Underpinning the Interaction between NO2 and CuO Nanoplatelets at Room Temperature by Tailoring Synthesis Reaction Base and Time.

An approach to tailor the morphology and sensing characteristics of CuO nanoplatelets for selective detection of NO2 gas is of great significance and an important step toward achieving the challenge of improving air quality and in assuring the safety of mining operations. As a result, in this study, we report on the NO2 room temperature gas-sensing characteristics of CuO nanoplatelets and the underlying mechanism toward the gas-sensing performance by altering the synthesis reaction base and time. High sensitivity of ∼40 ppm-1 to NO2 gas at room temperature has been realized for gas sensors fabricated from CuO nanoplatelets, using NaOH as base for reaction times of 45 and 60 min, respectively at 75 °C. In both cases, the crystallite size, surface area, and hole concentration of the respective materials influenced the selectivity and sensitivity of the NO2 gas sensors. The mechanism underpinning the superior NO2 gas sensing are thoroughly discussed in terms of the crystallite size, hole concentration, and surface area as active sites for gas adsorption.

Introduction

The development and implementation of sustainable gas detection systems, that can potentially safeguard the environment, have received a great deal of attention. Gases such as hydrocarbons, carbon monoxide (CO), ammonia (NH3), hydrogen (H2), and sulphur dioxide (SO2) are detrimental to health, an explosion hazard and can even be lethal. Nitrogen dioxide (NO2) is one of the most common air pollutants from automotive and combustion emissions.[1] It is toxic and harmful to human health in lower concentrations causing serious diseases such as chronic bronchitis, emphysema, and respiratory irritation.[1b,1c] Recently, Greenpeace Africa has announced that the largest NO2 hotspot, across the six continents, is currently located in South Africa. This has reportedly led to numerous deaths and restricted workdays resulting in a total loss greater than US$2.3 billion per year.[2] Therefore, it is crucial to develop high-performance sensors to detect NO2 gas swiftly and reliably.

For over five decades solid-state metal oxide (MOX)-based gas sensors have been developed and applied because they are small, highly sensitive (detection of different target gases in the ppm to ppb range), and cost-effective compared to classical instrumental analysis.[3] The sensing performance of MOX-based gas sensors strongly depends on the electrical conduction, which is affected by the number of charge carriers and active sites available on the surface, of the sensing material.[4] Large specific surface area and point defects, which promotes adsorption and desorption of oxygen ions and gas atoms from the sensor materials, as well as morphology and crystallographic structures, are factors reported to have an influence on the gas-sensing performance.[5] Most of the gas sensor research thus far was directed toward using n-type MOX semiconductors, such as CeO2, SnO2, WO3, or ZnO for gas sensors.[6] More recently, the focus has shifted to p-type semiconducting and MOX nanocomposite gas sensors.[7] Compared to n-type materials, some p-type gas-sensitive materials have been reported to have significant surface reactivity to oxidizing and reducing gases at lower operating temperatures, which might lead to future advances in low energy consumption gas-sensing devices.[7a,7b,8] Cupric oxide (CuO) has attracted a great deal of attention over the last decade because of its wide range of potential applications,[9] including gas sensors.[7f,10] Gas sensors based on CuO nanostructures have exhibited superior sensing performances.[7f,10f,10g] Enhanced sensor response to NH3, H2S, and NO2 gases have previously been reported for CuO nanostructures below an operating temperature of 200 °C, while sensitivity to volatile organic compound (VOC’s) was found above 200 °C.[10e,11]

Previously, we showed that the sensing performance of CuO nanomaterials is dependent on the preparation conditions, that is, reaction temperature and copper ion intermediate. By adjusting these conditions, the material properties including point defects and surface area were advertently changed, resulting in the enhanced sensitivity to NO2 gas at room temperature.[7f] In this paper, we report on the effect of prolonged reaction times on the subsequent gas-sensing performance of the CuO nanoplatelets. A thorough study on the gas-sensing performance of the nanoplatelet CuO-based gas sensors to various gases, including NH3, CO, CH4, NO2, H2S, and isopropanol (IPA), at room temperature is investigated. We demonstrate the underlying mechanism toward the enhancement of gas-sensing response by modifying the synthesis reaction base and time.

Results and Discussions

Characterization of the CuO Gas-Sensing Material

CuO nanoplatelets were prepared by the sonochemical method, and the purified CuO products were characterized by X-ray diffraction (XRD), scanning electron microscope (SEM), and high-resolution transmission electron microscopy (HR-TEM). As confirmed by the XRD patterns and indexed according to the JCPDS card no. 48-1548, all as-prepared CuO products obtained after various reaction times (i.e., 45 and 60 min) at 55 and 75 °C in the presence of various bases (C-NHOH-45 and C-NHOH-60 prepared from NH4OH at a reaction temperature of 55 °C; E-NaOH-45 and E-NaOH-60, and E-Urea-45 and E-Urea-60 prepared, at a reaction temperature of 75 °C, from NaOH and urea, respectively) were pure monoclinic-phase, as seen in Figure a–c. No characteristic peaks from the precursors and intermediates, such as Cu2O or Cu(OH)2, was been observed, suggesting that the Cu(NO3)2·3H2O and Cu2(OH)2CO3 precursors have been completely transformed into CuO. The average crystallite size of the synthesized CuO nanoplatelets was calculated using the Debye–Scherrer formula[12] and the lattice strain (ε), as shown in the Supporting Information, Table S1. The characteristic bands of CuO were observed in Raman spectra at ∼280, 330, and 610 cm–1, corresponding to the standard Ag (297 cm–1) and two Bg (345 and 632 cm–1) modes, respectively, further confirming the phase purity of the synthesized CuO nanoplatelets (see Figure S1a–c in the Supporting Information).

Figure 1

XRD patterns of the as-prepared CuO products, (a) E-NaOH-45 and E-NaOH-60, (b) E-Urea-45 and E-Urea-60, and (c) C-NHOH-45 and C-NHOH-60.

The growth mechanism of CuO nanoplatelet products can be explained on the basis of the chemical reactions and nucleation. As previously reported, three synthetic routes were followed to obtain the desired CuO products.[7f] In all three cases, the copper salt precursor, (Cu(NO3)2·3H2O), was dissolved in distilled water to form an octahedral copper intermediate complex, [Cu(H2O)6]2+. The first and the most simple synthetic route included the addition of a high concentration of NaOH solution to the reaction mixture containing [Cu(H2O)6]2+ producing square-planar complex [Cu(OH)4]2– ions (ratio of ∼2.5:1, OH– to Cu2+), which is easily transformed to CuO nuclei. During the second and third synthetic routes, the [Cu(H2O)6]2+ ion complex was transformed to [Cu(NH3)4(H2O)2]2+ with the addition of NH4OH, and Cu2CO3(OH)2 with the addition of urea, respectively. From these subsequent octahedral complex solutions, the high concentration of OH– anions present from the addition of NaOH mixture resulted in the formation of a [Cu(OH)4]2– ion complex which led to the formation of the CuO nuclei. Thus, during each of these three synthetic routes, NaOH operates as a pH buffer to control the pH value of the solution. When the critical value of the concentration of the OH– and Cu2+ ions are achieved, the precipitation of CuO nuclei begins. Initially, the CuO nuclei growth takes place by self-aggregation to produce smaller crystallite size CuO nucleus. Further aggregation leads to the formation of CuO nanoplatelets, corresponding to the growth mechanism provided by Cao et al., where the addition of hydrazine hydrate decreased the Gibbs free energy of the reaction system and transformed the 2D nanoleaf to the 3D spherical morphology.[11d]

Figures a–f and 3a–i show the SEM and transmission electron microscopy (TEM) images of the platelet-like CuO nanostructures. As shown in Figure , the size of the nanoplatelets decreased as the reaction mixture was exposed to a longer reaction time, as seen for E-NaOH-45 and E-NaOH-60, E-Urea-45 and E-Urea-60. The nanoplatelets structural units of C-NHOH-45 and C-NHOH-60 were arranged in flower-like structures in the presence of NH4OH (see Figure e,f). The length of the nanoplatelets was also affected (i.e., length decreases) with the increase in reaction times. As shown in the size distribution of E-NaOH-60 and E-Urea-60 (Figure a,d insets), the length and diameter of the nanoplatelets are affected by the change in base, displaying an average diameter and length of approximately 8–12 and 40–80 nm, respectively. More detailed crystallographic analyses on the CuO nanoplatelets were performed by selected area electron diffraction (SAED) and HR-TEM analysis. The d-spacing of the CuO lattice was found to be 0.23 and 0.25 nm corresponding to the (1̅11) and (111) crystal planes of CuO crystal planes of CuO (see Figure b,e). The SAED patterns for E-NaOH-60 and E-Urea-60 show the diffused Debye rings, denoting that the diameter of the platelets is small (see Figure c,f). The (1̅11), (111), (110), and (200) crystal planes observed in the SAED pattern confirm the monoclinic phase of the CuO nanoplatelet products. Because of the large nature of the nanoplatelets of C-NHOH-60, the SAED pattern predominantly reveals brighter Debye rings with discrete spots, denoting that the CuO flower-like structures are more polycrystalline in nature, as seen in Figure g–i. These observations are consistent with the XRD patterns showed in Figure a–c.

Figure 2

(a–f) SEM micrographs at high resolution for of the six the CuO products prepared from different reaction bases and different reaction times.

Figure 3

(a–f) High-resolution TEM images show the platelet-like nature (inset is the size distribution of the platelet diameter and length), the lattice fringes, and the corresponding SAED patterns of E-NaOH-60 and E-Urea-60. (g,f) Typical low- and high-magnification TEM images with the corresponding SAED pattern (i) of C-NHOH-60.

By varying the reaction times, the surface areas and pore volumes as determined by BET surface analyses were affected. Analyses confirmed the surface area and pore volume of E-NaOH-60 was the largest, compared to the other as-prepared CuO products (see Figure S2 and Table S2 in the Supporting Information). As seen in Figure a–c, there is an inverse correlation between the BET surface area and average crystallite sizes determined from the peaks associated with the (111) and (1̅11) XRD planes, as a function of reactions time. Experimentally, it was observed that extending the reaction times to above 60 min led to a decrease in the surface area.

Figure 4

BET surface area and average crystallite size of the as-prepared CuO products as a function of reaction time. Note the compounds marked with the asterisk (*) are from our related work[7f]E-NaOH: 75 °C, E-Urea: 75 °C, and C-NHOH: 55 °C and the line visible in (a–c) are not experimental data; it is used to guide the reader. “Adapted and reprinted from Sensors and Actuators B: Chemical, Vol. 266, D.N. Oosthuizen, D.E. Motaung, H.C. Swart, in depth study on the notable room-temperature NO2 gas sensor based on CuO nanoplatelets prepared by sonochemical method: comparison of various bases, Pages 761-772., Copyright (2019), with permission from Elsevier.”

X-ray photoelectron spectroscopy (XPS) analyses were employed to understand the chemical state and shift of the oxygen and copper atoms in the CuO products, as seen in Figures and 6. The spectra, Figure a–f, reveal doublet peaks located at around ∼933.1 and ∼953.3 eV which can be assigned to the binding energies of Cu 2p3/2 and 2p1/2, characteristic to the values associated to Cu2+. The two overlapping satellite shake-up peaks are at ∼941.4 and ∼943.9 eV; and a strong satellite shake-up peak is at ∼962.4 eV, further confirming the presence of Cu2+.[13] Both CuO and Cu2O can be distinguished by the binding energy of Cu 2p3/2 and Cu 2p1/2 in the Cu 2p XPS spectra because of the difference of ∼1 eV for CuO and Cu2O.[13a,13b,14] The O 1s core level spectra (see Figure a–f) of the six samples contain three peaks: a main peak and smaller peak (blue and light blue), Oi, at ∼529.3 and Oii, at 531 eV, where Oi is the O2– binding with Cu2+ in the crystal lattice, the Oii peak is caused by weakly adsorbed species, which include OH– or O22– species from the environment and the smallest peak, and Oiii (light pink) at ∼532 eV is associated with weakly adsorbed oxygen species.[15] The Oiv peak (light green) found at ∼533 eV is associated with the compensation of deficiencies in the subsurface or oxygen-vacancy species.[15]

Figure 5

(a–f) Chemical shift of the copper peaks of in XPS analyses.

Figure 6

(a–f) Chemical shift associated with the oxygen peak found in the CuO products.

Photoluminescence (PL) studies were conducted at room temperature, with an excitation wavelength of 325 nm, to determine the amount of point defects in the CuO products. The overlaid PL spectra of the all six as-prepared products are shown in Figure a1–c1, with the convoluted spectra of E-NaOH-60, E-Urea-60 and C-NHOH-45 shown in Figure a–c2. The band at 410 nm (i.e., ∼3.0 eV) corresponds to the band-edge emission. The defects peaks associated with possible transition vacancy of oxygen (Vo) and interstitial oxygen (Oi) in the blue-violet region (440–475 nm or 2.61–2.82 eV) of the spectra are visible in the PL spectra, as shown in Figure a–c2. The defect peaks related to singly ionized oxygen vacancy (Vo+) and both O and Cu interstitial (Oi and Cui) deficiencies are visible in the green region (between 475 and 575 nm or 2.15 and 2.82 eV) of the PL spectra. The peak located at the red region between 575 and 700 nm (i.e., 2.15 and 1.77 eV) is associated with doubly ionized oxygen vacancies (Vo++), for E-Urea-60.[16]

Figure 7

(a1–c1) Overlaid PL spectra of all of the CuO products prepared in the presence of NaOH, NH4OH and urea, with the convoluted spectra of E-NaOH-75, C-NHOH-45, and E-Urea-60 (a2–c2).

To further study the magnetic properties of the CuO products, we carried out magnetization measurements at 10 K under a maximum applied magnetic field of 10 000 Oe with a by vibrating sample magnetometer (VSM). Figure displays the magnetization versus magnetic field (M–H) curves for the CuO nanostructures. The small hysteresis loop observed at lower field indicates the E-NaOH-45 containing both ferromagnetic (FM) and paramagnetic (PM) features (see Figure a). Such behavior is also observed for the C-NHOH-60 (see Figure d). Although the FM contribution exists for the E-NaOH-60, a superimposed negative contribution turns out to be dominant (see Figure b). This diamagnetic temperature-independent involvement may originate from filled electronic d-shell of the Cu2O layer. No peaks related to Cu2O were observed during XRD analyses (see Figure a); however, the presence of Cu2O was confirmed by XPS analyses (see Figure b). This diamagnetic contribution to the total magnetization becomes dominant at higher fields. A clear hysteresis loop is witnessed for C-NHOH-45 (see Figure c). Strong PM signals are evident for E-urea-45 and E-urea-60 (see Figure e,f); corresponding to the XPS analyses where large amounts of Cu2+ ions are observed in the products (see Figure c,d). C-NHOH-45 showed the largest saturation magnetization of (i.e., 0.1 emu/g), while saturation magnetization for the E-urea-60 was too small to detect.

Figure 8

Magnetization versus magnetic field (M–H) curves of CuO nanoplatelets measured at (a–e) 10 and (f) 300 K.

Gas Sensing Analyses of CuO Nanomaterials

CuO Sensing Performance

The performance testing further include exposing the sensors to various concentrations of preselected gases [test gases including NH3, CO, CH4, NO2, H2S and IPA] in dry air and humidity conditions, at room temperature. These gases were selected as cross-sensing gases because recent reports showed that South Africa is currently one of the largest emitters of NO, while other gaseous emissions of concern include the presence of NH3, CO, and other sulphur-related compounds. Even though VOC’s emissions in the South African atmosphere are currently limited, the sensors were exposed to one VOC gas (i.e., IPA gas) for future reference. The calculated response, S, is defined as ((Ra – Rg)/Rg) × 100%) in the oxidizing gas (i.e., NO2) atmosphere, where Ra and Rg are resistance of the sensor in air and in testing gas atmosphere, respectively.[17] For the initial gas selectivity screening, each gas concentration was pulsed for 5 min, followed by synthetic air for 5 min.

Selectivity is a vital parameter for gas-sensing application. Such selectivity between NO2 and other interferences gases such as IPA, NH3, H2S, CO, and CH4 remains the fundamental challenges for practical NO2 MOX-based nanosensors. The gas-sensing responses of the CuO nanostructures prepared using various bases at different reaction times were tested toward various gases at room temperature, and the results are presented in Figure . As displayed in the histogram, the E-NaOH-45, E-NaOH-60, and E-urea-60 show improved response to NO2 gases compared to other gases. Sensors E-NaOH-45 and E-NaOH-60 showed response of values of approximately 2465 and 2498% to 40 ppm NO2 gas, respectively, at room temperature in dry air conditions. These findings clearly show that the sensor response to NO2 gas is reaction base and time dependent. None of the sensors showed a response to H2S or CO gas. All relevant sensors’ performance data related to CH4 sensing are showed in the Supporting Information (see Figures S7 and S8). To be specific, it is important to note that large drifts in sensor current (i.e., sensor instability) were observed in the presence of CH4 gas, which in turn made it difficult to determine the maximum response values. By ignoring the drift, the maximum response values of between 1700 and 2100 was estimated for sensors E-NaOH-45 and E-NaOH-60, to 40 ppm CH4.

Figure 9

Selectivity histogram of as-prepared CuO nanostructures exposed to 40 ppm CH4, H2S, CO, NH3, and NO2 gas in dry air at room temperature.

NO2 Sensing in Various Atmospheres

The dynamic curves are provided in terms of resistance and are displayed in logarithmic scale in order to allow for a better comparison between the changes of the sensor signals (relative change of the resistance). The high response of E-NaOH-45-, E-NaOH-60-, and E-urea-60-based gas sensors to NO2 gas is attributed to the presence of weakly adsorbed oxygen species and the higher relative concentration of point defects (confirmed by XPS and PL analyses, see Figures and 7, respectively) which act as favorable adsorption sites for chemisorbed oxygen species to participate in the reaction than that of other CuO nanomaterials. Our previous work and that of other researchers have shown that defects have a direct influence on the sensing performance of the materials.[7f,18] The E-NaOH-60-based gas sensor is considered as the superior performing NO2 gas-sensitive material because it showed a response to lower concentrations (i.e., 5 ppm) of NO2 gas, compared to that of E-urea-60, as seen in Figures and 11. As expected, p-type conductivity was observed for the E-NaOH-45-, E-NaOH-60-, and E-urea-60-based gas sensor products in NO2 gas conditions (see Figures and 11). A large drift in sensor resistance (see Figure a,b) was observed, which was can normally attributed to low rate of diffusion of gas molecules away from the surface of the sensor at low operating temperatures. However, at higher NO2 concentrations, the sensors recover quicker than at lower concentrations, irrespective of the operating temperature. The trend occurs irrespective of the presence of humidity (see Figure b,c). The relationship between the sensor drift and sensor response of E-NaOH-60 over a period of 12 h can be seen in Figure S4, Supporting Information, thus leading us to conclude that at lower concentration the NO2 molecules chemically bond to the active sites on the surface of the sensing layer. At higher concentrations, however, a combination of chemi- and physisorption occurs because the NO2 molecules are competing for the same active sites. Prolonging the duration of the gas cycles (see Figure S5, Supporting Information) also proved to be ineffective. The interaction on the gas–solid interface is dependent on the competition between the oxidizing gas and the water vapor for the preadsorbed oxygen species.

Figure 10

(a) Dynamic resistance curves of the E-NaOH-45 and E-NaOH-60, extracted from different NO2 concentration cycles and (b,c) the dynamic resistance curve and calculated response (S) variation of E-NaOH-60 to NO2 concentrations ranging from 20 to 100 ppm in various relative humidity (%) at room temperature. *Note: the lines visible in (c) are not experimental data; it is used to guide the reader.

Figure 11

(a) Dynamic resistance curves of E-urea-60, extracted from different NO2 concentration cycles and (b,c) 40 ppm NO2 gas cycles in dry synthetic air and various relative humidity (%) conditions at room temperature. (d) Calculated response, S, of E-urea-60 to 40 ppm NO2 in various (RH %) conditions at room temperature. *Note: the lines visible in (d) are not experimental data; it is used to guide the reader.

During the exposure of sensor E-urea-60 to four cycles of 40 ppm NO2, limited drift in sensor resistance is observed compared to that of the sensor in varying concentrations (see Figure a,b). This can be ascribed to the prolonged duration of the gas cycles which allowed the gas molecules efficient time to diffuse away from the surface of the sensor. A decline in sensor resistance was observed after the sensor was exposed to NO2 in humid conditions (as seen in Figure c). The drift in resistance of the E-urea-60 sensor in the presence of humidity can be attributed to the gradual formation of stable chemisorbed OH-groups at the oxide surface for prolonged exposure.[19] Compared to dry air condition, a decrease in the sensor response for E-urea-60 was observed in testing conditions with more moisture present (see Figures and 11d).

Stability of Sensors in the NO2 Atmosphere

When exposing the E-NaOH-45, E-NaOH-60, and E-urea-60 sensors to NO2 in dry air and humidity conditions after aging for a year, more complex gas-sensing performances are observed (see Figures and S7 in the Supporting Information). An increase in initial sensor resistance (i.e., before the target gas was introduced) was observed for the sensors (see Figure a,b compared to Figure a,b), which can be ascribed to the presence of additional adsorbed species, like H2O and O–, on the surface of the sensors leading to overall increase in sensor conductivity. Under dry air conditions, a more pronounced drift in both the sensors’ resistance was observed (see Figure a). The NO2 species is most likely to now bond to the additional active sites on the surface of the sensing layers. In the scenario where the sensing performance is tested in dry air conditions (see Figure a), prolonging duration of the gas cycles even more (i.e., 20 and 20 min: gas in and gas out) or heating the sensors to above 100 °C between gas NO2 gas cycles may remove the additional species, thus regenerating (restore) the sensors to their original condition. However, enhanced sensing performance for E-NaOH-45 and E-NaOH-60 was observed in the presence of humidity (see Figure b,c). Because of the likelihood of additional adsorbed species present on the surface of the sensing layer, the most likely scenario is the reaction of preadsorbed oxygen, namely O–, with water resulting in the formation of terminal hydroxyl groups (i.e., the cause of more distinct the sensor drift in more humid atmosphere). Thus, the interaction mechanism is now dependent on the competition between the oxidizing gas and the water for the additional preadsorbed oxygen species.

Figure 12

(a) Dynamic resistance curves of E-NaOH-45 and E-NaOH-60, extracted from different NO2 concentration cycles and (b) 40 ppm NO2 cycles in dry synthetic air and various RH (%) conditions at room temperature. (c) Calculated response of E-NaOH-60 to 40 ppm NO2 gas at various RH (%).*Note: the lines visible in (c) are not experimental data; it is used to guide the reader.

Gas-Sensing Mechanism Arising from the Material Properties

The manner in which the conduction occurs in the MOX sensing layer is important for the magnitude of the sensor signal (sensor response).[20] The general sensing mechanisms of the semiconductor gas sensors are based on the ionosorption and oxygen-vacancy models.[5b] The adsorption and desorption of test gas on the surface of the CuO-based gas sensor leads to the changes of electrical resistance and directly affect the response to the test gas. In general, the gas-sensing response of MOX-based gas sensors can be described as followswhere a and b are constants, S is the response of sensor, and C is the concentration of target gas. The value of b is normally from 0.5 to 1 which depends on the charge of the surface species and stoichiometry of the elementary reactions on the sensor surface. The b value is 0.5 when the adsorbed surface oxygen ion is O2–, and the b value is between 0.5 and 1 or just above 1 when the adsorbed surface oxygen ion is O–. Therefore, the oxygen species adsorbed on the surface of the gas sensor can be deduced by the b value. At a certain working temperature, eq can be rewritten asThere is a linear relation between log(S – 1) and log(C), and b is the slope of the linear function. Figure a shows the relationship between the response of E-NaOH-45 and E-NaOH-60- and E-NaOH-60-based sensors and the NO concentration, and the corresponding curve of log(S – 1) ≈ log(C) is shown in Figure b. The calculated b is 1.0 and 1.139, respectively, which signifies that O– species are adsorbed on the surface of the sensing layers.

Figure 13

Relationship between (a) calculated response, S, of sensors E-NaOH-45 and E-NaOH-60 and the NO2 concentration at room temperature (22 °C) and (b) corresponding log(S – 1)–log(C) curves.

It is suggested that in the case of these CuO sensors, the NO2 firstly reacts with the available adsorbed O– species forming NO2–(ads) species. Then, a hole-accumulation layer is created at the boundaries of the CuO nanograins by chemisorption of NO2. The generation of holes makes the p-type CuO nanoplatelets more conductive.

When the flow of NO2 gas is stopped, the adsorbed NO2–(ads) species evaporates, leaving behind the captured electrons in the p-type CuO and decreasing the hole density by electron–hole compensation, resulting in the change in resistance. If however the NO2–(ads) species does not evaporate or diffuse away from the surface, no electron is donated back into the CuO bulk; thus, the hole density is not decreased (i.e., drift in sensor resistance is observed).

For the aged sensors, the surface chemisorbed O–(ads) and NO2–(ads), as well as their interactions, have a significant impact on the sensing properties because there is more O–(ads), as illustrated by the following reactions.

As is known, the fundamental principle of gas-sensing MOX materials depends on the change in the resistance (or conductivity) in the sensor’s circuit when exposing it to different atmospheres (i.e., different gases). It is established that the gas-sensing performance is influenced by many factors, such as the charge carrier concentration on the surface of the sensing materials, morphology, and large surface-to-volume ratio of the sensing materials.[20,21] In the present study, the variations in charge carrier concentrations and morphology were found to change the gas-sensing characteristics of the CuO nanostructures. To get a further indication of the factors that influenced the NO2 gas-sensing performance of E-urea-60, E-NaOH-45, and E-NaOH-60, in dry air conditions, their calculated NO2 response at 40 ppm was compared to the crystallite size, surface area, carrier concentration (as seen in the Supporting Information, Figure S3: Hall-effect analyses) and defects (determined by PL and XPS analyses) as seen in Figure a–d. From this comparison, it is apparent that the enhanced response of E-urea-60 and E-NaOH-45, to NO2 gas, respectively, was mostly influenced by the crystallite size, surface area, and the carrier concentration of the sensing materials. To be specific, the largest surface area of 67.6 m2 g–1 by E-NaOH-60 showed an enhanced NO2 gas sensitivity with a response of 2498% at 40 ppm, followed by E-NaOH-45 with a response of 2465% and a surface area of 64.8 m2 g–1. As the crystallite sizes decreased, the gas-sensing response toward NO2 gas decreases, denoting that in this case (taking all of the material properties into account) the optimum grain size contributing is 9.7 nm. A previous report has disclosed that the performance of the sensing can be enhanced by manipulating the size of the nanostructures to a critical value of approximately 14 nm.[22] The gas-sensing performance of similar CuO nanostructures, such as nanoleaves and platelet-like building blocks arranged as hierarchical structures, were reported to be selective to VOC’s at higher operating temperatures.[10f,11d] Selectivity to NO2 gas was reported for chromium-doped CuO nanorods at 250 °C, by Lee et al.[23] Limited cases were reported where undoped CuO nanostructure-based gas sensors showed selectivity to NO2 gas at room temperature;[7f,24] however, other studies include the enhanced NO2 response of NiO nanosheets[5a,8b] and at room temperature.[8b]

Figure 14

Relationship between the crystallite size, surface area, carrier concentration, defects, and the NO2 gas-sensing response of E-NaOH-45, E-NaOH-60 and E-urea-60 (a–d) *Note: the blue and black lines visible in (a,b) are not experimental data; it is used to guide the reader.

Conclusions

Various CuO nanoplatelets were successfully prepared and characterized, before being subjected to gas-sensing performance analyses at room temperature (22 °C) in dry and humid conditions. Both CuO products, prepared at 75 °C for 45 and 60 min in the presence of NaOH, and urea (E-NaOH-45, E-NaOH-60 and E-urea-60), showed promise as NO2 gas sensors, respectively. E-NaOH-60, with an average crystalline size of 9.7 nm and largest surface area of 67.6 m2 g–1 and highest carrier concentration, showed an enhanced response of 2498%, while E-urea-60 (i.e., crystallite size of 7.7 nm and surface area of 56.1 m2 g–1) showed a smaller response of 469% to 40 ppm of NO2 gas at room temperature in dry air. This could be attributed to the much larger specific area, the more active sites and the more abundant hole concentration provided by the E-NaOH-60 nanoplatelets. This finding contributes to the research and development of NO2 selective p-type MOX-based gas sensor operating at room temperature.

Experimental Section

Materials

All reagents were of analytical grade and used as purchased from Sigma-Aldrich. The NH3, CO, H2S, NO2 and CH4 gases were purchased from AFROX, South Africa.

Preparation of CuO Nanostructures

As shown in ref (7f), the CuO powders, E-NaOH: 75 °C, C-NHOH: 55 °C, and E-Urea: 75 °C, were prepared under specific reaction conditions and exhibited the best gas-sensing performance. During this study, the same preparation method used in ref (7f) was followed, under the relevant reaction temperatures, that is, 55 and 75 °C; however, the reaction times were prolonged to 45 and 60 min in this case. Forgoing the chosen notation, resulting CuO products were denoted as E-NaOH-45 and E-NaOH-60, according to the reaction temperature (75 °C), time, and the base used (NaOH). Thus, in the cases where NH4OH was used in the preparation of CuO nanostructures, the products were denoted as C-NHOH-45 and C-NHOH-60, while the products prepared in the presence of urea were denoted E-urea-45 and E-urea-60, respectively.

Characterization

The structural features and phase purity of the synthesized CuO nanoparticles were investigated using an X-ray diffractometer (PANalytical X′-pert PRO-PW 3040/60) fitted with a Cu Kα (λ = 0.1541 nm) radiation source,a HORIBA Jobin-Yvon HR800 Raman microscope at room temperature and transmission electron microscope (Tecnai F20). XPS analyses were carried out using a PHI 5000 Versaprobe-scanning ESCA microprobe. A low-energy Ar+ ion gun and low-energy neutralizer electron gun were used to minimize charging on the surface. Monochromatic Al Kα radiation (hν = 1486.6 eV) was used as the excitation source. The morphology was characterized using a SEM (Zeiss). The optical emission properties were characterized at room temperature using a pulsed laser-PL spectrometer, equipped with a Kimmon He–Cd laser. The surface area and pore size distribution were analyzed using a N2 adsorption–desorption isotherm (Micromeritics TRISTAR 300 surface area analyzer, USA). For magnetization measurements at low temperature, the VSM platform of a Dynacool facility was used (Quantum Design, USA). This platform enables measurement of magnetic moment to an accuracy of ±0.5% and with a noise rejection rate of better than 6H10–7 emu.

Fabrication and Analyses of Sensing Layer

The gas-sensing properties of the as-prepared CuO platelets powder were analyzed using a KSGAS6S gas testing station (Kenosistec, Italy). Figure shows the schematic diagram of the gas testing station.

Figure 15

Schematic diagram of the gas testing station used during the gas-sensing analyses.

As we reported previously, prior to gas-sensing analyses, the sensing layers were fabricated by dispersing the as-prepared CuO powders in ethanol before it was drop-coated onto alumina substrates (containing a heater on the bottom side) and dried in an oven at 100 °C for 1 h. The average thickness of the sensing layers range between 15 and 20 μm (see the Supporting Information, Figure S3a,b).

The fabricated sensors were placed in an airtight chamber and tested for sensitivity toward NH3, CO, CH4, NO2, H2S, and IPA gases at room temperature. The fabricated sensors were exposed to gas concentrations ranging from 5, 10 to 100 ppm with 10 ppm intervals. Dry air atmosphere (79% nitrogen and 21% oxygen, i.e., 0% relative humidity) and relative humidity (RH) of 10–60% conditions were used to conduct the measurements. The resistance of the device was measured using a Keithley 6487 picoammeter/voltage source meter.