Plasmonic Spin Meron Lattices with Height-Sensitive Topology Evolution Journal: opticajournal Publication type: Research Article Anand Hegde, 1 Komal Gupta, 1 and Chen-Bin Huang 1,* Abstract We demonstrate height-controlled topological switching of plasmonic spin meron lattices above a metallic square coupling structure under circularly polarized illumination. Near the interface, an evanescent surface plasmon polariton (SPP) channel yields a Néel-type meron lattice with ± 1 2 \pm\tfrac{1}{2} -like effective site charges. At larger heights, diffracted fields from the square edges dominate and convert the lattice into a Bloch-type configuration. Over a range of intermediate heights, crossover between the evanescent SPP and edge diffraction gives rise to rich rapid topology evolutions. The switching is accompanied by nucleation of off-boundary vortex-anti vortex pairs in the in-plane spin phase, producing height-dependent fractional site charges. Our findings are analytically formulated by linear superposition of SPPs in the plasmonic regime and Stratton-Chu model in diffraction regime and confirmed via full-wave finite-difference time-domain simulations. 1 Introduction Recent progress in structured light and nanophotonics has shown that vectorial optical fields can host spatially organized optical-spin textures. At metal interfaces, surface plasmon polaritons (SPPs) are surface-bound electromagnetic modes with evanescent decay away from the interface; they carry transverse optical spin and exhibit spin–momentum locking, enabling helicity-dependent near-field responses in the immediate vicinity of the surface [ 5 , 1 , 26 , 12 , 10 , 9 ] . However, SPP-mediated near-field contributions decay with distance from the interface, whereas geometry-induced diffraction and radiative components persist to larger heights. Consequently, the local optical-spin texture above a finite plasmonic element is inherently height-dependent, and classifications based on a single observation plane can obscure the near-to-far evolution pathways [ 24 ] . A height-resolved description is therefore needed to separate textures dominated by near-field (evanescent) contributions from those reshaped by propagating components above finite apertures or elements. In the near field above a metal interface, SPP-associated fields provide a canonical setting in which transverse optical spin arises from the coupled in-plane and surface-normal field components and their fixed phase relation [ 27 ] . This structure leads to transverse optical spin whose sign is locked to the in-plane propagation direction of the surface mode [ 11 ] . In this work, we classify a meron texture as Néel-type when the in-plane optical-spin component aligns parallel or antiparallel to the local radial direction defined around an out-of-plane spin extremum, and as Bloch-type when the in-plane component circulates tangentially (perpendicular to that radial direction) [ 22 ] . In SPP-associated near-field regions, the in-plane optical-spin orientation is set by the local propagation direction and the surface normal, and the resulting meron helicity is generally consistent with the Néel-type definition [ 13 ] . Diffraction from a finite element can produce propagating field components with strong transverse gradients and a nonzero longitudinal electric-field component [ 7 , 4 ] . In the nonparaxial regime, spin–orbit interaction couples incident helicity to spatial structure, yielding structured phase and polarization distributions [ 6 ] . Such fields can host phase and polarization singularities that organize the local optical-spin texture in the transverse plane [ 20 ] . In particular, singularity-mediated structure can promote azimuthal circulation of the in-plane optical-spin component around out-of-plane spin extrema, consistent with the Bloch-type meron definition adopted here [ 29 , 30 ] . Recent free-space studies have constructed optical-spin meron lattices from the spin angular momentum vector field of diffracted beams and reported meron charges near \pm\tfrac{1}{2} , with the sign set by the incident helicity [ 14 ] . Both Néel-type and Bloch-type meron textures can arise above a finite plasmonic structure because the relative influence of near-field (evanescent) and propagating field components changes with observation height [ 8 ] . A height-resolved analysis can therefore reveal a crossover interval in which the local optical-spin texture evolves between these two limits. Here, we investigate this evolution for a metallic square coupling structure under circularly polarized illumination, and we track the optical-spin vector field on monitor planes along z. We observe a transition from a near-field Néel-type plasmonic meron lattice to a far-field Bloch-type meron lattice, with the crossover occurring over a finite height interval that is influenced by the lateral size of the structure in our geometry. The transition is accompanied by the nucleation of opposite-vorticity vortex–antivortex pairs in the in-plane optical-spin phase, which perturbs the phase-singularity ordering of the lattice [ 17 , 18 ] . Consistent with this defect-mediated reorganization, the site-resolved effective meron charges deviate from values near \pm\tfrac{1}{2} and take height-dependent fractional values [ 19 ] . 2 Conceptual framework We consider a square coupling structure on a metallic film illuminated by a normally incident circularly polarized plane wave, and we evaluate the scattered field in the upper half-space on monitor planes at increasing height z as shown in Figure 1 . In the following, z denotes the out-of-plane coordinate, and the discrete monitor-plane heights used in Figure 2 are written as h=z . The total field for z>0 in angular-spectrum representation is, \mathbf{E}(x,y,z)=\iint\tilde{\mathbf{E}}(k_{x},k_{y})\,e^{i(k_{x}x+k_{y}y)}\,e^{ik_{z}z}\,dk_{x}\,dk_{y},\qquad k_{z}=\sqrt{k_{0}^{2}-k_{x}^{2}-k_{y}^{2}}, (1) where \mathbf{E}(x,y,z) is the complex electric field at the observation point (x,y,z) , \tilde{\mathbf{E}}(k_{x},k_{y}) is the angular-spectrum amplitude, (k_{x},k_{y}) are transverse wave-vector components, and k_{z} is the longitudinal wave-vector component in the homogeneous half-space z>0 . The free-space wave number is k_{0}=2\pi/\lambda , where \lambda is the free-space wavelength. We define the transverse wave number as k_{\parallel}=\sqrt{k_{x}^{2}+k_{y}^{2}} . For k_{\parallel}>k_{0} , the longitudinal component is purely imaginary, k_{z}=i\alpha with \alpha=\sqrt{k_{\parallel}^{2}-k_{0}^{2}}>0 , and the corresponding angular-spectrum components decay as e^{-\alpha z} . For k_{\parallel}\leq k_{0} , k_{z} is real and the corresponding components propagate as e^{ik_{z}z} . This yields a spectrum-defined split of the total field into an evanescent sector and a propagating sector, \displaystyle\mathbf{E}(x,y,z) \displaystyle=\underbrace{\iint_{k_{\parallel}>k_{0}}\tilde{\mathbf{E}}(k_{x},k_{y})\,e^{i(k_{x}x+k_{y}y)}\,e^{-\alpha z}\,dk_{x}\,dk_{y}}_{\mathbf{E}_{\mathrm{evan}}(x,y,z)} (2) \displaystyle+\underbrace{\iint_{k_{\parallel}\leq k_{0}}\tilde{\mathbf{E}}(k_{x},k_{y})\,e^{i(k_{x}x+k_{y}y)}\,e^{ik_{z}z}\,dk_{x}\,dk_{y}}_{\mathbf{E}_{\mathrm{prop}}(x,y,z)}, Figure 1: Schematic of height-dependent topological switching of plasmonic spin meron lattices. Circularly polarized illumination of a square slit in an Ag film excites SPPs and generates a diffracted field. Spin textures are evaluated on monitor planes at increasing heights above the metal-dielectric interface. In the near field, the SPP-dominant regime corresponds to a Néel-type meron lattice associated with SPP standing-wave interference and spin-momentum locking. At intermediate heights, coexistence of residual SPPs and diffracted fields yields a twisted meron lattice. In the far field, where diffraction dominates, the spin texture evolves into a Bloch-type meron lattice, indicating height-controlled topological switching. where \alpha=\alpha(k_{\parallel}) depends on the transverse spectral coordinate. \mathbf{E}_{\mathrm{evan}} labels the evanescent-sector contribution in Eq. ( 2 ), which contains the SPP-near-field content above the metal interface, and \mathbf{E}_{\mathrm{prop}} labels the propagating radiative contribution dominated by diffraction from the finite square. In this work, we consider \mathbf{E}_{\mathrm{SPP}} for evanscent sector and \mathbf{E}_{\mathrm{diff}} for the propagating sector. The height coordinate controls the balance between \mathbf{E}_{\mathrm{SPP}} and \mathbf{E}_{\mathrm{diff}} and produces three regimes in Figure 1 . For the square geometry, we represent the SPP-dominant near-field content by a polygonal superposition of four edge-launched surface-wave components with polarization fixed by the SPP dispersion, \mathbf{E}_{\mathrm{SPP}}(x,y,z)=e^{-\kappa z}\sum_{n=1}^{4}E_{n}(x,y)\begin{pmatrix}i\,\dfrac{\kappa}{|k_{\mathrm{SPP}}|}\cos\theta_{n}\\[6.0pt] i\,\dfrac{\kappa}{|k_{\mathrm{SPP}}|}\sin\theta_{n}\\[6.0pt] 1\end{pmatrix}e^{i\phi_{n}},\qquad\theta_{n}=\frac{2\pi(n-1)}{4}. (3) Here k_{\mathrm{SPP}} is the complex in-plane SPP wave number at the metal interface, \kappa is the SPP out-of-plane decay constant in the region z>0 , E_{n}(x,y) is the complex in-plane envelope of the n th edge-launched contribution, and the phase set \{\phi_{n}\} satisfies \phi_{n}=\phi_{n}(\sigma) under the excitation convention \sigma=\pm 1 . We represent the propagating radiative contribution can be modeled using vectorial scattering theories [ 25 ] . We utilize an edge-only Stratton-Chu construction [ 14 ] over the four square edges to describe the total field in the radiative region, \mathbf{E}_{\mathrm{diff}}(x,y,z)=\frac{1}{4\pi}\sum_{m=1}^{4}\int_{\Gamma_{m}}\Big[i\omega\mu_{0}\,G(\mathbf{r},\mathbf{r}^{\prime})(\mathbf{n}\times\mathbf{H}_{\mathrm{incident}})+G(\mathbf{r},\mathbf{r}^{\prime})(\mathbf{n}\times\mathbf{E}_{\mathrm{incident}})\times\hat{\boldsymbol{\ell}}_{m}\Big]\,d\ell^{\prime}, (4) where Green function G(\mathbf{r},\mathbf{r}^{\prime})=\frac{e^{ik_{0}|\mathbf{r}\mathbf{r}^{\prime}|}}{|\mathbf{r}-\mathbf{r}^{\prime}|}, with \mathbf{r}^{\prime}=(x^{\prime},y^{\prime},0) , \mathbf{r}=(x,y,z) , \mathbf{n} the surface normal, \hat{\boldsymbol{\ell}}_{m} the unit tangent, \Gamma_{m} the m th edge. The additional details about the formalism are provided in Supplementary Section 1. We quantify the spin texture using the optical spin angular momentum density constructed from the complex fields, \mathbf{s}=\frac{1}{4\omega}\,\mathrm{Im}\!\left(\varepsilon\,\mathbf{E}^{\ast}\times\mathbf{E}+\mu\,\mathbf{H}^{\ast}\times\mathbf{H}\right),\qquad\mathbf{S}=\frac{\mathbf{s}}{|\mathbf{s}|}, (5) where \varepsilon and \mu denote the permittivity and permeability of the medium in the region z>0 .The near-interface regime is SPP-dominant and inherits spin-momentum locking of the surface wave [ 27 , 10 , 13 ] . The far-field regime is diffraction-dominant and inherits spin-orbit interaction of the propagating radiative field generated by the finite square [ 6 , 4 , 14 ] . The intermediate regime contains non-zero \mathbf{E}_{\mathrm{SPP}} and \mathbf{E}_{\mathrm{diff}} , and the total spin texture is set by their interference. Equations ( 3 ) and ( 4 ) provide a geometry-linked interpretation of Figure 1 : the SPP-dominant regime supports a Néel-type meron lattice near the interface, the diffraction-dominant regime supports a Bloch-type meron lattice at larger heights, and the meron texture switches between these limits across a finite intermediate interval where both channels contribute. We validate the conceptual framework using full-wave finite-difference time-domain (FDTD) simulations. Specifically, we record the complex fields \mathbf{E} and \mathbf{H} on a set of monitor planes at successive heights z, evaluate the local optical spin density \mathbf{s} from these complex fields, and obtain the normalized optical-spin field \mathbf{S} on each plane. To promote well-ordered meron-lattice formation within a finite square aperture, we choose the side length L to be in alignment with two characteristic in-plane periods: the SPP-associated interference scale set by Figure 2: Height-dependent evolution of spin structure in the XY plane. (a-c) Normalized out-of-plane spin component S_{z} (color scale) with overlaid in-plane spin vectors (S_{x},S_{y}) , evaluated at three monitor heights above the Ag surface: (a) h=100\,\mathrm{nm} (SPP-dominant), (b) h=2\lambda (intermediate), and (c) h=10\lambda (diffraction-dominant). (d-f) Line profiles along y=0 quantify the redistribution of spin components with height. The upper panels compare S_{z} and the in-plane spin magnitude |S_{\perp}|=\sqrt{S_{x}^{2}+S_{y}^{2}} , while the lower panels show S_{x} and S_{y} , capturing the progression from Néel-like to Bloch-like in-plane rotation as the observation height increases. \lambda_{\mathrm{SPP}}=2\pi/\mathrm{Re}(k_{\mathrm{SPP}}) [ 13 ] and the propagating diffractive scale set by the free-space wavelength \lambda [ 14 ] . Deviations from this square geometry can introduce boundary distortions and reduce lattice regularity. We therefore select L to satisfy an Least Common Integer-Multiple (LCIM) constraint, L=m\lambda_{\mathrm{SPP}}=n\lambda,\qquad m,n\in\mathbb{N}, (6) which promotes coherent superposition of SPP-associated near-field interference and edge-diffracted propagating contributions across the finite domain used for height-resolved characterization. Here, simulations are performed for silver (Ag) using the optical constants in Ref. [ 15 ] at a free-space wavelength of \lambda=550\,\mathrm{nm} . We choose the square side length such that L=55\,\lambda_{\mathrm{SPP}} (i.e., m=55 and n=53 in Eq. 6 ); unless stated otherwise, all reported numerical values in the manuscript correspond to these parameters. Further simulation parameters are provided in Supplementary Section S2, and Supplementary Section S3 discusses examples of dimensions that does not satisfy Eq. 6 . Figure 2 compares the spin components that quantify the height-driven conversion of the meron-lattice spin texture. Figures 2 (a-c) plot the normalized out-of-plane component S_{z} with overlaid in-plane vectors (S_{x},S_{y}) . In Figure 2 (a) at h=100\,\mathrm{nm} , S_{z} forms a square-symmetric meron lattice and the in-plane vectors connect adjacent extrema without azimuthal circulation around the S_{z} cores. In Figure 2 (b) at h=2\lambda , the S_{z} extrema broaden and the in-plane vectors rotate away from the edge-to-edge alignment, indicating a change in in-plane orientation within the same lattice period. In Figure 2 (c) at h=10\lambda , S_{z} extrema become more circularly symmetric and the in-plane vectors circulate around the S_{z} cores, consistent with a Bloch-type rotation pattern in the transverse plane. In Figures 2 (d-f), the upper panels compare S_{z} and the in-plane magnitude |S_{\perp}|=\sqrt{S_{x}^{2}+S_{y}^{2}} along y=0 , where peaks in |S_{\perp}| mark the lattice-boundary regions and peaks in S_{z} mark the lattice-core regions. In Figure 2 (d), the |S_{\perp}| maxima occur where S_{z} approaches zero, and the lower panel shows S_{y}\approx 0 along the cut with finite S_{x} , so the in-plane vectors on the cut are predominantly collinear, which corresponds to a Néel-type meron along this section. In Figure 2 (e), S_{z} remains non-zero over the boundary interval where |S_{\perp}| is large, and the lower panel shows finite S_{x} and S_{y} with unequal amplitudes and shifted extrema, which corresponds to a twisted meron configuration. In Figure 2 (f), the boundary peaks of |S_{\perp}| overlap with non-zero S_{z} , and the lower panel shows comparable S_{x} and S_{y} magnitudes with alternating signs across the period, which corresponds to a Bloch-type meron configuration. 3 Defect-mediated mechanism for height-controlled texture switching Figure 3: Defect nucleation in the in-plane spin phase across the crossover height. (a-c) Spin-dominance indicator \chi in the XY plane, with \chi\approx 0 contours delineating the evolving lattice boundaries at (a) h=100\,\mathrm{nm} , (b) h=3\lambda , and (c) h=10\lambda . (d-f) In-plane spin phase \psi=\arg(S_{x}+iS_{y}) at the same heights with overlaid in-plane Poynting vector, showing an ordered phase-singularity lattice in the near field (d), the emergence of off-boundary vortex-antivortex pairs in the intermediate regime (f), and the reconfigured singularity distribution in the diffraction-dominant regime (e). Insets label clockwise (red) and counterclockwise (blue) winding around representative singularities. (g) \psi(x,z) map at y=0 showing the onset of an additional singularity within the crossover height range and its persistence for larger z . The conversion of transverse spin texture with increasing height must be associated with an underlying mechanism that reorganizes the in-plane winding and drives the switching between the near-field and far-field lattice classes. We quantify the phenomenon using a spin-dominance factor \chi that compares the squared magnitude of the spin component normal to a given plane with the squared magnitude of the spin components within that plane [ 2 ] , \chi=|S_{\mathrm{out}}|^{2}-S_{\mathrm{in}}^{2},\quad S_{\mathrm{in}}^{2}=\sum_{i\in\mathrm{plane}}S_{i}^{2},\quad|S_{\mathrm{out}}|^{2}=|S_{j}|^{2},\ \ j\perp\mathrm{plane}. (7) Figure 3 (a-c) plots the \chi in the XY plane and uses the \chi\approx 0 contours to delineate the evolving lattice boundaries across height. In Figure 3 (a), the boundaries form a square-symmetric network that encloses well-separated S_{z} cores. As shown in Figure 2 (a), at the height h\sim 2\lambda though the SPP fields are decayed, the topological lattice remains robust. However, in Figure 3 (b) , the boundary network deforms and the unit-cell edges become spatially nonuniform indicating a crossover height between 2\lambda and 3\lambda . In Figure 3 (c), the boundary morphology differentiates between axial and diagonal sites, where axial sites approach circular contours while diagonal sites adopt higher-order polygonal outlines, which indicates that the lattice boundary is reshaped during the crossover. While Figure 2 (a) shows that the meron cores are anchored at fixed x as height increases, it is an indication that the change from Néel-type to Bloch-type rotation does not originate from a lateral relocation associated with S_{z} extrema. We therefore probe the in-plane winding directly using the in-plane spin phase \psi=\arg(S_{x}+iS_{y}) , shown in Figure 3 (d-f) inspired by the previous works [ 23 , 28 ] . The vorticity of each phase singularity is defined by the phase winding number v=\frac{1}{2\pi}\oint_{\mathcal{C}}\nabla\psi\cdot d\boldsymbol{\ell}=\frac{1}{2\pi}\oint_{\mathcal{C}}d\psi,\qquad v\in\mathbb{Z}, (8) where \mathcal{C} is a closed loop encircling a single singularity with nonzero |S_{\perp}| away from the core and d\boldsymbol{\ell} is the tangential line element. In Figure 3 (d), the \psi map contains an ordered singularity pattern within each diamond unit cell (dotted outline), and the overlaid in-plane Poynting vectors exhibit circulation with opposite handedness around distinct singularities. The inset summarizes the corresponding distribution of clockwise and counterclockwise vorticities, where the sign of v is assigned by the direction of phase increase along \mathcal{C} . With the increasing Figure 3 (e) shows that the ordered singularity pattern is disrupted within the crossover height range and new singularities appear away from the \chi\approx 0 lattice boundaries. The intermediate-height map contains paired singularities of opposite vorticity that nucleate within a single unit cell, which is the generic reaction channel of phase dislocations in nonparaxial interfering fields [ 16 , 18 ] . These off-boundary vortex-antivortex pairs redistribute the in-plane winding between the core and boundary regions and reassign the local rotation sense of (S_{x},S_{y}) around the same S_{z} extrema. In Figure 3 (f), the singularity distribution reorganizes into a height-stabilized pattern in the diffraction-dominant regime [ 18 ] . The accompanied movie of \psi evolution further illustrates the stability of these nucleated defects at h=10\lambda through the pulse duration. The vertical evolution is resolved in Figure 3 (g), where \psi(x,z) at y=0 shows the birth of an additional singularity at the previously identified crossover height and its persistence for larger z , which establishes defect nucleation in the in-plane spin phase as the mechanism underlying the height-controlled switching. We compare lattice site sizes and boundary deviations in the rapid conversion height zones compared to SPP and diffraction dominant zones in Supplementary S4. 3.1 Generalized charge accounting from polarity and vorticity We express the site-resolved meron charge as a sum of contributions from in-plane phase singularities weighted by the local out-of-plane polarity at their cores. For a singularity indexed by j , we define the polarity as p_{j}=\mathrm{sgn}\!\left[S_{z}(\mathbf{r}_{j})\right]~\mathrm{if}~\in\{-1,+1\}~\mathrm{|S_{z}|\neq 0}~\mathrm{else}~p_{j}=0 where the singularity position \mathbf{r}_{j} , and we use the vorticity v_{j}\in\mathbb{Z} from Eq. ( 8 ); the generalized per-site charge is then Q_{\mathrm{site}}=Q_{\mathrm{CW}}+Q_{\mathrm{CCW}}=\frac{1}{2}\sum_{j\in\mathrm{site}}p_{j}v_{j}, (9) where the partition into Q_{\mathrm{CW}} and Q_{\mathrm{CCW}} groups terms by the circulation sense identified in Figure 3 (d-f). In Figure 3 (d), each lattice site contains a single core-associated singularity with p_{j}\neq 0 , while the boundary singularities satisfy p_{j}=0 because S_{z} vanishes on the \chi\approx 0 contours, so Eq. ( 9 ) reduces to a half-integer site charge set by the core polarity and vorticity. In Figure 3 (e), the diffracted-field contribution perturbs the singularity ordering and drives a splitting of counterclockwise singularities into vortex-antivortex pairs with opposite v_{j} , followed by migration within the unit cell, so additional terms with p_{j}\neq 0 enter the sum in Eq. ( 9 ). This defect-mediated redistribution yields fractional Q_{\mathrm{site}} values whose magnitude varies with height, and the height dependence is quantified in the topology analysis where the site charges are extracted from the full skyrmion-density maps. The deviation from half integer values are also reported in literature in certain non-photonic systems (for example: [ 21 , 3 ] ). 3.2 A note on controllability of the critical height and sample size The critical height region is dependent on the coupling structure size. However, it is worth noting that the topological structure associated with SPPs persists even after SPP fields have diminished to near-zero values unless the some perturbation forces the change. At every given height, the strength of the perturbation due to the encroaching diffracted field at the monitor plane depends on the lateral size of the sample. For the coupling structure studied the critical height of conversion is at 2\lambda