A two-step synthesis of nanosheet-covered fibers based on α-Fe2O3/NiO composites towards enhanced acetone sensing

A novel hierarchical heterostructures based on α-Fe2O3/NiO nanosheet-covered fibers were synthesized using a simple two-step process named the electrospinning and hydrothermal techniques. A high density of α-Fe2O3 nanosheets were uniformly and epitaxially deposited on a NiO nanofibers. The crystallinity, morphological structure and surface composition of nanostructured based on α-Fe2O3/NiO composites were investigated by XRD, SEM, TEM, EDX, XPS and BET analysis. The extremely branched α-Fe2O3/NiO nanosheet-covered fibers delivered an extremely porous atmosphere with huge specific surface area essential for superior gas sensors. Different nanostructured based on α-Fe2O3/NiO composites were also explored by adjusting the volume ratio of the precursors. The as-prepared samples based on α-Fe2O3/NiO nanocomposite sensors display apparently enhanced sensing characteristics, including higher sensing response, quick response with recovery speed and better selectivity towards acetone gas at lower operating temperature as compared to bare NiO nanofibers. The sensing response of S-2 based α-Fe2O3/NiO nanosheet-covered fibers were 18.24 to 100 ppm acetone gas at 169 °C, which was about 6.9 times higher than that of bare NiO nanofibers. The upgraded gas sensing performance of composites based on α-Fe2O3/NiO nanosheet-covered fibers might be ascribed to the exclusive morphologies with large surface area, p-n heterojunctions and the synergetic performance of α-Fe2O3 and NiO.


Results and Discussion
Crystal, morphological structure and compositional features of the as-prepared bare NiO and the nanocomposites based on α-Fe 2 O 3 /NiO samples were described by using XRD, SEM, TEM, EDS, XPS and BET examination. The composite NiO-PVP nanofibers were turned into pure NiO nanofibers after annealing at 600 °C in air. The XRD pattern of the NiO nanofibers display face-centered cubic phase geometry, which matched well with the standard card NiO (JCPDS Card No. , as shown in Fig. 1. We implemented a hydrothermal method to grow α-Fe 2 O 3 nanosheets on the surface of electrospum NiO nanofibers and the diffraction pattern with phase purity of as-produced composites based on α-Fe 2 O 3 /NiO nanosheet-covered fibers were presented in Fig. 1. All the XRD peaks present in the composites based on α-Fe 2 O 3 /NiO corresponds well with the standard crystallographic patterns of the face-centered cubic phase of NiO (JCPDS Card No.  and the rhombohedral phase of α-Fe 2 O 3 (JCPDS Card No. 33-0664) without any extra peaks, which indicates the high purity of α-Fe 2 O 3 /NiO heterostructures. With the increasing of Fe contents, the intensity of diffraction peaks decreases and became broader with a clear phase separation in sample S-2 and S-3 as compare to S-1 based α-Fe 2 O 3 /NiO nanocomposites, which results in reduction of the crystallite size as shown in Table 1. These facts demonstrate that ions can be isolated at the junction and the microstrain established in the composites nanosheet-covered fibers increases while the particle size decreases 46 .
The SEM images of the pure NiO sample before and after calcination comprised of ultralong nanofibers with uniform diameter having a smooth surface shown in Fig. 2a-d. The spreading of the nanofibers were enough random with no clear pattern. The average length and diameter range of NiO nanofibers were about ~13-15μm long and ~0.5-0.6 μm thick. Moreover, the transmission electron microscopy images of NiO nanofibers specifies that the NiO nanofibers were composed of clear smooth surfaces shown in Fig. 2e and f. The nanofibers were ultralong and continuous with random alignment. Typically, a single sample S-2 based α-Fe 2 O 3 /NiO composites were chosen for examination under scanning electronic microscope. Figure 3a-d display the SEM images with low and high-magnification of the sample S-2 based nanocomposites before and after calcination, which shows that the electrospum ultralong nanofibers architectures of NiO were maintained but, only growth of α-Fe 2 O 3 nanosheets as branches over the stem of NiO nanofibers after the hydrothermal treatment. The average length and diameter of S-2 samples were ~13 μm and a diameter of ~0.4 μm thick. The α-Fe 2 O 3 nanosheets grew outwardly from the surfaces of NiO nanofibers to form the branch heterostructure. The typical SEM images of the samples S-1 and S-3 based α-Fe 2 O 3 / NiO nanocomposites were shown in Fig. S1. The electrospum nanosheet-covered fibers were uniform in diameter having a rough surface with random alignment. Further investigation of morphology of the sample S-2 based nanocomposites was made by TEM, a similar distinct nanosheet-covered fibers structure could be clearly identified as shown in  Fig. 3e and f, which clearly matches with SEM images. It can be realized that the nanosheet-covered fibers have a coarser morphology and are composed of inter-associated grains. All the samples based on α-Fe 2 O 3 /NiO nanocomposites, the nanofibers were too long, continuous and random pattern with readily growth of α-Fe 2 O 3 nanosheet as branches. The increase of Fe content leads to the amendments in the nanosheet-covered fibers surface with the creation of porous hollow structure among the inter-related nanoparticles. These porous structures in the nanocomposites were also related to the putrefaction of organic phase 47 . Thus, these porous morphologies leads to the improved specific surface area and it benefit towards superior sensing features of the sensors. The elemental composition of bare NiO nanofibers and α-Fe 2 O 3 /NiO nanosheet-covered fibers with different Fe concentrations was performed using EDX's analysis was shown in Fig. S2. All the spectra of as-synthesized samples designate the presence of Ni, Fe and O elements in the samples. The atomic percentages of Fe in α-Fe 2 O 3 / NiO nanosheet-covered fibers were 2.6 at% (S-1), 4.4 at% (S-2) and 8.9 at% (S-3), respectively. The rise of oxygen content in composites samples with Fe amounts exposed the existence of oxide phase like α-Fe 2 O 3 as observed in XRD pattern. Obviously, the stem was NiO nanofibers and the outgrowth branches were α-Fe 2 O 3 nanosheets according to the spatial spreading of Ni and Fe signals in the EDS spectrum.
To further characterize the composition of the S-2 based α-Fe 2 O 3 /NiO composites, X-ray photoelectron spectroscopy (XPS) was evaluated. It can be realized from the XPS wide spectrum that the S-2 based nanocomposites contained Ni, Fe, O and C elements as shown in Fig. S3. The C signal could be ascribed to adventitious hydrocarbon. The complete XPS spectrums of the Fe-2p range and the Ni-2p range were presented in Fig. 4a and b. In case of Fe-2p spectrum, the Fe-2p peak could be distributed into two signals as presented in Fig. 4a. Fe-2p 3/2 and Fe-2p 1/2 signals were focused at 709 eV and 722.6 eV, respectively, which agreed with the electronic state of α-Fe 2 O 3 48 . In case of Ni-2p spectrum, the Ni-2p peak could be distributed into four signals as presented in Fig. 4b. The binding energy peaks at 852 eV and 859 eV were ascribed to the Ni-2p 3/2 and its satellite peak while, the peaks at 870 eV and 877 eV were ascribed to Ni-2p 1/2 and its satellite. The Ni-2p 3/2 peaks were credited to Ni 3+ while, Ni-2p1/2 peaks were ascribed to Ni 2+ 49 . The XPS results further specified that the ultimate products were composed of α-Fe 2 O 3 and NiO.
Likewise, the BET with N 2 adsorption-desorption analysis were explored for further examination of specific surface area and pore size distribution of the bare NiO nanofibers and the sample S-1, S-2 and S-3 based α-Fe 2 O 3 /NiO nanosheet-covered fibers. Figure 5a presents the N 2 adsorption-desorption isotherms of all prepared samples, which expose a type IV adsorption isotherm with a H3-type hysteresis loop at different relative pressure ranges. The consistent specific surface areas of bare NiO nanofibers and the sample S-1, S-2 and S-3 based α-Fe 2 O 3 /NiO nanosheet-covered fibers were 30 m 2 g −1 ; 41 m 2 g −1 ; 47 m 2 g −1 and 49 m 2 g −1 , respectively. From Fig. 5b, the adsorption pore size distribution curve of all prepared samples indicates the mesopores with broad size distribution, which clearly match with adsorption-desorption isotherms (Fig. 5a). The desorption distribution curve inset of Fig. 5b shows sharp pore size distribution which might be false peaks. The spreading of the pore diameter of S-2 based α-Fe 2 O 3 /NiO nanocomposites are more concentrated with a large number of mesopores (peak pore at ca. 2~10 nm) and a small number of large mesopores and macropores. The great improvements in specific surface area of nanocomposites based on α-Fe 2 O 3 /NiO samples may arise from the voids among close-packing nanosheet-covered fibers and mesoporous voids between the nano-crystallites 50 .
Gas-sensing properties. We fabricated gas sensors based on bare NiO nanofibers and the sample S-1, S-2 and S-3 based α-Fe 2 O 3 /NiO nanosheet-covered fibers, in order to study the gas sensing properties towards various toxic gases. The transient sensing response characteristics of as-synthesized sensors towards acetone, ethanol, methanol, xylene, toluene and benzene were investigated. The electric resistance of the as-prepared sensors increased sharply on the injection of target gas and then decreased promptly and recovered to its original value after the target gas was released, which displays the sensing performance of p/n type semiconducting sensors.  First of all, the maximum transient gas-sensing responses of the sensors based on bare NiO nanofibers and the sample S-1, S-2 and S-3 based α-Fe 2 O 3 /NiO nanosheet-covered fibers towards acetone gas were explored at diverse functioning temperatures from 125 to 250 °C to examine the optimal Fe doping amount into α-Fe 2 O 3 / NiO nanocomposites as well as the correlation between gas response and functioning temperature, as shown in Fig. 6a. Apparently, the volcano-shaped connection between gas responses and functioning temperature was perceived for all the samples, and the best functioning temperature of each sample was 169 °C. Meanwhile, the gas response was greatly improved due to increasing Fe concentration into α-Fe 2 O 3 /NiO nanocomposites. The sensing responses of the sensors based on bare NiO nanofibers and the sample S-1, S-2 and S-3 based α-Fe 2 O 3 / NiO nanosheet-covered fibers to 100 ppm acetone at 169 °C were 2.64, 7.32, 18.24 and 12.41, respectively. The result exposed that the sensor based on S-2 based α-Fe 2 O 3 /NiO nanosheet-covered fibers presented the maximum response to 100 ppm acetone and the value was about 6.9 times superior to that of bare NiO nanofibers.
Subsequently, the gas responses of sensors based on bare NiO nanofibers and the sample S-2 based α-Fe 2 O 3 / NiO nanosheet-covered fibers to 100 ppm of various target gases at 169 °C were examined. The target gases comprised acetone, ethanol, methanol, xylene, toluene and benzene. As shown in Fig. 6b, the sensor S-2 based α-Fe 2 O 3 /NiO nanosheet-covered fibers exposed superior response for all tested gases compared to that of bare NiO nanofibers. Moreover, the response of both sensors to acetone was obviously better than that to other gases. The sensing selectivity of obtained sensor S-2 based α-Fe 2 O 3 /NiO nanocomposites for acetone detection can be explained by the relatively low bond dissociation energy of acetone molecules (393 kj/mol) as compare to other volatile organic compounds (VOCs) gases 51 . Meanwhile, large number of oxygen contents (O C and O V ) and the formation of p-n junction on the surface of the Fe 2 O 3 /NiO nanocomposites make it release more electrons during redox reaction between acetone molecules and chemically adsorbed oxygen ions, which reflects its selective detection.
The real-time sensing response of as-synthesized gas sensors based on bare NiO nanofibers and the sample S-1, S-2 and S-3 based α-Fe 2 O 3 /NiO nanosheet-covered fibers at 169 °C towards different acetone concentration ranges from 5-100 ppm shown in Fig. 6c. As all the sensors based on sample S-1, S-2 and S-3 based α-Fe 2 O 3 /NiO nanosheet-covered fibers displayed enhanced sensing response for acetone gas compared to the bare NiO nanofibers. It has been perceived that the response increased with increasing acetone concentration from 5 to 100 ppm for all the prepared sensors and the growth gradually slowed down with increasing acetone than 100 ppm. Among the whole fabricated series, the response of the sensor based on the sample S-2 α-Fe 2 O 3 /NiO nanosheet-covered fibers were apparently higher than that of the other samples to various acetone concentrations we tested, shown in the Fig. 6d. Moreover, we could find that the response of the sensor S-2 based on α-Fe 2 O 3 /NiO nanocomposites did not tend to saturation gradually when the acetone concentration was raised to 100 ppm, although the increasing trend slowed down with the further increase of the acetone concentration. The result exposed that all the sensors presented tremendous response and recovery features with respect to diverse acetone concentrations ranging from 5 to 100 ppm. The response and recovery features of all the synthesized sensors to 100 ppm acetone at 169 °C were presented in Fig. S4. Consequently, the sensor S-2 based α-Fe 2 O 3 /NiO composites have displayed quick response and recovery times of 26 s and 37 s, respectively, to 100 ppm acetone at 169 °C. The relatively rapid response and recovery of S-2 based nanocomposites contribute to the real-time detection of acetone gas.
The acetone sensitivity of sensor based on NiO nanofibers and S-2 based α-Fe 2 O 3 /NiO nanosheet-covered fibers were explored in an operating temperature of 169 °C for various known value relative humidity 15, 30, 45 and 60% HR, as plotted in Fig. 6e. According to the two direct interaction mechanisms proposed by Heiland and Kohl 52 , water vapour provides the essential conditions for oxygen adsorption, electrons and oxygen vacancies. At a certain extent, the water adsorption could accelerate the oxygen adsorption. The variation in the total concentration of adsorption sites [S t ] is shown as: where [S t0 ] is intrinsic concentration of adsorption sites, k 0 is adsorption constant for water vapour and ρ H O 2 is the partial pressure of water vapour, respectively. The partial pressure of water vapour ( H O 2 ρ ) is proportional to the mass of relative humidity (RH). With the rises of relative humidity, the total concentration of adsorption sites [S t ] increases, and the rate of coverage of hydroxyl groups and oxygen species altered. Thus, the relative humidity sensitivity of sensors based on NiO and S-2 based composites depends on the relative surface distribution, coverage of hydroxyl groups and oxygen species 53 . When the relative humidity is 15 and 30% HR, the sensitivity of both NiO and S-2 sensor is low, due to less number of hydroxyl groups and oxygen species on the surface of sensing materials. The best sensitivity response to acetone is in 45%, which is contributed to the small coverage of hydroxyl groups that cannot constrain the acetone adsorption and the coverage of oxygen species increase. When the relative humidity increases up to 60% RH, the sensitivity response starts to decrease, which might be due to large coverage of hydroxyl groups and the adsorption of oxygen can be limited. The superior humidity sensing response of sensors based on S-2 to that of NiO might be due to the impact of porosity that boosted the diffusion of water vapors.
In order to reflect the enhanced sensing performance of as-synthesized sensors, the results obtained in this study were quantitatively compared with those stated by many research groups about n-type α-Fe 2 O 3 and p-type NiO based sensors towards acetone, listed in Table 2. The gas response of S-2 based α-Fe 2 O 3 /NiO nanosheet-covered fibers was the highest at low acetone concentration and relatively lower operating temperature of 169 °C, showing comparatively higher efficiency than those reported in the given literature 16,[54][55][56] . Therefore, it is believed that novel S-2 based α-Fe 2 O 3 /NiO nanosheet-covered fibers prepared in this study provide great interest for the further analysis in the field of gas-sensing application.
Gas-sensing mechanism. In order to better understanding of the enhanced gas sensing features of sensors based on α-Fe 2 O 3 /NiO nanosheet-covered fibers, the gas sensing mechanism of SMOs is presented first. The simple sensing mechanism of n-type or p-type SMOs broadly involves the variation in electrical conductivity/ resistivity due to the chemical reaction of gas molecules with the surface involving gas adsorption, surface reaction, and desorption processes, which can be well implicit by the depletion layer or space-charge model 57,58 . Generally, the adsorption and desorption of target gas molecules on the surface of SMO-based sensing materials can lead to the reaction process of electron exchanges, which are transmitted by the surface adsorbed oxygen 59,60 . Such electron shifting causes variations in the resistance of sensing devices by making changes in the thickness of their depletion layers. Thus, the sensing properties can be upgraded by enhancing its resistance deviation. The graphic illustration of the acetone gas sensing mechanism on the surface of sensors based on heterostructure α-Fe 2 O 3 /NiO nanosheet-covered fibers in the presence of air and target gas was illustrated in Fig. 7a.
The enhanced sensing properties of sensors based on heterostructure α-Fe 2 O 3 /NiO nanosheet-covered fibers as compare to bare NiO fibers may be attributed to the creation of p-n junction between p-type NiO nanofibers and n-type α-Fe 2 O 3 nanosheets. First, we used (α) 2 (hʋ) 2~h ʋ relation curve to compute the energy band gap of NiO nanofibers and α-Fe 2 O 3 nanosheet from their UV-vis spectrum, as shown in Fig. S5. The band gaps calculated for NiO and α-Fe 2 O 3 were 3.37 and 2.05 eV, respectively. Then we mentioned some literatures to get the Fermi level (NiO: 5.0 eV, α-Fe 2 O 3 : 4.39 eV), valence band of NiO (5.5 eV) and conduction band of α-Fe 2 O 3 (4.09 eV) [61][62][63][64][65] . According to this data, the energy band configuration of NiO and α-Fe 2 O 3 in air before combination had been drawn and presented graphically in Fig. 7b. After the formation of heterojunction, the energy band configuration of the composites based on α-Fe 2 O 3 /NiO in acetone had been drawn and presented graphically in Fig. 7c. A relative opposite motion of charge carriers took place at the junction of p-type NiO and n-type α-Fe 2 O 3 to obtain equalization of Fermi levels, which results in band bending. As a consequence of generation of an electric field in the space charge region nearby the p-n interfaces, the energy bands in the side of NiO bend downwards and the energy bands in the side of α-Fe 2 O 3 bend upwards. Instantaneously, an hole accumulation layer had been formed on the side of α-Fe 2 O 3 while electronic depletion layer formed on the side of NiO. The oxygen molecules from air adsorbed on the surface of sensors based on α-Fe 2 O 3 /NiO nanocomposites will capture electrons from the conduction bands of both α-Fe 2 O 3 and NiO and produce oxygen ions O δ− adsorbed on the surface, which results in the further widening of electron depletion layer and hole depletion layer on the surface of α-Fe 2 O 3 and NiO metal oxides, respectively. Upon exposure to acetone gas, the acetone molecules react with chemically adsorbed oxygen ions O δ− and breakdown into CO 2 and H 2 O with release of the captured electrons back to Ni vacancies, thus result in the upsurge of the resistance of the composites α-Fe 2 O 3 / NiO nanosheet-covered fibers. Consequently, the sensitivity response of the S-1, S-2 and S-3 based α-Fe 2 O 3 / NiO nanosheet-covered fibers compared to bare NiO nanofibers toward acetone test gas clearly amplified due to the p-n heterojunction effect which results in the increment of the initial resistance and the degeneration in the equivalent hole concentration of the α-Fe 2 O 3 /NiO nanocomposites.
As the differences in sensing properties among samples is associated to the surface chemical composition and microstructure of the as-prepared sensors. Consequently, the O 1 s peaks of bare NiO, S-2 and S-3 based α-Fe 2 O 3 / NiO nanosheet-covered fibers were asymmetrically fitted into three dissimilar components 66  on its surface, which provides comparatively more active sites and stronger adsorption capability for acetone gas to release more electrons during redox reaction as compare to bare NiO and S-3 based sensor, which suggests qualitatively that S-2 based α-Fe 2 O 3 /NiO nanocomposites display an optimal sensing behavior with fast response-recovery times 16,66 . Thus, the enhanced gas-sensing behavior of sensors based on α-Fe 2 O 3 /NiO nanocomposites might be attributed due to the small size effect of α-Fe 2 O 3 nanosheets, large surface morphology with numerous oxygen components and the synergetic effect prompted by the compact interfacial interaction between α-Fe 2 O 3 and NiO heterojunctions.

Conclusion
The ultralong NiO nanofibers were synthesized by a facile electrospum technique and then further functionalized by decorating α-Fe 2 O 3 nanosheets using an easy hydrothermal strategy. The electron microscopic images exposed that the α-Fe 2 O 3 nanosheets were deposited epitaxially on the surfaces of NiO nanofibers to produce α-Fe 2 O 3 / NiO nanosheet-covered fibers. The gas sensors were fabricated from the as-produced samples based on α-Fe 2 O 3 /
NiO nanocomposites, which exhibited enhanced sensing response and excellent selectivity toward acetone gas at relatively low temperature compared to bare NiO nanofibers. The sensing response of S-2 based α-Fe 2 O 3 /NiO nanosheet-covered fibers were 18.24 to 100 ppm acetone gas at 169 °C, which was about 6.9 times higher than that of bare NiO nanofibers. The upgraded gas sensing performance of nanocomposites based on α-Fe 2 O 3 /NiO nanosheet-covered fibers could be attributed to the unique large surface morphology, p-n heterojunctions and the synergetic performance of α-Fe 2 O 3 and NiO.  Preparation of the electrospum NiO nanofibers. The pure NiO nanofibers were synthesized by electrospinning procedure followed by annealing. In this synthesis, 1 g NiCl 2 •6H 2 O was mixed with 4 g DMF and 6 g ethanol under stirring for 20 minutes. Afterward, 1.5 g of PVP was added into the above solution under vigorous stirring for 4 hour. The precursor's solution was placed statically for one day to eliminate the gas bubbles before electrospinning. Then, the precursors solution was loaded into a glass syringe with a stainless steel needle connected to a diverse high voltage power range upto 30 kV. A smooth steel plate covered with aluminum foil was used as collector to collect the fibers. A 13 kV was provided between anode (needle) and cathode (collector) at a distance of 20 cm with feeding rate of 0.06 mm/min. Finally, the as-spum PVP-NiO nanofibers were calcined at 600 °C for 3 hours in air in order to eliminate PVP to get pure NiO nanofibers.  Gas-sensors fabrication and measurement. The comprehensive of gas-sensors fabrication and measurement can be found in our earlier report 67 . The gas sensors were made-up of the thick films obtained from the powder suspension of the as-produced bare NiO and nanocomposites based on α-Fe 2 O 3 /NiO samples. Each sample was mixed in separate beaker containing ethanol and ultra-sonicated into slurry, and then it was pasted onto an Al 2 O 3 ceramic tube by using brush to form a thick film between Au electrodes on the alumina ceramic tube. The thickness of prepared films is about 20-30 μm with the diameter of the tube is about 1.2 mm and the space between two parallel electrodes is ∼6 mm. In our case, the gas-sensing measurements were carried-out on an intelligent gas sensing analysis system (CGS-1TP, Beijing Elite Tech Co., Ltd, China). The saturated target gas was introduced into the test chamber of 18 liters by a micro-injector using a rubber plug. Hence, the sensors resistance were collected and explored by the system in real time. The sensor temperature adjusted conductively with an accuracy of 1 °C by the analysis system presented an external temperature control (from room temperature to 500 °C) and the relative humidity in the system could be controlled by a dehumidifier. The sensing response (Sr) towards target gas was defined as R g /R a , where R a and R g are the resistance of the sensor in the presence of air and target gas, respectively. Herein, the gas-sensing measurements were carried out at a functioning temperature of 169 °C with optimized relative humidity of 45%.