1. INTRODUCTION
Nitrogen oxide(IV) (NO2) is a highly oxidizing gas commonly produced by fossil fuel burning, forest fires,
and industrial and motor vehicles[1], leading to air pollution[2], acid rain, and smog[3,4]. Exposure to this gas causes eye irritation, headache, hearing damage, respiratory
disorders[5,6], and even premature deaths[7]. Also, asthma in children could be related to NO2 pollution[8]. Furthermore, there is a direct relationship between the NO2 amount and COVID-19 infection severity[9]. On the other hand, NO2 is a biomarker for lung infection[10]. Therefore, from different perspectives, its detection is of importance.
Resistive gas sensors, based on the resistance change upon exposure to gas, are among
the most popular types of gas sensors for NO2 detection[11], due to their high response and stability, fast dynamics, simple fabrication, and
low price[12]. They are mostly realized from metal oxide semiconductors, with ZnO being highly
popular among them. ZnO is an n-type semiconductor with an Eg of 3.37 eV, high electron mobility, and good thermal and chemical stability. Furthermore,
it is highly available with low cost, no toxicity, and can be easily synthesized[13]. However, it has high sensing temperature and poor selectivity in its pristine form[14]. There are various strategies, such as doping[15,16], composite making[17,18], noble metal decoration[19,20], and high-energy irradiation[21,22], to improve its performance in NO2 gas sensing. In this regard, irradiation of the sensing material with high-energy
beams has such advantages as the possibility to apply after the material synthesis,
generation of structural defects in it, and high control over the energy and duration
of the irradiation beams[23].
Among various irradiation techniques, microwave (MW) irradiation is highly available,
cheap, and needs no vacuum[24]. Microwaves are electromagnetic waves in the frequency range of 300 MHz to 300 GHz.
They can internally heat metal oxides[25,26]. MW irradiation, which delivers high energy over a short duration, can be used during
or after the synthesis of material[27,28]. Therefore, MW irradiation time is critically important, which can induce different
amonuts of defects such as oxygen vacancies[29].
Previously, Gui et al.[30] prepared ZnO nanoparticles (NPs) in an MW chemical reactor for NO2 gas sensing at various temperatures, and the sensor fabricated at 180 °C showed a
response to 4-5 ppm NO2 at 120 °C. Kim et al.[31] applied MW irradiation for 1 min on a ZnO-graphene composite and reported a response
of ~12.5 to only 1 ppm NO2 at 300 °C. However, as far as we know, there are few studies on the impact of MW
irradiation time on the NO2 response of ZnO NPs.
Therefore, ZnO NPs were irradiated with MWs for 1, 3, 5, and 7 min to study the impact
of the irradiation time on the NO2 response of fabricated gas sensors. Various characterization techniques were applied
to the irradiated samples, and the one irradiated for 5 min revealed most structural
defects. Also, the sensor MW-irradiated for 5 min had the highest response of 11.25
to 2 ppm NO2 and 34.63 to 10 ppm NO2. The improved response of the optimal sensor was related to the generation of structural
defects and the formation of ZnO-ZnO homojunctions and ZnO-Zn(OH)2 heterojunctions.
3. RESULTS AND DISCUSSIONS
3.1. characterization studies
Figure 2 (a) displays SEM views of the pristine ZnO NPs, and the inset presents the lower-magnification
SEM image. In general, ZnO NPs have a rectangular-like morphology with widths of ~100-150
nm and lengths of ~200-400 nm. Even though they are somewhat agglomerated together,
it seems that most ZnO NPs have access to air, and therefore, gas molecules can be
easily adsorbed on them[32,33]. Figure 2 (b-e) displays SEM images of the ZnO NPs MW-irradiated for 1, 3, 5, and 7 min, respectively.
The insets display the corresponding lower-magnification SEM images. Overall, the
morphology and size of ZnO NPs were not changed after the MW irradiation, and all
samples revealed almost similar morphology.
Figure 3 (a) gives XRD patterns of the pristine and MW-irradiated ZnO NPs at different times.
All NPs showed well-crystalline diffraction peaks, matching with JCPDS File No. 36-1451,
which confirms the presence of crystalline ZnO with wurtzite structure[34]. No peaks related to impurity phases were detected, confirming the high purity of
both the pristine and MW-irradiated ZnO NPs. After the irradiation, the ZnO phase
and relative intensity of the peaks were not changed.
Figure 3 (b) indicates the O 1s core-level XPS plot of the pristine ZnO NPs and those MW-irradiated
at different times. For the sample irradiated for 5 min, there is a small shift (0.1
eV) to lower binding energies. This is due to defects such as surface oxygen vacancies
induced by MW irradiation[35].
Figure 3 (c) presents the UPS spectra of the pristine and MW-irradiated ZnO NPs at different times.
Based on the Ecut-off values in high kinetic energy regions and using the procedure described in[36], work function values for the pristine ZnO NPs and those MW-irradiated for 1, 3,
5, and 7 min were obtained to be 4.05, 4.05, 4.1, 4.2, and 4.15 eV, respectively (Figure 3 (d)). Therefore, the sample irradiated for 5 min showed the largest work function.
Figure 3 (e) indicates the PL spectra of the pristine and MW-irradiated ZnO NPs. The PL spectra
of the ZnO NPs indicated two emission bands in the UV (389 nm) and visible (400-650
nm) regions. The first peak corresponds to the near band-edge emission of ZnO, which
was related to the radiative recombination of free excitons and involved free excitons,
donor/acceptor pairs, and excitons bound to acceptors, donors, and their electron
satellites[37,38]. Also, the deep-level emission at ~520nm (2.44eV) was due to the oxygen vacancies
in ZnO[38]. Figure 3 (f) displays the normalized intensity of the PL spectra of samples at a wavelength of
~520 nm. Higher intensity means more defects. With MW irradiation time, the intensity
of peaks increased, and the sample irradiated for 5 min showed the highest intensity
and thus most defects.
Figure 4 (a) displays a high-magnification TEM view of the pristine ZnO NPs, and Figure 4 (b) offers the corresponding selected area electron diffraction (SAED) pattern. It was
taken along the [111] zone and shows a spotty pattern, confirming the single crystalline
nature of the ZnO NPs. Also, we indexed the spots related to (
1
¯
01), (
1
¯
10) and (01
1
¯
) crystalline planes of ZnO. Figure 4 (c) gives a high-resolution TEM (HR-TEM) view of ZnO, in which the spacings of 2.87 and
2.51 Å between the parallel fringes can be related to the spacing between crystalline
planes of (
1
¯
10) and (
1
¯
01), respectively. The inset in Figure 4 (c) manifests the fast Fourier transform (FFT) pattern of the crystalline ZnO. Figure 4 (d) displays a high-magnification TEM image of the ZnO NPs MW-irradiated for 5 min, and
Figure 4 (e) shows the relevant SAED pattern taken along the [110] zone. The pattern is spotty,
which again confirms the single crystalline nature of the ZnO NPs. We indexed the
spots related to the and crystalline planes of ZnO. Figure 4 (f) displays an HR-TEM image of ZnO NPs MW-irradiated for 5 min, where the spacings of
2.45 and 4.44 Å between the parallel fringes can be related to the crystalline planes
of and , respectively, which correspond to Zn(OH)2. The inset in Figure 4 (f) yields the FFT pattern of crystalline ZnO. It can be deduced that MW irradiation
partially converted ZnO to Zn(OH)2. This can be due to the high RH (~90%) in the MW chamber and water adsorption on
the ZnO. However, Zn(OH)2 was scarce, so it was not detected by XRD.
Figure 5 (a-c) exhibits the HAADF color mapping of the pristine ZnO NPs, in which both Zn and O
elements are evenly dispersed over the entire surface of the ZnO NPs. Also, in Figure 5 (e), the HAADF elemental spectrum taken from Figure 5 (d) displays the weight and atomic percentages of Zn and O elements (17.25 and 82.75
wt.% for Zn and O, respectively). Figure 5 (f-h) exhibits the HAADF color mapping of the ZnO NPs MW-irradiated for 5 min. Similar
to the pristine ZnO NPs, both Zn and O are uniformly distributed over the surface
of the ZnO NPs. Also, in Figure 5 (j), the HAADF elemental spectrum taken from Figure 5 (i) confirms that the weight percentages of Zn and O were 10.69 and 89.31, respectively.
This demonstrates that after MW irradiation, the amount of oxygen decreased due to
evaporation, resulting in oxygen vacancy defects.
BET surface areas of pristine and MW-irradiated (5 min) sensors were obtained using
adsorption-desorption curves N2 adsorption-desorption curves (Figure 6 (a-b)). While pristine ZnO NPs showed a BET surface area of 5.461 m2/g, after 5 min MW irradiation it increased to 6.057 m2/g. This demonstrates higher amounts of adsorption sites for gas molecules after MW
irradiation.
3.2. Gas sensing studies
Figure 7 (a-e) presents the dynamic resistance curves of ZnO NPs with 5-minutes microwave irradiation
to 10 ppm of NO2 at 50-300 °C. At low temperatures, the base resistance was very high, in the range
of Giga ohm, and no meaningful response was obtained due to the high noise level,
as shown in Figure 7 (a, b). At 300 °C, even though the base resistance decreased, no noticeable sensing signal
was recorded-again, due to the high noise level. As shown in Figure 7 (c), at 150 °C, the base resistance was still very high with much noise. However, the
sensor showed a signal in the presence of NO2 from Figure 7 (d), at 200 °C, the base resistance decreased and a response was recorded, yet the sensor
still exhibited some noise in Figure 7 (c,d). Therefore, at all the above-mentioned temperatures, the sensor performance was not
acceptable. However, as shown in Figure 7 (e), the sensor exhibits a low noise level and a pronounced sensing response at 250 °C.
Figure 8 (a-e) presents the dynamic response curves of the pristine ZnO NPs and those MW-irradiated
for 1, 3, 5 and, 7 min to 2-10 ppm NO2 at 250 °C. Based on the sensing curves, we calculated the responses (Table 1) and plotted the calibration curves (Figure 8 (f)).
The responses of the pristine ZnO sensor and those fabricated from the ZnO NPs MW-irradiated
for 1, 3, 5, and 7 min to 2 ppm NO2 at 250 °C were 6.0, 5.5, 8.43, 11.25, and 7.8, respectively. Overall, the sensor
MW-irradiated for 5 min exhibited the highest output to all NO2 concentrations. In the next step, we obtained the response curves of the ZnO gas
sensor MW-irradiated for 5 min to 2 ppm H2S, H2, NH3, C2H5OH and C6H6 gases at 250 °C (Figure 9 (a-e)).
Based on these graphs, a selectivity histogram was constructed, as shown in Figure 9 (f). The responses of the optimal sensor to 2 ppm H2S, H2, NH3, C2H5OH, C6H6 and NO2 gases were 1.68, 1.25, 1.15, 1.27, 1.17 and 11.25. This confirmed the excellent selectivity
of the optimal gas sensor to NO2.
Several factors affected the selectivity to NO2 gas (i) NO2 gas has a high electron affinity (2.28 eV) and can directly adsorb electrons from
the sensor surface, while other gases need to react with adsorbed oxygen species to
generate sensing signal. (ii) the presence of nitrogen with one unpaired electron
can lead to participatation in bond formation with surface elements in ZnO[39].
Figure 10 (a-e) yields the sensing plots of fresh and 3-month-preserved gas sensors to NO2 at 250 °C. Based on these graphs, calibration curves were plotted in Figure 10 (f). The response is notably reduced after three months. This maybe due to the adsorption
of water molecules while keeping in the laboratoary.
Figure 11 (a-e) indicates the sensing plots of optimal gas sensor to various concentrations of NO2 at 250 °C and in the presence of 30% and 60% RH, respectively. Based on these graphs,
plotted the calibration curves, as shown in Figure 11 (f). Under humid conditions, the water molecules were adsorbed on the sensor surface
and reduced the number of available adsorption sites on it[40]. Hence, less gas could be adsorbed on the sensor in humid conditions.
To study the response of the optimal sensor to various levels of RH, we exposed the
pristine and optimal sensor to 10-60% RH at 250 °C, with the dynamic response curves
given in Figure 12 (a-f), respectively. The responses of the pristine ZnO sensor to 10, 20, 30, 40, 50, and
60% RH were 2.60, 2.90, 2.94, 3.01, 3.08, and 3.15, respectively. On the other hand,
the responses of the optimal sensor to the same RH levels were 3.08, 3.13, 3.55, 3.89,
3.65, and 4.02, respectively. Hence, compared to the original values, in both cases,
the response of sensors increased with growing humidity.
3.3 Sensing mechanism
In resistive gas sensors, variations of the sensor resistance occur due to the adsorption
of gases[41]. At first, oxygen molecules with high electron affinity become adsorbed on the sensor
surface in air and take electrons[42]:
Temperature is a desecive factor and at low temperatures O2– ions are dominant species, while at high temperatures O– ions are the dominant species. At 250 °C, the dominant species are O– ions[43]. The abstraction of electrons generates an electron depletion layer on the ZnO NPs.
In an NO2 atmosphere, NO2 molecules directly abstract electrons from the ZnO NPs[44]:
Furthermore, the reaction of NO2– (ads) with adsorbed oxygen species is likely[44]:
Hence, further abstraction of the electrons by NO2 gas leads to an increase of the resistance in the presence of NO2 gas.
Furthermore, in contact points between the ZnO NPs, ZnO-ZnO homojunctions are formed
in the air and upon exposing to NO2 gas, the height of homojunction barriers changes, leading to the sensor's resistance
modulation (Figure 13)[45].
However, the pristine ZnO sensor showed a lower response to NO2 compared to the optimal one. For the optimal gas sensor, in addition to the above
mechanisms, the oxygen vacancy defects generated from MW irradiation, play an important
role as favorable adsorption sites for gases[46]. Based on XPS and PL analyses, the sensor MW-irradiated for 5 min revealed most
oxygen vacancies, which eventually led to improved NO2 gas response. Furthermore, as evidenced by the HR-TEM image, some Zn(OH)2 was generated as a result of MW irradiation at the high RH. Therefore, ZnO-Zn(OH)2 heterojunctions were formed in air and upon injection of NO2 gas. These heterojunctions increased the baseline resistance (Ra) due to the formation of additional potential barriers at the interface. Upon exposure
to NO2, the modulation of these barrier heights contributed significantly to the sensing
signal (Figure 13). Besides, according to the UPS analysis, the work function of the optimal sensor
(4.2 eV) exceeded those of other sensors. This means that MW irradiation for 5 min
had the best effect on the electrical properties of this material, eventually leading
to enhanced NO2 response.