Dipolar Interfacial Excitons in Lateral Semiconductor Heterostructures

Abstract

One-dimensional (1D) quantum systems are a cornerstone of many-body physics. However, their realization in solids has traditionally relied on top-down methods, which are limited by structural disorder and coarse confinement. Here, we demonstrate a fundamentally distinct route: the emergence of 1D quantum matter at the atomically sharp interface between monolayer semiconductors. Using lateral MoSe2-WSe2 heterostructures, we identify interfacial excitonic quasiparticles that are bound to the crystal junction. Photoluminescence spectroscopy resolves these excitons into a ladder of discrete states, establishing nanoscopic 1D confinement at length scales of 3 nm. These excitons possess exceptional large permanent in-plane electric dipole moments exceeding e x 2 nm, and exhibit micron-scale, highly anisotropic diffusion confined to the interface. Crucially, the lateral geometry enables dynamic, in-situ reconfiguration of the exciton's internal structure. By introducing electrostatic doping, we demonstrate a collapse of the dipole moment and a 20-fold reduction in radiative lifetime. This structural tunability establishes lateral interfaces as a uniquely powerful platform for the 'bottom-up' engineering of 1D quantum matter. By enabling the dynamic tuning of wavefunctions within a single atomic monolayer, this work opens a scalable route toward 1D excitonic circuits and strongly correlated 1D bosonic phases.

Figures (4)

• Figure 1: Lateral interfaces and the interfacial dipolar exciton ($X_{\mathrm{LI}}$). a, Illustration of the 1D lateral interface between 2D semiconductors MoSe$_2$ and WSe$_2$, hosting interfacial excitons. b, Micrograph of CVD-grown monolayer lateral heterostructures (LHS) of MoSe$_2$ and WSe$_2$. c, High Angle Angular Dark-Field Scanning Transmission Electron Microscopy (HAADF-STEM) image confirming the atomic sharpness of the interface. d, Band edge diagram illustrating the Type-II alignment, confining the electron in MoSe$_2$ and the hole in WSe$_2$. $X_{\mathrm{Mo}}$ and $X_{\mathrm{W}}$ denote the bulk direct excitons. The predicted state $X_{\mathrm{LI}}$ is the interfacial dipolar exciton, where the separated charges remain Coulomb-bound at the junction. e, Schematic of the device architecture: h-BN (top and bottom are 25 nm thick) encapsulated monolayer LHS, featuring bottom (BG) and split top gates (SG1, SG2) for control over carrier density and in-plane electric field.
• Figure 2: The interfacial exciton in the lateral heterostructure. a, Normalized PL spectra at 4 K from the 2D bulk MoSe$_2$ (red) and from the MoSe$_2$-WSe$_2$ interface (blue). A new state, $X_{\mathrm{LI}}$, emerges at the interface at $\approx 1.53\,\mathrm{eV}$ ($\sim 100\,\mathrm{meV}$ below the 2D neutral exciton $X_{\mathrm{Mo}}$). b, Micrograph of the device: The LHS flake, interface, and gate electrodes (BG, SG) are outlined. c, Position-resolved PL integrated over $X_{\mathrm{LI}}$ demonstrates that the state exists exclusively at the 1D interface. d, Time-correlated single-photon counting (TCSPC) trace of the $X_{\mathrm{LI}}$ emission. Fitting the long-time decay yields a lifetime of $\tau = 14.6\,\mathrm{ns}$. e, Spatial PL diffusion map of the $X_{\mathrm{LI}}$ state obtained by wide-field imaging. The white contour is the Full Width at Half Maximum (FWHM) of the emission spot. f, Line cuts of the diffusion profile along the short ($x$) and long ($y$) axes. Gaussian fits confirm pronounced anisotropic exciton diffusion, yielding a diffusion length of $L_D \approx 0.7$ µm along $y$ and an estimated diffusion coefficient of $D_y \sim 0.4\,\mathrm{cm}^2\mathrm{s}^{-1}$.
• Figure 3: Quantized 1D Motion and Permanent Dipole Moment of the Interfacial Exciton. a, Low-power PL spectra showing the $X_{\mathrm{LI}}$ state resolved into a ladder of narrow, discrete quantum peaks -- an unambiguous signature of motional quantization. b, Energy of the discrete states ($E_n$) versus state index ($n$). A linear fit (blue dashed line), $E_n = E_0 + \hbar\omega_x n$, yields a harmonic level spacing of $\hbar\omega_x = 7.2 \pm 0.5\,\mathrm{meV}$, implying a transverse confinement length of $\ell_x \approx 2.8\,\mathrm{nm}$. c, Illustration of the exciton COM wavefunctions confined in a parabolic potential imposed by the band offset. d, PL spectra as a function of the in-plane electric field ($\vec{F}_x$) applied perpendicular to the interface via the split gates. The clear, systematic linear shift (Stark shift) of the peak energies is the definitive signature of a permanent in-plane dipole moment. Linear fits (dashed lines) yield the dipole length $d_\mathrm{e\text{-}h}^{(n)}$. e, Schematic illustrating the reduction of extracted dipole lengths with increasing energy, from $d_\mathrm{e\text{-}h} \approx 2.2\,\mathrm{nm}$ for the lowest state to $\approx 1.6\,\mathrm{nm}$ for higher COM states.
• Figure 4: Dynamic Control over Dipole Moment and Lifetime via Electrostatic Doping. a, PL spectra as a function of the bottom gate voltage ($V_\mathrm{{BG}}$) show a large, systematic blue shift ($\sim 70\,\mathrm{meV}$) of the $X_{\mathrm{LI}}$ state upon charge doping. b, Measured $X_{\mathrm{LI}}$ lifetime ($\tau$, blue circles, left axis) on the n-doping side reduces by more than an order of magnitude from $15\,\mathrm{ns}$ at neutrality to $800\,\mathrm{ps}$ at $V_{\mathrm{BG}}=2\,\mathrm{V}$. The corresponding dipole length ($d_\mathrm{e\text{-}h}$) extracted from $\tau$ (right axis) demonstrates dynamic tunability from $2\,\mathrm{nm}$ to $1.4\,\mathrm{nm}$. c, Extracted $X_{\mathrm{LI}}$ center energy (orange points) as a function of $V_\mathrm{{BG}}$. The blue curve represents the computed energy based on the DC Stark shift model, using the dipole lengths extracted independently from the lifetime data in (b) and simulated electric fields (e). The excellent qualitative agreement confirms the model linking lifetime, dipole collapse, and spectral energy. d, Schematic illustrating the Type-II alignment under doping: charge accumulation generates an in-plane electric field ($F_{\mathrm{x}}$) across the interface that opposes the permanent dipole moment ($\vec{p}$) of the exciton. e, Finite-element simulations confirming the concentration of strong in-plane electric fields near the interface upon doping, which drives the dipole collapse.