betaFeedback

WorksPhysics and Astronomy

Polariton tweezers

Abstract

The non-invasive, deterministic manipulation of charge carriers remains a fundamental challenge in solid-state quantum technologies. Here, we demonstrate the concept of dynamic polariton tweezers capable of trapping and transporting electrons across a two-dimensional electron system via the long-range electrostatic potential of dipolar (d-) exciton-polaritons. By driving a semiconductor microcavity with a specifically engineered, non-collinear three-beam optical pump with TE-polarization, we synthesize a macroscopic polariton lattice with hexagonal symmetry. The spatially periodic field of this d-polariton condensate acts as a reconfigurable trapping landscape for the electrons. We show that introducing targeted frequency detunings between the continuous-wave pumps strictly breaks the spatiotemporal symmetry of the system, converting the stationary macroscopic interference pattern into a dynamically moving conveyor belt. By precisely customizing these frequency detunings, the entire lattice - along with the electrostatically trapped electrons - can be rigidly driven in any arbitrary in-plane direction. This precise phase-matching mechanism bypasses the need for a net kinetic momentum transfer, establishing a robust, optically controlled platform for the long-range routing of charged particles and topological defects in driven-dissipative quantum fluids.

Introduction

Optical tweezers [1, 2] have become a versatile tool for the non-invasive manipulation of mesoscopic objects and isolated particles [3]. Their development has led to diverse specialized implementations, including holographic optical tweezers [4, 5] for complex multi-trap arrays, plasmonic tweezers [6, 7] that confine light below the diffraction limit, near-field photonic crystal traps [8], and optomechanical manipulation platforms [9, 10]. These techniques enable precise spatial control of dielectric particles, biological cells [11], and cold atoms [12]. Extending optical trapping to the direct capture and continuous spatial transport of individual electrons, however, remains challenging because the gradient-force mechanism underlying conventional optical tweezers couples inefficiently to free charge carriers. Their direct interaction with optical fields is instead governed by charge acceleration and ponderomotive forces, which generally require intense fields to produce substantial spatial manipulation [13, 14].

Parallel to the development of optical trapping [10], polaritonics has emerged as a robust solid-state platform for engineering light–matter interactions [15, 16, 17, 18]. Exciton-polaritons are hybrid light–matter quasiparticles formed in semiconductor microcavities through strong coupling between cavity photons and excitons hosted by active semiconductor layers, such as quantum wells (QWs), embedded within the cavity [15]. They can form macroscopically coherent states associated with Bose–Einstein condensation and exhibit strong optical nonlinearities [16, 17, 18]. The realization of dipolaritons (d-polaritons) in microcavities containing coupled QWs [19, 20] has introduced an electrically tunable permanent dipole moment into polaritonic systems [21]. As a result, d-polaritons combine the propagation characteristics inherited from their photonic component with long-range electrostatic interactions arising from spatially separated electron and hole components [20, 22].

Polariton condensates can be spatially controlled using optically induced potentials [23, 24, 25, 26], surface acoustic waves [27], and engineered confinement geometries [28, 29]. Conversely, coupling between polaritons and charge carriers has been explored in several contexts, including condensate-assisted photocurrent generation and electron–polariton drag phenomena [30, 31, 32]. In the latter case, motion of the electronic subsystem transfers momentum to polaritons and modifies their transport. However, to our knowledge, the use of a spatially structured polariton population itself as a moving electrostatic potential for the capture and transport of individual charge carriers has not been demonstrated.

A prerequisite for such a scheme is the ability to create and dynamically control spatially structured polariton density distributions. This capability is already well established through the generation of engineered polariton lattices [29, 33, 34] and the guiding of polariton flows along engineered channels [35, 36]. Furthermore, frequency-detuned optical excitation has been used to create traveling optical lattices [37] and dynamically controlled trapping potentials [38] for polariton condensates. In a dipolariton system, such spatial modulation of the polariton density can be directly translated into an electrostatic potential landscape acting on nearby charge carriers. Making this landscape dynamic therefore provides a route to optically controlled charge transport without relying on direct optical forces on the electron.

Here we theoretically propose a polariton tweezer architecture for the optical trapping and transport of individual electrons via the long-range electrostatic potential induced by a dynamically controlled d-polariton condensate. Using three laser pump beams with non-collinear in-plane wavevectors, we generate a macroscopically occupied d-polariton state in a double-QW microcavity. A controlled frequency detuning between the pump beams transforms the resulting stationary interference pattern into a traveling modulation of the d-polariton density. Owing to the permanent dipole moment of the d-polaritons, this density modulation produces a moving electrostatic potential in an adjacent electronic layer. We show that electrons initially confined within a potential minimum can remain entrained as the polariton lattice translates, providing a route to optically controlled charge transport and reconfigurable on-chip electronic routing.

Results

Dipolariton microcavity platform

The physical system under consideration is schematically shown in Figure 1. It consists of a planar semiconductor microcavity with a central cavity layer sandwiched between upper and lower distributed Bragg reflectors (DBRs), see Figure 1b. The central cavity layer contains two tunnel-coupled quantum wells (QW1 and QW2), supporting spatially indirect excitons (IXs) with electrons and holes residing in different wells. Such excitons are well established in structures with type-II band alignment [39] and in coupled QWs with electric-field-controlled excitonic states [40]. Through electron tunnelling, IXs hybridize with optically active direct excitons, enabling strong coupling to the cavity photon mode and the formation of dipolaritons [40]. The resulting dipolaritons inherit a permanent out-of-plane electric dipole moment from their IX component, with its magnitude determined by the admixture of this component.

Schematic of the dipolariton microcavity and optical excitation geometry

Figure 1. Schematic of the dipolariton microcavity and optical excitation geometry. a, Three-dimensional view of the system excited by three non-collinear coherent laser beams, depicted as semi-transparent cones with corresponding in-plane wavevectors k1,pumpk_{1,\mathrm{pump}}, k2,pumpk_{2,\mathrm{pump}}, and k3,pumpk_{3,\mathrm{pump}}. Their overlap produces a spatially modulated dipolariton condensate, forming an interference lattice in the central region. b, c, Enlarged cross-sectional views of two microcavity configurations, both incorporating tunnel-coupled quantum wells (QW1 and QW2) and an adjacent two-dimensional electron system (2DES). b, Conventional planar microcavity with a central cavity layer bounded by upper and lower DBRs. c, Alternative Tamm-plasmon configuration in which the upper DBR is replaced by a thin gold layer

The system also includes a separate electronic layer hosting a two-dimensional electron system (2DES), as previously considered in electron–polariton hybrid structures [41]. The electronic layer is positioned at distances lel_{\mathrm{e}} and lhl_{\mathrm{h}} from the electron and hole layers of the IX, respectively. An alternative Tamm-plasmon configuration (Figure 1c), in which the upper DBR is replaced by a thin gold layer, may also provide a suitable platform, although metal-induced optical losses can reduce the cavity quality [42, 43].

The optical excitation geometry is schematically illustrated in Figure 1a. The microcavity is driven by three continuous wave Gaussian laser beams (j=1,2,3j=1,2,3), with their spots arranged in a threefold symmetric configuration at a radial distance RR from the origin. The beams have non-collinear in-plane wavevectors kj,pump\mathbf{k}_{j,\mathrm{pump}} with a characteristic magnitude kpumpk_{\mathrm{pump}} and partially overlap in the central region. Their coherent superposition produces a spatially periodic modulation of the dipolariton condensate density, forming an interference lattice.

The polariton condensate is described by the spinor wavefunction ψ(r,t)=(ψ+,ψ−)T\psi(\mathbf{r},t)=(\psi_{+},\psi_{-})^{\mathrm{T}}, whose components correspond to right- and left-circular polarizations. Its dynamics is governed by the coupled driven-dissipative Gross–Pitaevskii equations, written in the frame rotating at the reference pump frequency ω0\omega_0, chosen to be resonant with the TE-polarized mode:

iℏ∂ψ±∂t=[−ℏ22m∗∇2−ℏω0−iℏγ2+α1∣ψ±∣2+α2∣ψ∓∣2]ψ±+Ω^±ψ∓+P±(r,t),(1)i\hbar\frac{\partial\psi_{\pm}}{\partial t}=\left[-\frac{\hbar^2}{2m^*}\nabla^2-\hbar\omega_0-i\frac{\hbar\gamma}{2}+\alpha_1|\psi_{\pm}|^2+\alpha_2|\psi_{\mp}|^2\right]\psi_{\pm}+\hat{\Omega}_{\pm}\psi_{\mp}+P_{\pm}(\mathbf{r},t), \tag*{(1)}

where m∗m^* is the effective polariton mass in the cavity plane, γ\gamma specifies the polariton decay rate, and α1,α2\alpha_1,\alpha_2 denote the interaction constants for polaritons with the same and opposite circular polarizations, respectively. The operators Ω^±=β(∂x±i∂y)2\hat{\Omega}_{\pm}=\beta(\partial_x\pm i\partial_y)^2 account for the TE–TM splitting of polariton modes [44], with strength β\beta.

To characterize the spatial polarization distribution of the condensate, we introduce the normalized Stokes vector S(r,t)=ψ†σψ/n\mathbf{S}(\mathbf{r},t)=\psi^\dagger\boldsymbol{\sigma}\psi/n, where σ=(σx,σy,σz)\boldsymbol{\sigma}=(\sigma_x,\sigma_y,\sigma_z) denotes the vector of Pauli matrices and n=ψ†ψ=∣ψ+∣2+∣ψ−∣2n=\psi^\dagger\psi=|\psi_+|^2+|\psi_-|^2 is the total polariton density. The vector S\mathbf{S} specifies the local polarization state on the Poincaré sphere.

The coherent driving term in (1) is represented as a superposition of three Gaussian pump beams with frequency detunings Δωj\Delta\omega_j relative to the reference frequency ω0\omega_0 and polarization-dependent phases θ1=0\theta_1 = 0, θ2=2π/3\theta_2 = 2\pi/3, and θ3=4π/3\theta_3 = 4\pi/3:

P±(r,t)=F0∑j=13exp⁡(−∣r−rj∣2σ2+ikj,pump⋅r−iΔωjt∓iθj)(2)P_{\pm}(\mathbf{r},t)=F_0\sum_{j=1}^{3}\exp\left(-\frac{|\mathbf{r}-\mathbf{r}_j|^2}{\sigma^2}+i\mathbf{k}_{j,\mathrm{pump}}\cdot\mathbf{r}-i\Delta\omega_jt\mp i\theta_j\right) \tag*{(2)}

where F0F_0 is the pump amplitude, σ\sigma characterizes the spatial width of the Gaussian beams, and rj\mathbf{r}_j and kj\mathbf{k}_j denote their center positions and in-plane wavevectors, respectively. The finite momentum spread of the Gaussian beams, of order 2/σ2/\sigma, affects the spectral selectivity of resonant excitation and the relative contributions of the TE- and TM-polarized modes. The opposite phase factors exp⁡(∓iθj)\exp(\mp i\theta_j) are chosen to match the linear polarization of each pump beam to the TE-polarized mode, thereby favoring resonant excitation of the corresponding polariton branch. Such a polarization configuration is consistent with the threefold symmetry of the excitation geometry, facilitating the formation and controlled manipulation of a symmetric interference lattice.

We consider three pump beams with independently adjustable frequency detunings Δωj\Delta\omega_j (j=1,2,3j=1,2,3) relative to the reference frequency ω0=(1+2m∗β/ℏ2)kpump2\omega_0=(1+2m^*\beta/\hbar^2)k_{\mathrm{pump}}^2. Within the parabolic approximation, the TE–TM splitting modifies the dispersion of the selected polarization branch, and the wavevector magnitudes KjK_j of the resonantly excited polariton modes are given by

Kj=kpump2+2m∗Δωj/ℏ1+2m∗β/ℏ2,j∈{1,2,3}.(3)K_j=\sqrt{k_{\mathrm{pump}}^2+\frac{2m^*\Delta\omega_j/\hbar}{1+2m^*\beta/\hbar^2}},\qquad j\in\{1,2,3\}. \tag*{(3)}

The in-plane wavevectors are expressed as K1=(0,−K1)\mathbf{K}_1=(0,-K_1), K2=(3K2/2,K2/2)\mathbf{K}_2=(\sqrt{3}K_2/2,K_2/2), and K3=(−3K3/2,K3/2)\mathbf{K}_3=(-\sqrt{3}K_3/2,K_3/2).

Spatiotemporal control of polariton lattices

The frequency detunings induce relative phase evolution between the three polariton fields, resulting in a traveling interference pattern. The lattice drift velocity V=(Vx,Vy)T\mathbf{V}=(V_x,V_y)^{\mathrm{T}} is determined by requiring the relative phases to remain time-independent in the frame moving with the lattice, r↦r+Vt\mathbf{r}\mapsto\mathbf{r}+\mathbf{V}t. This condition gives (Ki−Kj)⋅V=Δωij(\mathbf{K}_i-\mathbf{K}_j)\cdot\mathbf{V}=\Delta\omega_{ij}, with Δωij=Δωi−Δωj\Delta\omega_{ij}=\Delta\omega_i-\Delta\omega_j. For the wavevectors defined above, this yields

V=1D(13[Δω13(K2−K3)+Δω23(2K1+K3)]Δω23K3−Δω13(K2+K3))(4)\mathbf{V}=\frac{1}{D}\begin{pmatrix}\frac{1}{\sqrt{3}}\left[\Delta\omega_{13}(K_2-K_3)+\Delta\omega_{23}(2K_1+K_3)\right]\\ \Delta\omega_{23}K_3-\Delta\omega_{13}(K_2+K_3)\end{pmatrix} \tag*{(4)}

where the denominator is D=K1K2+K1K3+K2K3D=K_1K_2+K_1K_3+K_2K_3.

When all frequency detunings vanish (Δωj=0\Delta\omega_j=0), the lattice drift velocity becomes zero (V=0\mathbf{V}=0), resulting in a stationary interference pattern, as shown in Figure 2(a,Figure 2b). In this case, all in-plane wavevectors have the same magnitude K0K_0. The interference produces a regular hexagonal lattice where the horizontal spatial period is given by a0=4π/(3K0)a_0=4\pi/(\sqrt{3}K_0) (the distance between next-nearest neighbors along the xx-axis), while the nearest-neighbor distance is aNN=4π/(3K0)a_{\mathrm{NN}}=4\pi/(3K_0), as indicated by the white lines in Figure 2(b). The resulting density pattern exhibits C6C_6 symmetry and forms a hexagonal lattice. It is accompanied by spatial textures in linear polarization and topological defects, as illustrated in Figure 2(c). In particular, the polarization texture contains alternating left- and right-half-vortices with fractional winding numbers k,m=±1/2k,m=\pm1/2 [45] (see Supplementary Materials for details).

Figure 2. Stationary polariton interference lattice. a, Spatial distribution of the polariton density nn. White dashed circles indicate the positions and approximate spatial extent of the Gaussian pump spots. b, Enlarged view of the central interference lattice core. Solid white lines outline its analytically determined hexagonal geometry. c, Spatial distribution of the circular polarization degree S3S_3 of the polariton condensate in the region shown in b. The superimposed black arrows represents the in-plane Stokes components (S1,S2)(S_1,S_2), revealing local linear polarization textures and associated topological defects.

When one of the pump beams is frequency-detuned from the two others, Δω1≠Δω2,3=0\Delta\omega_1\ne\Delta\omega_{2,3}=0, the corresponding wavevectors satisfy K2=K3=K0=kpumpK_2=K_3=K_0=k_{\mathrm{pump}}, whereas K1K_1 varies with Δω1\Delta\omega_1. The resulting interference lattice exhibits reduced spatial symmetry, from C6C_6 to C2C_2, and drifts along the yy axis with velocity Vx=0V_x=0 and Vy=−Δω1/(K1+K0/2)V_y=-\Delta\omega_1/(K_1+K_0/2) (Figure 2(b,Figure 2c)). The sign of Δω1\Delta\omega_1 determines the direction of motion, while the resulting variation of K1K_1 makes the drift speed a nonlinear function of the detuning Δω1\Delta\omega_1. During this deformation, the horizontal spatial period a0=4π/(3K0)a_0=4\pi/(\sqrt{3}K_0) remains strictly unchanged. However, the nearest-neighbor distances become anisotropic: the vertical distance between adjacent density maxima along the yy-axis scales as avert=2π/(K1+K0/2)a_{\mathrm{vert}}=2\pi/(K_1+K_0/2), whereas the oblique nearest-neighbor distance varies according to adiag=π4/(3K02)+1/(K1+K0/2)2a_{\mathrm{diag}}=\pi\sqrt{4/(3K_0^2)+1/(K_1+K_0/2)^2}. The frequency detuning thus provides simultaneous control over the lattice drift velocity and its anisotropic deformation. The TE–TM splitting strength β\beta, determined by the microcavity structure, influences both quantities by modifying the relation between the frequency detuning and the polariton wavevector.

At a fixed reference pump frequency ω0\omega_0, increasing the TE–TM splitting strength β\beta reduces the resonant wavevectors K0K_0 and KjK_j, thereby increasing both the drift speed ∣Vy∣|V_y| and the characteristic lattice spacings a0a_0 and adiaga_{\mathrm{diag}}. Conversely, increasing the magnitude of the frequency detuning Δω1\Delta\omega_1 enhances the drift speed ∣Vy∣|V_y|, while positive (negative) detunings compress (expand) the diagonal lattice spacing adiaga_{\mathrm{diag}}, leaving the transverse period a0a_0 unchanged. These dependencies provide complementary means of controlling the motion and spatial geometry of the polariton interference lattice.

More generally, when multiple pump beams have nonzero frequency detunings, the interference pattern exhibits periodic temporal modulation if the independent frequency differences Δω13\Delta\omega_{13} and Δω23\Delta\omega_{23} are commensurate. Otherwise, the temporal modulation is quasiperiodic.

Electron transport by dipolariton lattices

The spatial separation of electrons and holes in QW1 and QW2 gives rise to permanent out-of-plane electric dipoles associated with the IX component of dipolaritons, characterized by the Hopfield fraction ηIX\eta_{\mathrm{IX}}. The spatial modulation of their density generates an electrostatic potential landscape Uc(r,t)U_c(\mathbf{r},t) acting on electrons in the adjacent 2DES. As illustrated in Figure 4, the resulting Coulomb potential exhibits local minima that can serve as electron trapping sites. The corresponding interaction kernel in momentum space is given by

Figure 4. Electron wave-packet transport under different initial conditions. a, Mean longitudinal position ⟨y⟩(t)\langle y\rangle(t) of the electron wave packets for six scenarios, compared with the trajectory YwellY_{\mathrm{well}} of the moving potential minimum. Scenarios 1, 2 and 5 closely follow the translating well, whereas Scenarios 3, 4 and 6 exhibit larger deviations. The inset shows the gapped Dirac dispersion of graphene, with the chemical potential μ=EF\mu= E_{\mathrm{F}} located inside the bandgap Δg\Delta_{\mathrm{g}}. b, Confinement probability Pwell(t)P_{\mathrm{well}}(t) within a circular region of radius 0.90 μ\mum centered on the moving potential minimum. High probability retention is observed for Scenarios 1, 2, and 5. The inset shows the confinement potential profile Uc(y,ti)−Umin⁡U_{\mathrm{c}}(y,t_{\mathrm{i}})-U_{\min} at the electron preparation time tit_{\mathrm{i}}, together with the initial electron probability distributions. The profiles illustrate the initial positions and spatial widths of the wave packets, while the arrows indicate their initial propagation directions and relative speeds.

V~0(q)=e22ε0ε∣q∣(e−∣q∣le−e−∣q∣lh),\widetilde{V}_0(\mathbf{q})=\frac{e^2}{2\varepsilon_0\varepsilon|\mathbf{q}|}\left(e^{-|\mathbf{q}|l_e}-e^{-|\mathbf{q}|l_h}\right),

where lel_e and lhl_h are the distances from the electron and hole layers of the IX to the 2DES, respectively, ee is the elementary charge, and ε0\varepsilon_0 and ε\varepsilon denote the vacuum and relative permittivities. Since the characteristic spatial scale of the polariton lattice a0a_0 considerably exceeds the interlayer distances, a0,avert,adiag≫le,lha_0,a_{\mathrm{vert}},a_{\mathrm{diag}}\gg l_e,l_h, the dominant Fourier components satisfy ∣q∣le,h≪1|\mathbf{q}|l_{e,h}\ll1. Expanding the exponentials to the first order, the momentum-independent interaction kernel is V~0≈e2(lh−le)2ε0ε\widetilde{V}_0\approx\frac{e^2(l_h-l_e)}{2\varepsilon_0\varepsilon}. The resulting electrostatic potential is therefore approximately local and proportional to the IX density nIX(r,t)=ηIXn(r,t)n_{\mathrm{IX}}(\mathbf{r},t)=\eta_{\mathrm{IX}}n(\mathbf{r},t):

Uc(r,t)=e2(lh−le)2ε0εnIX(r,t).(5)U_c(\mathbf{r},t)=\frac{e^2(l_h-l_e)}{2\varepsilon_0\varepsilon}n_{\mathrm{IX}}(\mathbf{r},t). \tag*{(5)}

For lh>lel_h > l_e, the potential is repulsive, and its minima coincide with the minima of the dipolariton density. For the parameters specified in Methods, the calculated potential modulation reaches 3.2 meV across the lattice (Figure 3).

Figure 3. Electron entrainment in a moving electrostatic lattice. a, Device schematic illustrating the electron trapping mechanism. b, Spatiotemporal evolution of the polariton-induced electrostatic potential Uc(x,t)U_c(x,t) along the cross-section at y=0y=0. c, Spatiotemporal evolution of Uc(y,t)U_c(y,t) along the cross-section at x=0x=0, overlaid with the mean longitudinal positions ⟨y⟩(t)\langle y\rangle(t) of the electron wave packets for six scenarios with different initial conditions. The coloured trajectories illustrate how the electron wave packets follow the moving potential minima under different initial conditions.

To examine electron confinement and transport, we solve the time-dependent Dirac equation for electron wave packets in the polariton-induced electrostatic potential (see Methods). The 2DES is modeled as gapped graphene, and six initial wave-packet configurations are considered to assess the effects of initial position, velocity, and spatial width on electron transport relative to a moving potential minimum. The same time-dependent electrostatic potential is used in all calculations, with the corresponding initial conditions specified in Table 1.

ScenarioΔyi\Delta y_i (μ\mum)wiw_i (μ\mum)uyu_y (μ\mum ps−1^{-1})
1 (comoving)00.12VyV_y
2 (at rest)00.120
3 (displaced)+0.400.140
4 (larger displacement)+0.800.160
5 (finite velocity)00.130.25
6 (higher velocity)00.130.50

Table 1. Initial conditions for the six electron transport scenarios. The displacement Δyi\Delta y_i is measured from the selected potential minimum along the direction of motion, wiw_i is the width of the initial Gaussian envelope, and uyu_y is the prescribed initial conduction-band group velocity. Here VyV_y denotes the lattice drift velocity.

The trajectories of electron wave packets in Figure 3 and Figure 4 show that, in Scenario 1 and 2, the wave packets follow the translating potential minimum, despite their different initial velocities. In Scenario 5, the packet also remains closely associated with the moving well, even when its initial speed exceeds the lattice drift speed. An initial displacement of 0.40 μ\mum (Scenario 3) or higher initial velocity (Scenario 6) leads to oscillations around the moving minimum, whereas the larger displacement (Scenario 4) results in more pronounced deviations from the well trajectory and reduced confinement.

Confinement is quantified by the probability Pwell(t)P_{\mathrm{well}}(t) integrated over a circular region of radius 0.90 μ\mum centered on the moving potential minimum (Figure 4b). High probability retention is observed in Scenarios 1, 2 and 5, while Scenarios 3 and 6 retain a substantial confined component. Scenario 4 exhibits more pronounced probability leakage from the well. These results demonstrate the transport of initially prepared electron wave packets without requiring exact matching of their initial velocities to the lattice drift velocity or electronic relaxation.

Discussion

We have demonstrated a concept for the dynamic spatial manipulation of electrons, using the electrostatic landscape generated by a driven condensate of dipolaritons. A defining advantage of this polaritonic tweezer architecture is its ultrafast transport capability, with estimated accessible lattice drift reaching 0.2 μm\mu\mathrm{m}/ps. In contrast to conventional solid-state charge-transport technologies, such as surface-acoustic-wave devices, limited by the speed of sound in the host material (e. g., ∼3×10−3\sim3 \times10^{-3} μm\mu\mathrm{m}/ps in GaAs), or charge-coupled devices constrained by finite charge-transfer times and electrical clocking, including RC delays, the polaritonic lattice can propagate at substantially higher velocities. The optically controlled motion of the polariton-induced potential thus offers a route to rapid in-plane transport of electrons in 2DES.

A key advantage of the proposed architecture lies in separating optical control from the physical mechanism of electron trapping. Conventional optical tweezers rely on spatial gradients of optical intensity, whereas direct optical manipulation of free electrons is challenging and generally requires intense electromagnetic fields. In contrast, the incident pump beams in our scheme generate and dynamically control a spatially structured dipolariton population, whose permanent electric dipoles produce a Coulomb potential acting on electrons in the adjacent 2DES. Electron confinement is therefore governed by electrostatic rather than direct optical forces, while the position and motion of the trapping wells remain optically controllable. Importantly, the spatial confinement of electrons is determined by the depth and curvature of the induced electrostatic potential rather than directly by the focusing of the incident optical fields.

Furthermore, the non-collinear pumping geometry allows the dipolariton lattice to operate as an optically controlled, two-dimensional charge conveyor. Unlike transport along predefined paths in gate-defined quantum dot arrays or surface-acoustic-wave channels, the lattice drift can be directed along an arbitrary in-plane orientation by adjusting the frequency detunings and resonant wavevectors of the incident beams. This provides control over both the direction and speed of transport without modifying the physical geometry of the device. The electrostatic potential depth can also be controlled through the dipolariton density and effective dipole moment, as well as the geometry of the hybrid heterostructure. Meanwhile, the lattice spacing is governed by the pump wavevectors and polariton dispersion, including the effective mass and TE–TM splitting.

Looking forward, the ability to transport individual electrons along optically controlled trajectories without predefined physical channels offers opportunities for reconfigurable charge routing in solid-state devices. Our results demonstrate that electron wave packets can follow the moving potential minima without requiring exact matching of their initial velocities to the lattice drift or electronic relaxation, although the degree of confinement depends on the initial conditions. A complementary development in electron quantum optics is the on-demand generation of minimal-excitation single-electron states known as levitons [46, 47], which provides a potential approach to preparing electronic excitations for subsequent controlled transport. Extending our approach to spin-based logic and quantum information processing would require additional control over electron loading, release, and spin coherence during transport. The proposed integration of dipolariton condensates with electronic systems thus provides a promising route toward optically programmable charge transport in integrated optoelectronic and spintronic architectures.

Methods

Modeling of electron wave-packet dynamics

The adjacent electronic layer is modeled as gapped graphene, described by a two-component wavefunction Φ=(ΦA,ΦB)T\Phi= (\Phi_A,\Phi_B)^{\mathrm{T}} corresponding to the two graphene sublattices. The energy gap Δg\Delta_g accounts for sublattice-symmetry breaking, as observed in graphene/hBN heterostructures [48, 49, 50]. Within the single-valley approximation, the electron dynamics in the polariton-induced potential Uc(r,t)U_c(\mathbf{r},t) is governed by the massive Dirac equation

iℏ∂tΦ=[H^D,0+Uc(r,t)I2]Φ(6)i\hbar\partial_t\Phi=\left[\hat{H}_{D,0}+U_c(\mathbf{r},t)\mathbb{I}_2\right]\Phi \tag*{(6)}

where H^D,0=ℏvF(σxk^x+σyk^y)+Δgσz/2\hat{H}_{D,0}=\hbar v_{F}(\sigma_x\hat{k}_x+\sigma_y\hat{k}_y)+\Delta_g\sigma_z/2 is the free massive Dirac Hamiltonian, k^=−i∇\hat{\mathbf{k}}=-i\nabla is the wavevector operator, and σx,y,z\sigma_{x,y,z} are the Pauli matrices.

For each initial condition, the electron wave packet is constructed from a Gaussian envelope Gi(r)=exp⁡[−∣r−ri∣2/4wi2+iki⋅r]G_i(\mathbf{r})=\exp[-|\mathbf{r}-\mathbf{r}_i|^2/4w_i^2+i\mathbf{k}_i\cdot\mathbf{r}], where ri\mathbf{r}_i, wiw_i, and ki\mathbf{k}_i specify its initial position, spatial width, and central wavevector, respectively. The initial state is projected onto the positive-energy branch of the free Dirac Hamiltonian:

Φ~i(k,ti)=NinΠ^+(k)(G~i(k),0)T,(7)\tilde{\boldsymbol{\Phi}}_i(\mathbf{k},t_i)=\mathcal{N}_{\mathrm{in}}\hat{\Pi}_{+}(\mathbf{k})(\tilde{G}_i(\mathbf{k}),0)^{\mathrm{T}}, \tag*{(7)}

where, G~i(k)\tilde{G}_i(\mathbf{k}) is the Fourier transform of Gi(r)G_i(\mathbf{r}), and Π^+(k)=[I2+HD,0(k)/ε+(k)]/2\hat{\Pi}_{+}(\mathbf{k})=[I_2+H_{D,0}(\mathbf{k})/\varepsilon_{+}(\mathbf{k})]/2, with HD,0(k)H_{D,0}(\mathbf{k}) denoting the momentum-space representation of H^D,0\hat{H}_{D,0} and ε+(k)=(Δg/2)2+(ℏvF∣k∣)2\varepsilon_{+}(\mathbf{k})=\sqrt{(\Delta_g/2)^2+(\hbar v_F|\mathbf{k}|)^2} its positive eigenvalue. The central wavevector is determined by the prescribed initial group velocity ui\mathbf{u}_i through ki=Δgui/[2ℏvF21−∣ui∣2/vF2]\mathbf{k}_i=\Delta_g\mathbf{u}_i/[2\hbar v_F^2\sqrt{1-|\mathbf{u}_i|^2/v_F^2}], while Nin\mathcal{N}_{\mathrm{in}} ensures normalization of the initial state.

The electron wave packet is initialized at ti=5t_i=5 ps, after the formation of the polariton lattice. The six initial configurations differ in the displacement from a selected potential minimum, spatial width, and group velocity, allowing us to examine their influence on electron confinement and transport. The resulting states are propagated under the time-dependent Dirac Hamiltonian. The confinement probability is calculated as Pwell(t)=∫∣r−Rw(t)∣≤aeΦ†Φ d2rP_{\mathrm{well}}(t)=\int_{|\mathbf{r}-\mathbf{R}_{w}(t)|\leq a_e}\boldsymbol{\Phi}^{\dagger}\boldsymbol{\Phi}\,d^2\mathbf{r}, where ae=0.90a_e=0.90 μ\mum is the radius of the integration region and Rw(t)\mathbf{R}_w(t) denotes the position of the tracked potential minimum. No renormalization is applied during propagation. Electronic relaxation and back-action on the polariton condensate are neglected.

Simulation parameters

The simulations use the following excitation parameters: radial displacement of the pump spots R=12R=12 μ\mum, in-plane pump wavenumber kpump=2.5k_{\mathrm{pump}}=2.5 μm−1\mu\mathrm{m}^{-1}, frequency detunings Δω1=−0.11\Delta\omega_1=-0.11 ps−1\mathrm{ps}^{-1} (unless otherwise stated) and Δω2,3=0\Delta\omega_{2,3}=0, Gaussian pump width σ=15\sigma=15 μ\mum and continuous-wave pump amplitude F0=2.4F_0=2.4 meV⋅μm−1\mathrm{meV}\cdot\mu\mathrm{m}^{-1}. The polariton parameters are m∗=10−4mem^*=10^{-4}m_e, lifetime τ=10\tau=10 ps (γ=0.1\gamma=0.1 ps−1\mathrm{ps}^{-1}), interaction constants α1=−10α2=1\alpha_1=-10\alpha_2=1 μeV⋅μm2\mu\mathrm{eV}\cdot\mu\mathrm{m}^2, and the TE–TM splitting constant β=11\beta=11 μeV⋅μm2\mu\mathrm{eV}\cdot\mu\mathrm{m}^2. The electronic parameters are vF=106v_F=10^6 m s−1\mathrm{m}\,\mathrm{s}^{-1} and Δg=30\Delta_g=30 meV, corresponding to a band-edge effective mass mD=2.6×10−3mem_D=2.6\times10^{-3}m_e. The distances between the 2DES and the electron and hole layers of IXs with ηIX=0.8\eta_{\mathrm{IX}}=0.8 are le=1.9l_e=1.9 nm and lh=3.5l_h=3.5 nm, respectively, corresponding to a dipole length d=lh−le=1.6d=l_h-l_e=1.6 nm. The Coulomb interaction is evaluated using a relative permittivity ε≃4\varepsilon\simeq4, the elementary charge e=1.602×10−19e=1.602\times10^{-19} C, and the vacuum permittivity ε0=8.854×10−12\varepsilon_0=8.854\times10^{-12} F m−1\mathrm{F}\,\mathrm{m}^{-1}.

Data availability

All data generated and analyzed during the current study are available from the corresponding author on reasonable request.

Acknowledgments

The support of the Ministry of Science and Higher Education of the Russian Federation (No. FSMG-2026-0012), Saint-Petersburg State University (research grant No. 125202803069-4), the state assignment in the field of scientific activity of the Ministry of Science and Higher Education of the Russian Federation (theme FZUN-2024-0019, state assignment of VISU) are acknowledged.

Conflict of interest

The authors declare that they have no conflict of interest.

Authors’ contributions

N. I. G. performed simulations, co-developed the theoretical framework, analyzed and interpreted the results, and wrote the manuscript with input from all co-authors. E. S. S. co-developed the theoretical framework, analyzed and interpreted the results, and critically revised the manuscript. A. V. K. initiated the study, developed the concept, and guided the work. All authors discussed the results and contributed to the manuscript.

References

  1. [1]Ashkin, A., Dziedzic, J. M., Bjorkholm, J. E. & Chu, S. Observation of a single-beam gradient force optical trap for dielectric particles. Optics letters 11, 288–290 (1986).
  2. [2]Grier, D. G. A revolution in optical manipulation. nature 424, 810–816 (2003).pubmed.ncbi.nlm.nih.gov/12917694
  3. [3]Moffitt, J. R., Chemla, Y. R., Smith, S. B. & Bustamante, C. Recent advances in optical tweezers. Annu. Rev. Biochem. 77, 205–228 (2008).DOI
  4. [4]Dholakia, K., Macdonald, M. & Spalding, G. Optical tweezers: the next generation. Physics world 15, 31–35 (2002).DOI
  5. [5]Padgett, M. & Di Leonardo, R. Holographic optical tweezers and their relevance to lab on chip devices. Lab on a Chip 11, 1196–1205 (2011).DOI
  6. [6]Juan, M. L., Righini, M. & Quidant, R. Plasmon nano-optical tweezers. Nature photonics 5, 349–356 (2011).DOI
  7. [7]Grigor enko, A., Roberts, N., Dickinson, M. & Zhang, Y. Nanometric optical tweezers based on nanostructured substrates. Nature Photonics 2, 365–370 (2008).
  8. [8]Rahmani, A. & Chaumet, P. C. Optical trapping near a photonic crystal. Optics express 14, 6353–6358 (2006).pubmed.ncbi.nlm.nih.gov/19516811
  9. [9]Marago, O. M., Jones, P. H., Gucciardi, P. G., Volpe, G. & Ferrari, A. C. Optical trapping and manipulation of nanostructures. Nature nanotechnology 8, 807–819 (2013).DOI
  10. [10]Giovanni, V. et al. Roadmap for optical tweezers. JPHYS PHOTONICS 5, 1–135 (2023).
  11. [11]Zhong, M.-C., Wei, X.-B., Zhou, J.-H., Wang, Z.-Q. & Li, Y.-M. Trapping red blood cells in living animals using optical tweezers. Nature communications 4, 1768 (2013).DOI
  12. [12]Grimm, R., Weidemuller, M. & Ovchinnikov, Y. B. Optical dipole traps for neutral atoms. In Advances in atomic, molecular, and optical physics, vol. 42, 95–170 (Elsevier, 2000).
  13. [13]Batelaan, H. Colloquium: Illuminating the kapitza-dirac effect with electron matter optics. Rev. Mod. Phys. 79, 929–941 (2007). URL https://link.aps.org/doi/10.1103/RevModPhys.79.929.DOI
  14. [14]Chirita Mihaila, M. C. et al. Transverse electron-beam shaping with light. Phys. Rev. X 12, 031043 (2022). URL https://link.aps.org/doi/10.1103/PhysRevX.12.031043.
  15. [15]Kavokin, A., Baumberg, J. J., Malpuech, G. & Laussy, F. P. Microcavities (Oxford university press, 2017).
  16. [16]Kasprzak, J. et al. Bose–einstein condensation of exciton polaritons. Nature 443, 409–414 (2006).pubmed.ncbi.nlm.nih.gov/17006506
  17. [17]Byrnes, T., Kim, N. Y. & Yamamoto, Y. Exciton–polariton condensates. Nature Physics 10, 803–813 (2014).DOI
  18. [18]Sanvitto, D. & Kéna-Cohen, S. The road towards polaritonic devices. Nature materials 15, 1061–1073 (2016).DOI
  19. [19]Christmann, G. et al. Oriented polaritons in strongly-coupled asymmetric double quantum well microcavities. Applied Physics Letters 98 (2011).DOI
  20. [20]Rosenberg, I. et al. Strongly interacting dipolar-polaritons. Science Advances 4, eaat8880 (2018).arxiv.org/abs/1802.01123
  21. [21]Nalitov, A., Solnyshkov, D., Gippius, N. & Malpuech, G. Voltage control of the spin-dependent interaction constants of dipolaritons and its application to optical parametric oscillators. Physical Review B 90, 235304 (2014).DOI
  22. [22]Kyriienko, O., Shelykh, I. A. & Liew, T. C. H. Tunable single-photon emission from dipolaritons. Physical Review A 90, 033807 (2014).DOI
  23. [23]Amo, A. et al. Light engineering of the polariton landscape in semiconductor microcavities. Physical Review B—Condensed Matter and Materials Physics 82, 081301 (2010).DOI
  24. [24]Askitopoulos, A. et al. Polariton condensation in an optically induced two-dimensional potential. Phys. Rev. B 88, 041308(R) (2013). URL https://link.aps.org/doi/10.1103/PhysRevB.88.041308.DOI
  25. [25]Askitopoulos, A. et al. All-optical quantum fluid spin beam splitter. Phys. Rev. B 97, 235303 (2018). URL https://link.aps.org/doi/10.1103/PhysRevB.97.235303.DOI
  26. [26]Aladinskаia, E. et al. Spatial quantization of exciton-polariton condensates in optically induced traps. Phys. Rev. B 107, 045302 (2023). URL https://link.aps.org/doi/10.1103/PhysRevB.107.045302.
  27. [27]Cerda-Méndez, E. et al. Polariton condensation in dynamic acoustic lattices. Physical review letters 105, 116402 (2010).
  28. [28]Lukoshkin, V. et al. Steady state oscillations of circular currents in concentric polariton condensates. Scientific Reports 13, 4607 (2023). URL https://doi.org/10.1038/s41598-023-31520-z.
  29. [29]Klembt, S. et al. Exciton-polariton topological insulator. Nature 562, 552–556 (2018).DOI
  30. [30]Myers, D. M. et al. Pushing photons with electrons: Observation of the polariton drag effect. Phys. Rev. Lett. 135, 146903 (2025). URL https://link.aps.org/doi/10.1103/fb2r-qdq7.DOI
  31. [31]Myers, D. M. et al. Superlinear increase of photocurrent due to stimulated scattering into a polariton condensate. Phys. Rev. B 98, 045301 (2018). URL https://link.aps.org/doi/10.1103/PhysRevB.98.045301.DOI
  32. [32]Cotleţ, O. et al. Transport of neutral optical excitations using electric fields. Phys. Rev. X 9, 041019 (2019). URL https://link.aps.org/doi/10.1103/PhysRevX.9.041019.DOI
  33. [33]Jacqmin, T. et al. Direct observation of dirac cones and a flatband in a honeycomb lattice for polaritons. Physical review letters 112, 116402 (2014).pubmed.ncbi.nlm.nih.gov/24702392
  34. [34]Liew, T., Rubo, Y. G. & Kavokin, A. Generation and dynamics of vortex lattices in coherent exciton-polariton fields. Physical review letters 101, 187401 (2008).pubmed.ncbi.nlm.nih.gov/18999864
  35. [35]Wertz, E. et al. Spontaneous formation and optical manipulation of extended polariton condensates. Nature physics 6, 860–864 (2010).DOI
  36. [36]Sich, M. et al. Spin domains in one-dimensional conservative polariton solitons. ACS Photonics 5, 5095–5102 (2018). URL https://doi.org/10.1021/acsphotonics.8b01410.
  37. [37]del Valle Inclan Redondo, Y. et al. Non-reciprocal band structures in an exciton–polariton floquet optical lattice. Nature Photonics 18, 548–553 (2024).
  38. [38]Gnusov, I. et al. Quantum vortex formation in the “rotating bucket” experiment with polariton condensates. Science Advances 9, eadd1299 (2023).DOI
  39. [39]van Kesteren, H. W., Cosman, E. C., van der Poel, W. A. J. A. & Foxon, C. T. Fine structure of excitons in type-ii gaas/alas quantum wells. Phys. Rev. B 41, 5283–5292 (1990). URL https://link.aps.org/doi/10.1103/PhysRevB.41.5283.DOI
  40. [40]Cristofolini, P. et al. Coupling quantum tunneling with cavity photons. Science 336, 704–707 (2012).DOI
  41. [41]Cotleţ, O., Zeytinoğlu, S., Sigrist, M., Demler, E. & İmamoğlu, A. Superconductivity and other collective phenomena in a hybrid bose-fermi mixture formed by a polariton condensate and an electron system in two dimensions. Physical Review B 93, 054510 (2016).DOI
  42. [42]Wurdack, M. et al. Observation of hybrid tamm-plasmon exciton-polaritons with gaas quantum wells and a mose2 monolayer. Nature communications 8, 259 (2017).DOI
  43. [43]Kaliteevskii, M. et al. Hybrid states of tamm plasmons and exciton polaritons. Applied Physics Letters 95 (2009).
  44. [44]Panzarini, G. et al. Exciton-light coupling in single and coupled semiconductor microcavities: Polariton dispersion and polarization splitting. Phys. Rev. B 59, 5082–5089 (1999). URL https://link.aps.org/doi/10.1103/PhysRevB.59.5082.DOI
  45. [45]Rubo, Y. G. Half vortices in exciton polariton condensates. Phys. Rev. Lett. 99, 106401 (2007). URL https://link.aps.org/doi/10.1103/PhysRevLett.99.106401.DOI
  46. [46]Keeling, J., Klich, I. & Levitov, L. S. Minimal excitation states of electrons in one-dimensional wires. Phys. Rev. Lett. 97, 116403 (2006). URL https://link.aps.org/doi/10.1103/PhysRevLett.97.116403.DOI
  47. [47]Dubois, J. et al. Minimal-excitation states for electron quantum optics using levitons. Nature 502, 659–663 (2013). URL https://doi.org/10.1038/nature12713.
  48. [48]Giovannetti, G., Khomyakov, P. A., Brocks, G., Kelly, P. J. & Van Den Brink, J. Substrate-induced band gap in graphene on hexagonal boron nitride: Ab initio density functional calculations. Physical Review B—Condensed Matter and Materials Physics 76, 073103 (2007).DOI
  49. [49]Hunt, B. et al. Massive dirac fermions and hofstadter butterfly in a van der waals heterostructure. Science 340, 1427–1430 (2013).DOI
  50. [50]Chen, Z.-G. et al. Observation of an intrinsic bandgap and landau level renormalization in graphene/boron-nitride heterostructures. Nature communications 5, 4461 (2014).DOI

Paper details

vlnc:W_y5rbkxd8
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents