Monitoring Gases Content in Modern Agriculture: A Density Functional Theory Study of the Adsorption Behavior and Sensing Properties of CO2 on MoS2 Doped GeSe Monolayer

The reasonable allocation and control of CO2 concentration in a greenhouse are very important for the optimal growth of crops. In this study, based on density functional theory (DFT), an MoS2–GeSe monolayer was proposed to unravel the issues of the lower selectivity, poorer sensitivity and non-recyclability of traditional nanomaterial gas sensors. The incorporation of MoS2 units greatly enhanced the sensitivity of the pure GeSe monolayer to CO2 and the high binding energy also demonstrated the thermal stability of the doped structures. The ideal adsorption energy, charge transfer and recovery time ensured that the MoS2–GeSe monolayer had a good adsorption and desorption ability. This paper aimed to solve the matter of recycling sensors within agriculture. This research could provide the theoretical basis for the establishment of a potentially new generation of gas sensors for the monitoring of crop growth.


Introduction
In modern agriculture, greenhouses are often used to artificially change the growth environment of crops in order to cope with random and changeable weather conditions and ensure the diversity of seasonal agricultural products [1][2][3]. In the growth process of plants, carbon is an important organic matter, mainly in the form of CO 2 [4,5]. CO 2 is a necessary material for plants, which plays a very important role in photosynthesis [6][7][8]. It directly affects the normal growth of plants and the abundance of nutrients within plants. In addition, the concentration of CO 2 can change the biological and abiotic conditions in soil to some extent [9,10]. Increases in the emissions of potent greenhouse gases from soil under increasing atmospheric CO 2 indirectly affect the growth of crops. Moreover, the demand for CO 2 is different in different growth stages of crops. For example, in this study, demand for CO 2 reached the maximum value of 1000-1200 ppm before sunrise and then decreased to about 100 ppm 2.5-3 h after sunrise, which was only about 30% of the atmospheric concentration. It did not begin to rise again until 2 h after sunrise and returned to the atmospheric level around 4 p.m. The concentration of carbon dioxide that is required for vegetables to grow is generally 1000~1500 ppm, so CO 2 deficiency in a greenhouse is quite serious and has become an important factor that affects the yield of greenhouse vegetables. Therefore, the reasonable allocation and control of CO 2 concentration in greenhouses are critical for the optimal growth of crops. Zhang et al. have conducted a lot of research on the study of density functional theory (DFT) [11][12][13][14][15][16]. They studied the microscopic interactions between two-dimensional (2D) materials and gas molecules to explore the application of these materials as gas adsorbents and sensors, such as the use of an Audoped graphene monolayer as a gas sensor [13,16]. Other relevant work has proved that the simulation calculation is very consistent with experimental results. Based on DFT, we
A 4 × 4 × 1 GeSe monolayer supercell was constructed and optimized using the highest precision setting in the material studio software. The k-point samples of the Monkhorst-Pack grid were sampled at 6 × 6 × 1 for geometry optimization and 10 × 10 × 1 for electronic optimization. The size of the vacuum space between two adjacent sheets was set at the value of 15 Å in order to avoid interactions between periodic images. A self-consistent loop energy of 10 −6 Ha, a global orbital cut-off radius of 5.0 Å and a smearing of 0.005 Ha were applied. In order to study the role of MoS 2 doping in the adsorption of CO 2 to the GeSe monolayer, we compared the most stable configurations of CO 2 that were adsorbed to the MoS 2 -GeSe monolayer. Our strategy was to calculate all typical adsorption structures and then the maximum adsorption energy would correspond to the most stable adsorption configuration.
where E MoS2-GeSe , E GeSe and E MoS2 represent the total energies of MoS 2 -GeSe, pure GeSe and MoS 2 units, respectively. E ads was calculated using Equation (2), as follows: where E CO 2 /MoS2-GeSe , E CO 2 and E MoS2-GeSe represent the energy of the total systems, CO 2 and the separated systems, respectively. The charge amount was obtained using the Milliken charge analysis and ∆Q was determined using Equation (3), as follows: where Q a and Q b represent the total charge of the gas molecules after and before adsorption, respectively.

Results
As shown in Figure 1d, CO 2 is a common compound in air that is composed of two oxygen atoms and a carbon atom, which are attached by a polar covalent bond. In Figure 1a,e, the ball-stick model structure of the pure GeSe monolayer is shown. GeSe monolayer are serrated black crystals that are similar to black phosphorus and belong to the orthogonal crystal system. After structural optimization, the Ge-Se bond lengths of the most stable GeSe monolayers were maintained at 2.641 and 2.544 Å. As shown in Figure 1b,f, the MoS 2 units were stably adsorbed to the GeSe monolayer surface and the E b value reached up to −5.994 eV, which indicated that a stable doping structure was formed with great thermal stability. In addition, the introduction of MoS 2 units caused the Ge atoms in the GeSe monolayer to shift upward, which was produced by the strong interaction between the MoS 2 units and the GeSe monolayer. As shown in Figure 1c,g, a stable structure with the maximum E ads value was finally selected after the calculation of eight different sites. For the metal oxide materials, we found the four intrinsic structures of ZnO, SnO 2 , TiO 2 and WO 3 through a literature review and compared and explored the gas adsorption characteristics of CO 2 to these common metal oxides. The specific parameters are shown in Table 1. It was obvious that the E ads values of several metal oxides were much lower than those of the MoS 2 -GeSe monolayer (−0.955 eV) using DFT calculations, which indicated that the MoS 2 -GeSe monolayer had excellent response characteristics for CO 2 detection. For pure ZnO, WO 3 had some positive E ads values, which indicated that the reaction could not be carried out spontaneously nor could that material be used as a sensor. adsorbed to the MoS2-GeSe monolayer. Our strategy was to calculate all typical adsorption structures and then the maximum adsorption energy would correspond to the most stable adsorption configuration. Eb was determined using Equation (1), as follows: where EMoS2-GeSe, EGeSe and EMoS2 represent the total energies of MoS2-GeSe, pure GeSe and MoS2 units, respectively. Eads was calculated using Equation (2), as follows: where ECO2/MoS2-GeSe, ECO2 and EMoS2-GeSe represent the energy of the total systems, CO2 and the separated systems, respectively. The charge amount was obtained using the Milliken charge analysis and ΔQ was determined using Equation (3), as follows: where Qa and Qb represent the total charge of the gas molecules after and before adsorption, respectively.

Results
As shown in Figure 1d, CO2 is a common compound in air that is composed of two oxygen atoms and a carbon atom, which are attached by a polar covalent bond. In Figure  1a,e, the ball-stick model structure of the pure GeSe monolayer is shown. GeSe monolayer are serrated black crystals that are similar to black phosphorus and belong to the orthogonal crystal system. After structural optimization, the Ge-Se bond lengths of the most stable GeSe monolayers were maintained at 2.641 and 2.544 Å. As shown in Figure 1b,f, the MoS2 units were stably adsorbed to the GeSe monolayer surface and the Eb value reached up to −5.994 eV, which indicated that a stable doping structure was formed with great thermal stability. In addition, the introduction of MoS2 units caused the Ge atoms in the GeSe monolayer to shift upward, which was produced by the strong interaction between the MoS2 units and the GeSe monolayer. As shown in Figure 1c,g, a stable structure with the maximum Eads value was finally selected after the calculation of eight different sites. For the metal oxide materials, we found the four intrinsic structures of ZnO, SnO2, TiO2 and WO3 through a literature review and compared and explored the gas adsorption characteristics of CO2 to these common metal oxides. The specific parameters are shown in Table 1. It was obvious that the Eads values of several metal oxides were much lower than those of the MoS2-GeSe monolayer (−0.955 eV) using DFT calculations, which indicated that the MoS2-GeSe monolayer had excellent response characteristics for CO2 detection. For pure ZnO, WO3 had some positive Eads values, which indicated that the reaction could not be carried out spontaneously nor could that material be used as a sensor.  CO 2 0.22 [53] As shown in Figure 2a, the overall TDOS of the MoS 2 -GeSe monolayer shifted to the right after doping with MoS 2 units, especially near the Fermi level, which indicated that the conductivity of the GeSe monolayer decreased. In addition, a new peak appeared at −6 eV, which lays the foundation for the gas adsorption reaction. After the CO 2 adsorption by the MoS 2 -GeSe monolayer, the peak amplitudes at −8.5 eV, −4 eV, −3 eV and 1.7 eV in the TDOS of the system increased slightly, which depended on the small regional state coating. This reflected the direct weak interactions between the two-dimensional materials and gases. We performed a PDOS analysis of the MoS 2 -GeSe monolayer decomposition ( Figure 2b). Obviously, the adsorbed CO 2 , doping MoS 2 units and modified GeSe had very obvious overlapping peaks at −8 eV and−4 eV, which indicated the existence of multi-electron orbital hybridization. After adsorption, we also found that the gas molecules did not produce new hybrid orbitals but rather shifted left overall. Therefore, we could conclude that the doped metal oxide had a tiny effect on the charge redistribution of the gas. In other words, the adsorbed CO 2 could not redistribute the surface charge of the MoS 2 -GeSe monolayer or the electrical conductivity of the system, which revealed the corresponding response mechanisms of the gas sensor. As shown in Figure 2d, all elements showed high levels of hybridization and the orbitals were distributed between −10 and 2 eV, where Ge-p, Se-p, O-p and C-p orbitals contributed greatly to the charge recombination. The insertion of the p orbitals of O atoms into the MoS 2 units caused electronic transitions in the gas molecules near the Fermi level, which formed new coupled states. In short, this indicated that there was an interaction between CO 2 , the MoS 2 units and the GeSe monolayer; however, it was generally a weak interaction, which conformed to the application prospects of the MoS 2 -GeSe monolayer as a gas sensor.   [53] As shown in Figure 2a, the overall TDOS of the MoS2-GeSe monolayer shifted to the right after doping with MoS2 units, especially near the Fermi level, which indicated that the conductivity of the GeSe monolayer decreased. In addition, a new peak appeared at −6 eV, which lays the foundation for the gas adsorption reaction. After the CO2 adsorption by the MoS2-GeSe monolayer, the peak amplitudes at −8.5 eV, −4 eV, −3 eV and 1.7 eV in the TDOS of the system increased slightly, which depended on the small regional state coating. This reflected the direct weak interactions between the two-dimensional materials and gases. We performed a PDOS analysis of the MoS2-GeSe monolayer decomposition ( Figure 2b). Obviously, the adsorbed CO2, doping MoS2 units and modified GeSe had very obvious overlapping peaks at −8 eV and−4 eV, which indicated the existence of multielectron orbital hybridization. After adsorption, we also found that the gas molecules did not produce new hybrid orbitals but rather shifted left overall. Therefore, we could conclude that the doped metal oxide had a tiny effect on the charge redistribution of the gas. In other words, the adsorbed CO2 could not redistribute the surface charge of the MoS2-GeSe monolayer or the electrical conductivity of the system, which revealed the corresponding response mechanisms of the gas sensor. As shown in Figure 2d, all elements showed high levels of hybridization and the orbitals were distributed between −10 and 2 eV, where Ge-p, Se-p, O-p and C-p orbitals contributed greatly to the charge recombination. The insertion of the p orbitals of O atoms into the MoS2 units caused electronic transitions in the gas molecules near the Fermi level, which formed new coupled states. In short, this indicated that there was an interaction between CO2, the MoS2 units and the GeSe monolayer; however, it was generally a weak interaction, which conformed to the application prospects of the MoS2-GeSe monolayer as a gas sensor. In Figure 3a,d, the gain and loss of electron density are indicated by the blue and red regions, respectively. From the DCD distribution and ∆Q values, the gas molecules contributed 0.266 e to the material surface as electron donors and the electron density changed significantly between the different distribution regions. The electron density of the O atoms in CO 2 increased greatly, while the region around the C atom lost electrons. The analysis indicated that the C atom of the CO 2 molecule provided electrons not only to the two O atoms, but also to the O atom in the doped MoS 2 units. It was evident from this transfer that the adsorption reaction was both exothermic and self-generated, which guaranteed stable and efficient adsorption effects. Figure 3b,e represents the redistribution of the charge density of CDD after adsorption. The adsorbed CO 2 was charged positively overall and had purple areas around it, which indicated strong electron drainage. The separated charge build-up and exhaustion also confirmed the weak interaction between the gas and the adsorbent, with no new chemical bonds taking shape. As shown in Figure 3c,f, the ELF values were between −1 and 1, where −1 and 1 stand for very low electron density and completely limited electron density, respectively. It was found that stable O-Ge and Mo-Se bonds were formed between the MoS 2 units and GeSe monolayer, which further confirmed the stability of the doping structure. In addition, there was no significant regional fusion between the CO 2 and the MoS 2 -GeSe monolayer, which suggested that the adsorption reaction was physical adsorption and showed the uniqueness of the MoS 2 -GeSe monolayer as a gas sensor that allows for a faster response and recovery. In Figure 3a,d, the gain and loss of electron density are indicated by the blue and red regions, respectively. From the DCD distribution and ΔQ values, the gas molecules contributed 0.266 e to the material surface as electron donors and the electron density changed significantly between the different distribution regions. The electron density of the O atoms in CO2 increased greatly, while the region around the C atom lost electrons. The analysis indicated that the C atom of the CO2 molecule provided electrons not only to the two O atoms, but also to the O atom in the doped MoS2 units. It was evident from this transfer that the adsorption reaction was both exothermic and self-generated, which guaranteed stable and efficient adsorption effects. Figure 3b,e represents the redistribution of the charge density of CDD after adsorption. The adsorbed CO2 was charged positively overall and had purple areas around it, which indicated strong electron drainage. The separated charge build-up and exhaustion also confirmed the weak interaction between the gas and the adsorbent, with no new chemical bonds taking shape. As shown in Figure 3c,f, the ELF values were between −1 and 1, where −1 and 1 stand for very low electron density and completely limited electron density, respectively. It was found that stable O-Ge and Mo-Se bonds were formed between the MoS2 units and GeSe monolayer, which further confirmed the stability of the doping structure. In addition, there was no significant regional fusion between the CO2 and the MoS2-GeSe monolayer, which suggested that the adsorption reaction was physical adsorption and showed the uniqueness of the MoS2-GeSe monolayer as a gas sensor that allows for a faster response and recovery. The energy of electrons in solids cannot be continuously valued, but some discontinuous energy bands can be valued. To conduct electricity, there must be free electrons or holes. The energy of free electrons is called the conduction band and the energy of free holes is called the valence band. When the bound electrons become free electrons or holes, they must obtain enough energy from the valence band for the conduction band. The minimum energy is Eg. In short, Eg refers to the energy transfer between the lowest level of the conduction band and the highest level of the valence band [54][55][56]: The energy of electrons in solids cannot be continuously valued, but some discontinuous energy bands can be valued. To conduct electricity, there must be free electrons or holes. The energy of free electrons is called the conduction band and the energy of free holes is called the valence band. When the bound electrons become free electrons or holes, they must obtain enough energy from the valence band for the conduction band. The minimum energy is E g . In short, E g refers to the energy transfer between the lowest level of the conduction band and the highest level of the valence band [54][55][56]: where σ is the conductivity, k is the Boltzmann constant (1.38 × 10 −23 J/K), A is a fixed parameter and T is the temperature. The change in E g after an adsorption reaction can reflect the change in system conductivity and can further reflect the change in resistance value in practical applications. The E g value of the MoS 2 -GeSe monolayer decreased after CO 2 adsorption, which indicated that the conductivity of the GeSe-based monolayer increased and the resistance value decreased. As shown in Figure 4a,c, the doping with MoS 2 units reduced the conductivity of the GeSe monolayer and increased the E g value from 0.725 eV to 0.786 eV, which indicated that the MoS 2 -GeSe monolayer maintained the excellent semiconductor characteristics and gas response characteristics of pure GeSe. However, the E g value of the pure GeSe monolayer after CO 2 adsorption changed very little (from 0.725 eV to 0.753 eV), which indicated that CO 2 had little effect on the electrical conductivity of the system (Figure 4b). In actual detection, it could be impossible to detect the resistance change of GeSe-based monolayer gas sensors using high precision instruments. The adsorption of CO 2 reduced the E g value of the MoS 2 -GeSe monolayer to 0.667 eV, which indicated that the difficulty of electron transition from the top of the valence band to the bottom of the conduction band was reduced (Figure 4d). This helped us to observe the resistance change that was caused by the gas concentration more clearly.
In contrast to the pure GeSe monolayer, the MoS 2 -GeSe monolayer could be used as a gas sensor for CO 2 detection.
where σ is the conductivity, k is the Boltzmann constant (1.38 × 10 −23 J/K), A is a fixed parameter and T is the temperature. The change in Eg after an adsorption reaction can reflect the change in system conductivity and can further reflect the change in resistance value in practical applications. The Eg value of the MoS2-GeSe monolayer decreased after CO2 adsorption, which indicated that the conductivity of the GeSe-based monolayer increased and the resistance value decreased. As shown in Figure 4a,c, the doping with MoS2 units reduced the conductivity of the GeSe monolayer and increased the Eg value from 0.725 eV to 0.786 eV, which indicated that the MoS2-GeSe monolayer maintained the excellent semiconductor characteristics and gas response characteristics of pure GeSe. However, the Eg value of the pure GeSe monolayer after CO2 adsorption changed very little (from 0.725 eV to 0.753 eV), which indicated that CO2 had little effect on the electrical conductivity of the system (Figure 4b). In actual detection, it could be impossible to detect the resistance change of GeSe-based monolayer gas sensors using high precision instruments. The adsorption of CO2 reduced the Eg value of the MoS2-GeSe monolayer to 0.667 eV, which indicated that the difficulty of electron transition from the top of the valence band to the bottom of the conduction band was reduced (Figure 4d). This helped us to observe the resistance change that was caused by the gas concentration more clearly. In contrast to the pure GeSe monolayer, the MoS2-GeSe monolayer could be used as a gas sensor for CO2 detection. WFs are a property of particular material surfaces, which indicate the minimum energy that is required for the surface to immediately move electrons to a point within a vacuum [57,58]. In gas interaction, this determines the aligned contact barrier between the gas molecules and specific surfaces. As shown in Figure 5a, the doping with MoS2 units increased the WFs of the GeSe monolayer to 4.653 eV, which hindered the electronic affinity of the system. This phenomenon further proved that the electronic transition of the GeSe monolayer was more difficult and that the chemical properties of the system were more stable after doping with MoS2 units. However, the introduction of CO2 reduced the WFs of the system to 4.544 eV, which indicated that after the CO2 was adsorbed, the difficulty of electron spillover from the vacuum reduced. This further guided the adsorption of the MoS2-GeSe monolayer to CO2. It also reflected the electrostatic attraction between the gas molecules and the adsorption system. It was this simultaneous attraction that led to the interaction between the CO2 and the MoS2-GeSe monolayer. In conclusion, gas adsorption could cause changes in WFs, which also proved the possibility of using a Kelvin probe microscope to detect WFs in the field of gas sensing. Zhang et al. also studied the WFs are a property of particular material surfaces, which indicate the minimum energy that is required for the surface to immediately move electrons to a point within a vacuum [57,58]. In gas interaction, this determines the aligned contact barrier between the gas molecules and specific surfaces. As shown in Figure 5a, the doping with MoS 2 units increased the WFs of the GeSe monolayer to 4.653 eV, which hindered the electronic affinity of the system. This phenomenon further proved that the electronic transition of the GeSe monolayer was more difficult and that the chemical properties of the system were more stable after doping with MoS 2 units. However, the introduction of CO 2 reduced the WFs of the system to 4.544 eV, which indicated that after the CO 2 was adsorbed, the difficulty of electron spillover from the vacuum reduced. This further guided the adsorption of the MoS 2 -GeSe monolayer to CO 2 . It also reflected the electrostatic attraction between the gas molecules and the adsorption system. It was this simultaneous attraction that led to the interaction between the CO 2 and the MoS 2 -GeSe monolayer. In conclusion, gas adsorption could cause changes in WFs, which also proved the possibility of using a Kelvin probe microscope to detect WFs in the field of gas sensing. Zhang et al. also studied the process of gas adsorption using WFs [33,53,[55][56][57][58][59][60][61]. Figure 5b presents the theoretical recovery time of CO 2 at different temperatures that were calculated. The recovery times of sensing materials are closely related to the exposed ambient temperature. The calculation formula that we used was as follows [62,63]: where k is the Boltzmann constant, τ is the recovery time and v 0 is the attempt frequency (10 12 s −1 ). The thermal environment offers more energy for the desorption process and fosters the further diffusion of molecules. At 298 K, the recovery time of CO 2 was 14400 s, which caused the sensor to become unrecyclable. At 398 K, the desorption time of CO 2 decreased to 1.25 s, which further illustrated the excellent performance of the MoS 2 -GeSe monolayer as a gas sensor. We inferred that 398 K could be used as the optimum temperature for the sensor, which would make it possible to apply the MoS 2 -GeSe monolayer in agriculture.
process of gas adsorption using WFs [33,53,[55][56][57][58][59][60][61]. Figure 5b presents the theoretical recovery time of CO2 at different temperatures that were calculated. The recovery times of sensing materials are closely related to the exposed ambient temperature. The calculation formula that we used was as follows [62,63]: where k is the Boltzmann constant, τ is the recovery time and v0 is the attempt frequency (10 12 s −1 ). The thermal environment offers more energy for the desorption process and fosters the further diffusion of molecules. At 298 K, the recovery time of CO2 was 14400 s, which caused the sensor to become unrecyclable. At 398 K, the desorption time of CO2 decreased to 1.25 s, which further illustrated the excellent performance of the MoS2-GeSe monolayer as a gas sensor. We inferred that 398 K could be used as the optimum temperature for the sensor, which would make it possible to apply the MoS2-GeSe monolayer in agriculture.

Conclusions
In this work, the sensing properties of an MoS2-GeSe monolayer for CO2 detection were systematically studied based on DFT. Compared to the pure GeSe monolayer, the MoS2-GeSe monolayer had a higher sensitivity to CO2 and the variation in Eg increased by 661% (0.119/0.018 × 100%). In practical applications, the MoS2-GeSe monolayer would have a greater change in resistance values as a chemical resistance sensor than during testing. The main parameters are shown in Table 2. The large Eb value (−5.994 eV) indicated that the MoS2 units and GeSe monolayer had excellent binding abilities. The thermal stability of the doped system was robust enough and the sensor could deal with changes in the ambient operating temperature. As can be seen from the DOS, the enriched orbital hybridization demonstrated the interaction between the CO2 and the active atomic groups. The ideal Eads (−0.955 eV), ΔQ (0.266 e) and recovery time (1.25 s in 398 K) not only ensured the detection ability of the MoS2-GeSe monolayer as a sensor, but also ensured its excellent recyclability. This is an advantage that most two-dimensional material gas sensors do not have. Based on this theoretical investigation, we will proceed to test flexible patch gas sensors for the monitoring of crop growth.

Conclusions
In this work, the sensing properties of an MoS 2 -GeSe monolayer for CO 2 detection were systematically studied based on DFT. Compared to the pure GeSe monolayer, the MoS 2 -GeSe monolayer had a higher sensitivity to CO 2 and the variation in E g increased by 661% (0.119/0.018 × 100%). In practical applications, the MoS 2 -GeSe monolayer would have a greater change in resistance values as a chemical resistance sensor than during testing. The main parameters are shown in Table 2. The large E b value (−5.994 eV) indicated that the MoS 2 units and GeSe monolayer had excellent binding abilities. The thermal stability of the doped system was robust enough and the sensor could deal with changes in the ambient operating temperature. As can be seen from the DOS, the enriched orbital hybridization demonstrated the interaction between the CO 2 and the active atomic groups. The ideal E ads (−0.955 eV), ∆Q (0.266 e) and recovery time (1.25 s in 398 K) not only ensured the detection ability of the MoS 2 -GeSe monolayer as a sensor, but also ensured its excellent recyclability. This is an advantage that most two-dimensional material gas sensors do not have. Based on this theoretical investigation, we will proceed to test flexible patch gas sensors for the monitoring of crop growth.