p-CuO/n-ZnO Heterojunction Structure for the Selective Detection of Hydrogen Sulphide and Sulphur Dioxide Gases: A Theoretical Approach

Abstract

DFT calculations at the B3LYP/LanL2DZ level of theory were utilized to investigate the adsorption of H₂S and SO₂ gases on the electronic properties of CuO-ZnO heterojunction structures. The results were demonstrated from the standpoint of adsorption energies (E_ads), the density of states (DOS), and NBO atomic charges. The obtained values of the adsorption energies indicated the chemisorption of the investigated gases on CuO-ZnO heterojunction. The adsorption of H₂S and SO₂ gases reduced the HOMO-LUMO gap in the Cu₂Zn₁₀O₁₂ cluster by 4.98% and 43.02%, respectively. This reveals that the Cu₂Zn₁₀O₁₂ cluster is more sensitive to the H₂S gas than the SO₂ gas. The E_ads values for SO₂ and H₂S were −2.64 and −1.58 eV, respectively. Therefore, the Cu₂Zn₁₀O₁₂ cluster exhibits a higher and faster response-recovery time to H₂S than SO₂. Accordingly, our results revealed that CuO-ZnO heterojunction structures are promising candidates for H₂S- and SO₂-sensing applications.

1. Introduction

Environmental pollution has recently reached an alarming level as a result of fast industrialization, which has resulted in an increase in the number of poisonous and dangerous gases and chemicals in the atmosphere [1,2,3,4]. Because of their colorless and odorless nature, most gases are breathed by humans without their knowledge, causing serious health problems and even death [5,6,7,8]. As a result, developing sensors and systems that can effectively identify such dangerous and hazardous pollutant gases is critical [9,10,11,12,13]. Solid electrolyte gas sensors, electrochemical gas sensors, metal oxide semiconductor sensors (MOS), catalytic combustion gas sensors, and other types of gas sensors have been documented in the literature [14,15,16,17]. MOS type sensors are among the most researched of the many types of gas sensors because they may be used to detect a variety of gases at lower concentrations merely by studying the change in resistance induced by the interaction between the sensing material and the target gas [18,19,20,21]. The operating concept of a gas sensor is based on changes in the sensing material’s electrical conductivity. Most target gas molecules are adsorbed by oxygen ions that have already been adsorbed at the metal-oxide surface due to the large number of surface sites. The charge carrier concentration can be modulated via interactions between adsorbed target species and oxygen ions, changing the material’s conductivity (or resistivity) [20].

Hydrogen sulphide (H₂S) and sulphur dioxide (SO₂) are extremely toxic, colorless, poisonous, and dangerous gases that do not only pollute the environment but also represent a major health risk to humans [22,23,24,25]. Inhaling such gases can cause serious health issues in humans, such as eye inflammation, respiratory tract infections, asthma, cardiac damage, cancer, and irreversible pulmonary impairment, and so on [26,27,28,29,30]. As a result, an efficient technique for detecting such hazardous gases is necessary.

Metal oxide semiconductor (MOS)-based resistive sensors have recently attracted a lot of interest because of their ease of manufacture and inexpensive cost, as well as their quick, programmable, sensitive, and selective response [27,28,29,30,31,32,33]. The performance of MOS-based sensors can be improved by using semiconductor metal oxide nanomaterials as sensing material because the nanomaterials have a higher specific surface area, which increases the interaction between the sensing material and the target gas and, thus, improves the sensing response [1,2,4,5]. Metal oxide nanomaterials such as zinc oxide (ZnO), copper oxide (CuO), tin oxide (SnO₂), tungsten oxide (WO₃), nickel oxide (NiO), cobalt oxide (Co₃O₄), iron oxide (Fe₂O₃), titanium oxide (TiO₂), indium oxide (In₂O₃), vanadium oxide (V₂O₅), molybdenum oxide (MoO₃), and others are used to fabricate efficient MOS sensors [28,29,30,31,32,33,34,35,36,37]. To explore the characteristics of both, n- and p-type semiconductor nanomaterials, MOS sensors have recently been made using p-n heterojunction nanostructures, which employ composites of p- and n-type semiconductors as functional materials to improve sensing characteristics. Thus, several p-n heterojunction nanostructure-based MOS sensors were fabricated and reported in the literature, for instance, SnO₂/Co₃O₄, CuO/ZnO, NiO/ZnO, CuO/CuFe₂O₄, SnO₂/CuO, WO₃/CuO, and so on [22,23,24,25,26,27,28,29,30,38,39,40,41]. CuO/ZnO nanoparticles demonstrated good selectivity towards H₂S gases among heterojunction nanomaterial-based MOS sensors due to the fact that the electron depletion layer on the surface of particles is enlarged by p-n heterojunctions, and the separation of electron-hole carriers increases the active sites of gas-solid reactions on the surface of the material [5,6,7,8,27]. Several experimental investigations on this p-n heterojunction nanomaterial for H₂S sensing have been conducted and reported, however, to the best of our knowledge, no theoretical work has been reported on this subject yet.

In this article, we report the theoretical investigations on the interaction and sensitivity between target sensing gases (H₂S and SO₂) and various quantum clusters of CuO-ZnO heterojunction structures. Density functional theory (DFT) computations in the Gaussian 09 software were used for theoretical research. The optimum compositions of the CuO-ZnO heterojunction structures were observed using a variety of geometric optimizations. Various computations relating to the gas-sensing characteristics were done to properly leverage the CuO-ZnO heterojunction structures.

2. Methods and Computational Details

Herein, density functional theory (DFT) calculations were performed to scrutinize the H₂S and SO₂ interaction with the CuO-ZnO heterojunction structures. The B3LYP hybrid functional contains exchange and correlation functionals and is based on the exact form of the Vosko–Wilk–Nusair correlation potential [42]. The functional B included the Slater exchange along with corrections involving the gradient of the density [43,44], whereas the correlation functional LYP includes both local and non-local terms [45,46]. Interestingly, the B3LYP/LanL2DZ has been utilized successfully to predict the geometric and electronic properties of various metal oxide (ZnO, CuO) surfaces [47,48,49]. In the previous studies, a twenty-four atom M₁₂O₁₂ (M=Zn, Cu) quantum cluster was utilized to simulate ZnO and CuO nanomaterials [50,51,52]. Therefore, in the present study, the CuO-ZnO heterojunction was simulated by adopting the Cu₂Zn₁₀O₁₂ quantum cluster. The B3LYP method was employed to perform all the calculations. The LanL2DZ basis set was utilized for Cu, Zn, S, O, and H atoms. The LanL2DZ pseudopotential was added to Cu and Zn atoms. The adsorption energies (E_ads) were estimated as:

$${\mathrm{E}}_{\mathrm{ads}}{=\mathrm{E}}_{\mathrm{adsorbate}/\mathrm{substrate}}-{ (\mathrm{E}}_{\mathrm{substrate}}{+\mathrm{E}}_{\mathrm{adsorbate}})$$

(1)

where, ${\mathrm{E}}_{\mathrm{adsorbate}/\mathrm{substrate}}$, ${\mathrm{E}}_{\mathrm{substrate}}$, and ${\mathrm{E}}_{\mathrm{adsorbate}}$ are the energies of the optimized adsorbate-substrate complexes, the Cu₂Zn₁₀O₁₂ cluster, and the free adsorbate gas, respectively. The negative adsorption values refer to an exothermic interaction. In other words, the more negative the value of the absorption energy, the more stable the adsorbate-substrate complex.

All the considered structures are fully optimized without any constraints under the following optimization conditions: the force acting on an atom, the root-mean-square of the force, the calculated displacement for the next step and the root-mean-square of the displacement for the next step must be below the cutoff values of 0.129 eV/Å, 0.086 eV/Å, 5.29 × 10⁻³ Å. and 3.53 × 10⁻³ Å, respectively, all in atomic units. All the calculations were performed using the Gaussian 09 program [53] and visualized by Gauss View 5, whereas the density of states (DOS) was represented by the Gauss Sum 3.0 program [54]. NBO version 3.1 [55] was used to estimate the atomic charge distribution.

Equation (2) [56,57,58] shows the dependence of the electrical conductivity (σ) on the E_g

$$\mathsf{\sigma } ={\mathrm{AT}}^{3/2}\exp \left({-\mathrm{E}}_{\mathrm{g}}/2\mathrm{kT}\right)$$

(2)

where A is a constant, k is the Boltzmann constant, and T is the temperature.

The sensor sensitivity is calculated by the following equation [58]:

$$\mathrm{S}= \left|\exp \left(\frac{{\mathrm{E}}_{\mathrm{g}}{\left(\mathrm{II}\right) -\mathrm{E}}_{\mathrm{g}}\left(\mathrm{I}\right)}{\mathrm{kT}}\right)\right|$$

(3)

where E_g (I) and E_g (II) are the energy gap value for the substrate and the adsorbate-substrate complex, respectively.

The recovery time ($\mathsf{\tau }$) is given as [59]:

$$\mathsf{\tau }={\mathsf{\nu }}_{\mathrm{o}}^{-1} \exp \left(\frac{{-\mathrm{E}}_{\mathrm{ads}}}{\mathrm{kT}}\right)$$

(4)

where ${\mathsf{\nu }}_{\mathrm{o}}$ is the attempt frequency, k is the Boltzmann constant and T is the temperature.

3. Results and Discussion

3.1. Geometric Optimization

Five isomeric structures of the Cu₂Zn₁₀O₁₂ cluster, labelled as a, b, c, d, and e (see Figure 1) that represent all potential locations for the copper atoms were studied in order to find the global minima structure of the investigated Cu₂Zn₁₀O₁₂ cluster. Interestingly, it was observed that structure (a) is the most energetically stable structure. Figure 2a shows the fully optimized structures at the optimum spin and represents the energy variations of the structures relative to the most stable structure. Therefore, structure (a) was considered for further calculations. Figure 2b illustrates the energy variations of structure (a) versus the spin of the cluster. Figure 3 demonstrates the density of states as well as the HOMO and LUMO orbital for the most energetically stable structure of the Cu₂Zn₁₀O₁₂ cluster. It is clear that the HOMO (α) orbital was localized on the O atoms whereas the LUMO (α) was localized on the Zn atoms of the cluster. This elucidated that the O atoms are electron-rich centers and the Zn atoms are electron-deficient centers.

Furthermore, the copper atoms contributed oxygen atoms to the HOMO (β), whereas the LUMO (β) was localized mainly on the Cu atoms. The contribution of Cu atoms in both HOMO and LUMO may be owing to the incompletely filled d orbitals. Moreover, the Cu₂Zn₁₀O₁₂ cluster was a semiconductor with a HOMO-LUMO energy gap (E_g) of 2.81 eV. The natural bond orbitals (NBO) showed that the copper atoms had positive charges of 0.96 |e|, the zinc atoms had positive charges ranging from 1.42 to 1.44 |e|, and the oxygen atoms carried negative charges ranging from −1.23 to −1.42 |e|. Moreover, the dipole moment of the Cu₂Zn₁₀O₁₂ was 0.46 Debye.

3.2. Interaction of Gases with the Cu₂Zn₁₀O₁₂ Structure

Additionally, the H₂S and SO₂ gases were optimized. For the H₂S molecule, the H-S bond was 1.377 Å and the H-S-H angle equaled 94.2°. The charges on the S and H atoms were −0.270 and 0.135 |e|, respectively, which were consistent with the reported literature [60,61,62]. For the SO₂ molecule, the S-O bond was 1.610 Å and the O-S-O angle equaled 112.8°. The S and O atoms carried charges of 1.32 and −0.66 |e|, respectively, which agreed with previous studies [63,64,65]. The dipole moment values for SO₂ and H₂S were 2.79 and 1.76 Debye. Because the high dipole moment of a molecule indicates a high reactivity with the surrounding medium [66], one can expect that the reactivity of the SO₂ molecule is higher than that of the H₂S molecule.

To understand the effect of the Cu₂Zn₁₀O₁₂ cluster sensing features on H₂S and SO₂ gases, the characteristics of adsorbate-substrate interaction were scrutinized. For that, the adsorbates (H₂S and SO₂), substrate (Cu₂Zn₁₀O₁₂), and the adsorbate-substrate complexes (H₂S/Cu₂Zn₁₀O₁₂ and SO₂/Cu₂Zn₁₀O₁₂) were fully optimized without any constraints. Since the adsorbed H₂S (or SO₂) molecule can interact via its S atom and H (or O) atom with the Cu, Zn, and O sites, six adsorption modes representing the previous possibilities were investigated (Figure 4). The adsorption energies for the investigated complexes were calculated by Equation (1).

Figure 5 and Figure 6 show the optimized H₂S/Cu₂Zn₁₀O₁₂ and SO₂/Cu₂Zn₁₀O₁₂ complexes, respectively, with adsorption energies. It is worth mentioning that the adsorption heat that is more negative than −0.2 eV specifies the chemisorptions [67,68].

It is clear that the adsorption for SO₂/Cu₂Zn₁₀O₁₂ and H₂S/Cu₂Zn₁₀O₁₂ is chemisorption for all the adsorption modes except for SO₂/Cu₂Zn₁₀O₁₂ mode II and H₂S/Cu₂Zn₁₀O₁₂ mode III. This indicates a strong interaction between the investigated gas and the cluster, which explains the sensitivity of the cluster to the considered gases. Assuming that the adsorption modes with the highest released adsorption energies are the most likely interaction, we will pay attention to SO₂/Cu₂Zn₁₀O₁₂ mode III and H₂S/Cu₂Zn₁₀O₁₂ mode V.

Table 1. The adsorption properties for SO₂/Cu₂Zn₁₀O₁₂ and H₂S/Cu₂Zn₁₀O₁₂. The adsorption energy (E_ads, eV), HOMO-LUMO gap (E_g, eV), percentage change in HOMO-LUMO gap (ΔE_g, %), NBO charges (Q, eV), and dipole moment (µ, Debye).

Adsorption Mode	Eads	Eg	ΔEg%	QS	QI	QII	Qgas	µ
SO2/Cu2Zn10O12	−2.64	2.67	−4.98	1.62	−0.96	−0.97	−0.31	3.35
H2S/Cu2Zn10O12	−1.58	1.60	−43.02	−0.73	0.54	0.14	−0.05	5.75

summarizes the adsorption properties, whereas Figure 7 represents the density of states and the optimized geometrical structures for the considered modes.

As clearly shown in Table 1 for the considered SO₂/Cu₂Zn₁₀O₁₂ cluster, the positive charge of the S atom of the SO₂ molecule rose to 1.62 |e| rather than 1.314 |e| for the free SO₂ molecule, whereas the negative charges of the O atoms increased to −0.96 and −0.97 |e| rather than −0.66 |e| for the free SO₂ molecule. This confirms that a charge transfer occurred from the S atom to the substrate and from the substrate to the O atoms of the SO₂ molecule. As a result, three bonds were formed between the SO₂ molecule and the cluster as shown in Figure 7a. These bonds are the Zn-O bond, S-O bond, and Cu-O bond, and they are dashed in Figure 7a. The first, second and the third bonds were formed between Zn, O, Cu atoms from the cluster and O(II), S, and O(I) atoms from the SO₂ molecule, respectively. This explains the high adsorption heat of the SO₂/Cu₂Zn₁₀O₁₂ cluster. To confirm the formation of such bonds, the Mulliken overlap population, as well as the bond order, was calculated for the suggested bonds and is tabulated in

Table 2. Overlap population and bond order analysis for the proposed bonds in the SO₂/Cu₂Zn₁₀O₁₂ complex.

Bond	Overlap Pop.	Bond Order
Zn O	−1.573	−0.974
S O	−9.963	−0.536
Cu O	−1.239	1.026

. The small value for the overlap population, close to zero, declares a non-significant interaction between the electronic populations of the two atoms, whereas the high value refers to a high degree of interaction [55,69]. It is noteworthy that high overlap population values of 1.57, 9.96 and 1.24 as well as high bond order values of 0.97, 0.54, and 1.03 were recorded for the Zn-O bond, S-O bond, and Cu-O bond, respectively. This indicates that a strong adsorbate-substrate interaction occurred.

On the other hand, for the considered H₂S/Cu₂Zn₁₀O₁₂ cluster, due to the adsorption, the H₂S dissociated into two fragments, HS⁻ and H⁺. The HS⁻ fragment was bounded to the Zn site of the cluster via the S atom. As a result, the negative charge of the S atom rose to −0.73 |e| rather than −0.27 |e| for the free H₂S molecule. This means a charge transfer occurred from the substrate cluster to the S atom. Although the H⁺ fragment was bounded to an O site of the cluster and a charge transfer occurred from the H to the substrate cluster, the charge of the positive hydrogen atom rose to 0.54 |e| rather than 0.14 |e| in the free H₂S molecule. In other words, there was charge donation, and charge back-donation occurred between the adsorbed gaseous molecule and the substrate for both SO₂/Cu₂Zn₁₀O₁₂ and H₂S/Cu₂Zn₁₀O₁₂. Consequently, charges redistribution occurred and as a result, the dipole moment rose to 3.35 and 5.75 Debye, respectively. This led to a decrease in the HOMO-LUMO gap (E_g) by 4.98% and 43.02% as shown in the density of the state spectra for SO₂/Cu₂Zn₁₀O₁₂ and H₂S/Cu₂Zn₁₀O₁₂, respectively, with regards to the bare cluster (see Figure 7). Since, the electrical conductivity (σ) depended on the E_g obeying Equation (2), the adsorption of SO₂ and H₂S gases decreased the electrical conductivity of the Cu₂Zn₁₀O₁₂ cluster. Therefore, the Cu₂Zn₁₀O₁₂ cluster was sensitive to the examined gases.

It is known that the sensor sensitivity was affected by the E_g as seen in Equation (3). Therefore, the sensitivity of the Cu₂Zn₁₀O₁₂ cluster toward the investigated gases is demonstrated in Figure 8. It is clear that the Cu₂Zn₁₀O₁₂ cluster was more sensitive to the H₂S gas than SO₂ gas. Furthermore, the sensor action was prominently dependent on the recovery time ($\mathsf{\tau }$). Equation (4) shows that the $\mathsf{\tau }$ increased as the released adsorption energy (E_ads) increased (more negative). Table 2 shows that the E_ads values for SO₂/Cu₂Zn₁₀O₁₂ and H₂S/Cu₂Zn₁₀O₁₂ were −2.640 and −1.576 eV, respectively. Therefore, the Cu₂Zn₁₀O₁₂ cluster exhibited a higher and faster response-recovery time to H₂S than SO₂.

4. Conclusions

In conclusion, the H₂S and SO₂ interaction with the CuO-ZnO heterojunction structures were investigated utilizing DFT calculations at the B3LYP/LanL2DZ level of theory. The CuO-ZnO heterojunction was simulated by the Cu₂Zn₁₀O₁₂ quantum cluster. Six adsorption modes were investigated. The calculated adsorption energies for the H₂S and SO₂ reached −1.57 and −2.64 eV, respectively, which revealed the strong interaction between the investigated gases and the Cu₂Zn₁₀O₁₂ cluster. Although the SO₂ molecule formed three bonds with the Cu₂Zn₁₀O₁₂ cluster, the H₂S molecule dissociated into HS⁻ and H⁺, which bound to the Zn and O sites of the Cu₂Zn₁₀O₁₂ cluster. This interaction led to a decrease in the HOMO-LUMO gap (E_g) of the Cu₂Zn₁₀O₁₂ cluster by 4.98% and 43.02% for SO₂ and H₂S, respectively. Therefore, the Cu₂Zn₁₀O₁₂ cluster was more sensitive to the H₂S gas than the SO₂ gas. Furthermore, the Cu₂Zn₁₀O₁₂ cluster exhibited higher and faster response-recovery time to H₂S than SO₂. Our present results indicate that the CuO-ZnO heterojunction can be a potential nanomaterial for the detection of H₂S and SO₂ gases.