (AO), benzoquinone (BQ), and tertiary butyl alcohol (t-BuOH) as h⁺, O ^·−₂ and OH^· scavengers, respectively (Wang et al. 2011; Yang et al. 2016; Zhiyong et al. 2015). About 0.07 g of solid AO, 0.09 g of BQ, and 0.48 mL of t-BuOH were added to 50 mL of congo red aqueous solution (50 mg L⁻¹) to make a scavenger concentration of 10 mM in each case. The photodegradation results after the introduction of scavengers are shown in Fig. S8. It can be seen from Fig. S8, when AO (holes scavenger) was added to the suspension, the congo red degradation is considerably suppressed as compared to that in the absence of AO, confirming h⁺ as active species in the degradation. On the addition of BQ (O ^·−₂ scavenger), the congo red degradation is again suppressed to a significant extent. However, the addition of t-BuOH (OH^· scavenger) does not affect the degradation much. On the basis of the above results, it is proposed that h⁺ and O ^·−₂ play important role in the photocatalytic degradation of congo red in an aqueous solution over ZnO@γ-Fe₂O₃ core–shell nanocomposites as the catalyst in the presence of sunlight.

Recently, many authors have reported photodegradation of congo red in the presence of different nanocomposites. Abdullah Mirzaie et al. have reported photodegradation of congo red in the presence of ZnO–Fe₂O₃ nanocomposites (Abdullah Mirzaie et al. 2012). The nanocomposite having (Fe³⁺/Zn²⁺) molar ratio of 1/100 shows the highest photocatalytic activity of about 93 % decolorization in 25 min compared to other nanocomposites (Fe³⁺/Zn²⁺ = 0.5/100, 5/100 and 10/100). Narayan et al. have reported photocatalytic degradation of congo red (25 mg L⁻¹) in the presence of Y³⁺ doped TiO₂ nanocomposites. They observed maximum rate constant of 0.027 min⁻¹ for nanocomposite TiO₂(Y₂O₃)_0.1 with 91 % of degradation in about 180 min (Narayan et al. 2009). Pouretedal and Keshavarz have reported photodegradation of congo red (5 mg L⁻¹) in the presence of Zn_1−x Cu_xS nanocomposites. They observed rate constant of 0.033 min⁻¹ for nanocomposite Zn_0.9Cu_0.1S with 98 % degradation in about 120 min (Pouretedal and Keshavarz 2010). Jiang et al. have reported photodegradation of congo red (20 mg L⁻¹) in the presence of γ-Fe₂O₃-CdS nanocomposite. They observed rate constant of 0.0023 min⁻¹ with 91.5 % of degradation in about 300 min (Jiang et al. 2014). Guo et al. have reported photodegradation of congo red (50 mg L⁻¹) in the presence of TiO₂-PANI nanocomposites. They observed maximum rate constant of 0.0007 min⁻¹ for nanocomposite (1:1) with 95 % photodegradation in about 160 min (Guo et al. 2014). Farbod et al. have reported photodegradation of congo red (20 mg L⁻¹) in the presence of Gd-TiO₂ nanocomposites. They observed rate constant of 0.034 min⁻¹ for the nanocomposite Gd(1.8 %)-TiO₂ with complete photodegradation in about 180 min (Farbod and Kajbafvala 2013). Lin et al. have reported photodegradation of congo red in the presence of ZnO-Zn nanocomposite with cellulose (ZnO/Zn)–C and starch (ZnO/Zn)–S. They observed rate constant of 0.0105 min⁻¹ for ( ZnO/Zn)–S with 98 % of photodegradation in about 3 h (Lin et al. 2014). Thiripuranthagan et al. have reported photodegradation of congo red (50 mg L⁻¹) in the presence of Ag impregnated TiO₂–SiO₂ composites. They observed rate constant of 0.0173 min⁻¹ for TiO₂–SiO₂ impregnated with 1 % Ag with 85 % of photodegradation in about 180 min (Thiripuranthagan et al. 2015). Liu et al. have reported photodegradation of congo red (25 mg L⁻¹) in the presence of Ag-ZnO nanocomposites. They observed a rate constant of 0.0312 min⁻¹ for 3 % Ag-ZnO nanocomposite with 97.3 % of photodegradation in about 50 min (Liu et al. 2015). Li et al. have reported photodegradation of congo red (10 mg L⁻¹) in the presence of TiO₂/PCNFs (porous carbon nanofibers). They observed a rate constant of 0.0158 min⁻¹ for nanocomposite TiO₂/PCNFs-4 (0.25 mL of TiCl₄) for complete photodegradation in about 60 min (Li et al. 2015). In the present work, complete photodegradation of congo red takes place in about 90 min with a rate constant of 0.022 min⁻¹ by the ZnO@γ-Fe₂O₃ nanocomposite ZF1.

Conclusions

ZnO@γ-Fe₂O₃ core–shell nanocomposites were synthesized by a novel thermal decomposition approach. The nanocomposites were characterized with various instrumental techniques. XRD results confirm the presence of wurtzite ZnO and studies on as-prepared and calcined samples suggest the phase of as-prepared iron oxide nanoparticles as γ-Fe₂O₃. FT-IR results on ZnO@γ-Fe₂O₃ core–shell nanocomposites show characteristic IR bands due to γ-Fe₂O₃ and FE-SEM images show formation of uniform iron oxide shell on the ZnO nanorods. Transmission electron microscopy studies on the ZnO@γ-Fe₂O₃ core–shell nanocomposites show uniform shell (10 nm-20 nm) of γ-Fe₂O₃ nanoparticles coated on the ZnO nanorods. HRTEM observations demonstrate that the constituents of the nanocomposites are crystalline with characteristics lattice spacing due to γ-Fe₂O₃ (0.48 nm) and ZnO (0.24 nm). Diffuse reflectance spectra of ZnO@γ-Fe₂O₃ core–shell nanocomposites indicate extended optical absorption in the visible range of 400–600 nm. The core–shell nanocomposites show red shift of band gap absorption due to iron oxide from 2.24 to 1.9 eV with an increase in the precursor concentration of γ-Fe₂O₃. The photoluminescence spectral studies indicate that the ZnO@γ-Fe₂O₃ nanocomposites exhibit enhanced defect emission and the ZnO@γ-Fe₂O₃ core–shell nanocomposites exhibit superparamagnetic behaviour at room temperature with maximum saturation magnetization of 2.1 emu/g. The ZnO@γ-Fe₂O₃ core–shell nanocomposites show better photocatalytic efficiency towards degradation of congo red in aqueous solutions compared to ZnO nanorods and γ-Fe₂O₃ nanoparticles. The enhancement in activity is attributed to stronger visible-light absorption and effective charge separation at the interface of ZnO@γ-Fe₂O₃ core–shell nanocomposites.