Monolayer MoSe₂: Properties & Applications

• Monolayer MoSe₂ is a single-layer molybdenum diselenide that exhibits direct-gap semiconductor behavior with pronounced excitonic and valley-coupled optical properties.
• Its diverse synthesis methods yield varied morphologies and defect profiles that critically influence exciton lifetimes, charge transport, and optical performance.
• The material’s environment-tunable interactions via charge doping, strain, and substrate effects enable applications in exciton-polaron physics, optomechanics, and quantum transport.

Monolayer MoSe$_2$ is the single-layer limit of molybdenum diselenide, a group-VI transition-metal dichalcogenide whose monolayer form is a direct-gap semiconductor with strong excitonic resonances, spin-valley-coupled optical selection rules, and pronounced sensitivity to dielectric environment, charge density, strain, and photonic boundary conditions. Across the literature, monolayer MoSe$_2$ has been studied as an excitonic model system, a charge-tunable many-body platform, an ultrathin optical reflector, a strong-coupling medium, a nanomechanical resonator, and a host for localized emitters and defect-mediated one-dimensional metallic states (Lundt et al., 2016, Back et al., 2017, Goldstein et al., 2020, Liu et al., 16 Feb 2025). Its reported behavior is notably sample- and platform-dependent, which is not incidental: substrate choice, encapsulation, growth protocol, gating geometry, and disorder level directly reorganize the balance among neutral excitons, trions, polarons, intervalley scattering, diffusion, and radiative coupling.

1. Crystal structure, synthesis, and morphology

Monolayer MoSe$_2$ has been realized by mechanical exfoliation, chemical vapor deposition, and molecular beam epitaxy, with markedly different structural outcomes. Exfoliated flakes on SiO$_2$/Si and sapphire underpin much of the spectroscopy and ultrafast literature, while large-area epitaxial growth has been pursued using both conventional and unconventional Mo sources. One epitaxial route used a homemade incandescent molybdenum source and reported large-area single-crystalline monolayer MoSe$_2$ supported by Raman, PL, RHEED, and ARPES; for the monolayer candidate on sapphire, the reported PL peak was $\sim 783$ nm with $\sim 14$ nm linewidth, and ARPES placed the valence-band apex at K $0.21$ eV above that at $\Gamma$ (Cheng et al., 2015). A later MBE strategy on exfoliated hBN used a two-step van der Waals epitaxy protocol, with low-temperature nucleation at $300^\circ$C followed by growth at $_2$0C, to obtain monolayer-thick grains $_2$1 nm to $_2$2m in size and low-temperature neutral-exciton PL linewidth down to $_2$3 meV at $_2$4 K, together with measurable room-temperature PL and reflectivity signatures of A, B, and charged excitons (Vergnaud et al., 2024).

Structural inhomogeneity is a recurrent theme. Tip-enhanced PL imaging on CVD-grown monolayer MoSe$_2$5 revealed edge-related blue shifts and strong correlation between emissive regions and AFM topography, while aligned particles at crystal boundaries were non-emissive in the MoSe$_2$6 PL window (Tang et al., 2017). Local strain can also generate qualitatively different optical objects: spectrally sharp quantum-dot-like emitters were found at wrinkles in exfoliated monolayer MoSe$_2$7, both on gold and in suspended devices, with measured linewidths typically $_2$8–$_2$9eV and local 2D neutral-exciton redshifts of about $_2$0 meV at the wrinkle sites (Branny et al., 2016).

Defect topology in epitaxial films can be even more drastic. MBE-grown monolayer MoSe$_2$1 on HOPG was shown to contain a dense triangular network of inversion domain boundaries, and STM/STS identified these boundaries as one-dimensional metallic mid-gap channels inside an otherwise semiconducting monolayer with an STS gap of about $_2$2 eV (Liu et al., 2014). This established that “monolayer MoSe$_2$3” is not, in practice, a single structural archetype: the same chemical monolayer can behave as a clean excitonic membrane, a strain-patterned localization landscape, or a 2D semiconductor perforated by an interconnected metallic defect network.

2. Excitonic spectrum and optical spectroscopy

The optical spectrum of monolayer MoSe$_2$4 is dominated by the neutral A exciton and negatively charged trion in n-type conditions, with reported low-temperature energies depending on substrate and dielectric environment. At $_2$5 K on $_2$6 nm SiO$_2$7/Si, the neutral exciton and trion were reported at $_2$8 eV and $_2$9 eV, respectively, and the neutral-exciton PL intensity oscillated with excitation energy with an average period $_2$0 meV, matching the LA(M) phonon energy (Chow et al., 2017). In high-resolution differential reflectivity at $_2$1 K, the neutral exciton and trion were identified at $_2$2 eV and $_2$3 eV, while ordinary PL and absorption linewidths remained on the meV scale (Schaibley et al., 2015). In encapsulated transport-quality samples, the A:1s exciton appears near $_2$4 eV, A:2s near $_2$5 eV, B:1s near $_2$6 eV, and B:2s near $_2$7 eV (Goldstein et al., 2020).

Temperature evolution is strong but not destructive. Reflectance measurements tracking the A exciton from cryogenic temperature to room temperature found a redshift described by the Varshni relation

$_2$8

with fitted parameters $_2$9 eV, $_2$0 eV/K, and $_2$1 K (Lundt et al., 2016). The same study reported linewidth broadening from about $_2$2 meV at $_2$3 K to about $_2$4 meV at $_2$5 K, together with an almost $_2$6 reduction in peak amplitude but only about $_2$7 reduction in the integrated absorption proxy used for oscillator strength (Lundt et al., 2016). This distinction matters because it separates thermal broadening from true loss of radiative weight.

Exciton relaxation spans an unusually wide dynamic range. Coherent nonlinear spectroscopy revealed population-pulsation resonances as narrow as $_2$8eV and below $_2$9eV, implying exciton-population dynamics of at least $\sim 783$0 ns and a long-lived state with lifetime $\sim 783$1 ns, even though PL and absorption linewidths are three orders of magnitude broader (Schaibley et al., 2015). By contrast, four-wave-mixing transient grating spectroscopy on exfoliated monolayers at room temperature found free-exciton population decay around $\sim 783$2 ps but intervalley scattering faster than the $\sim 783$3 fs time resolution, leading to negligible valley polarization in steady-state PL (Kuhn et al., 2019). Phonon-assisted relaxation studies further showed that on-phonon-resonance excitation narrows the neutral-exciton homogeneous linewidth and lengthens the measured lifetime, whereas off-resonant excitation generates a stronger bath of long-wavelength acoustic phonons, broadening the homogeneous linewidth and shortening the lifetime (Chow et al., 2017).

These measurements collectively exclude any single-timescale description. Monolayer MoSe$\sim 783$4 supports bright excitons, dark or trapped long-lived reservoirs, intervalley depolarization channels, and mode-specific exciton-phonon scattering, all within the same nominal monolayer.

3. Relaxation, diffusion, and electronic transport

Exciton transport and electronic transport probe different sectors of the same material. In hBN-encapsulated monolayer MoSe$\sim 783$5, PL imaging combined with numerical solutions of the 2D diffusion equation,

$\sim 783$6

yielded $\sim 783$7 at $\sim 783$8 K and an exciton mobility defined through $\sim 783$9 that exceeded $\sim 14$0 at low temperature and was $\sim 14$1 at room temperature (Hotta et al., 2020). The reported $\sim 14$2 dependence was interpreted as qualitatively different from GaAs quantum wells because the hBN/MoSe$\sim 14$3/hBN structure suppresses surface-roughness and charged-impurity scattering (Hotta et al., 2020).

On exfoliated monolayers at room temperature, transient-grating spectroscopy extracted a much smaller exciton diffusion constant, $\sim 14$4, after explicitly including exciton-exciton annihilation in the diffusion-reaction equation,

$\sim 14$5

with annihilation constant $\sim 14$6 (Kuhn et al., 2019). The contrast with the encapsulated value is not contradictory on its face; it reflects different sample classes, temperatures, and analysis frameworks. It does, however, imply that disorder and many-body loss channels remain first-order variables for transport.

Low-temperature charge transport has recently entered a much cleaner regime. A triple-gate monolayer MoSe$\sim 14$7 device with Bi/Au contacts reported contact resistance below k$\sim 14$8 down to $\sim 14$9, maximum Hall mobility of approximately $0.21$0, and quantum mobility of roughly $0.21$1 from SdH onset at $0.21$2–$0.21$3 T (Liu et al., 16 Feb 2025). The same study identified a metal-insulator transition near $0.21$4, single-fold Landau-level degeneracy at low density, two-fold degeneracy with odd/even sequence switching at intermediate density, and four-fold degeneracy above $0.21$5 when the upper spin-split conduction band begins to populate (Liu et al., 16 Feb 2025). Effective masses extracted from Lifshitz–Kosevich analysis changed from $0.21$6 at $0.21$7 to $0.21$8 at $0.21$9, while the ratio $\Gamma$0 scaled approximately as $\Gamma$1, indicating enhanced spin susceptibility at low density (Liu et al., 16 Feb 2025).

This suggests that monolayer MoSe$\Gamma$2 now spans two experimentally accessible limits: a clean excitonic diffusion system in encapsulated optics, and a low-density interacting 2D electron gas in gated transport.

4. Charge tuning, polarons, and ferroelectric electrostatic control

Charge density reorganizes the optical excitations of monolayer MoSe$\Gamma$3 far beyond a simple exciton-trion crossover. In hBN-encapsulated, hole-doped devices, both PL and differential reflection show that the A:1s exciton splits into repulsive and attractive branches identified as exciton polarons rather than simple few-body trions once a Fermi sea is established (Goldstein et al., 2020). For A:1s, the branch separation evolves approximately as

$\Gamma$4

and the repulsive-branch linewidth broadens approximately as $\Gamma$5 (Goldstein et al., 2020). The excited A:2s state is even more sensitive: the authors inferred 1s and 2s binding energies of $\Gamma$6 meV and $\Gamma$7 meV, respectively, and found

$\Gamma$8

together with linewidth broadening of about $\Gamma$9, signaling that the weak-coupling rigid-exciton approximation becomes inadequate for excited Rydberg states (Goldstein et al., 2020). This makes monolayer MoSe$300^\circ$0 a concrete setting in which exciton-polaron phenomenology extends into excited excitonic manifolds.

Electron doping under magnetic field exposes a different many-body sector. In a charge-tunable hBN-encapsulated monolayer, a perpendicular field $300^\circ$1 T drove near-complete occupation of one valley up to $300^\circ$2, with polarization-resolved PL showing $300^\circ$3 in the intermediate-density regime and reflection showing helicity-dependent transfer of oscillator strength from exciton to attractive Fermi polaron (Back et al., 2017). The neutral exciton at $300^\circ$4 had Zeeman splitting $300^\circ$5 meV, corresponding to $300^\circ$6, while at $300^\circ$7 the attractive polaron and repulsive-polaron/exciton branches yielded effective $300^\circ$8-factors of about $300^\circ$9 and $_2$00, respectively (Back et al., 2017). The paper interpreted this as interaction-enhanced giant paramagnetism rather than a single-particle valley Zeeman effect.

A nonvolatile electrostatic variant uses ferroelectric lithium niobate. Monolayer MoSe$_2$01 dry-transferred onto a $_2$02m-thick Z-cut LiNbO$_2$03 substrate with hexagonal domain-inverted inclusions experiences local doping from remnant polarization surface charge, producing strong PL contrast within a single continuous monolayer (Wen et al., 2019). The reported PL inside one ferroelectric domain was enhanced by about $_2$04-fold relative to the opposite domain, with the bright and dark regions aligned to the $_2$05/$_2$06 pattern (Wen et al., 2019). The interpretation is electrostatic rather than structural: the PL peak energy is almost the same on $_2$07 and $_2$08, and little change in peak energy and FWHM was found between MoSe$_2$09 on LN and on SiO$_2$10 (Wen et al., 2019). The relevant spectral logic is the exciton-trion relation

$_2$11

which connects local carrier density to neutral-exciton/trion balance. Temperature then tunes the effect through the pyroelectric response of LN: cooling from $_2$12 K to $_2$13 K quenched the PL in the relevant domain by nearly $_2$14, and the integrated intensities on $_2$15 and $_2$16 converged at low temperature (Wen et al., 2019). For MoSe$_2$17 in this geometry, the dominant evidence is PL and PSI rather than Raman-mode shifts, reinforcing the conclusion that the modulation is electronic in origin.

5. Valley, spin, and helicity-dependent phenomena

Valley response in monolayer MoSe$_2$18 is highly species- and protocol-dependent rather than uniformly large. Under circularly polarized PL excitation, the A exciton showed valley polarization $_2$19 whereas the B exciton reached about $_2$20 at $_2$21 K under near-resonant pumping at $_2$22 eV; the A$_2$23 trion displayed a small negative valley polarization, about $_2$24 at $_2$25 K under off-resonant excitation (Zhang et al., 2015). The analysis was organized by

$_2$26

so the contrast between A and B excitons was attributed mainly to different effective lifetimes rather than different optical selection rules (Zhang et al., 2015). This result directly counters a common overgeneralization from other TMDCs: in monolayer MoSe$_2$27, robust valley polarization is not automatic for the lowest bright exciton.

Electrical and photogalvanic measurements refine this picture. In an hBN-encapsulated MoSe$_2$28 phototransistor, the helicity-dependent photovoltage was fit by

$_2$29

with $_2$30 identifying the circular photocurrent contribution (Quereda et al., 2018). The CPC was strongest at the A-exciton resonance near $_2$31 eV and displayed two bias-dependent regimes: at low $_2$32 V the response was even under $_2$33 and approximately $_2$34, while at higher bias it reversed sign under $_2$35 and followed $_2$36 with $_2$37 (Quereda et al., 2018). The symmetry analysis concluded that reduced device symmetry is required and that Berry-curvature-induced CPGE is not a significant contributor in that geometry because the CPC vanishes at normal incidence (Quereda et al., 2018). This is an important corrective to valley-Hall-style interpretations: helicity-sensitive currents in monolayer MoSe$_2$38 are strongly shaped by excitons, trions, dissociation processes, and contact/device asymmetry.

Strong-coupling photonic environments can partially rescue valley contrast. A Tamm-plasmon structure with a monolayer MoSe$_2$39 on 10 nm GaInP, capped by 80 nm PMMA and 60 nm Au, established strong coupling between the Tamm mode and the trion resonance, with normal-mode splitting $_2$40 meV at positive detuning $_2$41 meV (Lundt et al., 2017). The lower trion-polariton at $_2$42 exhibited circular polarization degree $_2$43, contrasted in the paper with less than $_2$44 for bare MoSe$_2$45 under comparable non-resonant conditions, and the upconverted polariton emission preserved a circular Stokes component $_2$46 (Lundt et al., 2017). The interpretation was dynamical: faster polariton relaxation and depopulation leave less time for intervalley depolarization.

6. Strong light-matter coupling, nanophotonics, and optomechanics

Monolayer MoSe$_2$47 can function as an atomically thin resonant optical element. In a gated graphene/hBN/MoSe$_2$48/hBN/fused-silica heterostructure, the resonant exciton response produced extinction up to $_2$49 and maximum reflection $_2$50, with the simple input-output relations

$_2$51

where $_2$52 is the radiative decay rate and $_2$53 the total linewidth (Back et al., 2017). Reported linewidths were $_2$54–$_2$55 meV, and the inferred $_2$56 lay between $_2$57 and $_2$58 meV (Back et al., 2017). Gate voltage then redistributed oscillator strength among neutral exciton, repulsive polaron, and attractive polaron branches; at fixed photon energy the extinction changed from $_2$59 to $_2$60 when $_2$61 increased from $_2$62 V to $_2$63 V (Back et al., 2017). In this sense, monolayer MoSe$_2$64 operates as an electrically tunable mirror rather than merely an absorber.

The same excitonic robustness supports room-temperature polaritonics in sufficiently short cavities. Using experimentally extracted temperature-dependent exciton energies, linewidths, and oscillator strengths, a coupled-oscillator and transfer-matrix analysis predicted that single-monolayer MoSe$_2$65 in short monolithic dielectric cavities and Tamm structures should remain in the strong-coupling regime up to room temperature, with calculated Rabi splittings of $_2$66 meV at $_2$67 K and $_2$68 meV at $_2$69 K for the dielectric cavity, and $_2$70 meV at $_2$71 K and $_2$72 meV at $_2$73 K for the Tamm structure (Lundt et al., 2016). By contrast, an extended open cavity with low-temperature splitting $_2$74 meV was predicted to lose clearly resolvable strong coupling above about $_2$75 K because exciton broadening overwhelms visibility (Lundt et al., 2016). A plausible implication is that monolayer MoSe$_2$76 is not intrinsically limited to cryogenic polaritonics; the limiting variable is effective cavity length.

Exciton-mediated mechanics provides a third photonic regime. A suspended monolayer MoSe$_2$77 drum over a circular trench $_2$78 nm showed gate-tunable exciton-optomechanical coupling, with the neutral exciton redshifting under gate-induced tensile strain and the mechanical mode exhibiting optical damping, anti-damping, and spring shifts under optical pumping near the exciton resonance (Xie et al., 2021). The analysis used the photothermal-backaction expressions

$_2$79

and extracted $_2$80eV peak-to-peak at $_2$81 V for a $_2$82 mV peak-to-peak AC drive (Xie et al., 2021). The mechanical quality factor was $_2$83 at low gate voltage, decreasing with increasing $_2$84, while the optical resonance retained on-resonance reflection contrast approaching $_2$85 (Xie et al., 2021). This places monolayer MoSe$_2$86 in the small class of materials where excitonic optics and NEMS mechanics can be coupled without a conventional optical cavity.

7. Defects, localized states, and emerging directions

Localization in monolayer MoSe$_2$87 is not limited to weak disorder. Strain-correlated discrete emitters show Zeeman splittings of about $_2$88 meV at $_2$89 T and exciton $_2$90-factors $_2$91 and $_2$92, close to those of the host 2D excitons, supporting the view that these lines originate from confined direct-gap excitonic states rather than unrelated deep defects (Branny et al., 2016). In charge-tunable suspended devices, the same emitters display discrete spectral jumps of about $_2$93 meV as the voltage is swept, consistent with charge-state switching (Branny et al., 2016). At the opposite structural extreme, inversion domain boundaries in MBE films produce a dense 1D metallic network whose STS intensity undulates with fixed $_2$94 nm periodicity attributed to the moiré superlattice and with energy-dependent wavelengths attributed to finite-length quantum confinement (Liu et al., 2014).

Recent work has also extended monolayer MoSe$_2$95 into high-energy hot-carrier physics. Ultrafast transient absorption and spatiotemporal microscopy reported what the authors described as ideal carrier multiplication, with a quantum yield step from $_2$96 below threshold to $_2$97 above threshold at $_2$98, using $_2$99 eV from PL (Kim et al., 31 May 2026). The same study associated this with $_2$00 band nesting near K/K$_2$01, ballistic transport up to about $_2$02 ps, a maximum hot-carrier diffusion coefficient $_2$03, and experimentally inferred hot-carrier velocity $_2$04 m/s (Kim et al., 31 May 2026). Because this is framed as a threshold-ideal CM claim rather than a modest efficiency improvement, it should be read as a frontier result rather than a settled consensus. What it does establish, at minimum, is that monolayer MoSe$_2$05 has become a serious platform for testing whether 2D band structure and hot-carrier transport can jointly suppress carrier-lattice losses.

Taken together, these strands define monolayer MoSe$_2$06 less as a single canonical semiconductor than as a tunable excitonic and many-body material class. Clean, encapsulated monolayers support narrow exciton lines, high exciton mobility, and low-density quantum transport; ferroelectric, magnetic, or photonic environments convert the same monolayer into a nonvolatile optical modulator, a giant-paramagnetic valley system, an atomically thin mirror, a trion-polariton medium, or an exciton-optomechanical resonator; strain and inversion-domain defects generate quantum-dot-like emitters and one-dimensional metallic channels. The unifying feature is not one fixed observable but the unusually large leverage of external environment over the balance among excitonic, valley, transport, and photonic degrees of freedom in a single atomic layer.