Adsorption of NH₃ and NO₂ Molecules on Sn-Doped and Undoped ZnO (101) Surfaces Using Density Functional Theory

Abstract

The adsorption and interaction mechanisms of gaseous molecules on ZnO surfaces have received considerable attention because of their technological applications in gas sensing. The adsorption behavior of NH₃ and NO₂ molecules on undoped and Sn-doped ZnO (101) surfaces was investigated using density functional theory. The current findings revealed that both molecules adsorb via chemisorption rather than physisorption, with all the adsorption energy values found to be negative. The calculated adsorption energy revealed that the adsorption of the NH₃ molecule on the bare ZnO surface is more energetically favorable than the adsorption of the NO₂ molecule. However, a stable adsorption configuration was discovered for the NO₂ molecule on the surface of the Sn-doped ZnO surface. Furthermore, the adsorption on the undoped surface increased the work function, while the adsorption on the doped surface decreased. The charge density redistribution showed charge accumulation and depletion on both adsorbent and adsorbate. In addition, the density of states and band structures were studied to investigate the electronic behavior of NH₃ and NO₂ molecules adsorbed on undoped and Sn-doped ZnO (101) surfaces.

1. Introduction

Sensors assist in identifying the different types of pertinent information in the immediate environment and translate the obtained information into an information output such as an electrical signal. Gas sensors are among the common kinds of sensors and are generally used for the identification and detection of harmful or toxic gases. For example, ammonia (NH₃) is commonly used in industries such as cleaning and manufacturing chemicals, and as a refrigerant; exposure to NH₃ at 25 ppm or higher concentrations can cause irritation of the lung, skin, and eyes [1]. Nitrogen dioxide (NO₂) is an important material for the synthesis of nitric acid that is used in the production of fertilizers and explosives; however, exposure to NO₂ at a concentration greater than 1ppm can damage the respiration system [2]. In addition to these examples, gas sensors play very important roles in the food industry, automotive industry, and many other fields [3]. Therefore, it is important to understand their role in the respective applications and the mechanisms behind their performance.

To date, different gas sensors based on semiconductor metal oxide (SMO), including SnO₂, TiO₂, ZnO, and WO₃, have shown outstanding gas detection performance when in contact with oxidizing or reducing gases [4,5,6]. SMO materials operate based on changes in the electrical resistance of the sensing material upon contact with the analyte gas. The type of sensing mechanism depends on the type of material (p/n type), the type of gas (oxidizing/reducing), and environmental conditions such as temperature [7].

Among SMO-material-based sensors, ZnO-based gas sensors are considered to be the most potential candidates because of their physical, chemical, optical, and electrical properties. Extensive experimental research on the sensing performance of ZnO-based sensors has been reported on the sensing capabilities of CO [8], NH₃ [9], NO₂ [10], ethanol [11], and other gases [12,13]. However, ZnO has challenges, such as high-power consumption and poor selectivity, thus hindering its practical application. To overcome these challenges, strategic approaches such as noble metal incorporation [14,15], doping [16,17], and heterostructures [18,19] have been developed. Of the mentioned strategies, doping has been found to enhance the sensing performance of ZnO because of the microstructural changes induced by the dopant. For example, Patil et al. [20] observed enhanced NO₂ sensing performance for Al-doped ZnO nanorods [20]. Zhao et al. [21] reported on the enhanced ethanol sensing capabilities induced by the doping of ZnO nanowires with rare earth elements (Ce, Eu, and Er) [21]. These reports have also confirmed that the gas adsorption mechanisms that occur on the surface of the sensing material play vital roles in gas sensing.

Previously, the adsorption stability of NH₃, NO, and Co molecules on bare and (Ag, Au)-doped ZnO monolayers has been investigated using density functional theory (DFT) [22]. The adsorption stability of NH₃ molecules on a bare ZnO surface was reported to be higher. More importantly, it was stated that (Ag, Au) dopants improve adsorption strength, except for the NH₃ molecule. Liangruksa et al. reported the use of the first-principle approach to investigate the adsorption of gases CH₄, N₂O, NO, NH₃, and CO on a Pd-modified ZnO (0001) surface [23]. The adsorption stability of NH₃ gas was reported to be the most stable, compared with other gases, with an adsorption energy value of −1.06 eV. This previous study established an adsorption strength on the modified ZnO surface, indicating a significant sensitivity to these gases [23].

Therefore, keeping this in mind, we first investigated the sensing properties of pure and tin (Sn)-doped ZnO using DFT calculations. Doping ZnO with Sn can modify the structural, electrical, and optical properties of ZnO, which in turn can influence the gas sensing performance. In this work, the chemisorption and physisorption mechanisms were systematically considered, wherein the adsorption energy, work function, electronic properties, and charge density redistribution were calculated. Understanding the adsorption stability strength and electronic properties of NH₃ and NO₂ molecules on ZnO surfaces will elucidate their fundamental gas sensing mechanisms. Other works on the absorption of NO₂ and NH₃ have been reported before for other SMOs; for example, Prades et al. [24] reported on the adsorption of NO₂ and NO on a SnO₂ (110) surface [24]. The results showed that the NO₂ adsorption strength decreased as the surface area decreased. Furthermore, the bridging oxygen site was discovered to be the most stable adsorption site for both NO and NO₂ molecules. The interaction of NH₃ gas with the surface of kaolinite revealed a greater adsorption capacity at the hollow position than at the top and bridge positions via the N atom site [25]. The stable adsorption configuration of NH₃ was observed by Zhu et al. [26] on a Mn⁴⁺ site, whereas the adsorption energy of the NO molecule was highest at the O top site of the MnO₂ (110) surface through nitrogen coordination [26].

In this work, using the most stable ZnO (101) surface [27], we studied the changes in electronic band structure upon its exposure to a reducing gas compound, NH₃, and an oxidizing gas compound, NO₂. Our goal was to theoretically interpret the gas sensing mechanism involving the ZnO surface in the presence of reducing and oxidizing gas compounds and to also explain how Sn doping influences these mechanisms.

2. Materials and Methods

All the density functional theory calculations were carried out using a Cambridge Serial Total Energy Package (CASTEP) code [28], as implemented in Material Studio Software (MS2020) of BIOVIA Inc, USA, California, San Diego. The Perdew–Burke–Ernzerhof (PBE) exchange–correlation functional was employed within a generalized gradient approximation (GGA) [29]. All the calculations were optimized with the convergence tolerance accuracy and maximum force of 10⁻⁵ eV/atom and 0.03 eV/Å, respectively. We used a convergence plane-wave cutoff energy value of 500 eV and k-points gride of 6 × 2 × 5 using the Monkhorst–Pack approach [30] for surface and adsorption calculations. During geometry optimization, the limited-memory Broyden–Fetcher–Goldfarb–Shanno minimization scheme algorithm was used. The surface models were created from the optimized bulk structure of the ZnO crystal with a space group F-43m and lattice parameter a = b = c = 4.625 Å. The surface models were generated to have symmetric top and bottom layers to avoid dipole effects. The surface slabs were represented by five atomic layers separated by a vacuum region of 25 Å to avoid self-interaction. The inner layers were fixed, and the rest of the system was allowed to relax during structural optimization. To describe the reaction of gas molecules with the doped and undoped surfaces, the molecules were in position 2 Å on top of the surface. The adsorption energy strength was calculated using the following expression:

E_ads = E_mol-surf − (E_surf + E_mol)

(1)

where E_mol-Surf, E_surf, and E_mol describe the total energy values of the combined molecule–surface structure, the doped or undoped surface, and the free molecule, respectively. Both NH₃ and NO₂ molecules were placed above Zn and the dopant (Sn) on the surface, and full geometry optimization was performed.

3. Results and Discussion

3.1. Structural Analysis

Figure 1 depicts the atomic surface slabs for ZnO (101) and Sn-ZnO (101) to describe the surface plane. The surface slabs were constrained to have 5 atomic layers with 20 atoms. The equilibrium distance between the nearest surface layer ranged from 1.3 to 1.5 Å, with bond lengths d_Zn-O (1.871 Å) and d_Sn-O (2.083 Å).

In order to investigate the effects of NH₃ and NO₂ exposure, both molecules were placed above Zn and the dopant (Sn) on the ZnO (101) surface, and full geometry optimization was performed. Figure 2 presents the optimized atomic structure of NH₃ and NO₂ molecule adsorption on the doped and undoped surfaces, namely (a) NH₃/ZnO (101), (b) NO₂/ZnO (101), (c) NH₃/Sn-ZnO (101), (d) NO₂/Sn-ZnO (101), (e) chemisorption NH₃/Sn-ZnO (101), and (f) chemisorption NO₂/ Sn-ZnO (101) surfaces. The adsorption interaction configuration of both molecules with pure and doped ZnO (101) surfaces led by N atom bonding. It is worth noting that the equilibrium bond distance between the N atom of the molecule and the surface differed. The bond distance (d_N-Sn, Å) for NH₃ adsorption was found to be 2.785 Å, which was larger than d_N-Zn, with 2.163 Å. A similar observation was also noted on the NO₂ molecule adsorption, where the bond distance of d_N-Zn was smaller (2.361 Å) than d_N-Sn = 2.497 Å.

3.2. Adsorption Energy

The adsorption mechanism and exposure of the NH₃ and NO₂ gaseous molecules on the surface were investigated by calculating the adsorption energy using Equation 1.

Table 1. Calculated adsorption energy values (E_ads., eV) of NH₃ and NO₂ molecules on doped and undoped ZnO (101) surfaces and atomic bond distance (d_N-Zn, d_N-Sn, Å) for the nearest atoms.

		Eads (eV)
		Physisorption		Chemisorption
System	(dN-Zn, dN-Sn, Å)	1 Å	2 Å	Å
NH3/ZnO (101)	2.163–2.162	−1.042	−0.746	−0.746
NO2/ZnO (101)	2.361–2.360	−0.354	−0.280	−0.279
NH3/Sn-ZnO (101)	2.785–3.148	−0.339	−0.187	−0.357
NO2/Sn-ZnO (101)	2.493–2.522	−1.105	−0.436	−0.438

presents the adsorption energy rates (E_ads, eV) and equilibrium bond distance (d, Å). In this work, both chemisorption and physisorption approaches were considered to investigate the nature of adsorption. Generally, physisorption is adsorption without bonding, while chemisorption is adsorption with bonding. Upon a comparison of the adsorption energy, it was found that all the adsorption energy rates were negative, indicating thermodynamic stability. Strong adsorption was distinguished by the greatest negative value of adsorption energy. E_ads < 0 implies that the adsorption process occurred spontaneously due to attractive interaction and exothermic process. The chemisorption process was completed by placing the molecules in different positions (1 Å and 2 Å) above the surface. It was found that the calculated adsorption energy rates at 1 Å positioning were more preferential than those at 2 Å, suggesting a stronger interaction.

In addition, it was clearly observed with the adsorption energy results that the Sn-doped ZnO surface exhibited an interaction of the NO₂ molecule in both physisorption and chemisorption processes; however, the adsorption of NH₃ remained similar. The calculated adsorption energy of the NH₃ molecule on the ZnO (101) surface was −0.746 eV, which was negatively larger than the E_ads (−0.279 eV) of the NO₂ adsorption. Generally, a larger negative value of the adsorption energy suggests a stable configuration and stronger interaction. This implies that the NH₃ molecule adsorption was more favorable than the adsorption of NO₂ on the ZnO (101) surface. Furthermore, this indicates a weak interaction in the NO₂/ZnO (101) surface due to the large bond distance (d_N-Zn = 2.785 Å).

A significant adsorption energy value was observed on the doped ZnO (101) surface. The chemisorption and physisorption energy values for the NO₂ adsorption on the Sn-doped ZnO surface were −0.438 eV and −1.105 eV, respectively. These rates were relatively higher than those of the NO₂ adsorbed on the bare ZnO surface (see Table 1). Furthermore, it was found that the E_ads of the NH₃ molecule was −0.187 eV on the Sn-doped ZnO (101) surface, which was less than the E_ads value on the undoped surface.

The current findings revealed that the NO₂ molecule adsorption was more favorable on the doped surface than the NH₃ adsorption. The magnitude of the adsorption strength for NH₃ gas declines declined with Sn doping, whereas the interaction of the NO₂ molecule enhances. This indicates that NO₂ exhibited a stronger adsorption strength on the Sn-doped ZnO surface than NH₃. However, the NH₃ adsorption was more preferential than that of NO₂ on the bare ZnO (101) surface. This was also shown by the bond length interaction between the N atom (NO₂ and NH₃) and the surface atoms (Zn and Sn). This implies that the Sn-doped ZnO surface had a lower attraction of NH₃ gas while enhancing the stronger attraction of the NO₂ molecule. Previous work by Mhlongo et al. [31] demonstrated that, due to the high quantity of donor defects, the undoped ZnO-based surface responded to NH₃ gas more strongly than the doped ZnO, although transition-metal-doped sensors had faster responses and recovery periods than undoped ZnO.

3.3. Electronic Properties

Fundamentally, comprehending the classification of materials into the three groups of metals, semiconductors, and insulators depends on a system’s electronic properties. The width of the energy band gap between the conduction band (CB) and the valence band (VB) establishes the kind of material. Using the lattice parameters a = b = c = 4.625 Å, the band energy for bulk ZnO was estimated and found to be 0.597 eV, which was lower than the empirically reported value [32]. As it is well-known that the DFT calculations underestimate the lowest unoccupied level, this is a common issue with the band gap of semiconductors. However, this drawback had no impact on the examination of electronic structure in our work. When the findings of the gas molecules adsorbed on the ZnO surface were compared using the same system and calculation method, this calculation error could be disregarded.

The measured lattice parameters and the required high symmetry directions of the matching irreducible Brillouin zone were used to optimize the clean ZnO (101) surface. The calculated band structure revealed a direct energy band gap of 0.957 eV, as shown in Figure 3, which was located at the gamma (G) point. This value was higher than that of the bulk structure and indicates that the probability of distribution of the electrons was the greatest on the surface, i.e., the electron was constrained near the surface. The valence band of pure (101) ZnO surface, as seen from the density of states (DOS), had two peaks between 0 and 6.5 eV as well as between 15.9 and 17.0 eV. Doping (101) ZnO with Sn caused modifications in the electronic structure, as evidenced by the increase in the energy band gap from 0.957 to 1.03 eV and the addition of new peaks at about 20 eV in the DOS electronic structure.

Figure 4a–d show the electronic characteristics of NO₂ and NH₃ physisorption on ZnO (101) and Sn-doped ZnO (101) surfaces, where both NO₂ and NH₃ were adsorbed via a Van der Waals interaction to either Zn or Sn on the surface. The computed band gap values for NH₃/ZnO (101), NO₂/ZnO (101), NH₃/Sn-doped ZnO (101), and NO₂/Sn-doped ZnO (101) surfaces were 0.945 eV, 0.766 eV, 1.075 eV, and 0.634 eV, respectively. The predicted band gaps for the NO₂ adoption in both Sn-doped ZnO (101) and ZnO (101) surfaces were lower than the clean ZnO (101) surface, whereas the band gaps for the adsorbed NH₃ were larger. It was also discovered that the Fermi level of the NO₂/Sn-doped ZnO (101) surface introduced new states, resulting in the migration of both the conduction and valence bands, transforming the system from a p-type to an n-type semiconductor. DOS demonstrated the presence of molecular states in VB for all the structures, as seen by the creation of extra peaks near the bottom of the VB between −7 and 24 eV.

The band structures and DOS of NO₂ and NH₃ adsorbed on either the ZnO (101) or Sn-doped ZnO (101) surface through a chemical interaction are shown in Figure 5a–d. For NH₃/ZnO (101), NO₂/ZnO (101), NH₃/Sn-doped ZnO (101), and NO₂/Sn-ZnO (101), the calculated band-gap values were 0.937 eV, 0.991 eV, 0.911 eV, and 0.733 eV, respectively. For physisorption, NH₃/ZnO (101) and NH₃/Sn-doped ZnO (101) had larger band gaps than their equivalents, while NO₂/ZnO (101) and NO₂/Sn-doped ZnO (101) had band gaps lower than those obtained via chemisorption. This demonstrates that chemisorption and physisorption mechanisms had distinct effects on the band-gap values. However, we also made a similar observation on Fermi energy for NO₂/Sn-doped ZnO (101), where we noticed the emergence of new states at the bottom of the conduction band that overlapped the Fermi energy in the direction of the valence band, changing the system from a p-type to an n-type semiconductor. The emergence of additional peaks near the bottom of the VB between −7 and 24 eV in the case of DOS showed comparable observations to those in the case of chemisorbed mechanism, where we found the presence of molecular states in VB for all the structures.

3.4. Work Function

The minimal amount of energy necessary to remove or extract an electron from a crystal surface in vacuum is known as the work function, which is commonly referred to as electrostatic potential. This is the most fundamental crystal solid surface parameter for understanding a wide range of structural, physical, and chemical surface conditions. The electrostatic potential can be expressed as follows:

$$\Phi =E_{vac}-E_F$$

(2)

where $E_{vac}$ and $E_F$ represent the electrostatic potential energy of the vacuum and Fermi energy levels, respectively. In an adsorption system, the work function also plays a significant role in understanding the atomic interaction [33]. Figure 6 presents the work function plots for pure ZnO, Sn-doped ZnO, NH₃/ZnO, NO₂/ZnO, NH₃/Sn-doped ZnO, and NO₂/Sn-doped ZnO (101) surfaces. It appears that the work function of various surfaces significantly varied since it depended on the crystallographic orientation of the surface in most cases. The calculated work function of the ZnO (101) surface was 5.258 eV, which was comparable to the experiment work function value ranging from 5.4 to 6.5 eV [34]. It was observed that the computed work function (5.055 eV) of the Sn-doped ZnO (101) surface had a numerical value substantially smaller than that of the pure ZnO (101) surface. The value of the surface work function was reduced by 0.203 eV, compared with that of the undoped ZnO (101) surface; this is due to the charge rearrangement of electrons and ions at the surface. A surface with a lower work function easily transfers the electronic charge to any adsorbate with higher electronegativity, resulting in ionic bonding [35].

Furthermore, Figure 6 shows that the adsorption of NH₃ and NO₂ molecules changed the value of the work function compared with the pure and doped surfaces, indicating electron charge transfer. The adsorption of NH₃ and NO₂ on the surface of ZnO increased the value of the work function by 0.742 and 0.087 eV, respectively. The NH₃ molecule enhanced the surface work function larger than NO₂, which corresponded to a stronger adsorption energy strength. However, it was noted that adsorption on the Sn-doped ZnO surface reduced the surface work function. Previous researchers suggested that the induced work function is caused by a dipole involving the negative charge of the molecule [36].

3.5. Charge Density Distribution

The charge density distribution plots were examined in order to learn more about the nature of chemical bonding. The electronic hybridization between the molecule’s orbital and the surface, as well as the molecule’s adsorption and contact with the surface, caused the redistribution of charge density. Importantly, one of the attributes that facilitate the analysis of a chemical interaction is the changing density. It should be mentioned that charge density is influenced by crystal structure and can be used to comprehend a material’s electrical properties. The charge density distribution for the current system was described as follows:

∆ρ_system = ρ_mol/surf − ρ_surf − ρ_mol

(3)

where ρ_mol-surf is the charge density of the adsorbed surface, while ρ_surf and ρ_mol refer to the charge density for the pure surface and molecule, respectively. Figure 7 shows a 2D plot of the electronic density difference between the ZnO (101) and Sn-doped ZnO (101) surfaces. An electron-enriched (−q) area is shown by the blue region, while electron deduction is presented by the red regions. The plot of the electron charge difference presents a large blue isosurface on the O atom, while the red section is seen on both Zn and Sn atoms. This agrees with the Mulliken atomic charge values presented in Figure 7c,d.

Figure 8 presents the charge density distribution for NH₃/ZnO (101), NH₃/Sn-doped ZnO (101), NO₂/ZnO (101), and NO₂/Sn-doped ZnO (101) surfaces. As described in Equation 3, the charge density difference was calculated by subtracting the charge density of the pure or doped surface from that of the single (NH₃ and NO₂) molecule. The magnitude of the charges is represented by different colored regions, with yellow indicating depletion and blue indicating accumulation. The charges were primarily localized on the N-Zn and N-Sn bonding, according to the 3D isosurfaces of the charge redistribution, as seen in Figure 8. The figure shows that the charge distribution in the isosurface region significantly differed from one model to the other. A wide isosurface region indicates a greater adsorption strength and suggests an increased charge transfer rate. As shown in Figure 8a, a large isosurface region was observed on the NH₃/ZnO (101) adsorption, while a smaller volume was seen for the NO₂ adsorption. This is similar to the adsorption energy strength reported in Table 1 and the larger value of the work function in Figure 6.

Furthermore, the charge density distribution plot for the NH₃/ZnO (101) adsorption showed a covalent character, while for NO₂/ZnO (101), NO₂/Sn-doped ZnO (101), NH₃/Sn-doped ZnO surfaces, it showed the ionic character of the bonding. The yellow isosurface was discovered to be between the interactions of Zn-N and Sn-N (both N atoms from NH₃ and NO₂), indicating electron depletion from the surface atom. This means that the molecules serve as electron acceptors. A directional bonding of spherical shape was observed, which indicates ionic bonding. This was also reported by Tshwane et al. [36] on the adsorption of halogen ions and molecules. Furthermore, the yellow region for the NO₂/Sn-doped ZnO surface was larger than that observed on the NH₃/Sn-doped ZnO surface. This is due to the larger negative and stronger adsorption energy strength presented in Table 1.

4. Conclusions

In conclusion, the adsorption mechanisms of NH₃ and NO₂ molecules on ZnO and Sn-doped ZnO (101) surfaces were successfully investigated using density functional theory. The study considered both the chemisorption and physisorption phenomena. The calculated adsorption energy values were found to be negative, implying that the adsorption mechanism is thermodynamically favorable. It was discovered that both molecules adsorb on the surface via chemisorption rather than physisorption, which is due to the stronger interaction of N and Zn or Sn atoms. The adsorption strength of NH₃ on the ZnO surface was found to be more stable, whereas the NO₂ adsorption configuration on the Sn-doped ZnO surface was preferred. The relatively larger negative E_ads value indicated a stronger adsorbing strength. Furthermore, the adsorption of both molecules induced the charge density redistribution and changed the surface work function. This analysis revealed the charge accumulation and depletion patterns at the interface. Moreover, the results revealed that the surface work function increased on the doped ZnO surface while decreasing on the undoped ZnO surface.