The Effect of the Metal Impurities on the Stability, Chemical, and Sensing Properties of MoSe₂ Surfaces

Abstract

In this study, we present a comprehensive theoretical analysis of modifications in the physical and chemical properties of MoSe₂ upon the introduction of substitutional transition metal impurities, specifically, Ti, V, Cr, Fe, Co, Ni, Cu, W, Pd, and Pt. Wet systematically calculated the adsorption enthalpies for various representative analytes, including O₂, H₂, CO, CO₂, H₂O, NO₂, formaldehyde, and ethanol, and further evaluated their free energies across a range of temperatures. By employing the formula for probabilities, we accounted for the competition among molecules for active adsorption sites during simultaneous adsorption events. Our findings underscore the importance of integrating temperature effects and competitive adsorption dynamics to predict the performance of highly selective sensors accurately. Additionally, we investigated the influence of temperature and analyte concentration on sensor performance by analyzing the saturation of active sites for specific scenarios using Langmuir sorption theory. Building on our calculated adsorption energies, we screened the catalytic potential of doped MoSe₂ for CO₂-to-methanol conversion reactions. This paper also examines the correlations between the electronic structure of active sites and their associated sensing and catalytic capabilities, offering insights that can inform the design of advanced materials for sensors and catalytic applications.

1. Introduction

Molybdenum diselenide (MoSe₂) is a promising member of the emerging class of two-dimensional (2D) materials [1]. One of the key advantages of MoSe₂ is its exceptional chemical stability [2,3], especially in contrast to other diselenides, such as tin diselenide (SnSe₂) [4]. Among the various potential applications for this class of materials, chemical sensing is particularly significant [5]. Over the past few decades, extensive research has focused on exploring the sensing properties of MoSe₂, with results compiled in various comprehensive reviews [6,7,8,9,10].

Numerous studies have identified MoSe₂ as an effective sensor for nitrogen dioxide (NO₂) [11,12,13,14,15,16,17]. Additionally, MoSe₂ has shown promise as a sensor for various analytes, including water [18], hydrogen sulfide (H₂S) [19], methanol and ethanol [20], and carbon monoxide (CO) [21]. Notably, many investigations have reported that the sensing capabilities of MoSe₂ can be significantly enhanced through the incorporation of substitutional impurities. For instance, doping with platinum (Pt) or palladium (Pd) converts MoSe₂ into an efficient sensor for ammonia (NH₃) [22,23,24], while aluminum doping enhances its sensitivity to ketones [25]. Another approach to modifying the sensing properties of MoSe₂ involves intercalating oxygen [26].

These experimental findings have motivated density functional theory (DFT)-based simulations to investigate the effects of substitutional impurities on the sensing behavior of MoSe₂. Theoretical studies have predicted that the introduction of dopants such as copper (Cu), platinum (Pt), and gold (Au) can enhance sensitivity to sulfur dioxide (SO₂) [27,28,29], while nickel (Ni) doping can improve response to xylene and methanol [30], and palladium (Pd) doping can enhance sensitivity to hydrogen and CO [31]. Research by Jing Wang et al. has further demonstrated the ability to switch sensing properties through the doping of various 3d transition metals [32].

Traditionally, theoretical evaluations of sensing properties involve calculating the enthalpy of adsorption for various molecules and assessing charge transfer, as well as the impact of adsorbed species on the electronic structure of the system. However, this standard approach has certain limitations. Firstly, neglecting entropy contributions results in an inaccurate estimate of the temperature range for adsorption events. Furthermore, it overlooks the competition for active sites, which is particularly relevant in real air, where multiple gases can exhibit similar or identical adsorption enthalpies and thus potentially occupy the same active sites as the target analyte. Recent studies have shown that incorporating entropy considerations and competition for active sites yields better alignment with experimental results and improves predictive accuracy [33]. Accordingly, advancements are needed in theoretical approaches to predict sensing properties.

In this work, we present the results of simulations examining the effects of doping MoSe₂ with substitutional impurities such as various 3d transition metals (Ti, V, Cr, Fe, Co, Ni, Cu) and other transition metals (W, Pd, Pt) on Se-vacancy (v_Se) formation energies and the electronic structure and adsorption energies of different gases. Substitutional manganese and zinc were excluded due to difficulties in fabricating MoSe₂ doped with these species [34,35] and a lack of comprehensive experimental data on the properties of these systems. Tungsten has chosen as a dopant because it belongs to the same periodic table group as molybdenum, and tungsten dichalcogenides exhibit similar physical and chemical properties as MoSe₂ [36,37]. Platinum and palladium were considered to investigate how the ionic radii difference between impurities and molybdenum atoms [38] affects the electronic structure and sensing properties. We employed the Maxwell–Boltzmann formula for each substrate to calculate the adsorption probabilities, followed by assessments of active site saturation in air for the most significant cases. Additionally, the surfaces were screened for suitability in catalyzing CO₂ to methanol conversion at temperatures below 100 °C, to gain a deeper understanding of the reactive properties of doped MoSe₂.

2. Computational Method

The atomic structure and energetics of various configurations were studied using DFT with the QUANTUM-ESPRESSO code [39] and the GGA-PBE [40] method, taking into account van der Waals forces corrections [41]. For all calculations, we used ultrasoft pseudopotentials [42]. The values of energy cutoffs of 35 Ry and 400 Ry for the plane-wave expansion of the wave functions and the charge density, respectively.

The standard formula calculated physisorption enthalpies, as follows:

ΔH_phys = [E_host+mol − (E_host + E_mol)],

(1)

where E_host is the total energy of a pristine surface, and E_mol is the energy of the single molecules of selected species in the empty box. In the case of water adsorption, we considered only the gaseous phase. For the case of analyte (A) physisorption, we also evaluated the differential Gibbs free energy via the following formula:

ΔG_A = ΔH − TΔS

(2)

where T is the temperature and ΔS is the change of entropy of the adsorbed molecule, which was estimated considering the gas → liquid transition by the standard formula, as follows:

$$∆S={\displaystyle \frac{{∆H}_{vaporization}}{T}}$$

(3)

where ΔH_vaporization is the measured enthalpy of vaporization at the temperature T.

The Langmuir constant is defined as the equilibrium constant for adsorption on the active site, as described by the Maxwell–Boltzmann distribution. These constants were calculated using the standard formula, as follows:

$${K_A=e}^{-{\displaystyle \frac{{∆G}_A}{RT}}}$$

(4)

where R is the gas constant. The adsorption probabilities in the presence of other gases are defined as the Boltzmann distribution of the analyte divided by the sum of the Maxwell–Boltzmann distributions of all considered molecules.

$$P_A={\displaystyle \frac{K_A}{\sum _i{K}_i}}$$

(5)

Saturation of active sites by analyte with concentration C_A in the presence of oxygen was calculated using the following formula:

$$q_A={\displaystyle \frac{K_A{C}_A}{1+K_A{C}_A+K_{O_2}{C}_{O_2}}}$$

(6)

where K_O2 and C_O2 are the Langmuir constant and the concentration of oxygen.

The energies of CO₂ conversion are defined as the differences between the total energies of products and reactants. Since the hydrogenation of CO₂ is realized by hydrogen atoms from the solution, each reaction step consists of a gradual barrierless approach of the proton to the active site.

3. Results and Discussions

3.1. Lattice Stability and Electronic Structure

For modeling MoSe₂, a slab constructed from a 3 × 3 × 2 unit cell was used. This slab consisted of four layers. The top two layers are shown in Figure 1. To estimate the effect of the substitutional impurities on the lattice stability, we compared the energy cost of the formation of a vacancy in pure MoSe₂ (E0_form(v_Se)) and in the vicinity of the impurity (E_form(v_Se)). Calculated numbers are shown in the right-hand column of

Table 1. The Se-vacancy (v_Se) energy formation changes after incorporating the impurities, enthalpies, and the Gibbs free energies for physical adsorption on pure MoSe₂ and defect sites at different gas temperatures.

Dopant	Efom(vSe) − E0form(vSe), kJ mol−1	Analyte	ΔH, kJ mol−1	ΔG, kJ mol−1
				400 °C	800 °C
Pure MoSe2	0.0	O2	−127.4	−94.9	−10.7
		CO2	−157.1	−139.8	−93.4
		H2O	−35.1	−2.5	+114.6
		CO	−27.3	−7.9	+42.5
		H2	−67.9	−59.7	−38.4
		NO2	−195.2	−162.1	−76.0
		CH2O	−43.0	−12.3	+70.1
		C2H5OH	−150.1	−117.3	−32.0
Ti	+46.5	O2	−132.8	−100.4	−16.1
		CO2	−241.2	−223.9	−177.5
		H2O	−74.3	−41.7	+75.4
		CO	−22.2	−2.8	+47.6
		H2	−194.8	−186.6	−165.3
		NO2	−282.8	−249.7	−163.6
		CH2O	−55.1	−24.4	+58.0
		C2H5OH	−79.5	−46.7	+38.6
V	+109.4	O2	−126.8	−94.4	−10.1
		CO2	−48.0	−30.7	+15.7
		H2O	−43.1	−10.5	+106.6
		CO	−20.0	−0.6	+49.8
		H2	−20.8	−12.6	+8.7
		NO2	−74.7	−41.6	+44.5
		CH2O	−55.4	−24.7	+57.7
		C2H5OH	−155.9	−123.1	−37.8
Cr	+149.1	O2	−117.0	−84.6	−0.3
		CO2	−176.4	−159.1	−112.7
		H2O	−28.4	+4.2	+121.3
		CO	−25.9	−6.5	+43.9
		H2	−21.5	−13.3	+8.0
		NO2	−81.3	−48.2	+37.9
		CH2O	−46.7	−16.0	+66.4
		C2H5OH	−230.4	−197.6	−112.3
Fe	+88.9	O2	−72.0	−39.6	+44.7
		CO2	−39.9	−22.6	+23.8
		H2O	−33.0	−0.4	+116.7
		CO	−42.9	−23.5	+26.9
		H2	−37.6	−29.4	−8.1
		NO2	−73.2	−40.1	+46.0
		CH2O	−38.7	−8.0	+74.4
		C2H5OH	−43.0	−10.2	+75.1
Co	+150.1	O2	−31.9	+0.5	+84.8
		CO2	−38.6	−21.3	+25.1
		H2O	−23.0	+9.6	+126.7
		CO	−28.5	−9.1	+41.3
		H2	−29.5	−21.3	0.0
		NO2	−63.6	−30.5	+55.6
		CH2O	−38.9	−8.2	+74.2
		C2H5OH	−60.7	−27.9	+57.4
Ni	+220.3	O2	−166.2	−133.8	−49.5
		CO2	−43.8	−26.5	+19.9
		H2O	−32.5	+0.1	+117.2
		CO	−56.1	−36.7	+13.7
		H2	−61.5	−53.3	−32.0
		NO2	−153.9	−120.8	−34.7
		CH2O	−255.5	−224.8	−142.4
		C2H5OH	−271.0	−238.2	−152.9
Cu	+166.7	O2	−28.5	+3.9	+88.2
		CO2	−89.3	−72.0	−25.6
		H2O	−67.9	−35.3	+81.8
		CO	−20.9	−1.5	+48.9
		H2	−64.1	−55.9	−34.6
		NO2	−113.8	−80.7	+5.4
		CH2O	−30.3	+0.4	+82.8
		C2H5OH	−125.7	−92.9	−7.6
W	+20.1	O2	−119.3	−86.9	−2.6
		CO2	−51.7	−34.4	+12.0
		H2O	−26.8	+5.8	+122.9
		CO	−23.8	−4.4	+46.0
		H2	−38.2	−30.0	−8.7
		NO2	−79.7	−46.6	+39.5
		CH2O	−46.9	−16.2	+66.2
		C2H5OH	−195.7	−162.9	−77.6
Pd	+190.6	O2	−43.6	−11.2	+73.1
		CO2	−95.4	−78.1	−31.7
		H2O	−64.7	−32.1	+85.0
		CO	−247.1	−227.7	−177.3
		H2	−259.8	−251.6	−230.3
		NO2	−113.1	−80.0	+6.1
		CH2O	−189.6	−158.9	−76.5
		C2H5OH	−110.7	−77.9	+7.4
Pt	+206.8	O2	−147.1	−114.7	−30.4
		CO2	−222.3	−205.0	−158.6
		H2O	−93.7	−61.1	+56.0
		CO	−50.8	−31.4	+19.0
		H2	−258.1	−249.9	−228.6
		NO2	−106.5	−73.4	+12.7
		CH2O	−283.5	−252.8	−170.4
		C2H5OH	−138.4	−105.6	−20.3

. These results demonstrate that incorporating the impurities increases the energy required to form a selenium vacancy. Nickel and platinum demonstrated the most significant stabilizing effect on the MoSe₂ lattice, corresponding to an increase in the energy cost of vacancy formation of 220.3 kJ mol⁻¹ and 206.8 kJ mol⁻¹, respectively. The causes of this improvement in stability were the local distortions of the lattice near the impurity (see Figure 1b). The formation of the vacancy led to further increases in lattice distortions, thereby raising the energy cost of the defect. Thus, we excluded consideration of the combination of substitutional defects and anionic vacancies from further consideration.

The effects of embedded substitutional impurities on electronic structures can be divided into three groups. The first group consists of Ti, V, Cr, and W. Incorporation of tungsten atom does not lead to visible bandgap changes, vanadium impurity corresponds with a shift up of the valence band maximum, titanium and chromium defect shifts down the conduction band (see Figure 2a, since Ti⁴⁺ have almost zero occupancy of the valence band, we omitted its on these picture). Thus, Cr and V dopants can be used to manipulate the position of the energy levels in MoSe₂, which is essential for making efficient photo-catalysts.

Cu, Pd, and Pt impurities corresponded with the appearance of the defect states inside the bandgap (see Figure 2b). States near the Fermi level or on the Fermi level are typically associated with high catalytic activity in materials. This is discussed in the conclusions of our study. The third group consisted of iron, cobalt, and nickel. Incorporation of these impurities corresponded with a half-metallic spin-polarized electronic structure, corresponding with the absence of the states on Fermi level only for one spin, shown in Figure 2c. Magnetic moments on Cr, Fe, Co, and Ni impurities were 0.11, 0.59, 0.48, and 0.38 μ_B, respectively. These relatively small values are attributable to the significant local distortion effect in systems with substantial entangled 3d orbitals of metal and 4p orbitals of selenium, as demonstrated for VSe₂ [43]. Thus, these impurities can be used for MoSe₂-based spintronic devices [44]. The difference in the electronic structure of these impurities can be attributed to the variations in the typical oxidation states of the impurities. For Ti, V, Cr, and W, the 4+ oxidation state is usual in similar layered compounds [45,46,47]. Note that all these materials form stable 2D systems with atomic structures similar to those of MoSe₂ [48,49,50]. On the contrary, for other impurities, this oxidation state is unusual, even if these metals participate in the formation of dichalcogenides [51]. Since the entire MoSe₂ bath defines the oxidation state of Se atoms, the impurities also tend to have a 4+ oxidation state. This mismatch leads to the appearance of some states inside the bandgap on the Fermi level.

3.2. Sensing Properties

The adsorption of different gases was simulated to evaluate the effect of impurities on the sensing properties of MoSe₂ in real-life conditions, where water, hydrogen, various volatile organic compounds (VOCs), carbon monoxide, and carbon dioxide compete for adsorption on the active site. Ethanol (C₂H₅OH) and formaldehyde (CH₂O) were chosen to represent the most common families of volatile organic compounds (VOCs). The calculations demonstrate the favorability of the adsorption of the molecules on metal sites (see inset in Figure 1a). To illustrate the dramatic effect of temperature on adsorption patterns, relatively high temperatures, usually discussed for oxide-based sensors, were considered [51,52]. Calculated enthalpies of the physical adsorption (ΔH), summarized in Table 1, demonstrate that the dopants affected the energetics of the adsorption only quantitatively, without a significant change in the adsorption patterns. However, considering the contribution from entropy using Equation (2), these quantitative changes were transformed into qualitative changes even at 400 °C. A further increase in temperature corresponded to a switch in the sign of the free energies for most analytes. The different effects of temperature on the energetics of adsorption were caused by the different values of the enthalpy of vaporization in Equation (3). Larger molecules usually have higher values of vaporization enthalpy; therefore, the favorability of stable adsorption of larger molecules decreases more rapidly.

The Maxwell–Boltzmann probability of the analyte’s adsorption on active sites in the presence of all other competitors listed in Figure 3 was calculated using Equation (5). The results of the calculations are shown in Figure 3. These results demonstrate the outstanding sensitivity of pure MoSe₂ to NO₂. Doping by V, Cr, Cu, and W turns MoSe₂ into a highly selective sensor for ethanol. Doping with Ni makes MoSe₂ a good material for small organic molecules. In contrast, doping with Ti, Fe, Co, Pd, and Pt worsens the selectivity of the adsorption, since different molecules can be adsorbed at this temperature with a similar probability of adsorption. Thus, considering the contribution from temperature and competitive adsorption at the given temperature is essential for adequately describing the sensing properties of any materials. This conclusion suggests that collecting a larger number of experimental and theoretical results for further application of machine learning-based methods could be a solution for revealing the connection between sensing-related properties of impurities and features in electronic structure, as in recent works [53,54]. On the other hand, checking the possible effect of other analytes that can be present in real air is an integral part of the experimental evaluation of sensing properties. The comparison of the adsorption probabilities with the electronic structure patterns in Figure 2 demonstrates the absence of any connection between the electronic structure and adsorptive properties. For example, doping with tungsten does not lead to visible changes in the electronic structure, but it alters the selectivity from NO₂ to ethanol. Another example is the difference in the absorption selectivity for the groups shown in Figure 2b (Cu, Pd, and Pt) or Figure 2c (Fe, Co, and Ni). Thus, the large magnitude of the adsorption enthalpy (at 0 K temperature) obtained in DFT-based calculations does not correspond with the stable adsorption at working temperatures in the presence of other molecules in actual air.

The next step after recognizing the highly selective sensors was the standard check of the influence of the adsorption on the electronic structure of the substrate. For this purpose, the changes in the total densities of states were visualized (see Figure 4a,b). Note that in contrast to the partial densities of states shown in Figure 2, in this case, the total density of states for the entire supercell of more than a hundred atoms is depicted. As can be seen, the adsorption of analytes leads to minor, however visible, changes in the electronic structure. The charge transfer of 0.23 e⁻ from the substrate to NO₂ depicted in Figure 4c corresponded with the appearance of distinct states below the Fermi level (see Figure 4a). This significant charge transfer led to a decrease in the magnetic moment of Cr, Fe, Co, and Ni impurities, which were 0.09, 0.19, 0.16, and 0.13 μ_B, respectively. Thus, magnetic impurities in the MoSe₂ matrix can be utilized to create highly sensitive sensors. The changes in the electronic structure above the Fermi level caused by the adsorption of ethanol on Cu-doped MoSe₂ (Figure 4b) corresponded with the transfer of 0.41 e⁻ from analyte to substrate, as shown in Figure 4d. The relatively large charge transferred in both discussed cases caused the formation of robust bonds between the analytes and substrate (for more details about relationships between chemical bonds and charge transfer, see work by G.V. Pushkarev et al. [55]). Note that in both cases, the redistribution of charge density does not localize at the adsorption site but is distributed over the entire surface layer. Thus, pure MoSe₂ and Cu-doped MoSe₂ can be considered efficient sensors for NO₂ and ethanol, respectively. The reported number of transferred electrons can be used to calculate the change in resistivity for a given concentration of impurities and charge density in the samples.

The final step in evaluating the sensing properties was to estimate the effect of temperature and concentration on sensing efficiency. For this purpose, we calculated the saturation of active sites as a function of temperature and concentration using Equation (6). Since we did not know the concentration of other gases in the actual air, we considered only the competition between oxygen and analytes. Therefore, the results presented in Figure 4e,f can be regarded as estimations only. However, based on these results, we can propose that pure MoSe₂ is a highly efficient NO₂ sensor at temperatures below 100 °C, and Cu-doped MoSe₂ can detect even trace amounts of NO₂ up to 200 °C. These results agree with experimentally observed high performance of MoSe₂-based sensors at room temperature above [1,2,3,4]. However, some works have reported efficient sensing up to 200 °C [18].

3.3. Catalytic Properties

The firm adhesion of the molecules to the substrate is also essential for catalysis. Recent works (see, for example, the paper by X. Ren et al., [56]) demonstrate the high potential of MoSe₂-based structures for electrochemical applications. In contrast to gas sensing in actual air, a mix of reactants consists only of a few species (usually solvents and reactants). Thus, contention for the active sites occurs only between chemical species and molecules in solvents. For the reaction in liquid media, the contribution from entropy is minor (see Section 2). Therefore, the substrate can be proposed as potentially suitable for the catalysis only if the enthalpy of the adsorption of species is lower than the enthalpy of the adsorption of water. The results reported in Table 1 demonstrate that adsorption of CO₂ was substantially more favorable than water on the Ti, Cr, Cu, Pt, and Pd sites in the MoSe₂ matrix. Previous studies have highlighted the importance of the midgap states for CO₂ conversion [57,58]; therefore, we excluded Ti and V from the following consideration. Firm sticking of carbon monoxide with the Pd-site in MoSe₂ is associated with the possibility of CO poisoning of the active site. Thus, only Cu- and Pt-MoSe₂ substrates were selected for further analysis.

Figure 5 illustrates the reaction steps for the gradual hydrogenation of carbon dioxide in the liquid media by protons on both substrates. Both Cu- and Pt-doped substrates demonstrated a moderate (below 100 kJ mol⁻¹) energy cost for the rate-determining step. Considering semi-empirical relationships between calculated energies and reaction temperatures, this reaction requires temperatures below 100 °C [59]. The energy costs of the following steps do not exceed this value. Thus, both copper and platinum-doped MoSe₂ are prospective catalysts for CO₂ conversion.

Note that carbon dioxide adsorption on copper sites is significantly less favorable than on Pt. Thus, active site poisoning in Cu-doped MoSe₂ is much less probable than in Pt-doped MoSe₂. Considering this fact, along with the lower cost and higher abundance of copper, we propose Cu-doped MoSe₂ as a more promising catalyst, despite the slightly higher energy cost of one intermediate step. In contrast to the absence of a direct connection between electronic structure and sensing properties, we can establish a correlation between distinct states on the Fermi level (see Figure 2b) and exceptional catalytic performance. Note that chemical adsorption of the reaction intermediates occurred not on metal impurity but on one of the nearest Se sites for both copper and platinum impurities (see Figure 1b and insets in Figure 5). The cause of this peculiarity is the substitution of Mo⁴⁺ by Cu²⁺ or Pt²⁺. This leads to the reconstruction of Mo–Se bonds in the vicinity of the impurity, corresponding to the increase in chemical activity of the surrounding Se sites.

4. Conclusions

In summary, our systematic studies of MoSe₂ have elucidated significant relationships between the type of substitutional impurity and its electronic structure, sensing properties, and catalytic performance. Our findings indicate that the incorporation of various impurities affects the energy required for selenium vacancy formation and dramatically influences the material’s electronic characteristics. While tungsten doping maintains the bandgap, other transition metals, such as titanium, vanadium, and chromium, lead to its reduction. In contrast, palladium, platinum, and copper introduce defect states within the bandgap. Furthermore, introducing iron, cobalt, and nickel promotes a half-metallic and spin-polarized regime within MoSe₂. Based on these results, we can conclude that the influence of the dopants on the electronic structure depends on the combination of the oxidation states of the host and dopants, as well as the difference in their ionic radii.

The computed free energies of adsorption at 400 °C enable us to assess the material’s selectivity toward various analytes. Our results demonstrate the importance of considering the temperature for the proper description of competitive adsorption of analytes. Specifically, our results highlight that undoped MoSe₂ is particularly effective as a sensor for NO₂, while doping with vanadium, chromium, copper, and tungsten enhances selectivity for ethanol sensing. Conversely, embedding other impurities tends to detract from the sensing selectivity, underscoring the importance of controlled chemical composition for optimizing MoSe₂ sensing capabilities. Through targeted calculations, we explored the adsorption dynamics of NO₂ and ethanol on differently doped MoSe₂, providing a comprehensive understanding of active site saturation under varying concentrations of analytes. Our approach provides a framework for determining the temperature limits that affect sensing performance.

Moreover, this investigation revealed a complex interplay between electronic structure features and the material’s stability and selectivity for adsorption processes. This highlights the need for future experimental and theoretical investigations, particularly leveraging machine learning-based studies to uncover further connections between chemical structure and sensing capacity.

Our exploration into the catalytic potential of doped MoSe₂, particularly in the context of CO₂ conversion to methanol, identified Cu- and Pt-doped MoSe₂ as favorable candidates. These findings are attributed to their advantageous electronic structure and the reduced affinity of CO adsorption in copper centers, positioning Cu-doped MoSe₂ as a promising subject for subsequent experimental validation.

In conclusion, the insights garnered from this study pave the way for the rational design of MoSe₂-based materials for both sensing and catalytic applications, encouraging further exploration into the utility of various impurities for tailored performance outcomes.