Individual Gas Molecules Detection Using Zinc Oxide–Graphene Hybrid Nanosensor: A DFT Study

Abstract

Surface modification is a reliable method to enhance the sensing properties of pristine graphene by increasing active sites on its surface. Herein, we investigate the interactions of the gas molecules such as NH₃, NO, NO₂, H₂O, and H₂S with a zinc oxide (ZnO)–graphene hybrid nanostructure. Using first-principles density functional theory (DFT), the effects of gas adsorption on the electronic and transport properties of the sensor are examined. The computations show that the sensitivity of the pristine graphene to the above gas molecules is considerably improved after hybridization with zinc oxide. The sensor shows low sensitivity to the NH₃ and H₂O because of the hydrogen-bonding interactions between the gas molecules and the sensor. Owing to observable alterations in the conductance, large charge transfer, and high adsorption energy; the sensor possesses extraordinary potential for NO and NO₂ detection. Interestingly, the H₂S gas is totally dissociated through the adsorption process, and a large number of electrons are transferred from the molecule to the sensor, resulting in a substantial change in the conductance of the sensor. As a result, the ZnO–graphene nanosensor might be an auspicious catalyst for H₂S dissociation. Our findings open new doors for environment and energy research applications at the nanoscale.

1. Introduction

There are presently great efforts being made in developing novel gas sensors centered around new nanomaterials for the detection of toxic gases since they offer higher sensitivity, selectivity, and reliability; as well as immediate response and recovery at low cost. One of the most significant advantages that nanostructure-based sensors have over conventional microsensors is the high surface-to-volume ratio. This is a crucial parameter since it defines the sensitivity of the gas sensors. Moreover, the nanostructured materials provide a greater detection area leading to better adsorption of gas species and an increase in sensing abilities.

Graphene, the first discovered two-dimensional (2D) material, has become a promising material for different types of applications [1,2]. Graphene’s amazing properties have attracted considerable interest in the area of gas sensing since they enable the improvement of miniaturized sensors at low cost and power consumption [3,4,5]. Consequently, graphene has been extensively studied theoretically [6,7,8,9,10,11,12,13,14] and experimentally [15,16,17,18,19,20]. On the downside, because of the weak interactions between π electrons of graphene’s surface and the gas molecules, pristine graphene has shown limitations for the recognition of individual gas particles [9,10], exhibiting low sensitivity to common gases such as CO, CO₂, CH₄, N₂, NO₂, NH₃, H₂, and H₂O [6,7,8]. Therefore, physical or chemical modification of graphene’s surface has been considered to boost its performance. Hydrogen plasma [21], reduction reagents [22], and electron beam irradiation [23] are the physical methods that have been applied to modify the surface of the graphene by reducing the graphene oxide (GO). Besides the physical modification, chemical methods such as functionalization, impurity atom doping, and defects have been reported to successfully alter the electrical and magnetic properties of nanomaterials [24,25,26,27,28,29,30,31,32]. Techniques, where dopants or defects are introduced, have shown an improvement in the sensitivity of the graphene-based gas sensors [7,8,11,12,13,14,28,29,30,33]. Zhang et al. showed small gas molecules such as NO₂, CO, NO, and NH₃ display strong reactions with B-doped, N-doped, and defective graphene [7]. Additionally, when graphene is doped with B, N, and Si, it displayed boosted interactions with other common gases like N₂, SO₂, CO₂, O₂, H₂, and H₂O compared to pristine graphene [8]. Doped graphene with transition metals (Fe, Co, Ni, Ru, Rh, Pd, Os, Ir, and Pt) has also been demonstrated to have high sensitivity to O₂ adsorption [12]. Borisova et al. noticed that when defects are induced on graphene, it improved the interactions of H₂S with C atoms, showing much better response in comparison with pristine graphene [13]. Density functional theory (DFT) calculation on Eu-decorated single- and double-sided graphene sheets exhibited that each Eu could firmly bound to six H₂ molecules [14]. Recently, the sensitivity of other 2D nanomaterials such as MoS₂ [34,35,36], phosphorene [37,38], hexagonal boron carbide (BC₃) [39,40], WS₂ [41,42], silicene [43,44,45], and germanene [46] to several gases have been inspected.

Metal oxides like SnO₂, WO₃, etc. have been commonly integrated on gas sensors to improve their performance, specifically their sensitivity and selectivity [47,48,49,50,51,52]. However, these types of sensors have a significant constraint since they need to operate at high temperatures, between 200–600 °C, limiting their application and integration as well as presenting energy consumption issues and risks of gas explosion. Recently, metal oxide gas sensors (including Cu₂O, In₂O₃, Co₃O₄) have been hybridized with graphene to boost their ability of gas detection at room temperature [53,54,55,56,57,58,59,60,61,62,63,64].

Zinc oxide (ZnO) is a rising and significant semiconductor due to its large direct band gap of 0.37 eV and large exciton binding energy of 60 mV. ZnO nanoparticles are very popular to use in graphene hybrid systems because it is easy to control their size, morphology, and material properties. We have recently reported decoration of graphene with Zn₁₂O₁₂ nanoclusters to open a band gap of 14.5 meV around the Fermi level at K-point [65]. Graphene decorated with ZnO nanoparticles has been proven to heighten the sensitivity towards common toxic gases like CO, NH₃, NO [55], methane [66], NO₂ [67,68], formaldehyde [69,70], H₂ [63], and also to humidity [71].

The objective of this study is employing DFT in order to analyze the electronic and transport properties of the zinc oxide–graphene hybrid nanosensor when in close proximity to the gas molecules such as NH₃, NO, NO₂, H₂O, and H₂S. Variations in the electronic transport properties of the ZnO–graphene sensors are correlated with the sensitivity capability of the device. Our results reveal the promising future of ZnO–graphene hybrid nanostructure in the development of ultra-high sensitive gas sensors.

2. Computational Methods

This paper presents first-principles calculations based on DFT in combination with the non-equilibrium Green’s Function (NEGF) method performed in the Atomistix ToolKit (ATK) software package [72,73,74]. The Perdew, Burke, and Ernzerh (PBE) approach of generalized gradient approximation (GGA) is used for the exchange-correlation function. To expand the electron wave function, a double-ζ polarized basis set is employed. The Grimme van der Waals (vdW) correction (PBE-D2) [75] is utilized in order to take into account the long-range vdW interactions in the adsorption systems [76]. The electronic temperature is kept constant at 300 K, and the plane wave mesh cut-off energy is set to 150 Rydberg.

To mimic a free-standing graphene sheet, we consider a periodic (3 × 10) supercell of graphene in y and z dimensions. To avoid any image interactions, a vacuum spacing of 15 Å is imposed on the plane in the x-direction in which the sheet is not periodic. By employing the conjugate gradient method, the supercell is allowed to optimize fully until all the final atomic forces are less than 0.01 eV/Å. The first Brillouin zones are sampled using 1 × 21 × 11 k-points in the x, y, and z directions, respectively. In the next step, zinc oxide is placed in the center of the supercell in the vicinity of the C atoms, and then the whole system is optimized again. For the calculations, we increase the k-points to 1 × 121 × 111 in the x, y, and z directions, respectively. The adsorption energy gives us a quantitative explanation of the interaction strength between adsorbent and adsorbate. The adsorption energy of zinc oxide onto the graphene sheet is computed from:

$$E_{ad(ZnO)}=E_{ZnO-Graphene}-E_{Graphene}-E_{ZnO}$$

(1)

Here, $E_{ZnO-Graphene}$, $E_{Graphene}$, and $E_{ZnO}$ are the total energies of the ZnO–graphene system, pristine graphene sheet, and the isolated ZnO, respectively. In the last step, the gas molecules including NH₃, NO, NO₂, H₂O, and H₂S are added to the system by placing them in the vicinity of the ZnO. The gas molecule–ZnO–graphene complexes are completely relaxed. The adsorption energies of the gas molecules on the ZnO–graphene system are obtained by:

$$E_{ad(Molecule)}=E_{Molecule-ZnO-Graphene}-E_{ZnO-Graphene}-E_{Molecule}$$

(2)

where $E_{Molecule-ZnO-Graphene}$ is the total energy of the molecule–ZnO–graphene compound and $E_{Molecule}$ is the energy of an isolated gas molecule. Mulliken population analysis is engaged to study the charge transfer between the graphene and ZnO and between the gas molecules and the ZnO–graphene system. To simulate the gas sensing system and analyze its electronic and transport properties, the structure is converted to a two-point probe sensor, as shown in Figure 1. The sensor consists of 120 atoms of C which is functionalized with the zinc oxide. It also includes a scattering region (12.8 × 24.6 Ų) and 4.92 Å electrodes on each side. It should be noted that the left and right electrodes (the black boxes) are semi-infinite until cells and are repeated along the z-axis. Moreover, the scattering region is periodic along the y-axis. We describe the quantum conductance of the system via the transmission coefficient at the Fermi level:

$$C(E)=G_{0}T(E)$$

(3)

where $G_0=2e^2/h$ is the quantum conductance constant (e is the electron charge and h is Planck’s constant). The transmission is calculated along the z-direction. Moreover, transmission coefficients at energy E can be expressed based on the retarded Green’s function $G^{†}(E)$ and the advanced Green’s functions $G(E)$ of the scattering region.

$$T(E)=G(E){\mathsf{\Gamma }}^L(E)G^{†}(E){\mathsf{\Gamma }}^R(E)$$

(4)

Here, ${\mathsf{\Gamma }}^{L(R)}(E)$ is the coupling function of the left (right) electrode which is:

$${\mathsf{\Gamma }}^{L(R)}=\frac{1}{i}({\displaystyle \sum {}^{L(R)}}-{({\displaystyle \sum {}^{L(R)}})}^{{}^{†}})$$

(5)

where ${\displaystyle \sum {}^{L(R)}}$ is the electrode self-energy of the left (right) electrode.

3. Results and Discussion

We first optimize the ZnO–graphene hybrid nanostructure in order to obtain the most energetically stable geometry of the sensor. ZnO can adhere to the graphene at different positions including the atom site (above C atom), bridge site (above C–C bond), and hollow site (center of C–C hexagons). Additionally, concerning the surface of the graphene sheet, a ZnO can be aligned parallel or upright at these specific positions. The most stable configuration is found by full relaxation of all the possible structures and comparing their final adsorption energies. The most energetically favorable adsorption configuration, which is the arrangement with the minimum adsorption energy, is shown in Figure 2a. Our DFT calculations show that ZnO prefers to place itself perpendicular with respect to the graphene plane, while the Zn atom heading to the C atoms and the O atom way around. The Zn atom is located above the middle of the C₁–C₂ bond and is bonded to C atoms on the graphene sheet to form a Zn–C₁–C₂ triangle with Zn–C₁, Zn–C₂, and C₁–C₂ side lengths of 2.24, 2.20, and 1.44 Å, respectively. The C₁–Zn–C₂ angle is 37.95°. The Zn–O bond length in isolated ZnO (2.00 Å) is reduced to 1.70 Å after interaction with C atoms of the graphene sheet. For ZnO–graphene nanostructure, an adsorption energy of −0.68 eV is obtained that could be indicative of moderate chemisorption. This chemisorption accompanies with an electron transfer of 0.105 e from ZnO to graphene. These findings can be confirmed by comparing the band structures of the pristine graphene sheet and the ZnO–graphene hybrid nanostructure, as depicted in Figure 2b. The dashed green curves show the reference results for a pristine graphene sheet. As we know, graphene has no band gap. This means the valence and conduction bands touch at the Dirac points. Upon exposure to zinc oxide, the induced electrons shifted down the Dirac point of the graphene towards negative energies and moved the Fermi level to the conduction band, as shown in Figure 2b in solid black curves. This suggests that the n-type doping effect of graphene has taken place. The Dirac point of the graphene is distorted after interaction with the ZnO due to breakage of bond symmetry in graphene. Interestingly, a gap of 26 meV is observable between the bands c (shifted conduction band) and v (shifted valence band) of the ZnO–graphene hybrid.

Next, the adsorption mechanisms of the NH₃, NO, NO₂, H₂S, and H₂O gas molecules onto a graphene–ZnO nanostructure sheet are addressed. In the first step, each gas molecule is placed at a distance of 2 Å above the O atom of the ZnO–graphene nanostructure. Some initial configurations are considered. The gas molecules can be aligned in parallel or perpendicular concerning the surface of the ZnO–graphene nanostructure. Further, the N atom of NO with a diatomic molecular structure can point up or down to the ZnO–graphene nanostructure. For the triatomic molecules (NO₂, H₂O, and H₂S) and tetratomic molecule (NH₃), two orientations are examined. In the first orientation, the N atom (in NH₃ and NO₂), O atom (in H₂O), and S atom (in H₂S), point to the O atom of the ZnO–graphene nanostructure, while, in the second one, the O atom (in NO₂) and H atom (in NH₃, H₂O, and H₂S) point down to the surface of the sensor. In the second step, all the input structures are allowed to fully optimize. Once improved, the structural stability is studied regarding the adsorption energy. The most stable configurations which have the most negative adsorption energies are chosen for further analysis and are depicted in Figure 3. The conventional sensing mechanism is that charge transfer between the adsorbed gas molecules and the surface of the 2D nanomaterials is expected to affect the conductivity of the system upon exposure to the target gas molecules, while the chemical structure of the sensing layer is preserved. The adsorption energies of the gas molecules on the ZnO–graphene nanostructure, the minimum binding distances, and the net charge transfer on molecules using Mulliken population analysis are summarized in

Table 1. The calculated adsorption energy (E_ad), binding distance (D) which is the shortest atom to atom distance between the molecule and the sensor, the charge transfer (Q), and the evaluated value of the sensitivity of the sensor with respect to each gas molecule (S). The positive value of charge indicates a charge transfer from the molecule to the sensor.

System	Ead (eV)	D (Å)	Q (e)	S (%)
NH3-ZnO-G	−0.55	1.72	+0.056	11.8
NO-ZnO-G	−2.21	1.28	+0.155	188.2
NO2-ZnO-G	−1.85	1.33	+0.141	141.1
H2O-ZnO-G	−0.32	1.60	+0.031	6.4
H2S-ZnO-G	−1.01	1.04	+0.121	76.47

.

We first scrutinize the interactions of the NH₃ molecule and the ZnO–graphene compound. As can be seen in Figure 3, H atom of the NH₃ molecule interacts with the O atom of the ZnO–graphene nanostructure. The interaction distance H-O is found to be 1.72 Å which is larger than the sum of the covalent radii of the H and O atoms (0.95 Å) [77]. Furthermore, the adsorption energy is calculated to be −0.55 eV that is large enough to form a hydrogen bond between the NH₃ and the sensor (H–O). A hydrogen bond is a weak interaction whose strength can be placed between the chemical bond and the vdW force [78]. One can say that the NH₃ molecule is only physically adsorbed on the surface while maintaining its molecular form. The C₁–C₂ bond length remains unchanged while both Zn–C₁ and Zn–C₂ bond lengths are slightly altered to 2.21 Å. The C₁–Zn–C₂ angle is reduced from 37.95° to 14.86°. The bond lengths of N–H in an isolated NH₃ (1.03 Å) are found to be the same after adsorption onto the ZnO–graphene. These small variations substantiate that the influence of NH₃ adsorption onto the sensor is trivial. This observation can be further verified by the low amount of charge (0.056 e) that is transferred from the NH₃ to the graphene-based sensor. Excitingly, the calculated adsorption energy for the NH₃–ZnO–graphene complex is greater than that obtained for NH₃ adsorption onto pristine graphene (−0.11 eV) [7] and is less than the ammonia adsorption energy on the ZnO ($10\overline{1}0$) (−1.32 eV) [79], ZnO ($0001$) (−2.21 eV) [80], and ZnO (6,0) single-walled nanotube (−0.82 eV) [81] surfaces, revealing that hybridization of pristine graphene with ZnO increases its adsorption capability toward NH₃ significantly. This is in accordance with the experimental data. It was reported that sensors based on ZnO-decorated graphene oxide sheets can detect NH₃ for concentrations as low as 1 ppm at room temperature [55].

Next, the interactions of the NO and NO₂ gas molecules with the ZnO–graphene hybrid are inspected. The relaxed structures are provided in Figure 3. The adsorption energies of NO and NO₂ are −1.21 and −1.85 eV, respectively. One can conclude that there exists a strong chemisorption characteristic between NO and NO₂ gas molecules and the sensor. Upon adsorption of NO on the ZnO–graphene compound, the C₁–C₂ bond length is elongated to 1.45 Å, while Zn–C₁ and Zn–C₂ bond lengths are shortened to 2.13 and 2.18 Å, respectively. Additionally, the C₁–Zn–C₂ angle is slightly increased to 39.58°. The N and O atoms of the NO interact with O and Zn atoms of ZnO, respectively, to form a kite where its four sides can be grouped into two pairs of equal-length sides (N–O and Zn–O) that are adjacent to each other. The length of N–O is 1.28 Å and Zn–O equals to 2.06 Å. The N–O bond length is stretched from 1.17 Å in an isolated NO molecule to 1.28 Å in the NO–ZnO–graphene system. The O–N–O and O–Zn–O angles of the kite are 113.10° and 92.46°, respectively. The interaction of NO₂ and ZnO–graphene bears many similarities to that of the NO and ZnO–graphene. In much the same manner, the interaction between the NO part of the NO₂ molecule and ZnO results in the Zn–O–N–O kite. In this geometry, the lengths of the N–O sides equal to 1.33 Å and the Zn–O sides have the length of 2.06 Å. The O–N–O and O–Zn–O angles of the kite are 113.92° and 91.16°, respectively. In comparison with the length of N–O bonds in the isolated NO₂ (1.21 Å), the N–O bond of the NO₂ in the kite gets extended to 1.33 Å, while the second N–O bond of the molecule remains almost unchanged. It should be added that the O–N–O bond angle of isolated NO₂ decreases from 133.06° to 122.38° after complexation with ZnO–graphene compound. Despite the resemblances above between NO and NO₂ adsorption on the ZnO–graphene sensor, the Zn atom is shifted from above the middle of C₁–C₂ bond to top of the C₂ atom, where both Zn–C₁ and Zn–C₂ bond lengths are 2.50 and 2.03 Å, respectively. Interestingly, after exposure to NO₂, the C₁–C₂ bond length is stretched to 1.46 Å. Additionally, the C₂ atom protrudes outwards after NO₂ adsorption with a buckling distance of ~0.302 Å. These noticeable changes in the structure of the ZnO–graphene sensor are indicative of the chemical adsorption of NO and NO₂ and agree well with the calculated adsorption energies for the gas molecules on the sensor. The obtained adsorption energies for NO (−1.21 eV) and NO₂ (−1.85 eV) on the ZnO–graphene are considerably higher compared to those reported for NO (−0.30 eV) and NO₂ (−0.48 eV) adsorption onto pristine graphene [7] and those obtained for NO (−0.28 eV) and NO₂ (−0.36 eV) adsorption onto the ZnO ($10\overline{1}0$) surface [82] and for NO (−0.40 eV) and NO₂ (−0.31 eV) adsorption onto the ZnO ($2\overline{1}\overline{1}0$) surface [83]. It should be noted that the gas adsorption does not change the chemical structure of the sensing layer. As mentioned earlier, the sensing mechanism of the gas molecules is due to the charge transfer between the target gas molecules and the sensing layer. A significant amount of charge transfer is observed from NO (0.155 e) and NO₂ (0.141 e) to the sensor. The large electron transfer in these systems can be attributed to the strong chemisorption of the NO and NO₂ gas molecules on the sensor. Such a large charge transfer is expected to induce considerable changes in the conductivity of the system which will be discussed later. The experimental data support our theoretical findings. Singh et al. discovered that ZnO decorated graphene sheet sensor responses to 5 ppm NO gas in dry nitrogen at room temperature [55]. Furthermore, Liu et al. reported that the introduction of ZnO nanoparticles into reduced graphene oxide matrix enhances its sensing performance for NO₂ at room temperature [67].

We also evaluate the sensitivity of the ZnO–graphene sensor to the H₂O and H₂S gas molecules. Figure 3 depicts the most stable configuration of the H₂O and H₂S–ZnO–graphene complexes. For H₂O, our computational results indicate that it is weakly physisorbed on the sensor with an adsorption energy of −0.32 eV which is higher than its adsorption energy on pristine graphene (−0.027 eV) [84] and the pristine ZnO ($0001$) surface (−0.28 eV) [85], while it is less than its adsorption energy on the pristine ZnO ($10\overline{1}0$) surface (−0.94 eV) [86]. Hence, the ZnO–graphene sensor demonstrates promise to work under humid air. A charge of 0.031 e is transferred from the water molecule to the sensor. The minimum distance between the H₂O and the sensor is 1.60 Å and this occurred between H of the H₂O and O of the ZnO which is larger than the sum of the covalent radii of the H and O atoms (0.95 Å) [77]. One can conclude that there exists a weak hydrogen-bonding interaction between H₂O and the sensor. Therefore, the sensor undergoes minor structural changes when is exposed to H₂O (also known as, humidity). Consequently, H₂O is only physisorbed on the surface while preserving its molecular form. The Zn–C₁ and Zn–C₂ are slightly reduced to 2.22 and 2.19 Å, respectively, while C₁–C₂ is kept unchanged. The C₁–Zn–C₂ angle is increased to 38.22°, and the Zn–O bond length is extended to 1.72 Å. These small variations in the sensor are in agreement with the results obtained for the adsorption energy and charge transfer, and confirm the weak physisorption of the water molecule by the sensor. The H₂S adsorption mechanism on ZnO–graphene is more complicated than the other molecules studied above. The H₂S molecule and the sensor experience substantial structural changes after adsorption process. The H–S bond lengths in isolated H₂S (1.35 Å) are extended to 1.36 and 1.95 Å, indicating that the molecule is fully dissociated during the adsorption process into the chemisorbed H and separated HS species. While the first HS bond dissociates, the second one does not. Therefore, dissociation of adsorbed H₂S to adsorbed H and HS occurs on the surface of ZnO according to the following process:

$${\mathrm{H}}_2{\mathrm{S}}_{(\mathrm{ada})}\to \mathrm{HS}+{\mathrm{H}}_{(\mathrm{ads})}$$

(6)

The H atom of the molecule forms a covalent bond with the O atom of the ZnO with a bond length of 1.75 Å. The H–S–H bond angle of the isolated H₂S increases from 91.29° to 92.63° after complexation with ZnO-graphene sensor. The Zn–C₁, Zn–C₂, and C₁–C₂ bond lengths are 2.24, 2.13, and 1.45 Å, respectively. Although H₂S and H₂O are quite similar molecules, the strength of the HS compared to HO bonds is significantly different. The much higher preference of H₂S to dissociate in comparison with H₂O can be related to the different acidity of the two molecules and the much lower pKa value of H₂S [87]. The adsorption energy of H₂S on the ZnO–graphene compound is −1.01 eV which is accompanied by a charge transfer of 0.121 e from the molecule to the sensor. H₂S prefers to dissociatively adsorb on the ZnO ($10\overline{1}0$) surface with an adsorption energy of −1.36 eV [88]. Moreover, the calculated E_ad of H₂S on the ZnO–graphene compound is higher than the values reported for its adsorption on the pristine graphene (−0.11 eV) [89], pristine ZnO ($0001$) (−0.91 eV) [90], and pristine hexagonal monolayer ZnO (−0.54 eV) [91] surfaces, proposing the great capability of the ZnO–graphene hybrid for H₂S dissociation. The experimental reported data confirm our DFT findings. Song et al. experimentally shown that the dispersion of ZnO nanoparticle on the reduced graphene sheet promotes the H₂S adsorption capacities [92]. Moreover, Cuong et al. demonstrated that a sensor based on vertically aligned ZnO nanorods on a graphene film effectively detects 2 ppm of H₂S in oxygen at room temperature [93].

In order to better understand the binding strength of the gas molecules adsorption, the electronic total charges density of ZnO–graphene is obtained and compared with those calculated for gas molecules–ZnO–graphene complexes, as shown in Figure 4. An orbital overlap can be observed between ZnO and the graphene sheet, confirming this theory that ZnO is chemically adsorbed on the graphene sheet. For NO and NO₂, an orbital overlap can be seen between the gas molecules and the ZnO, revealing the occurrence of the strong chemisorption. This strong orbital mixing is attributed to the large charge transfer to the sensor. The weak interactions of the NH₃ and H₂O with the ZnO–graphene sensor can be verified by studying the electronic total charge densities of the molecules-sensor compound. It can be seen that there is no orbital overlap between the molecules previously stated and the sensor. In consistent with our obtained results, while the electron density of the H atom of the H₂S is bonded with the O atom of the ZnO, the HS part of the gas molecule has no overlap with the sensor, confirming the full dissociation of the hydrogen sulfide molecule through the adsorption process.

It is worth mentioning that the electronic properties of the sensor undergo significant changes owing to orbital mixing and the charge transfer between the molecule and the sensor, which is favorable for gas sensing applications. To explore the effect of gas molecules adsorption on the electronic properties of the sensor, the band structure of the ZnO–graphene sensor before and after exposure are calculated, and plotted in Figure 5. As discussed earlier, the ZnO–graphene sensor shows n-type behavior due to the transferred electrons from the ZnO to graphene. This causes a downward shift of the graphene’s Dirac point to the negative energies and opens a gap of 26 meV between conduction (band c) and valence (band v) bands of the graphene, which are highlighted with green color in Figure 5. When the ZnO–graphene sensor is exposed to the gas molecules, different amount of electrons are transferred from the molecules to the sensor which dropped more the K-point of the hybrid structure. The band structure of the sensor changed a little upon adsorption of NH₃ and H₂O. However, the variation in the electronic properties of the sensor after NH₃ adsorption (mostly in the c band) is more pronounced compared to that upon H₂O adsorption. The gap of the hybrid sensor is slightly increased to 33 and 30 meV after NH₃ and H₂O adsorption. These small changes agree well with the small number of electrons transferred from NH₃ (+0.056 e) and H₂O (+0.031 e) to the sensor. Upon exposure to NO and NO₂, the c and v bands are shifted even more to the negative energies, implying that the sensor’s n-type behavior is enhanced. The reason for this phenomena is the large electron transfer after adsorption of NO (+0.155 e) and NO₂ (+0.141 e) onto the sensor. Additionally, the gap between the bands are widened to 68 and 171 meV after the interaction of the sensor with NO and NO₂ gas molecules, correspondingly. Finally, the band structure of the sensor undergoes substantial alterations after interacting with the H₂S gas molecule. The energy range of band c is extended towards the positive electron energies, while the energy range of band v is shrunken. Moreover, the gap between bands c and v at K-point increases to 66 meV, while its location is moved down. These changes can be attributed to the dissociation of the H₂S after interacting with the sensor inducing an electron transfer of +0.121 e from the molecule to the sensor.

The charge transfer caused by the gas molecules adsorption brings about alterations in the conductance of the sensor which is a key element for assessing the performance of the substrate material. To learn about the impacts of the gas molecule adsorption on the transport properties of the sensor, we calculate the quantum conductances of the two-point probe sensor (see Figure 1) with and without the gas molecules’ adsorption, as plotted in Figure 6. This also helps us understand better the ability of the sensor to detect the aforementioned gas molecules. As seen in Figure 6, pristine graphene shows semi-metallic behavior with zero band gap and zero conductance at the Fermi level. The ZnO adsorption significantly stimulates the transport property of graphene. The electronic states of the ZnO–graphene hybrid at the Fermi level are noticeably increased due to the electrons transfer, resulting in an enhancement to the conductance of the system. The quantum conductance of the sensor is found to be 0.34 2e²/h at the Fermi level. In addition, the metallic ZnO–graphene compound shows a peak at electron energy of −0.24 eV. The conductance of the compound drops to zero at electron energy of −0.64 eV, confirming the existence of a gap between c and v bands in its band structure, as discussed before. The conductances of the ZnO–graphene sensor before and after adsorption of NH₃ and H₂O molecules vary very little, suggesting low level of sensitivity to these molecules. However, the alterations of the quantum conductances are more prominent for the other complexes. Once NO (NO₂) is absorbed on the sensor, the quantum conductance of the sensor increases from 0.34 to 0.97 (0.81) 2e²/h. Furthermore, the sensor shows no conductivity for the electron energy range of −1.28 to −1.11 eV after NO₂ adsorption and −1.40 to −1.33 eV after interaction with NO, which concurs well with our previous findings for their corresponding band structures. The observable conductance changes, high adsorption energy, and large charge transfer indicate the promising applications of the ZnO–graphene as a sensitive chemical sensor for NO and NO₂ detection. One can also see that the interactions of the H₂S with the sensor increase the quantum conductance to 0.59 2e²/h. The quantum conductance of the sensor after adsorption of H₂S is zero in the electron energy between −1.02 to −0.95 eV, indicating a shift of 0.36 eV to the negative electron energies in comparison with that of an unexposed sensor. Consequently, the ZnO–graphene sensor might be a potential catalyst for dissociation of the H₂S gas molecule. These results are in good agreement with reported experimental results, showing that the ability of graphene-based sensors to detect NH₃ [55], NO [55], NO₂ [67,94], and H₂S [57,95] are significantly enhanced after adsorption of ZnO.

The sensing capability of the ZnO–graphene sensor to various gas molecules are also assessed by quantifying its sensitivity, which can be expressed as follows:

$$S(\%)=\frac{\left|C-C_0\right|}{C_0}$$

(7)

Here, C and C₀ are the zero bias conductances for the nanosensor with and without the gas molecules, respectively. The results obtained are listed in Table 1. The sensitivities of NH₃, NO, NO₂, H₂O, and H₂S gas molecules on the sensor are calculated to be 11.8%, 188.2%, 141.1%, 6.4%, and 76.47%, respectively. The sensitivities of the sensor to NO, NO₂, and H₂S are much larger than its sensitivity to H₂O, showing superior sensitivity and selectivity toward those gas molecules even in the presence of water vapor. Another important factor for evaluating the performance of a gas sensor is the recovery time (τ), which can be described using conventional transition state theory as follows:

$$\mathsf{\tau }\propto \exp (-E_{ad}/k_{B}T)$$

(8)

A small adsorption energy can result in a fast desorption process of the gases from the sensor. The recovery time is found to follow an ascending order of H₂O < NH₃ < H₂S < NO₂ < NO. The sensor shows a low level of sensitivity to the NH₃ in spite of the short recovery time. Interestingly, the sensor shows promise to work under humid air as a result of its low sensitivity and quick recovery time to water molecules. Since the chemisorption of NO₂ and NO is accompanied with high adsorption energies, much longer time is necessary to completely desorb the gas molecules. Hence, the ZnO–graphene sensor can be utilized as a disposable molecule sensor for specific NO and NO₂ detection. Lastly, the moderate adsorption energy for H₂S suggests that it is possible to desorb it from the sensor by heating at room temperature, revealing a great potential of the ZnO–graphene hybrid to be used as H₂S catalyst.

4. Conclusions

We performed detailed first-principles computational analysis on the geometric and electronic transport properties of the ZnO–graphene hybrid gas sensor upon interaction with NH₃, NO, NO₂, H₂O, and H₂S gas molecules. It was found that NH₃ and H₂O are adsorbed on the sensor with weak hydrogen bonds. These weak interactions caused small variations in the electronic transport properties of the sensor that can be attributed to the small charge transfer from NH₃ and H₂O to the sensor. However, the NO and NO₂ are chemisorbed with a high adsorption energy and large charge transfer, resulting in significant alterations in the conductance of the ZnO–graphene hybrid gas sensor. Interestingly, the H₂S gas molecule is chemisorbed through a dissociative mechanism which is accompanied by a large charge transfer from the molecule to the sensor, leaving noticeable changes in the electronic transport properties of the sensor. Consequently, ZnO–graphene is a potential material for NO and NO₂ detection and is a propitious catalyst for H₂S dissociation.