Synthesis and characterization of mesoporous α-Fe2O3 nanoparticles and investigation of electrical properties of fabricated thick films

In this work,α-Fe2O3 nanoparticles (NPs) have been synthesized by using a simple Pechini sol-gel method from iron nitrate, citric acid as complexing agent and ethylene glycol as polymerization agent. The calcined α-Fe2O3 NPs were fully characterized by different techniques. It was confirmed that ultrafine and highly crystalline αFe2O3 NPs with high purity and mesoporous nature can be obtained after calcination at 550 °C for 3 h. In addition, the results of electrical resistance measurements of the fabricated Fe2O3 thick films showed that αFe2O3 thick films have stable electrical properties which are beneficial for electrical applications such as gas sensing and field effect transistors.

Since most of these reported methods either demand sophisticated instrumentation or require elevated temperature, the preparation of α-Fe 2 O 3 NPs at low cost at the industrial scale is a challenge.Therefore, using the cheap starting materials, simple fabrication processes and suitable conditions of synthesis are the main requirement for low cost/facile synthesis procedures.In this context, it is practically significant to obtain α-Fe 2 O 3 NPs using a facile, cheap, and effective way such as Pechini sol-gel method.Pechini et al. [19] introduced a polymerizable complex route, referred to as Pechini method, to serve as an important supplement for the conventional sol-gel method and due to its cheapness, handiness, good stoichiometric control, low toxicity, low cost and insensitivity to water existence, this method has been used rather extensively in synthesis of different materials.Moreover the highly branched polymer can reduce the cation mobility during heat treatment, so that the product with a good dispersion could be prepared [20].In a typical Pechini process, common metal salts are dissolved in a solution using an α-hydroxy carboxylic acid, such as citric acid (CA) as chelating agents of metal ions, then a poly hydroxyl alcohol, such as ethylene glycol (EG) or polyethylene glycol (PEG) is introduced as a crosslinking agent to form a polymeric resin on molecular level, thus, segregation of particular metal ions is reduced and compositional homogeneity is guaranteed.
The goal of this work is to synthesize mesoporous iron oxide by a Pechini route using iron nitrate as precursor, citric acid as chelating agent and ethylene glycol as crossing-linking agent.Afterwards, the aim is to fully characterize synthesized powders using different techniques and finally to study electrical properties and effect of UV light on the electrical resistance of mesoporous α-Fe 2 O 3 NPs.

Synthesis of α-Fe 2 O 3 NPs
Hydrated iron nitrate (Fe(NO 3 ) 3 ×9 H 2 O), citric acid (C 6 H 8 O 7 ×H 2 O), poly(vinyl pyrrolidone) and ethylene glycol (C 2 H 6 O 2 ) were purchased from Merck.All the chemicals were used as received and without further purification.Double distilled water was used to prepare precursor solutions.First, Fe(NO 3 ) 3 ×9 H 2 O was dissolved in distilled water at 70 °C for 1 h under magnetic stirring to make 0.5 M Fe +3 solution.In a separate beaker citric acid (CA) was dissolved in distilled water at 70 °C for 30 min.Subsequently, the CA solution was added slowly to the Fe +3 solution with stirring ([CA]/[Fe +3 ] = 2).Then the esterification agent, i.e. ethylene glycol (EG), was added with a molar ratio [CA]/[EG] = 2, while stirring and heating the solution.The final solution was reflux at 120 °C for 2 h.The generated precursor resin was dried at 120 °C for 12 h to obtain the precursor powders.Finally, the amorphous powders were calcined at 550 °C in air atmosphere for 3 h using a muffle furnace to obtain iron oxide NPs.

Characterization of synthetized nanopowders
Characterization of α-Fe 2 O 3 NPs was achieved by different techniques.X-ray diffraction analysis (XRD) was performed using Philips X-Pert diffractometer operating with monochromatic CuKα (λ = 1.54056Å) radiation at 40 kV and 30 mA.The XRD pattern was recorded at room temperature at a scan rate 0.05°/sec and 2θ from 10°to 70°.Morphological analysis was carried out using scanning electron microscopy (SEM) by using JEOL 5600 LV instrument operating with a 20 kV accelerating voltage and equipped with EDS detector (Oxford instruments).Morphological studies of synthesized powders were performed with JEOL JEM 2010 electron microscope (LaB 6 electron gun) operat-ing at 200 kV and equipped with Gatan 794 Multi-Scan CCD camera.Room-temperature micro-Raman spectroscopy (µRS) was used to get information about the material quality, the phase and purity.Raman spectrum was recorded at room temperature using micro-Raman spectrometer (Renishaw, Great Britain) in the back scattering geometry with 514.5 nm Ar + laser as an excitation source and data were collected in the range of 150-1550 cm -1 .The molecular structure of iron oxide NPs was studied by using Fourier transform infrared spectrometer (Nicolet FTIR Impact 400D), with the frequency ranging from 400-3500 cm -1 .Optical absorbance of iron oxide was recorded with a 1 cm path length quartz cell using a UV-Vis spectrophotometer as a function of wavelength in the range from 200 to 800 nm.X-ray photoelectron spectroscopy (XPS) was employed to determine the surface component and composition of the prepared nanoparticles.XPS was performed on ESCALAB MKII X-ray photoelectron spectrometer (VG Instruments, CA, USA), using non-monochromatized Mg-Kα X-rays as the excitation source.The binding energies for the samples were calibrated by setting the measured binding energy of C1s to 284.60 eV.Thermal analysis was carried out on Mettler Toledo thermal gravimetric analyser and Perkin Elmer differential thermal analyser.The analysis was performed in air flow, in the temperature range between 25 and 600 °C, with heating rate of 10 °C/min.Brunauer-Emmett-Teller (BET) surface area and pore size distribution of the samples were measured by nitrogen adsorption-desorption analyser (Quantachrome Autosorb-1C).All the particles were degassed under vacuum at 150 °C for 2 h before the measurements, and the nitrogen gas adsorption at 77 K was measured using a continuous flow method with nitrogen and helium mixture as the carrier gas.Specific surface area (SBET) was calculated by using the BET method, and pore size distribution was determined by applying the Barrett-Joyne-Halenda (BJH) model.

Electrical measurements
For electrical measurements, α-Fe 2 O 3 thick films were prepared by drop cast method on the Pt-electroded alumina substrate equipped with Pt heater and electrical resistance was studied by means of Agilent 34970A multimeter using the two point method.

Mechanisms in Pechini sol-gel preparation
The Pechini method used in the present work allowed obtaining of the homogeneous metal-organic polymer in which the iron ions were randomly distributed along the backbone of the polymer.During the heat treatment at 120 °C, this polymeric precursor method goes through two main stages: the first one is the reaction mechanism of esterification of citric acid with ethylene glycol as indicated in Fig. 1a, and the second one, the chelation of the iron nitrate by the formed esters as shown in Fig. 1b.Citric acid and ethylene glycol generate a reaction of esterification, and the formed ester acts as complexing agent of the iron cations forming a polymer network that is resin-like or solid as viscosity increases.At this point, it is necessary for the exchange reaction of the iron nitrate with the acid group to proceed until its completion.If this exchange reaction is not driven to completion, precipitation of unreacted iron nitrate would occur in the resin on cooling and this may lead to gross segregation of the iron in the ceramic.In the considered system, the polymer solidified to a clear resin on cooling, implying that all of the iron nitrate had reacted with the polymer [21].Figure 2 shows Fe +3 -citric acid complex and polyesterification process.

Structural and morphological studies
Figure 3 shows thermo-gravimetric (TGA) and differential thermal analysis (DTA) of the dried iron oxide gel.According to TGA curve of amorphous iron oxide, decomposition proceeds with thermal treatment through three steps.Initially, at a temperature below 100 °C, a small weight loss is observed, probably due to the evaporation of humidity adsorbed during storage is observed.Two other large weight losses are observed at temperatures higher than 150 °C, starting at about 185 and 250 °C, respectively.These are due to the decomposition of organic species from citric acid and ethylene glycol.No further weight loss is observed above 350 °C, thus indicating that the organic material was completely On the basis of these experimental evidences and literature data, the peak at 220 °C is related to the transformation of amorphous phase to crystalline γ-Fe 2 O 3 , whereas the second peak at 350 °C is likely attributed to its transformation to α-Fe 2 O 3 [23].
Figure 4 shows the XRD pattern of the calcined iron oxide powder.All the diffraction peaks can be assigned to the hexagonal α-Fe 2 O 3 phase and are consistent with those reported in the literature and with the respective JCPDS card no.89-596.
The lattice constants a and c of α-Fe 2 O 3 were evaluated according to Bragg law: where n is the order of diffraction, λ is the X-ray wavelength, and d is the spacing between planes of given Miller indices h, k, and l.In the hexagonal structure of α-Fe 2 O 3 , the plane spacing is related to the lattice constant a, c, and Miller indices by the following formulas: 1 According to the above formulas, for the (110) plane with diffraction peak at 2θ = 35.64°, the lattice constant a was calculated by: and for (104) plane with diffraction peak at 2θ = 33.15°, the lattice constant c was calculated by: The lattice parameters for the hexagonal phase of α-Fe 2 O 3 using Eq. ( 5) have been calculated to be a = 0.50 nm and c = 1.39 nm, which are in close agreement with the reported data.
Figure 5a shows TEM micrograph of iron oxide powder calcined at 550 °C for 3 h.Very fine particles with sizes between 10-30 nm are observed.Figure 5b shows SEM micrograph of α-Fe 2 O 3 thick film on the alumina substrate.As clearly observed, it has quite porous structure that renders a high surface area for these particles.Porous structure is the result of the burning of polymeric species from ethylene glycol.From these spectral data, the absorption coefficient was calculated using the following relationship [24]: where α is absorption coefficient, A is an energy-independent constant, m is a constant which determines type of the optical transition (m = 1/2 for allowed direct transitions and m = 2 for allowed indirect transitions) and E Opt g is the optical band gap.It is evaluated that the optical band gap of the α-Fe 2 O 3 has a direct optical transition.The optical band gap (E Opt g ) could be obtained by the extrapolating method using the (α • h • ν) 1/2 vs. h • ν plot.Inset in Fig. 7 shows plot of (α • h • ν) 1/2 vs. h • ν for the α-Fe 2 O 3 .Value of E Opt g obtained by this method is 2.4 eV.
The Raman spectroscopy technique has been extensively used for structure, composition, and phase characterization of materials [25].A Raman analysis was also carried out.α-Fe 2 O 3 belongs to the crystal space group D 6 3d .Since the primitive cell contains six atoms, there are 18 modes of vibration divided into 15 optical mode and 3 acoustic modes.According to the group theory of hematite, the following normal vibrational modes are predicted and represented by the hematite's irreducible vibrational mode at the first Brillouin zone centre [26]: The acoustic 2A 1u and 2A 2u modes are optically silent because they are inactive in Raman.As for the optical modes, they are 2A 1g , 5E g , 3A 2g , 4E u , in which the 2A 1g , 5E g modes are active in Raman and the 3A 2g , 4E u modes are inactive [27].So it has seven characteristics Ramanactive vibrations.The Raman spectrum of the sample is given in Fig. 9.This spectrum exhibit seven spectral signatures that are indicative for hematite, and do not contain any peaks associated with any maghemite.
Figure 8 shows characteristic bands of hematite which can be assigned to 2A 1g (222 and 498 cm -1 ) and 5E g (243, 290, 408 and 608 cm -1 ).Besides the typical 2A 1g and 5E g symmetry phonon modes (below 620 cm -1 ), a very strong feature at 1315 cm -1 and a very weak peak at 657 cm -1 are detected.Hematite is an antiferromagnetic material and the collective spin movement can be excited in what is called a magnon.The intense feature at 1315 cm -1 is assigned to a twomagnon scattering which arises from the interaction of two magnons created on antiparallel close spin sites.The peak at 657 cm -1 , which is typical of nanocrystalline hematite, originates from the presence of surface defects and/or reduced grain size [28,29].
This result confirms that the calcined product is α-Fe 2 O 3 .No other iron oxides such as magnetite or maghemite were detected indicating high purity of the final product.Also the large line widths and the low wave number shifts of the Raman peaks could be due to the phonon confinement effects in the nanocrystals.
To obtain further information about surface area and the distribution of pores in the NPs, BET N 2 adsorptiondesorption analysis was performed.The adsorptiondesorption isotherm and the corresponding BJH pore size distribution plot (inset) of the α-Fe 2 O 3 NPs are shown in Fig. 9.According to the IUPAC classification, the loop observed is ascribed to type H3 loops and the BET surface area of the material was calculated to be ∼30 m 2 /g.BJH plot indicates the existence of mesopores (pores 2-50 nm in diameter) in the α-Fe 2 O 3 NPs.The size of mesopores in the architecture was not uniform and most of the pores fall into a size range from 5 to 30 nm in diameter.The primary difference between the EDS and XPS techniques is that EDS could be used to determine the composition of several different particles individually, then an average could be taken of the composition, while for XPS which is a surface sensitive technique (it can only detect the element no more than 10 nm in the depth [30]), a relatively large analysis area results in simultaneous analysis of several particles.
Figure 10 shows survey of XPS spectrum of calcined (550 °C/3 h) iron oxide.It shows the presence of Fe and O elements.The peak observed at 711.1 eV is attributed to Fe 2p 3/2 and the peak at 725.08 eV is attributed to Fe 2p 1/2 .The shake-up satellite line is characteristic of Fe +3 in Fe 2 O 3 .The Fe 2p spectra of α-Fe 2 O 3 and γ-Fe 2 O 3 are almost identical with each other.In both cases, the main peaks maximum of 2p 3/2 and 2p 1/2 have the shake-up structures at their higher binding energy side at about 8.0 eV.In the present case, this value is found to be ∼8.0 eV, which confirms the formation of Fe(III) oxide.However, the marked difference in these XPS spectra is that the binding energy of the main peak, 2p 3/2 is slightly higher in the case of α-Fe 2 O 3 (711.3eV) compared to γ-Fe 2 O 3 (710.7 eV) [31].Also appearance of Fe +3 cations (711.1 eV) rather than Fe +2 cations (709 eV) indicates the presence of the Fe 2 O 3 phase [32].
The FTIR spectroscopy was utilized to detect the presence of functional groups adsorbed on the surface of the synthesized α-Fe 2 O 3 particles.Figure 11 shows the FTIR spectrum of α-Fe 2 O 3 powder acquired from Pechini sol-gel route calcined at 550 °C for 3 h.The band at ∼3350 cm -1 is assigned to the stretching vibration of water, indicating the existence of a little water absorbed on the sample.Two absorption bands below 1000 cm -1 represent characteristic features of α-Fe 2 O 3 and are assigned to metal oxygen stretching frequencies.The high frequency band ∼560 cm -1 refers to Fe-O deformation in the octahedral and tetrahedral sites while the low frequency band ∼462 cm -1 is attributed to Fe-O deformation in the octahedral site of hematite [31].The intensity of absorption band at 560 cm -1 is stronger than that at 462 cm -1 .It gives further evidence for the formation of α-Fe 2 O 3 and this is in agreement with the XRD measurement.There is no peak at 2900 cm -1 indicating the C-H stretching band, which means all organic compounds are removed from the samples after calcinations at 550 °C.
The DC electrical resistivity of the iron oxide thick film was measured as a function of temperature in the range of 483-673 K using DC two-probe method.Because resistance of the film at room temperature was very high and exceeds the limit of the used instrument (120 MΩ), it could not be measured using the experimental setup.However at higher temperatures, the resistance fall within readable value, and the electrical properties of α-Fe 2 O 3 NPs were evaluated in dry air in the range of 483-673 K. Figure 12 shows electrical resistance change of the sensor during heating and cooling cycles.As it clearly could be seen, resistance of sensor on heating and cooling is approximately the same.This means that sensor resistance is very stable upon heating and cooling cycles.The relationship between the resistance and temperature is as follows: where R 0 is the pre-exponential factor, k is the Boltzmann's constant, E g is bulk electrical band gap.From the plot of ln R against 1/T the band bending of the sensing layer can be determined.The plot of resistance versus inverse temperature, shown in inset of Fig. 12, shows that the temperature dependent resistance values obey the Arrhenius relation (a linear trend) and a linear decrease of the ln R T with increase in the temperature is observed.The estimated value of electrical band gap was found to be ∼2.47 eV.It is worthy to note that for individual particles, there is a relationship between electrical band gap (E El g = E bulk,el g +E ee ) and optical band gap (E Opt g = E bulk,el g + E ee − E eh , where E ee and E eh are the electron-electron and electron-hole interaction energies, respectively [33]).However, for most inorganic semiconductors the values of E El g and E Opt g of particles are almost the same.
After heating to high temperatures, the resistance decreased strongly.The resistance of a layer with length l and cross section A is defined as: where ρ is the resistivity of the layer.Conductivity (inverse function of resistivity, σ = 1/ρ) can be expressed by the product of the charge q, the effective mobility µ of the layer and the electron concentration n b : Combining the latter three expressions one obtains: Thus, for a given material with constant l and A, if n b increases R will decrease.For semiconductor materials, there is an exponential relationship between n b and temperature: where N c and N v are the effective density of states in the conduction band and valence band, respectively.According to above equation as temperature increases, n b will increase.In case of α-Fe 2 O 3 as a semiconductor, with increase of temperature more electrons transfer to conduction band and consequently n b increases and the resistance will decrease [34].It is worthy to note that as Fe 2 O 3 is a narrow band gap semiconductor and mobility of ions at tested temperatures are much lower than the mobility of electrons and it is supposed that electrical conductivity is dominant in this material [35].
Temperature coefficient α(T ) can be written as follow [36]: If temperature change is not significant as in our case, it can approximately be written as: The variation of the electrical resistance as a function of the temperature was characterized by a negative temperature coefficient of electrical resistance.As with increase of temperature the resistance decreases, so α has a negative value, which is consistent with a semiconducting-type behaviour.
Figure 13 shows a schematic diagram of the photoexcitation in semiconductors.When UV radiation falls on the metal oxide, the absorption of a photon with enough energy promotes an electron in conduction band and it generates a hole in the valence band; the photogenerated holes recombine with electrons trapped on the surface causing desorption of the adsorbed oxygen ions.This results in the conductivity increase due to the increased free electrons density.
The required energy for band to band excitation depends on the band gap energy of semiconductors.According to band gap determined from Tauc plot band gap of α-Fe 2 O 3 NPs was about ∼2.4 eV.The photon energy is determined by light wavelength (E = h • c/λ).This means that the photon energy of 400 nm (3.1 eV) is sufficient for band to band excitation of α-Fe 2 O 3 NPs.Figure 14 shows effect of UV light on the resistance change of α-Fe 2 O 3 NPs.As it can be observed, UV light has little effect on the resistance of α-Fe 2 O 3 NPs.Before UV light irradiation, large amount of O 2 -is absorbed on the surface of the α-Fe 2 O 3 NPs, which induced the formation of depletion layer.Upon UV irradiation, photo generated holes and electrons appeared, and then they are separated under the effect of the built-in electric fields.Due to the built-in electric fields, electrons migrated inside of the particles while the holes remained on the surfaces or migrated to the surfaces and reacted with O 2 -, which induced the desorption of O 2 [38]: As a result of this reaction oxygen species are photodesorbed from the surface (decrease of resistance).On the other hand, the photo-generated electrons will also contribute to a decrease of resistance.Little effect of UV light on the final resistance of α-Fe 2 O 3 NPs is probably due to the limited number of extrinsic electrons in α-Fe 2 O 3 NPs.

IV. Conclusions
In this study, α-Fe 2 O 3 nanoparticles (NPs) were successfully synthesized by Pechini sol-gel method and were fully characterized by different techniques.TEM results confirmed the formation of ultrafine NPs and results of XRD and TGA/DTA demonstrated the forma-tion of hematite at 550 °C.Also XPS, FTIR and EDS results proved the formation of desired chemical composition and purity of synthesized NPs.Furthermore BET results showed a high surface area which originated from nanosize nature of synthesized NPs.Thick films of Fe 2 O 3 NPs prepared on the interdigitated alumina substrates and the electrical resistances at different temperatures were studied.It was found that the prepared α-Fe 2 O 3 thick films had very stable electrical properties which are especially good for gas sensing applications.Moreover, the effect of UV light on the electrical resistance of sensor was studied and it was found that UV light had no significant effect on the change of electrical resistance of Fe 2 O 3 thick films.

Figure 1 .Figure 2 .Figure 3 .
Figure 1.Reaction mechanism for esterification of citric acid with ethylene glycol (a) [21] and reaction mechanism of chelation of iron nitrate by polymer (b)

Figure 12 .
Figure 12.Heating-cooling curves showing electrical resistance of α-Fe 2 O 3 NPs as a function of temperature (inset -variations of ln R versus (1/T) for the α-Fe 2 O 3 NPs)