2D Janus Materials: Tailoring Asymmetry

• 2D Janus materials are atomically thin crystals with chemically distinct faces that break out-of-plane symmetry to create intrinsic dipoles and novel functionalities.
• They are engineered via asymmetric substitution in layered parent compounds and validated through phonon and AIMD stability tests, though experimental synthesis remains challenging.
• The inherent polarity drives tunable band gaps, enhanced spin–orbit interactions, and strong piezoelectric and photocatalytic responses, opening avenues for advanced device applications.

2D Janus materials are atomically thin crystals whose two faces are chemically different, so the layer loses out-of-plane mirror symmetry and acquires an intrinsic polarity, a built-in electrostatic potential difference, and symmetry-allowed responses absent in symmetric counterparts. In most cases this Janus character is created by replacing one outer atomic plane of a layered parent compound, as in $X$-$M$-$Y$ monolayers derived from $X$-$M$-$X$ parents, but the concept has broadened to include asymmetric multi-atomic-layer crystals, Li-substituted Janus transition-metal chalcogenides, and even purely carbon allotropes with a non-chemical Janus configuration (Guo et al., 2022, Yu et al., 2023, Li et al., 5 Feb 2025).

1. Structural definition and materials classes

The canonical Janus construction in 2D materials is top-bottom chemical inequivalence. In conventional Janus transition-metal dichalcogenides, one chalcogen layer in an $X$-$M$-$X$ monolayer is replaced by a different chalcogen to form $X$-$M$0-$M$1, breaking the $M$2 symmetry and creating an intrinsic dipole. That same logic has been extended to many other structural motifs. Group IV–V Janus monolayers can be generated from the symmetric four-atomic-layer SnSb monolayer with $M$3–$M$4–$M$5–$M$6 stacking by one-sided substitution, producing compounds such as Ge$M$7SbAs, Sn$M$8SbAs, GeSnSb$M$9, GeSnAs$Y$0, and SbGe–SnAs; these monolayers retain the four-layer motif but lose the equivalence of the two outer atomic planes (Lu et al., 2024). Septuple-atomic-layer Janus $Y$1 monolayers place an $Y$2 slab between unlike Si–N and Ge–N bilayers, lowering the parent symmetry from No. 187 to No. 156 (Guo et al., 2020). Janus Si dichalcogenides, derived from monolayer $Y$3, lower the symmetry from inversion-symmetric $Y$4 to polar $Y$5 and retain a four-layer Si-based framework (Guo et al., 2022).

The same Janus principle also operates in more chemically diverse families. VOXY monolayers are obtained from multiferroic VOX$Y$6 parents by replacing one halogen layer, lowering the symmetry from Pmm2 to Pm while preserving the in-plane V off-centering responsible for ferroelectricity (Mahajan et al., 2022). Hexagonal In$Y$7 monolayers adopt a triple-layer $Y$8-In-$Y$9 sequence with space group P3m1 and point group $X$0, but their lowest-energy hexagonal registry depends on composition: Phase I is favored for InSCl, InSBr, InSI, InSeCl, InSeBr, and InSeI, whereas Phase II is favored for InTeCl, InTeBr, and InTeI (Liu et al., 31 Aug 2025). Janus nickel dichalcogenides NiSSe, NiSTe, and NiSeTe preserve the 2H hexagonal lattice of their NiX$X$1 parents but develop inequivalent Ni–X and Ni–Y bonds, with NiSSe showing $X$2 Å, $X$3 Å, and thickness $X$4 Å (Sengupta, 2021). Janus $X$5-PdXY monolayers are structurally distinct again: rather than a hexagonal TMD-like network, they consist of connected four- and six-membered rings arranged parallel to the $X$6-axis as a helical chain, which makes $X$7 and produces pronounced in-plane anisotropy (Jakhar et al., 2023).

A broader point is that Janus asymmetry is not restricted to chemically substituted trilayers. Hexagonal Janus MoSeLi replaces one Se layer of MoSe$X$8 by Li and remains a genuine Janus monolayer in the paper’s formulation, even though it is also described as Li-decorated (Seeyangnok et al., 2024). Janus-graphene goes further by proposing an intrinsic non-chemical Janus configuration in a purely sp$X$9-hybridized carbon monolayer: the two faces are chemically identical, yet out-of-plane symmetry is broken by unilateral geometric protrusion generated by a 4–6–8 ring topology in a $M$0 primitive cell with 12 carbon atoms (Yu et al., 2023). This expansion of the concept makes “2D Janus material” a symmetry-based category rather than a narrowly chemical one.

2. Stability, disorder, and realizability

A recurrent theme in the literature is that Janus symmetry breaking does not guarantee that the corresponding monolayer is the most stable arrangement of a given composition. Many works establish local stability through phonons, elastic criteria, or AIMD. Janus Si dichalcogenides satisfy the 2D Born criteria $M$1 and $M$2, show no imaginary phonons for the representative Si$M$3SeTe monolayer, and remain intact in AIMD up to 1000 K over 8 ps (Guo et al., 2022). Cr$M$4TeX and V$M$5AsP satisfy the adopted 2D mechanical stability conditions and have no imaginary phonons (Ma et al., 2022, Ma et al., 2022). NiSSe has no negative phonon bands, whereas NiSTe and NiSeTe are dynamically unstable (Sengupta, 2021). In the In$M$6 family, all members except InSeCl are dynamically stable; InSeCl shows an imaginary mode at K (Liu et al., 31 Aug 2025).

A common misconception is that phonon stability is sufficient to establish practical realizability. Direct comparison with competing atomic arrangements shows otherwise. For MoSSe, SnSSe, PtSSe, In$M$7SSe, and GaInSe$M$8, the ideal ordered Janus monolayer is less energetically favorable than less ordered allotropes of the same composition, because different metal–S and metal–Se bond lengths produce structural frustration inside the layer (Boukhvalov, 2022). In 2H-MoSSe, the Janus monolayer lies 35.91 meV/unit above the hf1 ground-state arrangement, and 6×6×1 checks still place disordered configurations 32.4 meV and 30.8 meV lower than Janus per formula unit (Boukhvalov, 2022). The same study estimates that the decrease in configurational entropy on going from hf1/hf2 to Janus is approximately 0.5 meV/K per formula unit, giving an entropic contribution of approximately 148 meV/unit at room temperature for 18-unit supercells and slightly higher than 184 meV/unit for 36-unit supercells (Boukhvalov, 2022). This is why the discrepancy between bright theoretical property predictions and limited experimental realization cannot be reduced to synthetic imperfection alone.

The electronic-structure origin of stability has also been analyzed across a much wider chemical space. For 84 candidate 1H-phase Janus transition metal dichalcogenides, only 45 are stable, and the stability depends strongly on the transition-metal group. Group-VIB-based monolayers are the most robust, with MoXY and WXY more stable than CrXY, consistent with the successful synthesis of MoSSe and WSSe (Li et al., 5 Feb 2025). The paper attributes this group dependence to competition between metal–ligand ionic bonding, ligand–ligand covalent bonding, and high-energy $M$9-electron orbital splitting, and organizes stable systems by $X$0, $X$1, and $X$2 frontier-electron configurations stabilized by molecular orbital splitting, spin polarization splitting, or crystal-field splitting, respectively (Li et al., 5 Feb 2025). It further proposes an electron-compensation strategy: hydrogenation stabilizes H-ScOS, H-ScOSe, H-ScOTe, H-ZrOTe, H-VOS, and H-VOTe by shifting the occupancy of frontier bonding and antibonding states into one of the stable configurations (Li et al., 5 Feb 2025).

Substrates add another layer to the stability problem. In a dataset of 1147 Janus 2D–bulk heterostructures built from 51 Janus monolayers and 19 elemental metallic cubic substrates, 828 configurations are thermodynamically stable and 318 of the 438 Janus–substrate pairs have negative adsorption energy (Boland et al., 7 Feb 2025). The stabilization criterion is written as

$X$3

with $X$4 indicating support-assisted stabilization (Boland et al., 7 Feb 2025). Feature importance from random-forest models shows that bulk properties dominate both binding energy and interfacial $X$5-separation; bulk electronegativity is the top feature for binding energy, and bulk surface energy is the top feature for $X$6-separation (Boland et al., 7 Feb 2025). This suggests that practical Janus realization is often an interface design problem rather than a freestanding-monolayer problem.

3. Internal electrostatics and electronic structure

Broken out-of-plane symmetry generates intrinsic electrostatics that are unusually strong in 2D Janus systems. In the four-atomic-layer group IV–V Janus monolayers derived from SnSb, the intrinsic dipole moments range from about 0.034 to 0.113 D per unit cell, whereas the parent SnSb has no such intrinsic asymmetry (Lu et al., 2024). In Janus Si dichalcogenides, the effect is even larger: for Si$X$7SeTe, the planar-averaged electrostatic potential gives $X$8 eV and an intrinsic polar field $X$9 V/Å, more than twice the $X$0 V/Å quoted for monolayer MoSSe (Guo et al., 2022). In NiSSe, ELF and total-potential maps directly show stronger electron localization around Se than around S, and more negative total potential around Ni and Se than around S, indicating top-bottom electrostatic asymmetry even though an explicit dipole moment is not reported (Sengupta, 2021).

These internal fields reshape the band edges and, in some families, enable exceptionally strong field tunability. For Janus group IV–V monolayers, the band-gap Stark shift is governed by the vertical separation $X$1 between the charge centers of the VBM and CBM states. The paper first writes

$X$2

and then introduces an induced-field correction through $X$3 with $X$4, yielding a screened effective field for quantitative agreement with HSE06 (Lu et al., 2024). Physically, Janus polarity spatially decouples the VBM and CBM across the layer thickness; in the studied field range of $X$5 to $X$6 V/Å, Sn$X$7SbAs shows a band-gap variation of 134 meV, about 30 times larger than the 4 meV variation in symmetric SnSb (Lu et al., 2024). The Janus semiconductors generally maintain $X$8 Å across the field range, whereas SnSb remains below about 0.5 Å and changes sign (Lu et al., 2024). The microscopic message is that strong Stark tunability is not a generic response to broken symmetry; it requires a large built-in real-space separation of band-edge states.

Spin-orbit-coupled electronic structure also changes qualitatively under Janus asymmetry. In Janus Si dichalcogenides, the conduction-band doublet near $X$9 splits by about 170 meV in Si$M$0SeTe, and the spin texture contains a sizable out-of-plane $M$1 component that cannot be captured by the linear Rashba Hamiltonian alone; the $M$2 model therefore contains a cubic $M$3 term associated with hexagonal warping (Guo et al., 2022). In Janus $M$4 monolayers, the intrinsic vertical electric field activates Rashba-type splitting in the valence bands around $M$5, with $M$6 eVÅ for $M$7 and $M$8 eVÅ for $M$9, while SOC together with inversion-symmetry breaking also produces valley polarization at the conduction-band edge (Guo et al., 2020). In the broader 1H Janus dichalcogenide design study, ScBrI is predicted to show ferromagnetic valley properties with $X$0 meV (Li et al., 5 Feb 2025).

The electronic outcomes are not uniform across families. Some Janus monolayers are indirect-gap semiconductors, such as $X$1, $X$2, Janus Si dichalcogenides, and $X$3-PdXY (Guo et al., 2020, Guo et al., 2022, Jakhar et al., 2023). Others are direct-gap or nearly direct-gap under specific compositions or SOC conditions, such as InTeCl, InTeBr, and InTeI, or Si$X$4SeTe after SOC shifts the CBM from $X$5 to $X$6 (Liu et al., 31 Aug 2025, Guo et al., 2022). There are also metallic or semi-metallic Janus systems, including NiSSe, MoSeLi, and the metallic SbSn–GeAs case in the group IV–V set (Sengupta, 2021, Seeyangnok et al., 2024, Lu et al., 2024). The common thread is not a particular band topology, but the electrostatic and symmetry reorganization caused by two inequivalent faces.

4. Ferroic, spintronic, and superconducting phases

Janus asymmetry has been combined with intrinsic magnetism most clearly in the Cr$X$7TeX, V$X$8AsP, and VOXY families. Janus Cr$X$9TeP, Cr$X$0TeAs, and Cr$X$1TeSb are predicted intrinsic ferromagnetic half-metals, with spin-down gaps of 2.84, 2.96, and 2.64 eV at PBE and HSE06-corrected values of 3.70, 3.84, and 3.53 eV; the half-metallic gaps are 1.25, 1.01, and 0.56 eV (Ma et al., 2022). Monte Carlo simulations based on a Heisenberg model give Curie temperatures of approximately 583, 608, and 597 K, and the ferromagnetic half-metallicity persists under biaxial strain from $X$2 to $X$3 (Ma et al., 2022). V$X$4AsP is another intrinsic ferromagnetic half-metal, with a half-metallic gap of 0.38 eV, a spin gap of 1.34 eV, MAE of 177.39 $X$5eV, and $X$6 K at zero strain; compressive strain raises $X$7 to 130 K at $X$8, whereas 4.9% tensile strain drives a ferromagnetic-to-antiferromagnetic transition (Ma et al., 2022).

VOXY monolayers demonstrate that Janus engineering can be applied to already multiferroic parents rather than only to nonmagnetic semiconductors. Derived from VOX$X$9, the Janus VOXY family preserves the in-plane V off-centering and therefore the in-plane ferroelectric polarization, while adding a finite out-of-plane static polarization and Janus-enabled out-of-plane piezoelectricity (Mahajan et al., 2022). In the collinear DFT treatment, all Janus VOXY monolayers except VOClBr are ferromagnetic; VOClBr is AFM3 (Mahajan et al., 2022). Their in-plane spontaneous polarizations span $M$00–$M$01 C/m, and the switching path FE $M$02 AFE $M$03 FE is lower in energy than the FE $M$04 PE route throughout the family (Mahajan et al., 2022). This is one of the clearest examples of Janus asymmetry augmenting a preexisting ferroic order rather than creating one from a symmetric parent.

Superconductivity enters the Janus field through chemically unconventional derivatives. Hexagonal 2H-MoSeLi, formed by replacing one Se layer of MoSe$M$05 with Li, is metallic and dynamically stable, with total electron-phonon coupling $M$06, $M$07 meV, and $M$08 K from Allen–Dynes (Seeyangnok et al., 2024). Anisotropic Migdal–Eliashberg calculations raise the gap-closing temperature to 5.7 K and show two-gap superconductivity: at 4 K the larger gap spans 0.69–0.80 meV on the outer Fermi-surface pockets, while the smaller gap is approximately 0.47–0.53 meV on the inner pockets (Seeyangnok et al., 2024). The multigap state is tied directly to the distinct orbital character of the two Fermi-surface sheets, so in this case Janus asymmetry is not merely a polar perturbation but part of a route to a multisheet metallic electronic structure.

Taken together, these families show that Janus asymmetry is compatible with half-metallicity, multiferroicity, valley polarization, and phonon-mediated superconductivity. This suggests that Janus design is best viewed as a symmetry-lowering platform that can coexist with very different many-body ground states, provided the underlying electron count and structural stability are appropriate.

5. Electromechanical, optical, excitonic, and photocatalytic functionality

Piezoelectricity is one of the most systematic consequences of Janus symmetry breaking. In Janus Si dichalcogenides, the broken inversion symmetry activates both in-plane and out-of-plane strain coefficients, with

$M$09

The calculated values are $M$10, 62.36, and 40.52 pm/V and $M$11, 0.310, and 0.182 pm/V for Si$M$12SSe, Si$M$13STe, and Si$M$14SeTe, respectively (Guo et al., 2022). In the In$M$15 family the same symmetry logic yields even larger coefficients, with $M$16–155.27 pm/V and $M$17–0.65 pm/V across the stable monolayers (Liu et al., 31 Aug 2025). Their large strain coefficients are assisted by very small 2D Young’s moduli, for example 12.20 N/m for InSeBr (Liu et al., 31 Aug 2025). In VOXY, Janus symmetry activates out-of-plane piezoelectric stress coefficients $M$18 and $M$19 that are absent in VOX$M$20; VOFI reaches $M$21 C/m (Mahajan et al., 2022). A general lesson is that Janus asymmetry does not merely add a dipole: it introduces new tensor elements coupling in-plane strain to out-of-plane polarization.

Optical and excitonic responses are similarly Janus-specific. In MoSSe and WSSe, real-time many-body calculations show that the shift current is strongly suppressed at the A/B excitons around 2 eV but strongly enhanced at the C excitons around 3 eV (Mao et al., 19 Jun 2025). The microscopic reason is spatial rather than purely spectral: A/B excitons involve electron and hole localized around the same atom, whereas the C excitons place the electron and hole on different atomic sites, producing a larger charge-center shift after photoexcitation and hence a stronger photocurrent (Mao et al., 19 Jun 2025). The paper also emphasizes that nonzero in-plane tensor components do not guarantee a net in-plane shift current under unpolarized light, because $M$22 symmetry still imposes cancellations unless the symmetry is further broken by strain, substrate effects, heterostructuring, or nanotube formation (Mao et al., 19 Jun 2025). This is an important corrective to the notion that Janus symmetry automatically removes all optical selection constraints.

Photocatalytic design in Janus systems relies on the interplay of intrinsic dipoles, band offsets, and carrier transport. In Janus-TMDC/TMDC heterobilayers, the sign and magnitude of the work-function difference $M$23 determine whether the system behaves as Type I, Type II, or direct Z-scheme. The screening study of 20 Janus–TMDC heterobilayers identifies several Z-scheme candidates; among the best are WSe$M$24-SWSe, WSe$M$25-TeWSe, and WS$M$26-SMoSe, with calculated solar-to-hydrogen efficiencies of 30.71%, 30.72%, and 33.24%, respectively (Bao et al., 2024). The same work also shows that Janus polarization improves charge separation but can simultaneously enhance Fröhlich scattering from polar optical phonons, so mobility and internal fields must be optimized together rather than independently (Bao et al., 2024).

The hexagonal In$M$27 monolayers provide a complementary monolayer route to piezo-photocatalysis. All stable In$M$28 members have band edges straddling the water redox potentials at pH = 0, intrinsic vertical polarization that produces an intralayer polarization field $M$29, and exciton binding energies of 0.44–0.78 eV from $M$30+BSE (Liu et al., 31 Aug 2025). Their electron mobility spans 101–899 cm$M$31/V/s, and the authors argue that the strong vertical piezoelectric response can further enhance $M$32 under in-plane stress to facilitate separation of photogenerated carriers (Liu et al., 31 Aug 2025). Heterojunction engineering then combines these traits: the vdW InSI/InSeBr bilayer is reported to show a direct Z-scheme charge-transfer pathway, enhanced redox ability, strong visible-light absorption, high carrier mobility, and excellent photocorrosion resistance (Liu et al., 31 Aug 2025). A plausible implication is that Janus asymmetry is especially powerful when paired with interface engineering, because intralayer and interlayer fields can be designed cooperatively.

6. Transport, applications, and open problems

Transport phenomena in Janus materials are unusually diverse, ranging from high-mobility piezoelectric semiconductors to highly anisotropic thermoelectrics. In $M$33, deformation-potential theory gives electron mobilities of 5205.14 and 6573.25 cm$M$34V$M$35s$M$36 for $M$37 and 7046.80 and 8767.94 cm$M$38V$M$39s$M$40 for $M$41 along $M$42 and $M$43, while the Janus asymmetry activates a small but finite out-of-plane piezoelectric response $M$44 and $M$45 pm/V absent in the symmetric parents (Guo et al., 2020). In the thermoelectric 2H-ISbTe monolayer, moderate biaxial tensile strain up to 4% preserves phonon stability, increases the PBE band gap from 1.04 to 1.23 eV at 2% strain according to the internal evidence of the paper, and raises the room-temperature p-type $M$46 from approximately 0.87 to approximately 1.31 at 2% tensile strain (Kumari et al., 2024). In anisotropic Janus $M$47-PdSSe, $M$48-PdSeTe, and $M$49-PdSTe, the $M$50-direction lattice thermal conductivity is only 0.80, 0.94, and 0.77 W m$M$51 K$M$52 at 300 K, and the corresponding p-type $M$53 values reach 0.68, 0.86, and 0.68 at 300 K and 2.21, 4.09, and 3.63 at 800 K (Jakhar et al., 2023). In these systems the electronic advantage comes from degenerate top valence bands, whereas the phononic advantage comes from low group velocities and strong scattering phase space in a low-symmetry β framework (Jakhar et al., 2023).

At the application level, Janus materials are now studied not only as isolated monolayers but also as components of heterostructures. The Janus 2D–bulk heterostructure study shows that substrate choice is the primary design variable for interface stabilization, with bulk electronegativity dominating binding energy and bulk surface energy dominating equilibrium separation (Boland et al., 7 Feb 2025). This is consistent with the broader lesson from disorder studies: the realizable Janus phase is often the one that is stabilized by its environment, whether that environment is a substrate, a multilayer stack, or a chemical compensant (Boukhvalov, 2022, Boland et al., 7 Feb 2025). This suggests that future Janus-device design should co-optimize asymmetry, substrate chemistry, disorder tolerance, and field screening rather than treating them as separable modules.

Several open problems remain systematic across the literature. Experimental realization still lags far behind theoretical enumeration, especially outside the benchmark MoSSe- and WSSe-like compositions (Boukhvalov, 2022, Li et al., 5 Feb 2025). Many calculations remain at the PBE or HSE06 level, omit SOC in heavy-element systems, or rely on idealized models such as the constant scattering time approximation, rigid-band doping, or defect-free freestanding monolayers (Kumari et al., 2024, Jakhar et al., 2023). Field strengths commonly used in Stark-effect or gating studies, such as $M$54 V/Å, are standard in theory but may be difficult in some device geometries (Lu et al., 2024). In photocatalysis, pristine basal planes often remain poor HER surfaces even when band alignment is favorable, as illustrated by the large positive $M$55 values in Janus-TMDC heterobilayers (Bao et al., 2024). The broader implication is that Janus asymmetry is a powerful enabling condition, but not a sufficient descriptor of functionality by itself.

Taken as a whole, the field now points toward a set of convergent design rules. Strong out-of-plane chemical or geometric asymmetry can generate intrinsic electric fields, new piezoelectric tensor components, and large Stark or shift-current responses. Frontier-state localization on opposite sides of the layer enhances electrostatic tunability. Stable realization often requires either electronically favorable $M$56, $M$57, or $M$58 frontier configurations, compensation chemistry, or substrate-assisted stabilization. And the most effective Janus architectures appear to be those in which asymmetry, phase stability, transport, and interface environment are engineered together rather than sequentially.