---
title: 'Janus Monolayer TMDs: Asymmetric 2D Materials'
url: https://www.emergentmind.com/topics/janus-monolayer-transition-metal-dichalcogenides-tmds
type: topic
---

# Janus Monolayer TMDs: Asymmetric 2D Materials

Searching arXiv for recent and foundational work on Janus monolayer TMDs to ground the article in the current literature.
Janus monolayer transition metal dichalcogenides (TMDs) are two-dimensional crystals of the form \(MXY\), in which a transition-metal plane is sandwiched between two inequivalent chalcogen planes \(X\) and \(Y\). In these monolayers, replacing one chalcogen layer of a parent \(MX_2\) sheet breaks the out-of-plane mirror symmetry that characterizes conventional symmetric TMDs and lowers the point-group symmetry to \(C_{3v}\) in the common 1H family, thereby generating a built-in out-of-plane dipole, Rashba-type spin splitting, and modified optical, transport, catalytic, and electromechanical responses [1704.06389] [2502.02806]. The field encompasses several structural polytypes and chemistries, including 1H group-VIB systems such as MoSSe and WSSe, 1T′-derived Janus monolayers such as WSeTe and WSTe, and lower-symmetry variants such as Janus rhenium dichalcogenides and vanadium dichalcogenides, whose reduced or altered symmetries enlarge the accessible phase space of band, spin, and collective-order engineering [2201.01467] [2109.00071] [2406.12180].

## 1. Crystal chemistry, symmetry, and defining structural motifs

Janus monolayer TMDs extend the family of two-dimensional materials by replacing one of the two chalcogen layers in an \(MX_2\) sheet with a different chalcogen, yielding \(MXY\) [2502.02806]. In the 1H phase, the structure is trigonal-prismatic, with a plane of transition-metal atoms sandwiched between two inequivalent chalcogen planes \(X\) and \(Y\); because \(X \neq Y\), out-of-plane mirror symmetry is broken, but in-plane \(C_3\) symmetry remains [2502.02806]. In conventional monolayer TMDs the point group is \(D_{3h}\), whereas Janus \(MXY\) lowers the point-group symmetry to \(C_{3v}\) by replacing the top chalcogen by a different species [2403.06165]. This broken mirror symmetry produces a built-in dipole \(p_z\) per unit cell and an internal polarization field \(E_{\rm polar}\) along \(z\), with the details of the dipole set by the inequivalent metal-chalcogen bond lengths [2403.06165].

The same symmetry logic generalizes beyond the 1H family. In 1T′-derived Janus monolayers obtained from \( {\rm WTe_2} \), replacing one Te layer by Se or S breaks out-of-plane mirror symmetry \(M_z\) and inversion \(P\), leaving only the in-plane mirror \(M_x\) [2201.01467]. In Janus VTeSe, replacing only the top Te with Se removes inversion and changes the point group from \(D_{3d}\) to \(C_{3v}\) [2406.12180]. In low-symmetry rhenium dichalcogenides, in-plane anisotropy generates a much larger number of inequivalent Janus monolayers than in hexagonal TMDs, including 29 distinct \({\rm ReX_{2-x}Y_x}\) structures [2109.00071].

Several structural metrics recur across the literature. Ideal Janus \(MXY\) monolayers with \(M={\rm Mo,W}\) and \(X/Y={\rm Se,S}\) adopt a hexagonal lattice with \(a \simeq 3.3\,\text{\AA}\), \(d_{M-X} \simeq 2.52\,\text{\AA}\), and \(d_{M-Y} \simeq 2.42\,\text{\AA}\) [2403.06165]. For specific materials, DFT gives \(a({\rm MoSe_2}) = 3.322\,\text{\AA}\), \(a({\rm SMoSe}) = 3.252\,\text{\AA}\), and \(a({\rm MoS_2}) = 3.185\,\text{\AA}\) [1704.06389], while optimized values for freestanding WSe\(_2\) and WSSe are \(a_{\rm WSe_2} = 3.318\,\text{\AA}\) and \(a_{\rm WSSe} = 3.254\,\text{\AA}\) [2510.04113]. In Janus SeMoS grown on Au, the AFM step height is \(0.8 \pm 0.2\,{\rm nm}\), consistent with a single layer [2205.04751].

A frequent misconception is that all Janus TMDs are simply “polar versions” of ordinary hexagonal TMDs with otherwise unchanged physics. The literature instead shows that the consequences of symmetry lowering depend strongly on polytype and orbital character. In 1H WSSe, the dominant modification is a Rashba-type splitting and an upward shift of the \(\Gamma\)-derived valence band, while the \(K\)-valley valence bands remain nearly unchanged [2510.04113]. In 1T′-derived WSeTe and WSTe, by contrast, Janus asymmetry removes the parent quantum-spin-Hall inversion and produces a topologically trivial but strongly nonlinear-Hall-active state [2201.01467].

## 2. Synthesis routes and structural identification

Experimentally, Janus monolayers have been realized primarily through selective chalcogen exchange. The foundational demonstration of Janus SMoSe used controlled sulfurization of monolayer MoSe\(_2\): as-grown MoSe\(_2\) was placed in the furnace center at \(800\,^\circ{\rm C}\), sulfur powder was heated to \(150\,^\circ{\rm C}\), Ar at \(100\,{\rm sccm}\) carried \( {\rm S_8} \) vapor to the MoSe\(_2\) surface, the reaction time was 30 min, and the selective reaction was maintained in the \(750\!-\!850\,^\circ{\rm C}\) window, below which no top-layer exchange occurs and above which full Se\(\rightarrow\)S substitution happens in both layers [1704.06389]. The net reaction was written as
\[
\tfrac12 S_8(g)+{\rm MoSe_2(s)}\longrightarrow {\rm SMoSe(s)}+\tfrac32\,Se_2(g).
\]

A distinct scalable route is one-pot chemical vapor deposition on Au foils. In that method, MoSe\(_2\) is first grown on polycrystalline Au foil and then sulfurized at \(700\,^\circ{\rm C}\), with the net exchange reaction
\[
{\rm MoSe_2\,(ML)} + \tfrac12 S_2(g)\rightarrow {\rm Se\!-\!Mo\!-\!S\,(Janus\,ML)}+\tfrac12 Se_2(g),
\]
or more generally \( {\rm MoSe_2 + xS \rightarrow SeMoS + ySe} \), with DFT-computed reaction energy \(\Delta E_{\rm replace} \simeq -0.25\,{\rm eV}\) per formula [2205.04751]. DFT on an \(8\times8\) MoSe\(_2\)/\(9\times9\) Au(111) supercell showed that \(S_2\) dissociation on Au(111) is exothermic with \(E_{\rm ads} \simeq -3.5\,{\rm eV}\) per S atom, S adatoms diffuse with \(E_{\rm mig}<0.4\,{\rm eV}\), removing a bottom-layer Se atom and binding it to Au costs only \(\simeq +1.7\,{\rm eV}\), whereas removing a top-layer Se costs \(>+5\,{\rm eV}\), and the net replacement energy for bottom-layer Se by S is slightly exothermic \((\Delta E \simeq -0.25\,{\rm eV})\) [2205.04751]. These energetics explain why S atoms diffuse beneath the MoSe\(_2\) on Au and selectively exchange the bottom Se layer.

Room-temperature plasma or atomic-replacement strategies have also proved important. In monolayer WSSe, in-situ \( {\rm H_2} \) plasma treatment at room temperature replaces the top Se layer with S, transforming Se–W–Se into S–W–Se; during the reaction the A-exciton PL peak shifts from \(\sim 1.6\,{\rm eV}\) in WSe\(_2\) to \(\sim 1.8\,{\rm eV}\) in WSSe over \(\sim 10\) min of plasma exposure [2510.04113]. In charge-tuneable Janus WSeS, the “Selective Epitaxial Atomic Replacement” (SEAR) method converts exfoliated 1L-WSe\(_2\) into Janus WSeS in situ at room temperature under controlled atmosphere, with time-resolved Raman spectroscopy used to stop conversion upon appearance of Janus signature modes [2210.06930]. Remote \( {\rm H_2} \)-plasma-assisted atomic-layer substitution has likewise been used to convert MoSe\(_2\) to MoSSe and MoS\(_2\) to MoSeS at room temperature within 5–15 min [2503.22861].

Structural identification relies on spectroscopy and microscopy. In SMoSe, Raman-active Janus signatures include an out-of-plane S–Mo–Se mode at \(290\,{\rm cm^{-1}}\), an in-plane S–Mo–Se mode at \(350.8\,{\rm cm^{-1}}\), and an emergent \(B_{2g}\) mode at \(436.4\,{\rm cm^{-1}}\), while TOF-SIMS depth profiling established the knockout sequence S\(\rightarrow\)Mo\(\rightarrow\)Se\(\rightarrow\)SiO\(_2\), directly proving S–Mo–Se ordering [1704.06389]. In one-pot-grown SeMoS, Raman spectroscopy shows new peaks at \(A_1^{\prime 1}\simeq 290\,{\rm cm^{-1}}\) and \(E_{2g}\simeq 353\,{\rm cm^{-1}}\), with \({\rm FWHM}(A_1^{\prime 1}) = 4.5 \pm 0.2\,{\rm cm^{-1}}\), and angle-resolved XPS directly confirms Se on top, Mo in the middle, and S at the bottom [2205.04751]. In WSeS, pristine WSe\(_2\) shows the 1L-WSe\(_2\) \(E' + A_1'\) mode at \(254\,{\rm cm^{-1}}\), whereas fully converted Janus WSeS exhibits a single \(A_1^1\) mode at \(284\,{\rm cm^{-1}}\), with disappearance of WS\(_2\)/WSe\(_2\) alloy peaks [2210.06930]. A later “Raman digital twin” library generalized these fingerprints: in 2H Janus TMDs the broken \(\sigma_h\) activates four first-order Raman modes, two \(E\) and two \(A_1\), and in Janus WSSe the predicted frequencies are \(E(1)=205\,{\rm cm^{-1}}\), \(E(2)=335\,{\rm cm^{-1}}\), \(A_1(1)=284\,{\rm cm^{-1}}\), and \(A_1(2)=423\,{\rm cm^{-1}}\) [2510.09839].

## 3. Electronic structure, orbital asymmetry, and Rashba physics

The electronic hallmark of Janus monolayer TMDs is that broken out-of-plane mirror symmetry modifies bands in an orbital-selective manner. Direct micro-focused ARPES on an identical sample transformed from monolayer WSe\(_2\) to Janus WSSe provided a particularly clear case: both WSe\(_2\) and WSSe show nearly identical spin-orbit-split valence bands at \(K\), with the valence-band maximum pinned at \(E_B = 0.82\,{\rm eV}\) below \(E_F\), whereas at \(\Gamma\) monolayer WSe\(_2\) has a single spin-degenerate valence-band maximum at \(E_B(\Gamma)=1.35\,{\rm eV}\) and Janus WSSe exhibits two split branches with extrema at \(E_B(\Gamma)=1.19\,{\rm eV}\) [2510.04113]. The upward shift is therefore
\[
E_B^{\rm WSe_2}(\Gamma)=1.35\,{\rm eV}\rightarrow E_B^{\rm WSSe}(\Gamma)=1.19\,{\rm eV},
\]
that is, \(\Delta E_\Gamma = +0.16\,{\rm eV}\) \((160\,{\rm meV})\) [2510.04113].

Near \(\Gamma\), Janus WSSe realizes a Rashba-type splitting because the \(C_{3v}\) crystal lacks horizontal mirror symmetry. The effective Hamiltonian is
\[
H_R=\alpha_R (\sigma \times k)\cdot \hat z,
\]
leading to
\[
\Delta E(k)=E_+(k)-E_-(k)=2\alpha_R |k|.
\]
From \(\mu\)-ARPES along \(\Gamma\)-M, the splitting at \(k_0 \simeq 0.35\,\text{\AA}^{-1}\) is \(\Delta E(k_0)\simeq 270\,{\rm meV}\), which yields \(\alpha_R \simeq 0.39\,{\rm eV\cdot \AA}\) [2510.04113]. Fully relativistic PBE-GGA calculations reproduce a Rashba splitting of \(\sim 200\,{\rm meV}\) at \(\Gamma\) in WSSe, no Rashba splitting in WSe\(_2\), and an upward shift of \(\sim 0.22\,{\rm eV}\) at \(\Gamma\) in WSSe versus WSe\(_2\), traced to the larger S \(3p_z\) versus Se \(4p_z\) weight, with projected \( {\rm W}\,5d_{z^2}:{\rm Se}\,4p_z:{\rm S}\,3p_z \simeq 12:1:4 \) [2510.04113]. This is the paper’s “chalcogen-orbital engineering.”

The persistence of the \(K\)-valley bands alongside strong \(\Gamma\)-point reconstruction is not accidental. In WSSe, the \(K\)-valley invariance is attributed to the dominant in-plane W \(5d(x^2-y^2,xy)\) orbital character at \(K\), which is insensitive to \(\sigma_h\) breaking, together with Fermi-level pinning by the graphite substrate [2510.04113]. This orbital selectivity recurs in model treatments. A low-energy tilted massive Dirac Hamiltonian has been used to describe Janus TMD monolayers lacking inversion symmetry,
\[
H(k)=t\,k_x\,\sigma_0 + v(k_y\sigma_x+\eta k_x\sigma_y)+\left(\frac m2-\alpha |k|^2\right)\sigma_z,
\]
with the tilt parameter \(t\) breaking inversion symmetry but preserving time reversal [2201.01467] [2206.03416]. In the excitonic analysis of inversion-broken TMDC monolayers, once \(t\neq 0\) the conduction-band minimum and valence-band maximum shift in opposite directions in \(k\)-space, making the gap indirect, and for \(t\ge v\) the spectrum becomes semimetallic with electron and/or hole pockets [2206.03416].

Broken symmetry also reshapes orbital and spin textures. In non-Janus monolayers, the wavefunctions at the valence and conduction bands are dominated by \(|x^2-y^2\rangle\), \(|xy\rangle\), and \(|z^2\rangle\) orbitals, but in Janus systems the internal electric field intermixes these states with \(|xz\rangle\) and \(|yz\rangle\), producing a robust orbital texture around the valleys \(\Gamma\), \(K\), and \(K'\) [2511.09974]. Near each valley \(\eta\), the in-plane orbital texture is parametrized by
\[
m_{L,x}^\eta(q)=-\alpha_L^\eta q_y,\qquad m_{L,y}^\eta(q)=\alpha_L^\eta q_x,
\]
and spin-orbit coupling not only generates a chiral spin texture but also reverses the chirality of one orbital branch [2511.09974]. In WSeTe, a six-band tight-binding model further gives a Rashba term with \(\alpha_0 = 0.045\,{\rm eV}\) and conventional Ising SOC \(\lambda_{\rm SOC}=0.228\,{\rm eV}\) [2505.06622].

A plausible implication is that “Janus band engineering” is best understood not as a uniform shift of all band edges, but as symmetry- and orbital-selective perturbation: weak at valleys dominated by in-plane \(d\) orbitals, strong where \(p_z\) and \(d_{z^2}\) admixtures are substantial, and especially consequential near avoided crossings or \(\Gamma\)-derived states.

## 4. Optical transitions, excitons, and magneto-optical response

Janus monolayer TMDs retain direct-gap optical transitions in several experimentally relevant cases while simultaneously introducing dipolar and spin-split excitonic structure. In charge-tuneable Janus WSeS, DFT including SOC confirms that the system is direct-bandgap at the \(K/K'\) points, with both the conduction-band minimum and valence-band maximum at \(K\) [2210.06930]. The measured neutral exciton appears in both PL and reflectance contrast at
\[
E_{X^0}=1.893\,{\rm eV},
\]
with the narrowest PL full width at half maximum \(\Delta E \simeq 5.9\,{\rm meV}\) and average \({\rm FWHM}\simeq 8.4(4)\,{\rm meV}\) across the device [2210.06930]. Gate tuning resolves two negatively charged trions: \(X^-_{\rm inter}\) at \(1.864\,{\rm eV}\), \(32\,{\rm meV}\) below \(X^0\), and \(X^-_{\rm intra}\) at \(1.857\,{\rm eV}\), \(39\,{\rm meV}\) below \(X^0\), giving an exchange splitting \(\Delta E_{\rm exch}\simeq 7\,{\rm meV}\) [2210.06930]. At high \(n\)-doping \(>10^{13}\,{\rm cm^{-2}}\), a lower-energy feature \(X^{-\prime}\) appears at \(\simeq 1.845\,{\rm eV}\) and red-shifts by \(\simeq 10\,{\rm meV}\) as carrier density increases [2210.06930].

The same WSeS work attributes the out-of-plane dipolar character of neutral excitons to the asymmetric chalcogen environment. DFT predicts for the neutral exciton in WSeS a dipole moment \(p \approx 0.24\,{\rm D}\), with an estimated Stark shift
\[
\Delta E_{\rm Stark} = -pF,\qquad \Delta E_{\rm Stark}\approx 5\,{\rm meV}\ {\rm at}\ F=1\,{\rm V/nm}
\]
[2210.06930]. Magneto-optic measurements indicate that monolayer WSeS has a direct bandgap at the \(K\) points, and at 4 K the measured exciton \(g\)-factors are \(g\simeq 4.14(6)\) for \(X^0\) in PL, \(4.5(2)\) in reflectance contrast, \(5.7(7)\) for \(X^-_{\rm inter}\), \(5.4(6)\) for \(X^-_{\rm intra}\), and \(4.1(4)\) for \(X^{-\prime}\) [2210.06930].

One-pot-grown SeMoS monolayers exhibit similarly strong optical quality. Low-temperature PL on hBN-encapsulated SeMoS reveals a free-exciton peak at \(\sim 1.778\,{\rm eV}\) with \(\Gamma_0 \simeq 18\,{\rm meV}\) at 5 K, visible up to room temperature, and a lower-energy localized peak near \(\sim 1.5\,{\rm eV}\) appearing only below 150 K [2205.04751]. The exciton energy follows the O’Donnell–Chen form
\[
E_X(T)=E_X(0)-S\langle \hbar \omega \rangle \big[\coth (\langle \hbar \omega \rangle /2k_BT)-1\big],
\]
with \(E_X(0)=1.778\,{\rm eV}\), \(S=2.4\), and \(\langle \hbar \omega \rangle = 38.9 \pm 0.7\,{\rm meV}\), while the linewidth broadening is described by
\[
\Gamma(T)=\Gamma_0+aT+B[\exp(\hbar \omega/k_BT)-1]^{-1}
\]
with \(a = 24\,\mu{\rm eV/K}\) and \(B = 124\,{\rm meV}\), indicating stronger exciton-phonon coupling than in MoSe\(_2\) and MoS\(_2\) monolayers [2205.04751]. Under circularly polarized excitation, the free exciton shows \(P_c \simeq 25\%\) up to 100 K; in Faraday geometry up to 9 T, the Zeeman splitting is \(\Delta E_Z \simeq 1.7\,{\rm meV}\) at \(B=9\,{\rm T}\), corresponding to \(g_X \simeq -3.3\) [2205.04751].

Theory has pushed the excitonic perspective further. For inversion-broken TMDC monolayers described by a tilted massive Dirac Hamiltonian, solving the Bethe–Salpeter equation with a Rytova–Keldysh interaction gives \(1s\) binding energies of \(134\,{\rm meV}\) in hBN and \(56.6\,{\rm meV}\) in quartz for representative parameters \(m=0.2\,{\rm eV}\), \(v=1\,{\rm eV\,\AA}\), and \(\alpha=1\,{\rm eV\,\AA^2}\) [2206.03416]. An excitonic instability occurs when \(E_{1s}>E_g(t)\); numerically, the crossing occurs at \(t_c \approx 1.10\,v\) in hBN and \(t_c \approx 0.83\,v\) in quartz [2206.03416]. In the optical conductivity, only excitons with angular momentum \(\ell=0\) or \(\ell=\pm 2\) are active, and the \(\ell=0\) resonances carry oscillator strengths more than two orders of magnitude larger than the \(\ell=\pm2\) transitions [2206.03416].

These results clarify a second common misconception: broken mirror symmetry does not necessarily destroy the bright \(K/K'\) optical transitions of monolayer TMDs. In WSeS, magneto-optic data and charge-tunable spectroscopy are explicitly consistent with direct-gap transitions at \(K\) [2210.06930], while in SMoSe the gap becomes slightly indirect, with \(E_g^{\rm ind}(I\rightarrow K)=1.558\,{\rm eV}\) and \(E_g^{\rm dir}(K\rightarrow K)=1.568\,{\rm eV}\), producing PL quenching by \(\approx 10\times\) and a blue shift to \(\sim 1.68\,{\rm eV}\) [1704.06389]. The optical consequences are therefore material specific rather than universal.

## 5. Topology, Berry curvature, nonlinear Hall response, and collective ordering

In 1T′-derived Janus monolayers, inversion breaking can change not only spin splitting but also band topology. With SOC, monolayer WTe\(_2\) is a quantum-spin-Hall insulator with a direct gap \(\sim 0.14\,{\rm eV}\) and inverted band ordering near the \(\Gamma\)-X line, whereas WSeTe and WSTe remain gapped but lose band inversion because of the asymmetric crystal field introduced by the Janus structure [2201.01467]. The \(Z_2\) invariant determined from the Wannier charge-center flow is \(Z_2=1\) for WTe\(_2\) and \(Z_2=0\) for WSeTe and WSTe, and ribbon spectra confirm that only the parent WTe\(_2\) hosts helical edge states crossing the gap [2201.01467]. The direct gap along X–\(\Gamma\) decreases from \(\sim 143\,{\rm meV}\) in WTe\(_2\) to \(\sim 32\,{\rm meV}\) in WSeTe and \(\sim 6\,{\rm meV}\) in WSTe [2201.01467].

Despite being topologically trivial, these Janus monolayers exhibit large Berry curvature dipoles. For a 2D system,
\[
\Omega_z^n(k)=2\hbar^2 \sum_{m\neq n}\frac{{\rm Im}\big[\langle n|v_x|m\rangle \langle m|v_y|n\rangle \big]}{(\epsilon_n-\epsilon_m)^2},
\]
and the Berry curvature dipole is
\[
D_a\equiv \sum_n \int_{\rm BZ} f_n(k)\,\partial_a \Omega_z^n(k)\,d^2k .
\]
With only \(M_x\) symmetry, the only nonzero component is \(D_x\equiv D_{xz}\) [2201.01467]. First-principles calculations show that \(\Omega_z(k)\) concentrates near band anticrossings along \(\Gamma\)-X, its magnitude increases as the gap shrinks from WTe\(_2\) to WSeTe to WSTe, and the computed \(D_x(E_F)\) at the Fermi level is roughly ten times larger in WSTe than in WTe\(_2\) [2201.01467]. In time-reversal-invariant but inversion-broken crystals, this yields a second-order Hall-like response \(j_y \propto D_x E_x^2\) [2201.01467].

Janus asymmetry can also alter collective electronic order. In monolayer Janus VTeSe fabricated by surface selenization of VTe\(_2\), low-temperature STM at 4.5 K reveals a rhombic \(\sqrt{13}\times \sqrt{13}\) supercell with side length \(a_{\rm CDW}\simeq 1.26\,{\rm nm}\simeq \sqrt{13}\,a\) and cell angle \(\simeq 88^\circ\) [2406.12180]. The charge modulation can be written as
\[
\rho(r)=\rho_0+\sum_{i=1}^3 A_i \cos (q_i\cdot r+\phi_i),
\]
with \(q_i=(1/\sqrt{13})\,G_i\) [2406.12180]. DFT finds that Fermi-surface nesting and nonmagnetic phonon instabilities both point to a \(4\times4\) modulation rather than the observed \(\sqrt{13}\times \sqrt{13}\) state, whereas in the \(\sqrt{13}\times \sqrt{13}\) supercell the pronounced flat PDOS peak at \(E_F\) splits and shifts away, lowering the total energy by \(\sim 0.7\,{\rm meV}\) per unit cell [2406.12180]. The proposed mechanism is therefore charge modulation in the Janus VTeSe beyond the conventional electron-phonon picture [2406.12180].

Magnetic Janus TMDs add yet another sector. In vanadium-based Janus monolayers \( {\rm VXY} \), first-principles calculations predict that the built-in electric field \(E_z\) of order \(0.1\!-\!0.2\,{\rm V/\AA}\) generates Rashba-like spin-orbit effects, while heavier chalcogens enhance magnetic exchange and anisotropy [2007.07579]. The extracted Rashba coefficients are \(\alpha_R = 0.17\,{\rm eV\cdot \AA}\) for VSSe, \(0.24\,{\rm eV\cdot \AA}\) for VSeTe, and \(0.28\,{\rm eV\cdot \AA}\) for VTe\(_2\), with exchange \(J=50.8\,{\rm meV}\) and anisotropy \(K=0.241\,{\rm meV}\) per V in VSeTe [2007.07579]. Kubo–Bastin transport calculations on a \(40\times 40\) tight-binding supercell give an overall spin-orbit torque efficiency \(\eta_{\rm peak}\) up to \(\simeq 2.5\%\) and a zero-temperature critical switching current density \(J_{c0}\approx 7.5\times10^7\,{\rm A/cm^2}\) [2007.07579].

## 6. Stability, defects, and materials optimization

A central issue for Janus monolayer TMDs is why some compounds are realized experimentally while many predicted ones are not. A systematic electronic-structure analysis of 1H-phase JTMDs identifies a group dependence of stability, with Group-VIB-based monolayers exhibiting robust stability, as evidenced by the successful synthesized MoSSe and WSSe [2502.02806]. The underlying competition is between metal-ligand ionic bonding and ligand-ligand covalent bonding, together with high-energy \(d\)-electron orbital splitting near \(E_F\) [2502.02806]. In trigonal-prismatic \(C_{3v}\) coordination, the five \(d\) orbitals split into \(a_1'(d_{z^2})\), \(e'(d_{x^2-y^2},d_{xy})\), and \(e''(d_{xz},d_{yz})\), and the authors classify JTMDs using an \(N^n\) scheme based on occupied antibonding states and unoccupied bonding states [2502.02806]. Across 84 computed \(MXY\) systems, 45 are stable in 1H, including 32 from groups IVB and VIB [2502.02806]. Prototype values are \(a\approx 3.24\,\text{\AA}\) and \(E_{\rm form}\approx +0.47\,{\rm eV}\) per formula unit for MoSSe, and \(a\approx 3.25\,\text{\AA}\) and \(E_{\rm form}\approx +0.47\,{\rm eV}\) for WSSe, both dynamically stable [2502.02806].

High-throughput thermodynamic characterization similarly shows that stable Janus ordering at room temperature is not generic. In a survey of 72 TMD alloys using cluster expansion and Monte Carlo, the only alloys with stable out-of-plane Janus ordering at 300 K are the X-site mixed Ti- and Zr-based systems: T-TiSTe, T-TiSeTe, H-TiSTe, H-TiSeTe, T-ZrSTe, T-ZrSeTe, H-ZrSTe, and H-ZrSeTe [2204.11223]. Their mixing enthalpies at \(x=0.5\) are negative, ranging from \(-25\,{\rm meV}\) per MX\(_2\) for T-TiSTe to \(-5\,{\rm meV}\) for H-ZrSeTe [2204.11223]. This suggests that thermodynamic favorability of ordered Janus stacking is highly chemistry dependent even within the broader TMD family.

Defects are another decisive factor. High-resolution STEM, DFT, and cryogenic spectroscopy on Janus \(MXY\) monolayers identify the most energetically stable point defects as single chalcogen vacancies \(V_{\rm S}\) and \(V_{\rm Se}\), the double vacancy \(V_{\rm S}\!-\!V_{\rm Se}\), interstitial defects \(M_i\), and metal impurities \(M_W\) [2403.06165]. Under elemental chemical potentials, the neutral formation energies are \(E_f^0 \approx 2.1\,{\rm eV}\) for \(V_{\rm S}\), \(2.3\,{\rm eV}\) for \(V_{\rm Se}\), \(2.0\,{\rm eV}\) for \(V_{\rm S}\!-\!V_{\rm Se}\), and \(>3.5\,{\rm eV}\) for \(M_i\) and \(M_W\) [2403.06165]. The defect complex \(V_{\rm S}\!-\!V_{\rm Se}\) has a binding energy \(E_b \approx 50\,{\rm meV}\), making it slightly more favorable than two isolated single vacancies [2403.06165]. In-gap Kohn–Sham levels appear at \(\Delta E_{\rm KS}\approx 1.73\,{\rm eV}\) for \(V_{\rm S}\), \(1.65\,{\rm eV}\) for \(V_{\rm Se}\), and \(1.59\,{\rm eV}\) for \(V_{\rm S}\!-\!V_{\rm Se}\), with SOC splittings of 210–300 meV [2403.06165]. In hBN-encapsulated Janus WSeS at 4 K, these correlate with narrow bound-exciton lines at \(P1 \approx 1.72\,{\rm eV}\), \(P2 \approx 1.65\,{\rm eV}\), and \(P3 \approx 1.60\,{\rm eV}\) [2403.06165].

Synthesis-related nanoscale defects have also been resolved directly. Cross-correlated AFM and TERS imaging of MoSSe and MoSeS monolayers shows that bilayer nanoislands in the precursor crystals lead, after Janus conversion, to Janus/TMD vertical heterostructure islands of \(\sim 15\!-\!30\,{\rm nm}\) lateral size, while the precursor bilayer patches themselves are \(\sim 20\,{\rm nm}\) across [2503.22861]. Strain transfer from growth substrate and conversion pathway strongly affects morphology: for MoSe\(_2\rightarrow\)MoSSe on SiO\(_2\)/Si, the net strain is \(\varepsilon_{\rm net}\approx +2.4\%\) tensile, producing fragmentation into \(\sim 0.5\!-\!1\,\mu{\rm m}\) domains with a defect density of \(\sim 10^8\!-\!10^9\) cracks/cm\(^2\), whereas for MoS\(_2\rightarrow\)MoSeS on SiO\(_2\)/Si the net strain is \(\varepsilon_{\rm net}\approx -1.6\%\) compressive, producing \(\sim 5\,{\rm nm}\)-high wrinkles with 100–200 nm spacing but no fragmentation [2503.22861]. Switching to fused silica yields \(>90\%\) area free of wrinkles in Janus MoSeS [2503.22861].

Carrier transport adds another optimization axis. A Born effective charge (BEC) method proposed for Janus TMDs includes both acoustic and polar-optical phonon scattering, correcting the well-known overestimation of mobilities by deformation-potential theory in polar Janus structures [2203.03389]. The reported trend is \(\mu \propto |Z_B|^{-n}\) with \(n\approx 2\!-\!3\), and the lower the absolute BEC, the higher the electron or hole mobility [2203.03389]. Among pristine systems listed, H-WSSe has \(|Z_B|=0.89\), \(\mu_e=395\,{\rm cm^2/V\cdot s}\), and \(\mu_h=382\,{\rm cm^2/V\cdot s}\), higher than H-MoSSe with \(|Z_B|=1.48\), \(\mu_e=156\), and \(\mu_h=125\) [2203.03389]. The paper therefore recommends W–S–Se Janus monolayers for high-mobility 2D electronics [2203.03389].

## 7. Functional responses and application space

The combination of intrinsic asymmetry, strong SOC, tunable band edges, and chemically distinct surfaces gives Janus monolayer TMDs a broad application space that spans optoelectronics, spintronics, electromechanics, and catalysis. In band-engineered Janus WSSe, the elevation of \(\Gamma\)-derived valence-band states and the Rashba splitting are expected to open new direct and spin-forbidden optical channels, potentially observable in helicity-resolved PL or pump–probe studies, while the coexistence of out-of-plane \(K\)-valley spin locking and in-plane Rashba spins at \(\Gamma\) suggests richer spin textures for spintronic devices, including electric-field control of spin orientations [2510.04113]. In Janus WSeTe, optical irradiation can generate pure spin Hall currents; at hole doping \(E_F=-0.85\,{\rm eV}\), \({\rm Re}[\sigma_{xy}^{\rm spin(y)}(\omega)]\) peaks near \(0.12\, (e\hbar^2)/[4(2\pi)^2]\) at \(\hbar \omega \approx 1.8\,{\rm eV}\), \({\rm Re}[\sigma_{xy}^{\rm spin(z)}(\omega)]\) peaks near \(0.25\, (e\hbar^2)/[4(2\pi)^2]\) at \(\hbar \omega \approx 1.9\,{\rm eV}\), and the spin Hall angle reaches values of order \(0.05\!-\!0.2\) in the visible range [2505.06622].

Current-induced orbital and spin polarizations constitute a related direction. For Janus TMDs of the form \(MXX'\), a DC in-plane electric field produces both orbital and spin Edelstein effects, with the orbital response one order of magnitude larger [2511.09974]. For MoSSe hole doped by \(0.04\,{\rm eV}\) below the valence-band maximum, \(\chi^O_{yx}\approx 3.5\times 10^{-8}\,\mu_B\,{\rm m\,V^{-1}}\) and \(\chi^S_{yx}\approx 0.4\times 10^{-8}\,\mu_B\,{\rm m\,V^{-1}}\) [2511.09974]. Across the family, Te-based compounds such as MoSTe and WSTe show the largest \(\chi^O\), up to \(\sim 1.5\times 10^{-7}\,\mu_B\,{\rm m\,V^{-1}}\) [2511.09974]. This suggests that Janus asymmetry can act not only through spin textures but also through current-induced orbital accumulation, an increasingly important theme in orbitronics.

Electromechanical functionality is likewise enhanced by asymmetry. In a charge-dipole model combined with molecular dynamics, the bending flexoelectric constant \(\mu_{zyzy}\) of Janus TMDCs is positively correlated with the initial bond-length asymmetry \(\Delta l=l_2-l_1\) [2208.09640]. Reported values are \(0.117\,{\rm nC/m}\) for MoSSe, \(0.120\,{\rm nC/m}\) for MoSeTe, \(0.125\,{\rm nC/m}\) for MoSTe, \(0.089\,{\rm nC/m}\) for WSSe, \(0.092\,{\rm nC/m}\) for WSeTe, and \(0.114\,{\rm nC/m}\) for WSTe, compared with \(0.032\,{\rm nC/m}\) for MoS\(_2\) [2208.09640]. The enhancement is linked to stronger \(\sigma\!-\!\sigma\) and \(\pi\!-\!\sigma\) interactions as asymmetry increases [2208.09640].

Catalysis was one of the earliest experimentally demonstrated functional gains. For the hydrogen evolution reaction, basal-plane overpotentials at \(10\,{\rm mA/cm^2}\) were reported as \(465\,{\rm mV}\) for MoS\(_2\), \(355\,{\rm mV}\) for MoSe\(_2\), \(305\,{\rm mV}\) for SMoSe, and \(285\,{\rm mV}\) for SeMoS, with Tafel slopes of \(\approx 120\,{\rm mV/dec}\), \(\approx 105\,{\rm mV/dec}\), \(\approx 80\,{\rm mV/dec}\), and \(\approx 70\,{\rm mV/dec}\), respectively [1704.06389]. DFT \(\Delta G_H^\ast\) calculations on vacancy sites gave \(+0.161\,{\rm eV}\) for MoS\(_2\) S vacancies, \(+0.060\,{\rm eV}\) for SMoSe S vacancies, \(+0.063\,{\rm eV}\) for MoSe\(_2\) Se vacancies, and \(-0.007\,{\rm eV}\) for SeMoS Se vacancies, indicating that vacancies together with intrinsic strain and dipole modulate local electronic states toward \(\Delta G_H^\ast \approx 0\) [1704.06389].

Finally, work-function and band-edge engineering expand the heterostructure design space. In anisotropic Janus rhenium dichalcogenides, the work function can be tuned across \(\sim 0.8\,{\rm eV}\); for example, in the selenization series \( {\rm ReS_{2-x}Se_x} \) it varies from \(5.90\,{\rm eV}\) in ReS\(_2\) to \(5.14\,{\rm eV}\) in ReSSe [2109.00071]. In high-throughput studies of Ti- and Zr-based Janus systems, negative \(\Delta H_{\rm mix}\) at \(x=0.5\), \(T_c\) up to \(\approx 640\,{\rm K}\), and continuously tunable band edges were identified as a favorable combination for device design [2204.11223]. Across the literature, the recurring device themes are photovoltaic and photodetector architectures, piezoelectric and flexoelectric energy conversion, nonlinear Hall and spin-Hall transport, orbitronic and spin-orbit-torque platforms, valley-selective optics, and catalysis [2205.04751] [2201.01467] [2208.09640] [2007.07579].

Taken together, these results define Janus monolayer TMDs not as a single material class with one canonical behavior, but as a symmetry-engineered family in which the consequences of chalcogen asymmetry depend sensitively on polytype, orbital composition, defect landscape, and dielectric environment. The most robustly established principles are the lowering of symmetry from symmetric \(MX_2\) parents, the emergence of a built-in out-of-plane dipole, orbital-selective band reconstruction, and the resulting access to functionalities that are weak or absent in centrosymmetric or mirror-symmetric monolayer TMDs.

Source: https://www.emergentmind.com/topics/janus-monolayer-transition-metal-dichalcogenides-tmds