Abstract

The efficient monitoring of the environment is currently gaining a continuous growing interest in view of finding solutions for the global pollution issues and their associated climate change. In this sense, two-dimensional (2D) materials appear as one of highly attractive routes for the development of efficient sensing devices due, in particular, to the interesting blend of their superlative properties. For instance, graphene (Gr) and graphitic carbon nitride g-C3N4 (g-CN) have specifically attracted great attention in several domains of sensing applications owing to their excellent electronic and physical-chemical properties. Despite the high potential they offer in the development and fabrication of high-performance gas-sensing devices, an exhaustive comparison between Gr and g-CN is not well established yet regarding their electronic properties and their sensing performances such as sensitivity and selectivity. Hence, this work aims at providing a state-of-the-art overview of the latest experimental advances in the fabrication, characterization, development, and implementation of these 2D materials in gas-sensing applications. Then, the reported results are compared to our numerical simulations using density functional theory carried out on the interactions of Gr and g-CN with some selected hazardous gases’ molecules such as NO2, CO2, and HF. Our findings conform with the superior performances of the g-CN regarding HF detection, while both g-CN and Gr show comparable detection performances for the remaining considered gases. This allows suggesting an outlook regarding the future use of these 2D materials as high-performance gas sensors.

1. Introduction

A fast-paced race for devices’ miniaturization based on 2D materials is marking the particularly rapid evolvement in the research and development of carbon nanomaterials (CNM) in biomedical systems, electronic appliances, computing devices, and sensor manufacturing industries [15]. Advancements in these application fields depend on the ability to grow CNM on various substrates, with controlled morphologies and quality, while increasing their efficiency of getting assembled in complex architectures [6]. Since its first isolation in 2004, graphene (Gr) has constantly been the highlight superlative star of 2D materials. Formed by sp2-hybridized carbon atoms bonded covalently and packed densely to form honeycomb crystal lattice [7, 8], graphene is a single-atom-thick sheet semimetal with commendable electron transport properties and superior thermal and electrical conductivities in addition to superior mechanical properties [913]. In contrast to graphene, graphitic carbon nitride g-C3N4 (g-CN) is a semiconducting nanomaterial with a crystal structure analogous to graphite constructed from Van der Waals sheets of sp2 hybrid carbon and nitrogen atoms. g-CN is an abundant, inexpensive, and easy to manufacture nanomaterial at large scale [14]. This, added to its high electrical properties and good thermal and chemical stabilities, has recently created an increasing interest around it. With an appropriate bandgap and high electron mobility, g-CN also constitutes a potentially good candidate for various applications in energy storage, solar cells, gas sensing, and catalysis [1518].

Gas sensing is one of the important applications of Gr and associated 2D nanomaterials, owing mainly to several interesting properties such as (i) large surface area (for Gr 2630 m2/g), (ii) good conductivity, (iii) high gas sensitivity, (iv) and the possibility of tailored surface functionalization for a specific gas absorption. In Gr, the gas absorption is a result of a pull electron effect [19, 20]. In fact, the adsorbed molecules change the Fermi level of Gr and hence its conductivity, due to hybridization and coupling on its surface. They act as electron donors in case their valence band energy is higher than the Fermi level of Gr and as electron acceptors in the opposite case when the valence band energy is lower than Fermi level of Gr [21, 22].

Nowadays, highly performing portable sensors have become a key necessity for all technological advances. While metal-oxide-semiconductor based gas sensors are far commonly used, their performances remain very limited due to their short life, poor selectivity, temperature instabilities, and high sensitivity to environment’s perturbations. The use of nanostructured materials with higher gas-sensing capabilities has therefore become inevitable [5].

Besides, the development of g-CN based gas sensors, compared to those based on Gr or transition metal disulfide nanosheets, is becoming more attractive [23]. Many approaches have been adopted so far for the development of this material in view of its incorporation in gas-sensing applications. In this sense, researchers attempted to combine g-CN with other materials to improve its performances [24]. For instance, the use of porous g-CN in nanostructure designs as gas sensors appeared to greatly improve their selectivity and their sensitivity in ambient conditions. Similarly, hybridization of g-CN with other materials was found as a promising route for further enhancement in detection performances [23]. The resulting heterostructures made of metal oxide nanoparticles attached to the surface of 2D materials were found to exhibit higher detection properties [14]. For instance, g-CN/MOX heterostructures revealed much superior detection performances compared to a pure MOX sensor (MnO2, ZnO, and TiO2). However, some drawbacks remain, which prevent the direct application of these sensors, such as the high working temperature and the slow sensing response [25]. These achievements are attributed, on one hand, to the increased surface area obtained by the combination of 2D g-CN with the MOx leading to an increased number of gas-material interaction sites [20] and to the low operating temperature while improving the gas sensitivity and selectivity on the other hand [23]. Despite these interesting gas-sensing performances, g-CN presents some limitations when compared to Gr such as its moderate band gap that requires a high excitation voltage and its fast electron-hole pairs recombination leading to low responsivity. In addition, g-CN adsorbs most of gases and liquids, while Gr is known to be impermeable to most of fluids which reduces the operating temperature of graphene compared to g-CN.

In the following, we intend to provide a detailed overview of both CNM in terms of fabrication and characterization with special emphasis on their gas-sensing performances. Then theoretical calculations using density functional theory (DFT) will be carried out to determine the adsorption mechanisms of selected gas molecules, namely, CO2, NO2, and HF, and to examine the g-CN sensing performances to be compared to other graphitic nanomaterials such as Gr or reduced graphene oxide (r-GrO).

2. Synthesis of Graphene

Several methods are most commonly used for the fabrication of Gr layers including (i) micromechanical exfoliation, (ii) reduction of graphene oxide, (iii) chemical vapor deposition (CVD) [5, 26] or plasma-enhanced chemical vapor deposition (PECVD), and (iv) epitaxial growth on insulating materials. The choice of a fabrication route is mainly dependent upon the material quality and overall production cost required for given application.

The structural and vibrational properties of Gr are usually examined by Fourier-transform infrared spectroscopy (FTIR), X-ray diffraction (XRD), and Raman spectroscopy, respectively, and given in Figure 1 [27].

Transmission spectra (Figure 1(a)) show an obvious similarity for all three materials, mainly due to the C-C bond peaks present in the 2325–1981 cm−1 range. The XRD diagram (Figure 1(b)) shows a narrow diffraction peak at 26.5° (2θ) corresponding to the characteristic (002) plane of graphite. This peak is broadened and downshifted in graphene towards 23.9° representing an increased spacing of C-C layer in sp2 bonding. The Raman spectra (Figure 1(c)) show a characteristic vibration peak at 1601 cm−1 (G-band) related to the original graphitic structure for all three materials. However, a wider peak at 1352 cm−1 (D-band) appears for graphite oxide and Gr sheets related to the presence of defects in these structures. The ratio of these bands’ intensities (ID/IG) is a good indicator of the graphitization’s quality in these layers.

2.1. Mechanical Exfoliation

Mechanical exfoliation is the first process through which Gr was isolated by Novoselov et al. [28]. It is a top-down approach in which Gr layers are exfoliated from bulk graphite using adhesive tape. The procedure involves repeated pealing of Gr from flakes of graphite. The tape containing Gr is finally placed on an Si/SiO2 surface and gently pressed to get Gr attached to the Si substrate. This method allows investigating the basic properties of Gr. For in-depth investigations of Gr properties for technical applications, scalable methods are being developed [29].

2.2. Chemical Vapor Deposition

Chemical vapor deposition (CVD) is the most common method used to produce Gr on a substrate with commendable quality and tailored dimensions. Growth mechanism and number of layers grown on the substrate depend on the catalyst metal substrate used such as Cu [26], Ni [30, 31], Cu-Ni alloy [1], and Co. Figure 2 shows a schematic of CVD setup used to grow carbon nanomaterials.

For Gr grown on Ni substrate, the Ni foil is annealed in hydrogen atmosphere and then carbon source gas such as methane or acetylene is admitted into the chamber. Argon gas is used during the process to maintain an inert atmosphere. A controlled cooling rate is necessary to produce desired thicknesses of the Gr layers. Faster cooling rates yield thick graphite and slower cooling rates avoid segregation of carbon deposited on Ni foil [33]. If Ni foil grain growth is small (few minutes of preannealing time) and chamber pressure is maintained high during growth step, nanosized graphite films are formed [30, 34] rather than few-layer Gr [5, 35]. Processing Gr on Cu substrate is most usually used to produce large-scale monolayer Gr films. The Cu foil is heated up to 1000°C in the presence of hydrogen atmosphere at low pressure and then carbon-containing gases such as methane or acetylene are admitted into the CVD chamber for 20–30 minutes and then gradually cooled. Table 1 shows comparison of Gr obtained using several fabrication methods [35].

Table 1 shows comparison of various properties of Gr synthesized through various methods. Gr synthesized through CVD shows a good mobility compared to Gr fabricated by other methods. Processing costs are also reduced with high volume of production by CVD technique.

Moreover, plasma-enhanced chemical vapor deposition (PECVD) was used as a potentially industrially scalable route for Gr synthesis. Kaindl et al. [42] have synthesized Gr on Si (100), Si/SiO2, Ni, Cu, MgO (100), and NaCl and on standard glass substrates using a mixture of C2H2 and C2H2-Ar irradiated with Ar+. At a deposition temperature of 400°C and postdeposition annealing up to 800°C under Ar flow, crystal growth and graphitization are obtained on Cu and Ni foils and interestingly on NaCl. Recently, the growth of Gr grains on Cu substrate using the radio frequency PECVD (RF-PECVD) at different temperatures has been reported [43]. These authors demonstrated the fabrication of large polycrystalline Gr grains at high temperatures and a relatively long dwell time. Large area Gr sheets of several mm2 were synthesized on Cu substrates by this technique using a H2/CH4 gas mixture [44]. Using PECVD, Yang et al. [45] demonstrated the direct growth of Gr walls on dielectric substrates and Hussain et al. [46] fabricated vertically aligned Gr at lower temperature and low power, while Chugh et al. [47] compared the growth of Gr on arbitrary noncatalytic substrates at low temperature. To summarize the various conditions of the Gr growth using PECVD technique, we have drawn up Table 2.

2.3. Chemical and Green Synthesis

Oxidative exfoliation of graphite was also used to synthesize GrO, which in turn was reduced to obtain Gr sheets. Brodie, Staudenmaier, and Hummers are the major methods to obtain Gr from the oxidative exfoliation-reduction (Figure 3).

Oxidation of graphite in the presence of graphite intercalation compounds increases the distance between graphite layers. This weakens the interlayer Van der Waals forces, which allows their exfoliation by sonication in the presence of suitable solvents. Single-layer GrO, dual-layer GrO, and few-layer GrO are thus obtained. This is followed by the reduction of GrO into Gr using hydrazine (N2H4). However, due to its toxicity and high cost, alternate methods were developed to replace the use of hydrazine. Among those methods, we cite thermal reduction, hydrothermal reduction, and electrochemical reduction. Besides, Gürünlü et al. reported on a green synthesis method of Gr at the temperature range of 500–800°C, in which Gr powder was mixed with LiCl/KCl and subsequently heated under Ar or N2 atmosphere in a steel reactor. The thickness of the synthesized Gr films was found to be in the range of 5 to 9 nm, which was attributed to the presence of few layers of Gr [48].

3. Synthesis of g-CN

Various elaboration techniques have been reported in the literature for the fabrication of g-CN. Methods such as CVD, physical vapor deposition (PVD), Radio Frequency Magnetron- (RFM-) based sputtering method, and sol-gel-based spin-coating deposition have been used most often. The various characteristics are listed and compared in Table 3.

The spin-coating method is a simple and efficient technique. The stability of the dispersion depends on many parameters, such as the type of solvent, the viscosity of the solvent, and the size of the particles. To increase the performance and quality of the thin film, it is necessary to avoid the agglomeration of the powder and to ensure that it is deposited evenly on the substrate to prevent the formation of holes [54]. PVD is one of the effective techniques for fabricating thin films and nanostructures with high optical and electronic properties. In the PVD process, the crucial step is to produce the vapor phase of the solid precursor. PVD has been successfully applied in the elaboration of many materials, including carbon nitride for nanotubes [55]. CVD method can synthesize ultrathin films and can be used in the fabrication of heterojunctions [56]. Reactive RFM-based sputtering technique is a technique capable of making large area films. However, this is a multiparameter process. The reactive sputtering process involves the use of a high purity target and plasma medium [52].

3.1. Thermal Condensation Synthesis

g-CN bulk crystal is easily prepared at large scale by thermal condensation of precursors rich in carbon and nitrogen such as thiourea (CH4N2S) [57], melamine (C3H6N6) [58], dicyandiamide (C2H4N4) [23], and urea (CH₄N₂O) [59] as schematically depicted in Figure 4.

Single-layered or multilayered g-CN nanosheets (NSs) are subsequently obtained through exfoliation of the bulk g-CN resulting in large active surfaces [58]. The bulk g-CN presents dense and thick layers due to the agglomeration of flat sheets of g-CN to build a massive, layered structure. Table 4 summarizes the various conditions for manufacturing g-CN under atmosphere.

Ismael et al. synthesized bulk g-CN from melamine, thiourea, and urea [61] and reported that bulk g-CN elaborated by melamine and thiourea contains smooth, thin, and large sheets. Meanwhile the sample elaborated by urea consists of smaller sheets and porous sheets. The exfoliation of bulk g-CN results in the formation of large, irregular-shaped, porous thin sheets, with a variable degree of flexibility [56, 6264]. Iqbal et al. [65] synthesized mesoporous g-CN NSs by a simple one-step approach using glucose and NH4I together with the chosen precursor for exfoliation of bulk g-CN.

3.2. Chemical Vapor Deposition

For the synthesis of g-CN films by the CVD method, generally a precursor rich in C and N is placed at the inlet of an Ar gas stream in a quartz tube that is maintained at low pressure. The powder vapors penetrate inside the hot zone of the furnace (600°C), due to the flow of Ar gas with the low vacuum inside the quartz tube. Condensation of the precursor at high temperature produces the g-CN structure. After deposition, the furnace should be allowed to cool to room temperature automatically in Ar environment [66]. Table 5 summarizes the conditions for depositing g-CN using the CVD technique.

Similarly, plasma-enhanced chemical vapor deposition (PECVD) allows the deposition of thin layers of material on different types of substrates (glass, Si, etc.), after decomposition of the molecules of the reactive gases in a plasma. Table 6 shows the conditions for depositing g-CN using the PECVD technique.

4. Characterization of g-CN

The characterization of g-CN by molecular vibrational spectroscopy (FTIR) basically reveals specificities of g-CN. The transmission spectrum of bulk g-CN in the wavelength range of 500–4000 cm−1 is shown in Figure 5.

The peaks at 802 cm−1 and 889 cm−1 correspond to the presence of s-triazine units in g-CN, caused by the bending vibration of the triazine ring. The peaks in the range of 1200 to 1700 cm−1 are attributed to the stretching vibration of the aromatic heterocyclic units of heptazine (C6N7). The smaller peaks observed at 1250, 1320, 1419, and 1457 cm−1 are related to the stretching vibrations of the C-N bonds, while the peak appearing at 1612 cm−1 relates to the stretching vibration of the C=N bond in heptazine units. The large absorption band centered at 3124 cm−1 corresponds to the stretching vibration of the N–H bond to NH and NH2 groups present at the end of g-CN, indicating that g-CN was undergoing incomplete polycondensation [65, 71]. Guigoz et al. [72] reported that the FTIR spectrum of g-CN NSs exhibits the same distinctive main bands, signature of the g-CN bulk.

X-ray photoelectron spectroscopy (XPS) indicates the presence of two main elements (carbon, C, and nitrogen, N) with two predominant peaks of C1s and N1s (Figure 6).

At high-resolution XPS spectrum of C1s, the peak is deconvoluted into three different components with binding energies at 284.6, 286.0, and 288.1 eV corresponding to functional groups of graphitic carbon [74]. The first component is assigned to C-C or C=C bonds, and the second is identified as a carbon bonded to nitrogen atoms in aromatic rings [C- (N)3], while that at 288.1 eV is assigned to the (N-C=N) bonds. The high resolution of N1s spectrum exhibits three components around 398.3 eV, 399.9 eV, and 400.6 eV. The peak at 398.3 eV is attributed to N binding to carbon atoms involved in triazine units (C-N=C), while the peak at 399.9 eV is assigned to tertiary nitrogen groups N–©3. These last two peaks constitute the heterocyclic ring units of tri-s-triazine, constructing g-CN. The third peak at 400.6 eV indicates the presence of amino functions (C-N-H), resulting from the incomplete condensation of the structures of the melon [57].

The elemental composition, their atomic percentages, and the carbon/nitrogen atomic ratio (C/N) in g-CN are shown in Table 7 [75]. A smaller percentage of oxygen formed at the samples’ surfaces is found with the ratio of C/N often at 0.73 [75].

The molecular vibration properties of g-CN examined by Raman spectroscopy under UV illumination reveal several characteristic peaks corresponding to the vibrational modes of CN heterocycles (Figure 7(a)). Similar trends are obtained using the Cambridge Sequential Total Energy Package (CASTEP) code implemented using density functional theory (DFT) (Figure 7(b)) [76].

This confirms the results of experimental Raman spectroscopy. Interestingly, the comparison of the bulk and single layer of g-CN theoretical Raman spectra shows that the characteristic bands remain visible in both cases, indicating that the exfoliation treatment does not cause any damage to the g-tri-s-triazine-CN. Only a decrease in intensity for the nanosheets is observed, which is attributed to the low-dimensional character. Nonetheless, a small shift is observed for the N-C (sp2) bending vibration peak at 1224 cm−1 for the bulk sample, which is displaced to 1232 cm−1 in g-CN NSs. This originates from phonon confinement and strong quantum confinement effects [76]. The second characteristic is related to the peak height ratios of 583–478 cm−1 (I583/I478) and 761–704 cm−1 (I761/I704), equivalent to the layer-layer strain vibrations or the correlating vibrations induced by the layer-layer vibrations. It is found that these ratios tend to increase with the decreasing number of g-CN layers.

The XRD diagrams depicted in Figure 8 show two diffraction peaks at 13.0° and 27.2°. The large peak (002) at 2θ = 27.2° is the peak characteristic of the interlayer stacking of an aromatic cyclic system of graphitic type. The much smaller peak (100) at 2θ = 13.0° is characteristic of tri-s-triazine repeating units in melon-like materials [71, 77]. Its unchanged position in the g-CN NSs revealed that the chemical exfoliation has no impact on the aromatic cyclic structure of g-CN.

The intensity reduction of (002) peak observed in nanosheets is a signature of successful exfoliation of g-CN. It is worth noting that the diffraction peak at 13.0° is also decreased compared to the g-CN bulk, which is likely due to the reduced planar size after exfoliation step. The calculated interplanar distance of the aromatic series is d = 0.33 nm, indicating a layered-like structure similar to crystalline graphite (d = 0.335 nm) [78]. This result suggests a typical g-CN structure [79]. Both NSs and bulk g-CN show strong emission peaks, respectively, at 445 nm and 464 nm when exposed to visible light excitation using a photoluminescence experimental setup. These peaks correspond to the band gap energies of 2.78 eV and 2.67 eV for the NSs and bulk g-CN, respectively [54, 62, 74, 80].

The optical band deviation () was also estimated from absorbance measurements using the Tauc model [71, 81]:where α is the absorption coefficient, hν is the photon energy, the optical band gap, B is a constant, and m is an exponent that characterizes the optical absorption process. For the direct authorized transition m = 1/2, for the indirect authorized transition m = 2, for the direct forbidden transition m = 3/2, and for the indirect forbidden transition m = 3.

The plotting of [(αhν)1/m] as a function of the energy of the photon (hν) and the extrapolation of the linear part of the curve to [(αhν)1/m = 0] on the energy axis provides the value of the optical band gap (). The value of m = 1/2 is suitable for g-CN because it has a direct allowed band gap [74]. By using the Tauc model, Huang et al. [56] found the band gaps for bulk g-CN and g-CN NSs to be approximately 2.72 eV and 2.82 eV, respectively. They attributed this increase in the band gap energy from bulk to nanolayers by the effect of quantum confinement.

5. Gas-Sensing Investigations

5.1. Gas-Sensing Mechanisms

The physical principle of a gas sensor is based on charge transfer process, in which the detection material acts as a charge donor or a charge acceptor [5, 16, 34, 82]. Upon exposure to different gases, the charge transfer reaction occurs in different directions between the sensing material and the adsorbed gases, which leads to different changes in the resistance of the material [83]. In absence of adsorbed gas molecules, or upon reexposition to air, the original resistance of the material is recovered. Figure 9 shows a schematic that describes the change of resistance related to charge transfer processes from adsorbed gases. Considering an n-type material exposed to an electron donor gas (reducing gas), electrons are transferred from the gas molecules to the n-type material causing an increase in the density of majority of charge carriers (electrons). Consequently, the resistance of the n-type material decreases. Seemingly in the case of an n-type material exposed to an acceptor gas (oxidizing gas), the electrons are transferred from the material to the gas molecules resulting in a decrease of the majority of charge carriers’ density in the material. As a result, the resistance of the n-type material increases.

Similar processes occur in the case of a p-type material. Its exposure to a reducing gas causes electron to transfer from the gas molecules to the material, hence decreasing the density of the majority of charge carriers (holes) in the p-type material and resulting in a decrease of the p-type material’s resistance. In contrast, if the adsorbed gas is oxidizing, electrons are transferred from the material to the gas molecules, resulting in an increase of the charge carriers’ density in the material. As a result, the resistance of the p-type material decreases.

5.2. Graphene-Sensing Performance

Studies have proved that defects and dopants in Gr increase its gas sensitivity considerably [8487]. Modification of Gr with boron, nitrogen, and defect increases its adsorption energy with gases such as ammonia, as well as oxides of nitrogen and carbon [84]. Doping Gr with metals such as Mg, Co, Fe, and Ni increases sensitivity of Gr for gases such as H2S, SO2, and O2. Table 8 shows sensing characteristics of various dopants with Gr for various gases.

Microsensors made of mechanically exfoliated Gr are capable of sensing even single molecules of NO2 in high vacuum [21]. The adsorption of electron accepting molecules such as NO2 and H2S onto Gr leads to a decrease of its resistivity due to an increase in holes concentration. Meanwhile, electron donor molecules such as CO and NH3 increase resistivity of Gr when adsorbed to it, which is due to increase in electron concentration as shown in Figure 10.

A 3.5–5 nm thick mechanically exfoliated Gr used as a gas sensor has shown commendable sensitivity, reproducibility, and selectivity in response to NO2 at 100 ppm as shown in Figure 11. Sensors based on CVD made Gr show significant sensitivity to various gases at a minimum concentration of 1 ppq [90]. Schottky diode sensors made of Gr/n-si have proven to be potential H2S gas-sensing device [91].

Graphene oxides (GrO) are promising gas-sensing materials owing to their oxygen-rich functional groups. Single-layer to few-layer GrO (dimensions ranging from 27 to 500 µm) based sensors fabricated by drop casting show a p-type response in both oxidizing and reducing environments [92]. GrO based sensors fabricated by dielectrophoresis are used to sense hydrogen gas [93]. GrO with acoustic wave propagated surface is used to detect H2 and NO2 in the atmosphere [94, 95]. Dense functional groups of oxygen in GrO maintain its sensitivity even in harsh conditions such as high humidity and strong acidic or basic conditions and also sense harsh volatile content like chloroform [96].

Reduced graphene oxide (r-GrO) is widely used in sensing gases. r-GrO is preferred over Gr due to its low cost and tailorable structure and properties [22]. The enhancements in the r-GrO response to gases are associated with the carbon atoms, holes, or vacancies created due to thermal treatments. Sensors fabricated by r-GrO (made by annealing of GrO at 300°C) show conductance that is 4.3% more than that of mechanically exfoliated Gr [21]. Cost-effective gas sensors based on r-GrO were fabricated on paper substrates to sense NO2 at lower ppm levels. It is also noted that sensing behavior depends on both the thickness and size of the r-GrO flakes [97]. The poor selectivity of Gr constitutes one main drawback for its usage in gas-sensing devices. Nevertheless, its functionalization with metallic or metal oxide nanoparticles has been shown to enhance its selectivity [98102]. In fact, pristine Gr shows insignificant change in resistance when hydrogen gas molecules are adsorbed onto it. Meanwhile, nano-Pd decorated Gr shows considerable changes in resistance because Pd hydride on the surface of Gr donates more electrons [99]. r-GrO decorated with palladium can sense NO from 2 to 420 ppb. r-GrO functionalized with –SO3H and decorated with nano-Ag showed a superior performance in measuring NO2 and NH3 gases [102, 103]. Table 9 shows details of gases sensed by different Gr decorated by various nanoparticles.

Regardless of its comparison with g-CN, Gr remains an interesting 2D material suitable for gas sensing due to its large specific surface area, high carriers’ mobility, high sensitivity, fast response, low power consumption, and its operation under ambient conditions [120]. Moreover, when Gr is associated with perovskites [121], polymers [122], or organic molecules [123], its sensing performance was found to be improved at room temperature. Indeed, the composite form of Gr, whether with a polymer or a perovskite or organic molecules, provides specific reaction sites for gas adsorption, hence improving the sensor selectivity at room temperature. When Gr was used in conjunction with g-CN [124], g-CN showed superior performance compared to Gr in both gas sensing and catalysis, which was attributed to the increased interaction sites offered by 2D g-CN structure.

5.3. g-CN Sensing Performance

Several works have been reported on the influence of the g-CN content on its gas-sensing performances. Karthik et al. [103] reported that thin layers (elaborated by a spray pyrolysis) of TiO2 porous nanospheres decorated with g-CN showed excellent detection performance of CO2 gas with very fast response and recovery. The results show that 10 wt.% g-CN decorated TiO2 composite layer exhibits an exceptional detection response of 88% and a great stability for CO2 gas at 450°C under 1500 ppm. The enhanced detection response is attributed to the increased surface area of 108.5 m2/g due to the porous nature of g-CN (15.7 nm). Zhang et al. [58] showed that gas detection properties such as gas selectivity and sensitivity of SnO2 to acetic acid as well as the operating temperature were enhanced by the incorporation of g-CN into SnO2. Their results indicate that the operating temperature for the detection of acetic acid was decreased from 230°C to 185°C. In particular, the g-CN-SnO2 nanocomposite-based gas sensor containing 10% by weight of g-CN revealed the highest sensitivity of 87.7 to 1000 ppm acetic acid, the fastest response, lower recovery time, and the lowest detection limit down to 0.1 ppm. Besides, calcination temperatures, calcination times, and precursor ratios were also reported to significantly affect the performance of sensors to target gases. For example, Hu et al. [14] have examined the effect of elaboration conditions on the detection performance of g-CN-SnO2 nanocomposites using 500 ppm acetone as the target gas. The 200/100 ratio between melamine and SnCl2 · 2H2O and the calcination for 3 hours at 500°C appeared to offer better detection properties. For instance, these preparation conditions enhanced the detection sensitivity 22 times compared to pure SnO2 at a fast response and recovery time (7 s/8 s) while achieving very low detection limit of 67 ppb. In addition, the element ratios (melamine and SnCl2 · 2H2O) were also used to determine the important role of g-CN in the g-CN-SnO2 nanocomposite sensor. In parallel, Absalan et al. [25] have synthesized CO gas sensors based on Pd/SnO2/g-CN nanocomposites by the hydrothermal method at different Pd and g-CN contents. The nanocomposites made of 5% Pd/SnO2/5% g-CN presented remarkable CO detection capabilities such as good selectivity and stability, superior response, short response, and low recovery times at very low operating temperature of 125°C. The exceptional gas detection performances of these nanocomposites were attributed to the large specific surface area of g-CN. Similarly, Chu et al. [125] examined the g-CN-WO3 composite synthetized by hydrothermal treatment. Their results revealed that adding a suitable quantity of g-CN to WO3 was found to enhance the selectivity and sensitivity to acetone vapor. The sensor based on 2 wt% g-CN-WO3 composite, at an ideal operating temperature of 310°C, presented a response to 1000 ppm acetone of Ra/Rg = 58.2. Moreover, Zhang et al. [126] studied g-CN/MgFe2O4 composites synthesized by a one-step solvothermal process as acetone gas sensor. The gas detection properties of the models were determined and compared to a sensor based on pure MgFe2O4. The higher sensing performance was achived at 10 wt% g-CN content, where the sensitivity to acetone was enhanced by approximately 145 times, while the best possible operating temperature was 320°C. The MgFe2O4/g-CN-10wt% based sensor exhibits outstanding acetone detection performance, high sensitivity and selectivity, rapid response, and low recovery time and thus favorable stability.

Additionally, the gas sensors based on chemically exfoliated g-CN NSs were found to exhibit very low detection limits and ultrahigh sensitivity towards NO2 gas compared to Gr as reported by Hang et al. [127]. Precisely, 15 wt% of g-CN NSs associated with Gr showed a better response than the sole Gr based sensor towards NO2 gas. Similar gas-sensing enhancement was reported to occur for other compounds associated with g-CN such as ZnO/g-CN composite [128]. For instance, when ZnO/g-CN composite was irradiated with 460 nm light, it displayed the maximum response of 44.8 under 7 ppm of NO2, with a response and recovery times of 142 s and 190 s, respectively, and a detection limit of 38 ppb.

Table 10 summarizes the g-CN based nanocomposites for gas sensor applications including their characteristics, elaboration methods, and gas-sensing performances.

All aforementioned gas-sensing performances did not include specific sensing conditions that greatly influence the detection capabilities such as the relative humidity or mixed gases. In fact, humidity is a key factor that needs to be taken into account when analyzing the gas-sensing performances. For instance, the responsivity of g-CN gas sensors towards NO2 was found to be altered in the presence of high levels of water vapor (about 90%) as shown in Figure 12(a) [129]. The sensing performance was decreased by 3 times for 90% RH compared to 50% RH.

Hence, the presence of humidity might compromise the detection of target gases. This result has to be considered with special care as this limitation is commonly overcome by increasing the operating temperature; this implies that the gas sensors need to be stable at intermediate temperatures while keeping their gas sensitivity unchanged. As an example of the sensing selectivity, Figure 12(b) is provided to illustrate the gas selectivity of g-CN sensors towards several common gases, where the highest selectivity was obtained on NO2.

6. Simulated Gas-Sensing Performances

Theoretical calculations show the adsorption energies between selected gas molecules (CO2, NO2, and HF) and g-CN monolayer. The comparison with Gr and r-GrO is shown using HF as a selected adsorption gas. The first-principles calculations are based on the density functional theory (DFT) using generalized gradient approximation (GGA) with the Perdew-Burke-Ernzerhof (PBE) exchange-correlation functional [81, 130], implemented in Quantum Espresso code [131]. In the optimized process, the vacuum spaces of 15 Å were set for the three compounds to separate the interactions between the layers. A (3 × 3 × 1) k-point was used in the structure optimization and total energy calculations. To study the stability of g-CN, Gr, and r-GrO surfaces with considered gas molecules, the adsorption energy of system was computed using the following relation [132, 133]:where E(compound−gas), Ecompound, and Egas are the total energy of gas molecules adsorbed by the surface of the compound, compound, and gas molecule, respectively. A positive adsorption energy value shows that the adsorption is energetically unfavorable, and the reaction is endothermic and hence unstable. A negative value indicates that the reaction is exothermic and hence stable and energetically favorable. The position of the nitrogen atom (N) in g-CN according to [134] shows that the adsorption energies of CO2 gas on a nitrogen atom at two coordinates are more negative than those of other positions. The monolayer model was obtained by removing other layers in the unit cell of bulk.

The adsorbed CO2, NO2, and HF molecules were constructed by placing a single gas molecule parallel on the surface of compounds. The values of d for all compound-gas systems are 2.5 Å. The gas molecules CO2, NO2, and HF have bond lengths between atoms of 1.16 Å, 1.4 Å, and 0.86 Å, respectively. Top views of systems adsorbed by gas molecules are shown in Table 11. The five adsorbed models were geometrically optimized to obtain the adsorption energies and they were compared successively.

According to the optimization of the g-CN-gas system (see Table 12), the CO2 and NO2 molecules are moved away from the surface, while the HF molecule is brought closer to the surface. The distances between the surface of g-CN and the atoms of the molecules are 2.5 Å before optimization. After adsorption on the g-CN surface, the molecules also exhibit some deformation. The angles and bond lengths change, while the angles O-C-O and O-N-O in CO2 and NO2, respectively, are decreased. The N-O and H-F bond lengths increase and the C-O bond length decreases. The distance between the surface of g-CN and the atom closest to the surface in gas molecules is noted with d-g-CN-gas.

The DFT calculation results show that all the adsorption energies are negative, indicating that the adsorption of the three gases on g-CN is exothermic. It can be briefly summarized that g-CN exhibits a relatively high adsorption capacity for HF (−1.03 eV) and CO2 (−0.59 eV) and a low adsorption intensity for NO2 (−0.166 eV). Therefore, the adsorption behavior of g-CN corresponds to efficient adsorption of reagents, which is very beneficial for gas sensor performance The adsorption results of CO2, NO2, and HF gas molecules on g-CN showed that the Eads of these gases on virgin g-CN follows the order of NO2> CO2> HF. The results suggest that the g-CN nanolayer is more stable with HF than other gases with an adsorption energy of −1.03 eV. The results revealed that the g-CN system is a suitable candidate for HF gas sensors, due to their high adsorption capacity. In addition, the results indicate that the interaction between g-CN and HF gas is significantly higher than that of other gases. Table 13 shows other gases detected by g-CN and their adsorption energies.

For Gr and GrO, the DFT calculation results show that the interaction of HF gas with the latter two is relatively low compared to the adsorption of HF on g-CN. The optimized structures of HF adsorbed on Gr and r-GrO are shown in Table 14. The length of the HF bond is less modified compared to its initial state of a little near ∼0.94 Å. The adsorption energies of HF on Gr and r-GrO are, respectively, −0.04 eV and −0.74 eV, lower than those of HF on g-CN. This calculation shows that g-CN can be used for the detection of HF gas.

Gr-based gas sensors show commendable effect in various sensing gases. Pristine Gr shows insignificant effect in gas sensing and hence its sensitivity is enriched by either doping the Gr with suitable metal dopants such as Mg, Co, Fe, and Ni or hybridizing with other nanoparticles. Stability of the GrO sensors at higher temperatures is generally lower than that of g-CN based sensors. Ye et al. showed that TiO2 hybridized with GrO senses NH3 gas and the same TiO2 when hybridized with g-CN shows better sensitivity for CO2 even at 400°C. GrO based sensors are used in acidic or alkaline conditions and can sense high volatile chemicals such as chloroform. SnO2-Gr proves to be an excellent candidate for sensing humidity.

7. Conclusion

In this work, an overview of both Gr and g-CN low-dimensional materials was outlined in terms of elaboration and characterization with respect to their gas-sensing performances. Subsequently, theoretical calculations along with simulated gas-sensing performances of both nanomaterials were carried out to address the researcher’s paradigm regarding their optimization as potential gas sensors. This sensor would exhibit high sensitivity, low detection limit, fast response, and low operating temperature. Based on our investigation, g-CN, which reveals interesting gas-sensing performances over Gr by 10-folds, may offer new opportunities for the design and manufacturing of high-performance gas sensors. This is mainly related to its semiconducting character and its porosity that provides large adsorption sites for the target gases. Besides, the conductivity and the tailorable functional groups of Gr provide a great advantage as resistive sensors. Consequently, for NO2, NH3, and organic gases such as ethanol and acetone, Gr would show high performances. The main bottlenecks are large-scale production of Gr-based sensors and sensitivity of these sensors for specific gases. The specific gas sensitivity can be improved by (1) increasing the surface area of Gr and combination with other specific nanomaterials and (2) focusing on the morphology of various GrO with specific gas adsorption [138, 139].

Data Availability

Data collected from the literature can be consulted in the relevant articles; the authors’ data are available upon request to Professor Mustapha Jouiad at [email protected].

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Acknowledgments

This research was funded by a seed grant provided by MAScIR “2D–HF–Detect #240004055”.