Morphology-independent general-purpose optical surface tractor beam

When a laser beam is scattered by a particle, momentum exchange mediated by the Lorentz force results in optical forces. In specially prepared situations, the forward photon momentum increases upon scattering; by momentum conservation, the particle experiences a negative pulling force, often termed an optical tractor beam. Prior realizations are typically in free space and occasionally near a surface, but require careful tailoring of both beam and particle, limiting applicability. Recent schemes in waveguides and metamaterials partially alleviate particle constraints but do not fully solve the problem. To remedy this, the authors propose a general-purpose optical surface tractor beam (OSTB) that always pulls irrespective of particle composition and morphology, relying on a mechanism different from directional mode conversions or topological protection. Based on an analytical theory from the electromagnetic energy-momentum tensor, they consider a particle near a planar surface illuminated by a surface wave (SW) with a single well-defined Bloch k. The theory indicates the longitudinal force is always parallel to the canonical (Minkowski) momentum ħk; hence an SW propagating forward with negative k carries negative canonical momentum and exerts a pulling force. The optical force consists of an incident force (capture of incident photons) and a recoil force (re-emission). Unlike traditional tractor beams that tune scattering to reduce incident and increase recoil, here pulling is driven by a negative incident force from negative k, rendering it independent of scattering details and exceptionally robust. The OSTB is more convenient than waveguide/metamaterial schemes because its gradient force can automatically capture particles transversely, enabling stable transport. Owing to its ability to pull arbitrary particles or clusters, the OSTB is applicable to optofluidic or lab-on-a-chip systems to simultaneously pull collections of particles backward without concern for morphology or distribution.

Theory: The optical force on a particle near a planar surface illuminated by a single-Bloch-vector surface wave (SW) is formulated from the Maxwell (Minkowski in air) stress tensor. For deep subwavelength particles (dipole limit), the longitudinal force Fx ≈ Re[αx Einc (∂Einc/∂x)] = k0 |E0|^2 Im(α) = (k0/2) |E0|^2 σs, where σs > 0 for passive particles; thus Fx < 0 when the SW propagation constant k0 < 0. The canonical momentum density g = (k0 W)/ω (W ≥ 0), so a negative k0 implies negative canonical momentum.

General derivation: Consider a closed surface around the particle composed of S1 (cap in air) and planar surfaces S2 (just above interface) and S3 (just below). Using the continuity of tangential fields and the equivalence of Maxwell and Minkowski stress tensors in air, it is shown that Fx can be computed by integrating the Minkowski stress tensor over any closed surface enclosing the particle, establishing a link between optical force and canonical momentum flux. A rigorous transformation yields the key result: Fx = Wp − Wext − (1/c)[Wsca^SW (cosθi + cosθs) + Wsca^FPW (cosθi + cosθs)], which reduces to the standard free-space plane-wave force Fx = (1/c)[Wext − Wsca (cosθi + cosθs)] when only FPWs in air are involved. The first term represents momentum transfer from incident photons (incident force); the remaining terms are recoil forces from re-emission into SWs and FPWs. Without gain, Wext = Wsca^SW + Wsca^FPW + Wabs. For an SW with negative phase velocity vp = ω/kp (kp < 0), one obtains Fx = (1/vp) Wabs + (1/c)(1 − cosφ) Wsca^SW + (1/c)(1 − cosφ) Wsca^FPW ≤ 0, guaranteeing pulling for any passive particle. Conversely, kp > 0 yields pushing.

Metamaterial design (supporting OSTB): Conditions are derived for TE- and TM-polarized SWs with negative canonical momentum on a double-negative-index metamaterial (ε, μ < 0). For TE polarization, enforcing field continuity and positive total energy flux yields necessary and sufficient conditions −1 < ε < μ^−1 < 0 and propagating wavenumber kp = k0 √[ε/(ε − μ)] with kp chosen negative. By duality, analogous conditions and kp expression apply for TM polarization (kp = k0 √[με/(μ − ε)]).

Numerical validation: Full-wave finite-element simulations (COMSOL Multiphysics) compute optical forces by integrating time-averaged Maxwell stress tensor on a closed surface in air. Mesh: free triangular mesh with multiple refinements; solver: Direct. Example TE-OSTB on homogeneous metamaterial with ε = −0.9, μ = −1.2 at f = 100 THz (λ0 = 3 μm). SW excitation by phased line magnetic currents; Fourier transform of Ez at the interface confirms a single negative Bloch wavenumber |kp| = 1.19 k0. Forces are computed for cylinders (y-invariant) of various permittivities, sizes, and shapes.

Realistic structure: A photonic crystal (square lattice of air holes of radius r = 0.1 a in a high-index host ε = 120) realizes an effective double-negative metamaterial supporting a TM-OSTB. Band structures and effective constitutive parameters are obtained via boundary effective medium theory. Surface bands are identified; SWs excited by phased line sources embedded one lattice constant below the surface (13 sources, amplitude 1 mA, phase increment 0.2π) or by prism coupling (n = 1.69, ~45° incidence) for phase matching. Fourier analysis on the surface field confirms negative kp. Particle trajectories are computed by integrating m d^2r/dt^2 = Fopt − γ dr/dt with γ = 20 pN/μm^2 per 1 μm length (for air), particle density ρ = 2500 kg/m^3, and restitution coefficient 0.7 for collisions.

• A rigorous analytical framework links longitudinal optical force near an interface to the canonical momentum of light via Minkowski stress tensor integration. For a surface wave (SW) with a single Bloch vector, the longitudinal force is always parallel to the canonical momentum ħk.
• Any SW with negative propagation constant (kp < 0) carries negative canonical momentum and necessarily exerts a pulling longitudinal force (Fx ≤ 0) on any passive particle, independent of size, composition, or geometry. This arises because the negative incident force dominates the recoil force (Eq. (8)).
• Closed-form force expression (Eq. (6)) generalizes the free-space plane-wave result and shows that, near an interface, force can be engineered via phase speed in addition to scattering angles, enabling robust pulling without tailoring scattering pathways.
• Metamaterial conditions for supporting an OSTB are derived. For TE polarization, necessary and sufficient conditions are −1 < ε < μ^−1 < 0 with kp = k0 √[ε/(ε − μ)] (negative branch). By duality, analogous TM conditions yield kp = k0 √[με/(μ − ε)]. Example: ε = −0.9, μ = −1.2 gives |kp| = 1.19 k0, confirmed by Fourier analysis.
• Numerical simulations show consistently negative longitudinal forces across particle permittivities (transparent and lossy), sizes (Rayleigh to Mie), and shapes (circular, triangular, square). Loss generally enhances pulling by increasing absorption (incident force), except near resonances where extinction/scattering may decrease.
• The transverse (gradient) force attracts particles toward the interface, establishing stable equilibrium heights for certain parameters, enabling non-contact transport along the surface. Particle trajectories show capture and backward pulling toward the source.
• Realistic photonic crystal implementation (a = 100 nm, k0 = 0.527/a) supports a TM-OSTB. SWs are excited via phased line sources or prism coupling; Fourier spectra confirm negative kp. Despite partial excitation of bulk FPWs, the OSTB dominates far from sources.
• Multi-particle scenarios: Even with optical binding causing position-dependent forces (some particles may experience pushing at certain separations), the total longitudinal force on the cluster remains negative, enabling pulling of particle ensembles.
• Efficiency: Because SW phase speed vp < c, comparison of Eq. (6) with free-space force indicates SWs can generate larger forces for the same number of scattered photons.

The work addresses the central challenge in optical pulling—robust, morphology-independent backward forces—by exploiting a surface wave with negative canonical momentum. The derivation from first principles (stress-tensor formalism) shows the longitudinal force equals the rate of change of total canonical momentum and, for a single-Bloch-vector SW, aligns with ħk. Thus, selecting an SW with kp < 0 guarantees pulling through a negative incident force, rather than relying on enhanced recoil via carefully engineered scattering. This mechanism eliminates the need to fine-tune particle size, geometry, material, or to rely on mode conversions or topological protection, greatly broadening applicability. Simulations validate unconditional pulling across materials (dielectric, plasmonic, chiral), shapes, and sizes, and demonstrate practical excitation on homogeneous and realistic metamaterial surfaces. The gradient force naturally provides transverse trapping, enabling stable, low-friction transport. Comparisons indicate SWs can be more force-efficient than free-space beams for equal photon budgets, due to reduced phase speed. Overall, the results establish OSTBs as a robust platform for optical manipulation near surfaces and highlight canonical momentum engineering as a powerful design knob for optical forces.

The paper introduces a general-purpose optical surface tractor beam (OSTB) based on surface waves with negative canonical momentum that exerts pulling forces on any passive particle, irrespective of size, composition, or geometry. A rigorous stress-tensor-based theory derives a closed-form force expression linking force to canonical momentum and shows that, for kp < 0, the negative incident force outweighs recoil, ensuring pulling. The authors derive metamaterial conditions for TE/TM surface waves supporting negative-k propagation and demonstrate operation on both homogeneous double-negative metamaterials and realistic photonic crystal structures. Numerical simulations confirm robust, morphology-independent pulling, loss-enhanced forces, transverse trapping, and effective multi-particle pulling. Potential applications include particle transport and sorting in optofluidic and lab-on-a-chip systems and manipulation of particle clusters. Future work could explore optimizing excitation efficiency and range in lossy substrates, engineering dispersion to tailor phase speed and force magnitude, extending to three-dimensional particles and complex environments, and leveraging canonical momentum engineering to design bespoke force landscapes beyond pulling.

• The theoretical derivations assume a lossless homogeneous metamaterial substrate to simplify stress-tensor integrals; in practice, finite loss shortens propagation length, reducing pulling range, though pulling persists with moderate loss.
• The main result applies to surface waves characterized by a single well-defined Bloch wavevector. SWs with multiple Bloch components (e.g., backward surface waves on photonic crystals) carry effectively positive canonical momenta and fall outside this framework, making pulling more dependent on scattering features.
• The same stress-tensor treatment does not directly extend to the normal (transverse planar) force Fz due to discontinuity of Txz across the interface.
• In realistic periodic substrates, broken translational symmetry makes the longitudinal force weakly position-dependent along x, though this dependence is small for particles not too small relative to the lattice.
• Excitation efficiency can be reduced when surface states are not within a complete band gap, leading to concurrent excitation of bulk FPWs that decay but can impact near-source behavior.
• Experimental realization requires precise phase-matching for negative-k SW excitation (e.g., prism coupling or phased sources) and fabrication of suitable double-negative metamaterials or effective photonic crystals.