One-step sputtering of MoSSe metastable phase as thin film and predicted thermodynamic stability by computational methods

We present the fabrication of a MoS2−xSex thin film from a co-sputtering process using MoS2 and MoSe2 commercial targets with 99.9% purity. The sputtering of the MoS2 and MoSe2 was carried out using a straight and low-cost magnetron radio frequency sputtering recipe to achieve a MoS2−xSex phase with x = 1 and sharp interface formation as confirmed by Raman spectroscopy, time-of-flight secondary ion mass spectroscopy, and cross-sectional scanning electron microscopy. The sulfur and selenium atoms prefer to distribute randomly at the octahedral geometry of molybdenum inside the MoS2−xSex thin film, indicated by a blue shift in the A1g and E1g vibrational modes at 355 cm−1 and 255 cm−1, respectively. This work is complemented by computing the thermodynamic stability of a MoS2−xSex phase whereby density functional theory up to a maximum selenium concentration of 33.33 at.% in both a Janus-like and random distribution. Although the Janus-like and the random structures are in the same metastable state, the Janus-like structure is hindered by an energy barrier below selenium concentrations of 8 at.%. This research highlights the potential of transition metal dichalcogenides in mixed phases and the need for further exploration employing low-energy, large-scale methods to improve the materials’ fabrication and target latent applications of such structures.

From the co-deposition process of MoS 2−x Se x phase by RF-sputtering, we obtained a thin film of approximately 200 nm sandwiched between ITO layers (being approx.150 nm thick).ToF-SIMS measurements confirm that our multilayer array consists of four distinctly different regions with sharp interface formation (Fig. 1).The presence of sharp interfaces between the top and bottom ITO layers with the MoS 2-x Se x phase in between is specified by the signals for In 2+ , InSnO + , and SnO + ions are distinguished and decay exponentially as the Mo + signals start to appear as depicted in Fig. 1a implying no intermixing between the ITO and the MoS 2−x Se x phase.In terms of selenium and sulfur, these measurements detected a homogeneous distribution within the fabricated sample, denoted by the signals from S − and Se − ions (Fig. 1b).Cross-section scanning electron microscopy (SEM) images present visual aspects of the ITO/MoS 2−x Se x /ITO array (Fig. 1c), confirming a well-defined and sharp interface formation as previously reported 10,25 and of comparable quality to those obtained by CVD methods 26 .
The Raman spectroscopy reveals that the film is homogenous and presents a degree of crystallinity (Fig. 1d).The sample presents vibrational modes at 402 cm −1 and 374.8 cm −1 , which correspond to the A 1g and E 1 2g vibrational modes of MoS 2 and suggest the presence of pure Mo-S vibration in some regions (Supplemental Information Fig. S1).However, signals at 355 cm −1 and 255 cm −1 correspond to the blue-shifted E 1 and A 1 modes, respectively, due to the formation of the MoSSe phase 27,28 .The blue shift in the A 1 vibrational mode is attributed to the out-of-plane change in symmetry caused by the integration of Se into the MoS 2 matrix during the co-deposit, while the shift in E 1 modes is related to the expansion of the lattice 29 and will be discussed in the next section.Deconvolution of the Raman spectra shows a thin film with traces of MoSe 2 signals, detectable by the A 1g Raman mode before 250 cm −1 alongside the shifted A 1 modes of the MoSSe phase (Supplemental Information Fig. S2).Grazing incidence x-ray diffraction (GIXD) spectroscopy matches with our Raman analysis on the degree of crystallinity of the sample, indicating the presence of peaks at 13°, 33°, 54° and 56° which corresponds to the Bragg's angle (2θ) of the planes (002), (101), (008) and (100) of 2H MoS 2 and 2H MoSe 2 [30][31][32] with P6 3 /mmc-194 space group, as shown in Fig. S3 from the Supplemental Information.The average crystal size was calculated to be ~ 75 nm (Table S1) in agreement with previous reports 2 by using the Scherrer equation.Future studies can exploit low-energy methods like laser annealing 33 , annealing in a controlled atmosphere 15,26,34 , or ink printing 35 , integrated into the one-step process to fabricate mixed MoS 2−x Se x thin films with improved crystal quality for applications in energy harvesting and optoelectronics devices.
Compared to previous reports on RF-sputtering of MoS 2 and MoSe 2 , we report the fabrication of a co-sputtered MoSSe phase under a low-temperature and high-power conditions for an smooth and homogenous deposit knowing that high RF power tends to facilitate the crystallization of MoS 2

32
. The listed works in Table 1 also show a similar trend, indicating that for an optimal deposition of MoS 2 by RF-sputtering, and probably most TMDC, enough kinetic energy is require to achieve the Mo +4 state and with this, an homogeneous crystallization 36 .This former argument is supported by our computational analysis presented in the next section.In our work, we increased the RF power for high deposition rates, but we also increased the working distance (see "Methods" section) to avoid the impact of high-energy sputtered material on the substrate.Furthermore, the literature also points out that deposition time has a crucial role in determining the crystallinity of the MoS 2 thin film, as short deposition times (< 120 s) normally yield amorphous MoS x phases or minuscule crystalline structures 30 .Then, the current course in physical vapor deposition (PVD) of TMDC, either by RF or direct current (DC), is to ease the fabrication of thin films avoiding subsequent high-energy annealing and post-sputtered treatments, aiming for one-step direct processes.

Structural optimization of the MoS 2−x Se x phase
To obtain insights into the structural stability and possible formation routes, we considered a MoS 2−x Se x phase model using density functional theory (DFT) calculations.Details are provided in the "Methods" section.MoS 2 crystallizes in the hexagonal P6 3 /mmc space group, and its structure consists of two MoS 2 sheets oriented parallel to the (001) plane.From these calculations, it is found that as the concentration of selenium increases, the initial MoS 2 lattice needs to expand in the random and Janus-like situations to accommodate the foreign ions and reach the structural optimization convergence.In the MoS 2−x Se x random phase at 33.33 at.% of selenium, a relative volume expansion (ΔV) of almost 7% is noted, while in the Janus-like situation the ΔV results 5.8% (see Table 2).The random phase at 33.33 at.% of selenium approaches more closely to the lattice parameter of 0.329 nm for 2H MoSe 2 40,41 .The volume variation indicates a distortion from the pure MoS 2 to the MoS 2−x Se x phase in the random situation and contrasts with the volume variation in the Janus-like, which is less abrupt.The volume variation in the Janus-like model is in line with previous reports, where a low lattice distortion occurs in the experimental Janus MoSSe monolayer 42 .
We attribute the increase in the volume of the supercell as a direct consequence of the change in the bonding distance between molybdenum and the chalcogen ions.In the random phase, the Mo-S bond distance (d Mo-S ), presents a contraction from 0.2406 nm to 0.2405 nm in most cases, a minor relative contraction of 0.04%.The Mo-Se bond distance (d Mo-Se ) shows contractions up to 0.08%, recalling the d Mo-Se in pure MoSe 2 resulted in 0.25322 nm.In the Janus distribution, an expansion of 0.17% of d Mo-S resulted after the structural optimization, while the d Mo-Se presents a contraction close to 0.1% concerning the pure 2H phase in both cases.However, as the d Mo-Se is significantly larger than d Mo-S , about 4%, we found a significant change in the resulting volume of   Table 2. Formation energy (E form /eV atom −1 ), substitutional energy (E subst /eV atom − ), substitutional energy with vacancy defects (E subst-V /eV atom −1 ), resulting lattice constant (a), lattice parameters (a, b, and c), volume (V), and relative volume expansion (ΔV) of the random and Janus-type MoS www.nature.com/scientificreports/ the MoS 2−x Se x .This is attributed to the rearrangement of charge from disturbed Mo d orbitals and p orbitals as reported previously 6,43 and possible electron density change in the 4d orbital due to lower electronegativity of Se over S.This uneven charge distribution may lead to local piezoelectricity as reported in other studies on Janus heterojunctions 7,44,45 .This change in the bonding distance agrees to previous reports 27 on Janus MoSSe systems and may be related to the resulting blue shift in the E 1 and A 1 Raman modes, which relates with our experimental observations presented earlier.The A 1 Raman modes represent the out-of-plane vibrational modes, perpendicular to the (001) basal plane, but selenium and sulfur have different bonding distances with molybdenum, affecting the symmetry along such planes and thus, the resulting vibrational mode.Similarly, the in-plane E 1 vibrational modes (alongside the atomic plane of the S-Mo-S distribution) will have differences compared to the pure 2H phase due to the included selenium.Even though in our case we focused on bulk structures, we expect a comparable explanation for monolayers and two-dimensional (2D) systems.

Thermodynamic stability of the MoS 2−x Se x phase
The formation energy (E form ) -for MoS 2−x Se x phase for various selenium concentrations-indicates that at low and high selenium concentrations, the lattice can favorably accommodate these ions in the random distribution.Figure 2 describes the atomistic model used for the DFT computations.For the Janus-type phase, at low selenium concentrations, the system reaches a low entropy state, and the process becomes endothermic.Under these circumstances, the lattice shows a lattice expansion situation quantified by a ΔV of 1.78%.However, after surpassing a selenium concentration of 8 at.% the Janus-like lattice reaches favorable conditions of formation, expressed by the exothermic values of E form .The clear presence of a threshold value for the formation of the Janus-like phase indicates that it would not be a feasible process unless enough energy is provided to overcome this energy barrier.On the other hand, the random phase shows an exothermic tendency, regardless of the selenium concentration.This analysis indicates that a co-deposit process would favor the formation of the random MoS 2−x Se x phase.
In terms of the site specific substitution of S by Se-denoted by the substitutional energy, E subst -we observe that for the random phase, the process is endothermic in all ranges of selenium concentrations up to 33.3 at.% having energy demands below 0.3 eV, which is comparable to the proposed activation energy for MoS 2 crystallization of 0.7 eV 46 .In this situation, as the concentration of selenium increases, the energy demand also increases as the systems need to accommodate foreign Se atoms into the initial matrix.For the Janus-like phase, the process is not favorable until a threshold value of 8 at.% selenium concentration is reached.In the latter situation, the presence of an energy barrier becomes noticeable; for concentrations below 8 at.% of selenium, the system has no tendency to exchange sulfur with selenium with E subst values above 16 eV.After surpassing this threshold value, the process becomes favorable and behaves similarly to the random phase, with substitution energies below 0.3 eV.This suggests that subsequent substitution is self-maintained in the Janus-type phase.Previous reports have hinted towards the critical stability of Janus MoSSe monolayers between 700 °C and 800 °C47,48 .Lu et al. 27 report a randomized MoSSe phase achieved above 600 °C, which supports our statement that MoSSe Janustype structures are indeed low entropy systems in a metastable state.The Janus-like and random phases reach almost an equal value of equilibrium energy (Fig. 3a), meaning both phases are located at the same low entropy point, and with this, the same metastable state located between pure 2H MoS 2 and pure 2H MoSe 2 as depicted in Fig. 3b.Nevertheless, the low entropy state of the Janus-type is determined by the highly ordered distribution of selenium and sulfur ions, related to the observed increase in E subst as mentioned before.
Sulfur vacancies might be present in sputtered MoS 2 49 , consequently on MoSe 2 as well, due to the higher vapor pressure of the chalcogen species 46 .To explore the role of S vacancies as lattice defects in the substitutional process of Se into MoS 2 , we computed the values of E subst but added an energy contribution for sulfur vacancy formation.The computed formation energy of a sulfur vacancy (E V ) is 7.09 eV, meaning it is an endothermic process and is within the range of defect formation in MoS 2 37,50,51 .The latter was added to the previous E subst to get an energy of substitution with an S vacancy formation (E subst-V ) and expressed as where n V is the number of S vacancies needed to accommodate the targeted Se concentration.The difference in our computed energy of vacancy formation with previous studies reporting the effect of S vacancies in MoS 2 is attributed to the addition of long-range dispersion correction in our study and the fact that we are using a crystal model rather than monolayer models.Besides, using crystal models helped to reach lower entropy phases compared to monolayers, i.e., lower surface energy.From this analysis, a clear difference between the Janus and the random MoS 2−x Se x phase becomes noticeable.For the Janus-like phase, the inclusion of Se atoms in a MoS 2 lattice with n V sulfur vacancies is not favorable to occur below a Se concentration of 8.3 at.%, after this tipping point the process follows a self-maintained behavior (Table 2) and the energy demand for creating S vacancies is compensated.For the random MoS 2-x Se x phase, the contrary occurs, as the energy demand remains in the www.nature.com/scientificreports/endothermic regime but is doable in terms of energy supply.This indicates that the fabrication of Janus-like phases is locked or inhibited by this energy barrier and explains the experiments performed by Lu et al. 27 and Li et al. 52 and the critical part of the sulfur vacancy creation and post-selenization process (Fig. 3).

Electronic structure of the MoS 2−x Se x phase
To get insights into the type of bonding between the chalcogen atoms and the molybdenum, we computed the electronic structure for the random and the Janus-like distribution.First, the projected density of states (PDOS) indicates the strong metallic influence around the Fermi level, characterized by the high contribution of the molybdenum d orbitals in all situations like in the case of MoS 2 and MoSe 2 10,41,53 .Second, the contribution of the sulfur and selenium p orbitals in the random distribution has a pronounced overlap around the Fermi level, indicated by the close similarities in the sulfur and selenium DOS curves' distribution.This overlap is less pronounced in the Janus-like distribution compared to the latter, especially at -1 eV and after 2 eV (Fig. 4).The PDOS analysis suggests a higher degree of coupling between the sulfur and the selenium ions when the MoSSe phase is reached in the random distribution, correlating the resulting contraction in the d Mo-S and d Mo-Se as described in previous sections.Inside the Janus-like distribution, a strong peak at − 1.2 eV suggests a hybridization of sulfur and selenium p orbitals and molybdenum d orbitals, primordially attributed to the in-plane distribution of the chalcogen atoms and alignment of the p x , p y, and p z symmetries.Lastly, the computed band structure reveals that the MoSSe phase remains with its semiconducting properties and a reduced indirect band gap close to 0.9 eV in both situations.Such reduction in the band gap agrees with what was observed by Li and coworkers, where a graded MoS 2(1−x) Se 2x nanosheet showed a variable band gap ranging from 1.8 eV to 1.6 eV as the selenium concentration increased 48 .The latter potentially implies that the MoS 2−x Se x mixed phase could work as a research platform for bandgap engineering of TMDC thin films.

Conclusions
The co-deposited thin film exhibits blue-shifted A 1 and E 1 Raman modes at 355 cm −1 and 255 cm −1 , which correspond to the Raman modes of the MoSSe phase and are evidence of lattice expansion.Time-of-flight secondary mass spectrometry shows an even distribution of selenium and sulfur ions throughout the thin film and evidenced the fabrication of sharp interfaces with the encapsulated layers.Our density functional theory calculations demonstrate that the formation of the random MoS 2−x Se x phase is thermodynamically favorable compared to the Janus-like phase, the latter hindered by an energy barrier below 8 at.% selenium concentrations.Interestingly, both mixed phases reached similar metastable states, denoted by their corresponding equilibrium energy.However, they presented different lattice expansion, above 7% for the random distribution and 5% for the Janus-like, attributed to different orbital's reallocation in Mo-S and Mo-Se bonds.These results display outstanding ease of fabrication of MoS 2-x Se x mixed phase with a random distribution of chalcogens by RF-sputtering.We believe that this report provides critical insights that can enhance the fabrication methods of large-area and scalable MoS 2−x Se x mixed phase in the future.The co-deposition by RF magnetron sputtering of the MoS 2−x Se x phase was completed using a Kurt J. Lesker PVD-75 instrument, which is equipped with three holders and these can open simultaneously during the deposition process without breaking the vacuum.ITO, MoS 2 and MoSe 2 targets were mounted separately inside each holder.
The substrate was a thermally oxidized Si/SiO 2 with 25.4 mm in diameter mounted within a working distance of 20 cm and no heating was applied to the substrate during the co-deposition process.The system was ready for the deposition once a vacuum of 1 × 10 -9 bar was reached.After this, a flow or Ar + ions was introduced into the main chamber setting a working pressure of 4 × 10 -6 bar.A layer of indium-tin-oxide (ITO) was first deposited to enhance the adhesion of the MoS 2-x Se x phase to the substrate using a commercial target of In 2 Sn 2 O 7 (99.9%purity and 76.2 mm in diameter) for 1800 s at 145 W of RF power at a frequency of 13.56 MHz to achieve the ~ 150 nm layer thickness.Subsequently, the MoS 2−x Se x phase was deposited by simultaneous sputtering of MoS 2 and MoSe 2 targets (99.99% purity and 76.2 mm in diameter) for 1000 s at 275 W of RF power at a frequency of 13.56 MHz, aiming for a MoSSe phase with a layer thickness of approximately 200 nm.Finally, a layer of ITO was deposited on top of the MoSSe phase to protect the material from environmental degradation without breaking the vacuum with the same parameters described before.

Cross-section, time-of-flight secondary ion mass spectrometry, Grazing incidence x-ray diffraction, and Raman spectroscopy
Cross-section images were acquired using a scanning electron microscope ZEISS Auriga 60 high-resolution dual beam equipped with an ion gun to qualify the interface formation of our co-deposited thin film.Images were recorded using the InLens detector at 20 kV.A ToFSIMS5-100 (ION-TOF GmbH) instrument was used for the time-of-flight secondary ion mass spectrometry (ToF-SIMS) with the aim of resolving the elemental composition of the thin film.This spectrometer is equipped with a Bi cluster primary ion source (field emission from liquid Bi wetting a tungsten tip) and a reflectron type time-of-flight analyzer.Ultra-high vacuum (UHV) base pressure during analysis was < 3 × 10 -8 mbar.For high mass resolution the Bi source was operated in the "high current bunched" mode providing short Bi + primary ion pulses at 25 keV energy, a lateral resolution of approx.4 μm, and a target current of 1.4 pA.The short pulse length of 1 ns allowed for high mass resolution (8000 m/∆m).The primary ion beam was scanned across a 250 µm 2 × 250 µm 2 field of view on the sample, and 64 × 64 data points were recorded.For depth profiling a dual beam analysis in interlaced mode was performed.The sputter gun (operated with Cs + or O 2 + ions, 2 keV, scanned over a concentric field of 500 µm 2 × 500 µm 2 , and target current of 180 nA and 600 nA, respectively) was applied to erode the sample.No further sample preparation was required for this measurement.Grazing incidence x-ray diffraction (GIXD) spectroscopy was collected using a Panalytical X-Pert system with a CuK α radiation source at 40 kV and 35 mA at room temperature.The grazing incidence angle was set at 5° < θ < 60° and step size of 0.05° using a graphite flat crystal monochromator.Raman measurements were used to discern the crystallinity of the thin films and were run on a Renishaw inVia Raman microscope using a laser excitation wavelength (λ e ) of 532 nm, a laser power of approx. 2 mW, and a 100x NA0.85 objective lens.All Raman measurements were taken at room temperature as well.

Computational details
Our DFT calculations were carried out utilizing the Quantum Espresso 54,55 package, which solves the Kohn-Sham equations by a plane-wave method.For the exchange-correlations term, the generalized gradient approximation (GGA) and the Perdew-Burke-Ernzerhof (PBE) options were chosen.The electrons' distribution were described by the optimized Vanderbilt pseudopotentials 56 provided through the SSSP package in its 1.2.1 version 57 .To account for the long-range forces inherently present in layered materials, we included the long-range dispersion correction DFT-D3 as described by Grimme et al. 58 for an accurate description of the material.Structural visualization of the models was assisted by the VESTA 59 and XCrysden 60 codes.For the structural optimization calculations, the plane-wave cutoff energy was set to 470 eV, while the convergence criterion of ionic minimization was achieved when all forces were smaller than 5.1 × 10 -2 eV nm −1 and the total energy changes less than 1.3 × 10 -2 eV atom −1 in two consecutive self-consistent field steps.Additionally, a k-point set of 4 × 4 × 4 was used to sample in the Brillouin zone.During the structural optimization of the MoS 2−x Se x phases, all atoms were able to move freely, and the lattice dimensions were not fixed.For the electronic structure calculations, an increased plane-wave cutoff energy of 544 eV was used along with a denser k-point set of 10 × 10 × 10.
All models started with an optimized MoS 2 unit cell with space group P6 3 /mmc and lattice constant a = b = 0.3168 nm, c = 1.249067 nm, α = β = 90°, and γ = 120°, with a Mo-S bond distance (d Mo-S ) of 0.2406 nm.The ionic state of S and Se inside the lattice is as S 2− and Se 2− , both in a three-coordinate geometry bonded to three equivalent Mo 4+ ions, the latter remaining in trigonal prismatic coordination.We created a 4 × 4 × 1 supercell from this model containing 32 molybdenum atoms and 64 sulfur atoms.This ensured modeling a wide variety of selenium concentrations avoiding self-interactions and achieved a robust and reliable model.For this study, we considered the evolution of the MoS 2−x Se x phase to occur from a selenium concentration of 4 at.% up to a 33.3 at.%, corresponding to a 1:1 relation between S and Se.
Two types of MoS 2−x Se x phases were considered, the first in a random configuration and the second in a Janus-type configuration.For the random MoS 2−x Se x phase, the selenium atoms substituted sulfur atoms and were distributed randomly throughout the supercell model.For the Janus MoS 2-x Se x phase, the selenium atoms replaced sulfur atoms in only one atomic layer of the MoS 2 structure as depicted in Fig. 1.The stability analysis of these two atomic arrangements first considered the formation energy per atom, (E form /eV atom −1 ), of the different MoS 2−x Se x phases, mimicking a co-deposit process, where all the atomic species converged to the formation of the targeted MoS 2-x Se x phase.This was computed using the formula: where E MoS 2−x Se x is the resulting enthalpy of the phase from the structural optimization calculations, N is the total number of atoms, and µ i is the chemical potential of the i th atom in the mentioned phase; this latter parameter was considered as the bulk energy of the atom.This approach considers that the distribution of selenium and sulfur atoms have an equal probability and that no other type of defect is left at the supercell afterward.Secondly, we were interested in the case where a possible inclusion of Se might occur into the MoS 2 lattice considering initially a pristine MoS 2 supercell and ultimately derive in the formation of the targeted MoS 2-x Se x phase.For this scenario, the substitutional energy per atom (E seg /eV atom −1 ) was computed as follows: where E MoS2 is the energy of the 4 × 4 × 1 defect-free MoS 2 supercell.Here, µ r is the chemical potential of the ithatom that was substituted or released from its original position, and µ s is the chemical potential of the ith-atom that substitutes the released atoms.As in the previous case, we consider that the occupation of sulfur positions by selenium atoms has an equal probability and that no other type of defect exists at the supercell before or afterwards.

Figure 1 .
Figure 1.(a, b) Time-of-flight secondary ion mass spectrometry depth profile measurements for positive and negative ions, respectively.(c) Schematic description of the deposited multilayer arrangement and cross-section scanning electron microscopy image.(d) Raman spectra of the obtained MoSSe thin film showing the blueshifted A 1 and E 1 corresponding to the MoSSe phase.Inset displays a schematic of the A 1 and E 1 Raman modes of MoSSe with gray balls denoting molybdenum atoms, and orange and yellow selenium and sulfur atoms, respectively.
1g normal Raman modes at ~ 376 cm −1 and ~ 408 cm−1   No direct relation between substrate and resulting crystallinity Higher crystallinity with higher deposition time 30 MoS 2 RF sputtering (MoS 2 target of 99.95% purity) on amorphous SiO 2 and (002) oriented graphsputtering on SiO 2 /Si, quartz, and sapphire substrates Power = 25 W Dwell time = 1, 3, 5 and 15 min Working pressure = 1.33 × 10 -5 bar Substrate temperature = RT to 500 °C Working distance = N/A Bilayer to few layer MoS 2 domains Post-deposition annealing at 700 °C in a sulfur-rich atmosphere to improve crystallinity Improved carrier mobility 16 MoSe 2 DC sputtering on quartz and Si substrates Power = 75 W Dwell time = 4 min Working pressure = 6.7 × 10 -6 bar Substrate temperature = RT Working distance = 5 cm Thin film with wall-like structures Thickness = 325 nm Preferential growth along the c-axis E 1 2g and A 1g normal Raman modes at ~ 242 cm −1 and ~ 284 cm −1 31 MoSe 2 RF co-sputtering of Mo and Se targets with 99.99% purity on Si substrate Power = 15-45 W for Mo/15 to 25 W for Se target Dwell time = N/A Working pressure = 1.33 × 10 -1g normal Raman modes at ~ 239 cm −1 and ~ 301 cm −1 39

Figure 2 .
Figure 2. (a) Schematic of the 2H-MoS 2 supercell, (b) random MoS 2−x Se x supercell at x = 1, and (c) Janus-type MoS 2−x Se x supercell at x = 1.All MoS 2-x Se x models were created using a 2H-MoS 2 4 × 4 × 1 supercell containing 32 molybdenum atoms and 64 sulfur atoms.Color code: yellow is sulfur, orange is selenium and grey corresponds to molybdenum atoms.

Figure 3 .
Figure 3. (a) Computed equilibrium energy (E q ) for pure 2H MoS 2 , 2H MoSe 2 , Janus, and random MoS 2−x Se x at a Se concentration of 33.33 at.%.The random MoS 2-x Se x phase at x = 1 reaches almost the same entropy state as the Janus-like phase.E q computed by Quantum Espresso code includes a one-electron energy part (e elec ), the Hartree contribution (e H ), the exchange-correlation energy (e xc ), the Ewald contribution (e ewald ), and the dispersion correction part by DFT + 3 (e DFT+D ).(b) Estimated energy profile for the transition from pure MoS 2 to pure MoSe 2 having a random and Janus-like phase.The fabrication of MoSSe Janus-like would be limited by an energy barrier at 8 at.% of selenium.Arrows point to the location of their relative energy of formation.

Figure 4 .
Figure 4. (a) Computed projected density of states (PDOS) for the MoSSe in a random and Janus-like distribution.The Fermi level is located at 0 eV.The MoSSe phase having a random distribution presents a higher degree of orbital hybridization between the S p-and Se p-orbital, indicated by the similarities in the curvatures of the pDOS around the Fermi level.(b) Computed band structure for the random and Janus-like distribution in the Γ-Μ-Κ-Γ path.The estimated band gap of both situations is close to 0.9 eV. https://doi.org/10.1038/s41598-024-57243-3

Table 1 .
Comparison of different MoS 2 and MoSe 2 fabrication methods and the resulting characteristics of the material.

at. % E form /eV atom −1 E subst /eV atom −1 E susbt-V /eV atom −1 a/nm a/nm b/nm c/nm V/nm 3 ΔV
2−x Se x phases at different concentrations of Se.Negative values of E form , E seg , and E seg-V designate a thermodynamically favorable process.