CQDs@NiO: An E ﬃ cient Tool for CH 4 Sensing

Featured Application: The composite material made of CQDs on NiO surface can be used as sensing material in CH 4 detection. As show-case a home-made CH 4 sensor has been built, based on the CQDs@NiO. Abstract: A composite material based on carbon quantum dots (CQDs) and NiO was prepared and tested for methane sensing. The synthesis procedure is simple and foresees the preparation of the CQDs by citric acid pyrolysis and NiO by hydrothermal synthesis. A phase sonication and stirring procedure yielded the composite CQDs@NiO at di ﬀ erent loads. The composites were characterized by X-ray di ﬀ raction, ultraviolet–visible light (UV–Vis) spectroscopy, SEM microscopy, energy-dispersive spectroscopy (EDS) mapping, and surface area, porosity, and impedance measurements. A gas sensor was built in-house and used to probe the response of the synthesized samples to CH 4 detection, at constant environmental humidity. The CQDs@NiO at 1% weight load displayed excellent performances in terms of gas response both vs. temperature and vs. concentration, whereas higher loads resulted in CQD aggregation and diminished output. Response / recovery times of the 1%CQDs@NiO sample were good, as well as the selectivity and the stability over time and for variable environmental humidity. The estimated limit of detection was 0.1 ppm.


Introduction
Methane gas is extensively employed as fuel in houses and industries, and as a propellant for automotive vehicles [1][2][3]. As a chemical raw material, it also finds application in the production of carbon black, acetylene, carbon disulfide, and methanol [4], among others. CH 4 is the main gas component in coal mines, and it is extremely dangerous, since it easily explodes when it reaches a volume concentration between 4.9% and 15.4% in air. Because of the odorless and flammable nature of this asphyxiant gas, the development of sensitive detectors for real-time detection at concentrations lower than its explosive limit is in pressing need. Furthermore, low-cost fabrication of CH 4 sensors is an asset in order to make such devices easily accessible to the general public. In this framework, the ideal combination of sensors properties includes sensitivity, quick response, low-cost sensing material and sensor components. Platforms such as the electrochemical ones, which often proved to be effective toward multiple targets [5][6][7][8][9][10][11][12], were also applied to CH 4 sensing [13][14][15], which required the gas being trapped in liquids. It must be added that, for other gases such as CO 2 , potentiometric sensors were developed, without a need for trapping in liquid, albeit working at high temperatures and requiring a BaCO 3 layer for limiting the humidity interference [16]. Under these premises, we designed and probed a low-cost material to insert in a home-assembled sensor, built with an easily available set of components, for high-sensitivity CH 4 detection. The rationale in the approach for defining the sensing material is the best match of substrate and additional modifying material to enhance the sensing properties. We opted for NiO as a substrate, because of its good sensing this type of sensor has the advantage of a facile synthesis using low-cost materials. Conversely, since the performance is largely dependent on the amount of CQDs, it may require some effort to tune the synthesis toward the best working composition.

Synthetic Procedure
The NiO sample was prepared via a hydrothermal route, by adding 50 mL of 1.2 mol·L −1 (NH 2 ) 2 CO dropwise to 50 mL of 0.6 mol·L −1 Ni(NO 3 ) 2 , at room temperature, under vigorous stirring. The pH was adjusted to 8.0 using HNO 3 . The suspension was then transferred into a 200 mL Teflon-lined stainless-steel autoclave, carefully sealed, and heated at 180 • C for 24 h in a furnace. After gentle cooling, the powder was repeatedly washed with deionized water and dried at 60 • C overnight. The powder was then placed in a tubular oven and heated at a rate of 10 • C/min to 500 • C, and then calcined for 3 h. CQDs were synthesized by pyrolyzing 2 g of citric acid (CA) in a round-bottom flask, immersed in an oil bath at 200 • C for 30 min. The yellow-orange liquid supplemented with deionized water, up to a volume of 50 mL. Quotas of 0.5 mL, 1 mL, and 1.5 mL of the CQD solutions were each supplemented with 2 g of NiO and brought to 5 mL using deionized water. The dispersions of NiO in CQD aqueous solutions were subjected to phase sonication and stirring, i.e., cycles of 10 min sonication in a 40 kHz bath followed by room temperature stirring at 1000 rpm. The deposition of the yellow-orange CQDs from the aqueous solution to the NiO surface resulted in the depletion of the colored material from the solution, which was, as a consequence, discolored. Therefore, the phase sonication procedure was deemed complete when the discoloration of the solution took place. The samples were centrifuged at 3000 rpm and then dried overnight at 80 • C. They were then labeled 1%CQDs@NiO, 2%CQDs@NiO, and 3%CQDs@NiO, reflecting the ratio by weight of CA vs. NiO.

Materials and Equipment
Citric acid (CA), Ni(NO 3 ) 2 , (NH 2 ) 2 CO, ethanol, and acetone were purchased from Sigma-Aldrich (Sigma-Aldrich, St. Louis, Missouri, USA). All chemicals were of reagent grade and used without any further purification. The samples were characterized by X-ray diffraction (XRD), using an X' pert pro X-ray diffractometer by Philips (Almelo, The Netherlands), operated with CuK-Alpha radiation. Ultraviolet-visible light (UV-Vis) spectra were acquired using a Varian Cary 50 Scan spectrophotometer (Agilent, Santa Clara, USA). SEM images were acquired with a Zeiss Auriga field-emission scanning electron microscope instrument (Jena, Germany) operating at 7 kV. The energy-dispersive spectroscopy (EDS) maps were obtained by coupling the field-emission scanning electron microscope (SUPRA™ 35, Carl Zeiss SMT, Oberkochen, Germany) with energy-dispersive microanalysis (EDS/EDX, INCAx-sight, Model: 7426, Oxford Instruments, Abingdon, Oxfordshire, UK), operating at 20 kV. High-resolution transmission electron microscopy (HR-TEM) was performed on an FEI Tecnai G2 20 high-resolution transmission electron microscope (FEI Company, Hillsboro, USA). The specific surface area and porosity were measured on Micrometrics Instrument Corporation ASAP 2020 (Norcross, Georgia, USA) using N 2 adsorption-desorption isotherms. The methane sensor was assembled according to the scheme in Figure S1 (Supplementary Materials). It allows resistive gas response as a function of temperature and concentration, simultaneously keeping record of the ambient conditions (external temperature and humidity). CQDs@NiO powders were ground in an agate mortar, mixed with deionized water in a ratio of 5:1 until a paste was formed, and then coated onto the ceramic support, which was then connected to four Pt wires through two Au rings. The coated ceramic elements were then dried at 200 • C for 5 h in air for water to evaporate and to ensure stability. The thickness of the coating was 200 m. The sensor was placed in a chamber of 1 L of volume with its own heating system and temperature control. Synthetic air (Sigma-Aldrich), supplemented with saturated vapor of distilled water, was used both as reference and as diluting gas. All measurements of gas response were made at a relative humidity (RH) of 20%, unless stated otherwise. The resistance was then measured with a signal-to-noise ratio of about 4% at the minimum level and about 2% at the maximum level.

Results and Discussion
The achievement of the composite CQDs@NiO was evaluated in several ways. The synthesis of CQDs via CA pyrolysis yielded the typical UV-Vis spectrum with a maximum absorption at 352 nm [42], as reported in Figure 1A. The NiO sample and the composites were analyzed by X-ray diffraction, as shown in Figure 1B

Results and Discussion
The achievement of the composite CQDs@NiO was evaluated in several ways. The synthesis of CQDs via CA pyrolysis yielded the typical UV-Vis spectrum with a maximum absorption at 352 nm [42], as reported in Figure 1A. The NiO sample and the composites were analyzed by X-ray diffraction, as shown in Figure   The average particle size was estimated by analysis of the peak's full width at half maximum (FWHM) using the Scherrer formula D = K λ/βcos(θ), where K is a constant (ca. 0.9), λ is the X-ray wavelength used to collect the XRD patterns (i.e., 1.5418 Å), θ is the Bragg angle, and β is the pure diffraction broadening of a peak at half-height due to the crystallite dimensions, yielding an appraised value of 35 nm.
From the morphological point of view, NiO synthesized via urea and nickel nitrate and calcined at 500 °C has a peculiar texture of polygonal-shaped nanoparticles, with an average size of 40 nm, which tend to merge to yield a sort of bidimensional extended structure ( Figure 2(a1)).
The texture was retained by adding 1% CQDs, although the porosity appeared affected to some extent. The 2%CQDs, and 3%CQDs samples still retained the polygonal-shaped structure, but the deposition of the CQDs definitively helped the merging of the nanoparticles, resulting in an extended structure (Figures 2(b1-d1)).
This reflects the distribution of the carbon on the NiO surface as mapped by EDS and shown in Figures 2(b2-d2), which was quite homogeneous for the composite 1% CQDs and had scattered islands in the 2%CQDs and 3%CQDs composites. The typical TEM image of the synthesized NiO is shown in Figure 2(c1). The corresponding HRTEM image is presented in Figure 2(b1), where an interplanar spacing of 0.208 nm was measured from the fringe pattern, corresponding to the (200) plane. From Figures 2(b3-d3), it can be found that the CQDs were homogeneously distributed onto the NiO surface in the 1% CQDs sample (round dots embedded in the NiO structure), and that their density increased in the 2% CQDs sample and tended to merge in the 3% CQDs sample.
Surface area, average pore size, and total pore volume measurements were in line with the structure and morphology analysis (see Table 1). The surface and pore parameters of NiO were slightly modified by the presence of 1% CQDs, resulting in a small reduction of the average pore size and increase of the surface area. A higher load of QDs resulted in a partial pore occlusion and reduced surface area as expected for a more consistent deposition of the QDs into the pores and an aggregation of the quantum dots. A higher load resulted in a more pronounced effect. The average particle size was estimated by analysis of the peak's full width at half maximum (FWHM) using the Scherrer formula D = K λ/βcos(θ), where K is a constant (ca. 0.9), λ is the X-ray wavelength used to collect the XRD patterns (i.e., 1.5418 Å), θ is the Bragg angle, and β is the pure diffraction broadening of a peak at half-height due to the crystallite dimensions, yielding an appraised value of 35 nm.
From the morphological point of view, NiO synthesized via urea and nickel nitrate and calcined at 500 • C has a peculiar texture of polygonal-shaped nanoparticles, with an average size of 40 nm, which tend to merge to yield a sort of bidimensional extended structure (Figure 2(a1)).
The texture was retained by adding 1% CQDs, although the porosity appeared affected to some extent. The 2%CQDs, and 3%CQDs samples still retained the polygonal-shaped structure, but the deposition of the CQDs definitively helped the merging of the nanoparticles, resulting in an extended structure (Figure 2(b1-d1)).
This reflects the distribution of the carbon on the NiO surface as mapped by EDS and shown in Figure 2(b2-d2), which was quite homogeneous for the composite 1% CQDs and had scattered islands in the 2%CQDs and 3%CQDs composites. The typical TEM image of the synthesized NiO is shown in Figure 2(c1). The corresponding HRTEM image is presented in Figure 2(b1), where an interplanar spacing of 0.208 nm was measured from the fringe pattern, corresponding to the (200) plane. From Figure 2(b3-d3), it can be found that the CQDs were homogeneously distributed onto the NiO surface in the 1% CQDs sample (round dots embedded in the NiO structure), and that their density increased in the 2% CQDs sample and tended to merge in the 3% CQDs sample.
Surface area, average pore size, and total pore volume measurements were in line with the structure and morphology analysis (see Table 1). The surface and pore parameters of NiO were slightly modified by the presence of 1% CQDs, resulting in a small reduction of the average pore size and increase of the surface area. A higher load of QDs resulted in a partial pore occlusion and reduced surface area as expected for a more consistent deposition of the QDs into the pores and an aggregation of the quantum dots. A higher load resulted in a more pronounced effect.
Further analyses were carried out on the best-performing composite, i.e., the 1%CQDs@NiO. In Figure 3b, the gas response vs. concentration is reported for detection at 150 • C, in the range 5-50 ppm, showing a reasonably linear response (the R 2 coefficient of the linear regression is 0.99). In the inset, the error bars of the measurements are reported. The detection range is very good, since most of the proposed devices detect CH 4 in higher concentration ranges, whereas it is desired to be capable of detecting even small gas leaks. From the slope of the linear fit, it is possible to estimate to the limit of detection (LOD) as: where rms noise and slope are the root-mean-square error of the baseline and the slope value of the linear curve [27]. For the 1%CQDs@NiO sensor, we appraised an LOD of 0.1 ppm. Figure 3c demonstrates the dynamic response and recovery curve of the 1%CQDs@NiO to 20 and 30 ppm CH 4 at 150 • C. The sensor response increased dramatically when CH 4 gas was injected into the test chamber and was swiftly recovered to the initial resistance when the sensor was exposed to air. The time taken by the sensor to reach 90% of the total resistance change is defined as the response time in the case of gas adsorption, whereas the recovery is the time needed for the specular gas desorption process [54,55]. The response and recovery times of the NiO sensor toward 30 ppm CH 4 were calculated to be about 10 s and 14 s, respectively. The performance of the 1%CQDs@NiO composite-based sensor is compared to previously reported data in Table 2. The reproducibility of the 1%CQDs@NiO-based sensor measurements is reported in Figure S2 (Supplementary Materials).  [45] The stability of the 1%CQDs@NiO sensor over time was probed by testing the response to 30 ppm CH 4 at 150 • C in a time span of three weeks, revealing a rather constant behavior (Figure 3d). The effect of the relative humidity on the 1%CQDs@NiO sensors was probed by monitoring the response to 30 ppm CH 4 at 150 • C as a function of increasing RH (Figure 3d). The sensor maintained an excellent CH 4 -sensing performance with little response loss in a high-RH environment due to the increased oxidation reaction facilitated by the CDQ decoration of the NiO (Figure 3e). Finally, the selectivity of Appl. Sci. 2020, 10, 6251 7 of 11 the 1%CQDs@NiO sensor was evaluated by investigating the comparative response to four target gases (CH 4 , CO 2 , NH 3 , and acetone) at 30 ppm concentration, operating at 150 • C. The results are shown in Figure 3f and indicate a marked selectivity toward CH 4 , with poor responses to the other gases. materials is reported as a function of temperature. The optimal value was 175 °C for pristine NiO and 150 °C for the composites. Previous NiO-based sensors for methane detection worked at optimum temperatures of 200 °C (ultra-thin NiO [17]), 250 °C (rGO-NiO [28]), and 330 °C (SnO2-NiO [45]), which in all cases are higher than in the current work. This optimum temperature is also lower or, at best, equal to the working temperature of SnO2-based sensors [41][42][43]. The 1%CQDs@NiO and 2%CQDs@NiO composites also displayed a larger Rg/Ra ratio than pristine NiO, whereas the highest loaded composite 3%CQDs@NiO displayed the worst performance.  The detection mechanism is based on the reaction between CH 4 and adsorbed oxygen according to Equation (2).
The first step of the overall process is the adsorption of O 2 by reduction on the surface to O − , with formation of a hole accumulation layer (HAL) related to electrostatic interaction of opposite charges [56]. In this regard, the morphology plays a role, since the edges of the polygonal-shaped structure of the NiO nanoparticles may be efficient in capturing O 2 . The effects of the deposition of the CQDs on NiO can be reconducted to the formation of a hetero-junction potential barrier at the interface between CQDs and NiO, which greatly facilitates the adsorption and diffusion of the methane molecule [28,57]. More in detail, the deposition of CQDs on the NiO causes an increase in resistance following the electron-hole equilibration at the interface. On the other hand, the interface between CQDs and NiO junctions is a source of defects which are potential gas adsorption sites. Therefore, a larger number of gas molecules may adsorb and react at the interface, resulting in a decreased thickness of the HAL and higher resistance of the CQDs-NiO as compared to pure NiO sensor. By the time the chemical reaction is accomplished, electrons flow from the CQDs to NiO, which results in a further decrease of the hole concentration in NiO and even larger resistance [58,59]. Furthermore, an electrocatalytic effect may take place [60] through methane activation. The detection mechanism, then, occurs via a redox reaction between O − and CH 4 and injection of electrons into the solid, with a depletion of the HAL and an increase of the resistance (Figure 4) [29]. The efficiency of the detection mechanism is negatively affected by the CQD aggregation, with a consequent worsening of the detection performance at higher loads.
Appl. Sci. 2020, 10, x 8 of 11 The first step of the overall process is the adsorption of O2 by reduction on the surface to O − , with formation of a hole accumulation layer (HAL) related to electrostatic interaction of opposite charges [56]. In this regard, the morphology plays a role, since the edges of the polygonal-shaped structure of the NiO nanoparticles may be efficient in capturing O2. The effects of the deposition of the CQDs on NiO can be reconducted to the formation of a hetero-junction potential barrier at the interface between CQDs and NiO, which greatly facilitates the adsorption and diffusion of the methane molecule [28,57]. More in detail, the deposition of CQDs on the NiO causes an increase in resistance following the electron-hole equilibration at the interface. On the other hand, the interface between CQDs and NiO junctions is a source of defects which are potential gas adsorption sites. Therefore, a larger number of gas molecules may adsorb and react at the interface, resulting in a decreased thickness of the HAL and higher resistance of the CQDs-NiO as compared to pure NiO sensor. By the time the chemical reaction is accomplished, electrons flow from the CQDs to NiO, which results in a further decrease of the hole concentration in NiO and even larger resistance [58,59]. Furthermore, an electrocatalytic effect may take place [60] through methane activation. The detection mechanism, then, occurs via a redox reaction between O − and CH4 and injection of electrons into the solid, with a depletion of the HAL and an increase of the resistance (Figure 4) [29]. The efficiency of the detection mechanism is negatively affected by the CQD aggregation, with a consequent worsening of the detection performance at higher loads.

Conclusions
A new composite material was prepared using CQDs and NiO. The synthetic procedure is quite straightforward and foresees the separate achievement of the CQDs and of the NiO and the combination of the two by phase sonication and stirring. Composites at different loads were, then characterized, indicating a good dispersion of the CQDs on the NiO at the lowest load and partial reaggregation at higher loads. The sensing properties of the composite material toward CH4 were probed using a low-cost, home-made device and parallel the dispersion of the CQDs on NiO. The lowest loaded sample shows the best performance in terms of operational temperature. Furthermore, it has an estimated LOD of 0.1 ppm and a good response/recovery time, is stable over time, can operate in environments with high humidity (80%), and shows high selectivity.
Supplementary Materials: The following are available online at www.mdpi.com/xxx/s1: Figure S1. Schematic layout of the gas sensor measuring module; Figure S2. Reproducibility of the CH4 detection by the1%CQDs@NiO sensor at 30 ppm; Text. Description of the assembled device.
Funding: This research received no external funding.
Acknowledgments: René Bakker, currently working at CINECA, Rome Italy, is gratefully acknowledged for helping assemble the sensor device.

Conflicts of Interest:
The author declares no conflict of interest.

Conclusions
A new composite material was prepared using CQDs and NiO. The synthetic procedure is quite straightforward and foresees the separate achievement of the CQDs and of the NiO and the combination of the two by phase sonication and stirring. Composites at different loads were, then characterized, indicating a good dispersion of the CQDs on the NiO at the lowest load and partial reaggregation at higher loads. The sensing properties of the composite material toward CH 4 were probed using a low-cost, home-made device and parallel the dispersion of the CQDs on NiO. The lowest loaded sample shows the best performance in terms of operational temperature. Furthermore, it has an estimated LOD of 0.1 ppm and a good response/recovery time, is stable over time, can operate in environments with high humidity (80%), and shows high selectivity.
Funding: This research received no external funding.