Solvothermal synthesis of porous Fe3O4 nanoparticles for humidity sensor application

In this research, the effect of PVP on magnetic properties and morphology of Fe3O4 nanoparticle (Fe3O4-NPS) is investigated. Also, the sensitivity of the humidity of the selected Fe3O4-NPS is studied. X-ray diffraction (XRD), transmission electron microscope (TEM), and vibration sample magnetometer (VSM) were used to characterize the synthesized Fe3O4-NPs. The XRD and TEM results demonstrated that Fe3O4-NPs were crystallized in cubic structure with spherical pores morphology. Superparamagnetic behavior was seen in the samples prepared with the maximum saturation of approximately 10 emu g−1 for the sample synthesized using PVP:Fe(ac ac)3 ratio equal to 4. The outcomes of the humidity sensing of the selected sample revealed that the prepared Fe3O4-NPs with a porous structure is a good candidate to be used for humidity sensing.


Introduction
Nowadays, magnetic nanostructures (MNPs) have been considerably regarded by the researchers because of their important features and applications. A number of of the magnetic nanostructures can be used in medical applications due to their non-toxicity and biocompatibility [1][2][3]. Several studies have been performed into the use of the materials in drug delivery [4,5], magnetic resonance imaging (MRI) [6], hyperthermia [7,8], medical diagnosis and therapy [9,10] gas sensors [11][12][13], catalysts [14,15], and environmental remediation [16,17]. Several methods are used to synthesize magnetic material; e.g., sol-gel [18,19], reverse micelle [20], co-precipitation process [21,22], γ-ray irradiation [23,24], non-aqueous route [25], microwave plasma synthesis [26], and hydrothermal treatment [27,28]. Among these methods, solvothermal is a powerful technique to control the morphology and size of the nanostructure [29][30][31][32][33][34]. By this method, we can control the size, crystallinity, and the shape distribution of metal oxide nanostructures and synthesize well-crystallized and mono-dispersed ferrite nanostructures. In the solvothermal method, the morphology and size can be monitored by chemical parameters such as reaction time, surfactant, protective reactions, iron levels, etc For example, Huang et al [35] produced Fe 3 O 4 with a particle size between 450 and 750 nm by the solvothermal method and using NaAc·3H 2 O as a surfactant. Also, Aiguo et al [36] synthesized Fe 3 O 4 nanoparticles with particle size distribution of 15-190 nm using SDS and PEG as protective agents, using the solvothermal method and indicated that the magnetic properties of the magnetic material can be modified by controlling the synthesis conditions. Iron oxide has different phase and component as (α, β, and γ)-Fe 2 O 3 and Fe 3 O 4 . They have been widely used for sensing applications such as electrochemical sensors for glucose and dopamine detections . Also they were used for manufacturing gas sensors [37][38][39][40][41][42]. The humidity sensing of Fe 2 O 3 nanoparticles was studied in the literature but we could not find any study for Fe 3 O 4 nanoparticles [43].
In the present study, iron acetate (III), polyvinylpyrrolidone (PVP), and ethylene glycol were used respectively as a precursor, a size control polymer, and a solvent to prepare hollow Fe 3 O 4 nanoparticles by solvothermal method. The Fe 3 O 4 -NPs were synthesized through different PVP:Fe(ac ac) 3 ratios and then characterized to study their magnetic properties and morphology. Also, the humidity sensing of the selected sample was studied.

Synthesis of Fe 3 O 4 -NPs
The present research aims to study the effect of PVP:Fe(ac ac) 3 ratio on the morphology and size of Fe 3 O 4 -NPs. Therefore, three solutions were prepared with different PVP:Fe(ac ac) 3 ratios of 2, 4, and 6 that were named as samples (a), (b), and (c), respectively. To prepare these samples, first, 25 ml ethylene glycol was poured in a backer that is placed in an oil bath and the oil bath temperature was fixed at 60°C. Next, 0.55 g Fe (acac) 3 was added to the ethylene glycol to obtain a red color solution. Then, a proper amount of PVP was added to the solution slowly. The mixed solution was left for 1 h and then the prepared solution was poured into a Teflon vessel supported by a stainless still autoclave. The autoclave was placed in an oven and the oven temperature was kept at 180°C for 12 h. Deionized water using a centrifuge was used to wash the obtained dark brown precipitate  three times. This process was replicated for the samples with the other amounts of PVP. Finally, the products were oven dried at 60°C.

Sensor fabrication
RCA protocol was used for cleaning five Si-P silicon wafers (1×1 cm). The prepared samples dried properly using nitrogen gas. The S1813 photoresist was used for lithography technique to make electrode over the silicon surface. The photoresist was deposited on the silicon wafer using spin coater with a rotation speed of 2000 rpm during 30 s. The sample was baked in the oven at 110°C for 30 min. After the UV light exposure through the inter-digit (IDT) mask (10 s, power 100%), the light affected parts of photoresist were removed by acetone and followed by electrode deposition. Chromium and gold deposited on the wafer with thicknesses of 20 nm and 100 nm respectively using thermal evaporation technique. The residual photoresistance was removed from the surface in the final stage to achieve IDT (see figure 1).

Setup
A semi-automatic set up used to measure the IDT sensor response to the humidity. The chamber consists of sensor stage, glass jar, a sensor holder, humidity meter, thermometer, heater, fan, electric part. Weston bridge circle and data logger were utilized to connect the sensors and to collect data, respectively. The water was evaporated using the heater and the humidity was controlled by controlling the volume of a water drop. The humidity sensor recorded the humidity during the measuring process (relative humidity of 10%-70%). Prova 803 and 6485 Keithley were utilized to read and record the resistance and current, respectively.

Characterizations
The structure of the prepared sample was studied by x-ray diffraction method (XRD, using Philips, Xpert, Cu K α ) and the obtained data were analyzed by Size Strain Plot (SSP) method. TEM, CM120, Philips was used for morphology observations. The magnetic properties of the Fe3O4-NPs prepared with the different amount of PVP were examined by Vibrating Sample Magnetometer, VSM (Magnetic Daghigh Kavir, MDKB, Iran). Figure 2 shows the XRD patterns of the synthesized Fe 3 O 4 -NPs prepared by solvothermal methods. As mentioned earlier, different PVP:Fe(ac ac) 3 ratios of 2, 4, and 6 were tested, which were named as a samples  seen that the peak intensity of sample (b) is higher than that for the other samples, suggesting that this sample has better crystallinity than samples (a) and (c).

X-ray diffraction
The samples' crystallite size was determined by the SSP method through the g equation below: where K for spherical particles ¾. The term (d hkl β hkl cos θ) 2 is plotted against (d hkl 2 β hkl cos θ) for the detected peaks of Fe 3 O 4 -NPs from 2θ=25°to 2θ=75°. The slope of linearly fitted data estimated the particle size ( figure 3) [44]. The crystallite size of samples (a), (b), and (c) were obtained to be 65±2, 54±2, and 54±2 nm, respectively.

Magnetic properties
The magnetic properties of the prepared Fe 3 O 4 -NPs were studied by the vibrating samples magnetometer (VSM). Figure 4 presents the obtained results. All the samples show superparamagnetic behavior with the saturation magnetizations obtained to be 6, 10, and 2 emu g −1 for the sample (a), (b), and (c), respectively. These results are smaller than that for (Ms=100 emu g −1 ). It is indicated that sample (b) indicates a higher saturation than samples (a) and (c), probably because of the better crystallinity of sample (b). This result is in good consistency with the XRD results. Figure 5 presents the TEM micrograph of the prepared Fe 3 O 4 -NPspresented in. The results demonstrate that the morphology and size of the particles slightly affected by PVP quantity. The Fe 3 O 4 -NPs morphology was detected as spherical hallow shape for samples (a), (b), and (c) with a particle size of 86±22, 81±21, and 84±28 nm, respectively.

Sensitivity measurement
The sensitivity, S, is measured as the differential in sensor resistance (Rs) in the chamber humidity environment and its resistance in the air humidity (Ra) as a function of humidity, equation (1).
Fe 3 O 4 nanoparticles are porous and have oxygen atoms on the surface. These oxygen atoms arise due to the Fe 3 O 4 nanoparticles synthesis method. When humidity reaches to the surface and adsorbed by the nanoparticles, the resistivity of the samples decreased because of the charge carrier increased [45]. During the  adsorption of water by the surface, some of the hydrogen ions are dissociated and bounded with oxygen on the surface resulting to form the hydroxyl groups, equation (2) [46,47].
The formed hydroxyl groups bound with the iron atoms of the lattice and free electrons are obtained, equation (3) [48].
These free electrons increase the conductivity of the samples. Figure 6 presents the resistance variation of the sensor exposed to various humidity. Figure 7 indicate the humidity sensor sensitivity versus the room temperature humidity. As figure 7 shows, the association between the sensitivity and humidity has a linear behavior that can be fitted with a straight line using linear regression analysis. By applying the linear regression on the sensitivity-humidity curve, the slope of 0.508 with the coefficient of R 2 =0.96 was obtained.

Conclusion
Fe 3 O 4 -NPS were synthesized by a solvothermal method using different PVP:Fe(ac ac) 3 ratios. The XRD results indicated cubic structure for the as-synthesized samples. The crystalline size of the samples (a), (b), and (c) was calculated by the SSP method and obtained to be 65±2, 54±2, and 54±2 nm, respectively. The TEM micrograph demonstrated that the morphology of Fe 3 O 4 -NPs are hollow spherical with porous structure and size of 86±22, 81±21, and 84±28 nm for samples (a), (b), and (c), respectively. Also, it was observed that the best size distribution is associated with sample (b). In addition, it was revealed that the Fe 3 O 4 -NPs properties were not much affected by the PVP:Fe(ac ac) 3 ratio; however, sample (b) shows the best results due to the higher saturation of the magnetization as 10 emu g −1 than that of for sample (a) and (c). Therefore, the humiditysensing test was conducted on sample (b). The result shows that Fe3O4-NPs (sample (b)) have a good response to humidity, making them appropriate for this application.