Switchable Narrow Nonlocal Conducting Polymer Plasmonics

Article https://doi.org/10.1038/s41467-025-59764-5 Switchable narrow nonlocal conducting polymer plasmonics Dongqing Lin ,YulongDuan,PravallikaBandaru,PengliLi, Mohammad Shaad Ansari, Alexander Yu. Polyakov , Janna Wilhelmsen & Magnus P. Jonsson Dynamically switchable surface plasmons in conducting polymers constitute an emerging route towards intelligent metasurfaces, but polymer plasmons have so far suffered from weak resonances with low quality factors (Q<1-2). Here, we address this by nonlocal coupling of individual poly(3,4-ethylenedioxythiophene) (PEDOT) nanoantennas through collective lattice resonances (CLR) in periodic arrays (with resonance wavelengths around 2.0-4.5 μm). The results show that careful tuning of CLR matching conditions enables organic plasmonic resonances with Qup to 12. Angle-dependent extinction spectra connect the results to the enhancement of radiative coupling from diffractive lattice effects. Furthermore, the nonlocal coupling strength between nanoantenna units and lattice could be modulated via redox reactions, enabling the narrow CLRs to be reversibly switched with large modulation depth (between 7% and 45% extinction). By improving resonance strength and Q, the study circumvents previous limitations of conducting polymer plasmonics and shows feasibility for practical applications in active metasurfaces and nano-optics. Conducting polymer plasmonics1, where surface plasmon resonances originate from π-conjugated polymers, are emerging as a promising candidate in next-generation smart nano-optical devices with dynamic tuneability of light wavefronts2and spectral signals1,3. Conducting polymers such as poly(3,4-ethylenedioxythiophene) (PEDOT)4make organic plasmonics switchable by varying their intrinsic permittivities through the modulation of their redox states3, distinguished from gold or silver-based inorganic plasmonic materials5,6with fixed permittivities. In brief, redox-tuning controls the polymer doping level and thereby carrier density and mobility, offering electrical or chemical tuning between optically metallic and dielectric response1. Since 20201,7, investigations of polymer plasmonics cover both fundamental principles and device applications, including exploration of new polymer materials8,9, modulation of plasma frequencies and resonance wavelengths10–12, and applications in electrically switchable metasurfaces2,12,13. However, previous reports on conducting polymer plasmonics were limited to localized surface plasmonic resonances (LSPRs), which suffer from weak intensities and broad linewidths (full width at half maximum [FWHM] > 1–3μm). This puts restrictions and performance limits for many practical applications, including metalenses and nanosensors14–16. Likewise, the low quality factors (Q<2)of previous polymer plasmonics reflect low coherency and high optical losses, which is disadvantageous for nonlinear optics such as second/ third-harmonic generation17–19 or lasing actions20–22. By optimizing synthetic procedures23,24 or nanocrystallization processes25, conducting polymers may exhibit higher carrier mobilities, which should favor longer scattering times and provide narrower plasmon resonance linewidths10,26. Nevertheless, such improvement could be limited because of constraints related to hopping-like carrier transport in conjugated polymers4,27. Another strategy for high-Qplasmon resonances is to transform LSPRs into nonlocal collective lattice resonances (CLRs)6by coupling between single nanoantennas and diffractive modes of periodic arrays28–30. Collecting optical energy by adjacent nanoantennas Received: 12 February 2025 Accepted: 2 May 2025 Check for updates Laboratory of Organic Electronics, Department of Science and Technology (ITN), Linköping University, Norrköping, Sweden. e-mail: magnu[email protected] Nature Communications | (2025) 16:4484 1 1234567890():,; 1234567890():,; strengthens light-matter interactions31, enables higher local density of optical states20, enhances optical fields32, and elongates plasmon lifetimes33. These features suppress radiation and dissipation losses32, and can increase Qfrom 5-20 (LSPR) to 100–5000 in gold nanoantenna arrays6,15,32. Although PEDOT nanoantenna arrays have been demonstrated to observe broad and weak LSPRs, they did not exhibit clear narrow CLR features10,12. An essential aspect here is that the large permittivity difference between PEDOT and a conventional metal like gold significantly affects how to achieve CLR matching conditions between the polarizabilities of single nanoantennas (α iso ) and array factors (S)6,34,35. This leads to mismatch if simply exchanging gold for PEDOT under the same nanoantenna dimensions and lattice spacings. The critical development of conducting polymer CLRs for high-performance all-organic metasurfaces thereby remains. An important aspect in this respect is the potential to also modulate the coupling strength via the polymer’s redox state, aiming at offering CLR actions with reversible switching ability between “on”and “off”states. Herein, we utilized PEDOT nanodisk antenna arrays to evolve low-quality LSPRs into more intense CLRs (Fig. 1a) with up to ten times narrower linewidths. The PEDOT used in this work was acidtreated PEDOT:Tosylate (PEDOT:Tos) made through vapor phase polymerization as detailed in the Methods section and in previous works36. Periodic nanoantenna arrays made from this material were fabricated through electron-beam lithography (Fig. 1b) for accurate control of nanoantenna dimensions and periodicity (r) (Fig. 1c). This allowed careful tuning of coupling interactions between individual nanoantennas and lattices to maximize diffractive effects and minimize damping relaxation rates, with details revealed through the dependence of periodicity and incidence angle. Optimizing the geometric matching conditions between array periodicity and nanoantenna dimensions enabled narrow conducting polymer CLRs + -+ - + -+ - + - Localized surface plasmonic resonance (LSPR) Collective lattice resonance (CLR) + - + -+ - Nonlocal mode Oxidized state of PEDOT chains + - + - Switch on Switch off Redox reactions Dynamic switching a Substrate b Substrate PEDOT Substrate iii iii iv v c 2 μm On state Off state 123 CLR (with coupling) LSPR (without coupling) Extinction Wavelength (μm) S O O O O S S O O Fig. 1 | Concept and structure of switchable collective lattice resonances with PEDOT nanoantenna arrays. a The transformation from PEDOT nanoantenna LSPR to CLR,with dynamic switchability. The pink-color and red-color regions mark the weak (isolated nanoantenna) and intense collective (periodic array) optical response, respectively. Simulated example extinction spectra of an isolated PEDOT nanoantenna showing a broad LSPR peak (obtained without periodic boundaries) and an optimized PEDOT nanoantenna array showing CLR (obtained using periodic boundaries), are also provided in corresponding colors. The blue dashed line frame presents the molecular structure of PEDOT in its oxidized high conducting state. The black dashed frame further below illustrates the switch between “on”and “off” states via redox reactions. bThe preparation procedure of PEDOT-based periodic arrays consisting of nanodisk units. The procedure includes: (i) preparing a layer of PEDOT film; (ii) spin-coating a layer of positive resist; (iii) electron-beam lithography, including development; (iv) dry etching with oxygen plasma; (v) removing the positive resist on top of the nanodisks. cSEM image of a typical final periodic PEDOT nanoantenna array. Article https://doi.org/10.1038/s41467-025-59764-5 Nature Communications | (2025) 16:4484 2 (FWHM < 0.4–0.5 μm) across the mid-infrared wavelength range from 2.1 to 4.5 μm. We further study effects of modulating the redoxstate of the PEDOT, and demonstrate switchability with large modulation depth enabled by reversible tuning between matching and mismatching CLR conditions. With Qvalues over 10, the CLRs show a dramatic improvement compared to LSPRs from previous conducting polymer nanoantennas. The work thereby overcomes one of the main bottlenecks of conducting polymer plasmonics, showing promise for their integration toward practical applications in smart optical metasurfaces. Results Determining CLR matching conditions for PEDOT nanoantenna arrays In periodic arrays, the polarizability of a single nanoantenna is transformed from the isolated α iso to a periodic mode α p through latticeinduced coupled dipole interactions (S), according to6,34,37: αp=ð1=αiso SÞ1ð1Þ Squantifies the sum of coupled dipole interactions (including electrostatic and radiative couplings38,39) generated from other nanoantenna units in the lattice and can be evaluated using34,37: S=X N j expðikrjÞð1ikrjÞð3cos2θj1Þ r3 j +k2sin2θj rj "# ð2Þ where r j is the center-to-center distance between two nanoantennas, θ j is the angle between the electrical field direction and the lattice vector direction, and Nis the number of other nanoantennas along the specific direction. To achieve plasmon resonance with an intensified extinction σof the system (from σ∝|k|Im(α p ), where kis the wave vector with |k|=2πn s /λ, with wavelength λand refractive index of medium n s ), the CLR matching condition relates to minimizing the denominator (1/α iso –S) to maximize α p . At the resonance wavelength (λ r ), the CLR therefore satisfies Re(1/α iso )=S r ,whereRe(1/α iso )andS r are the real parts of 1/α iso and S, respectively6,37. Meanwhile, the difference between the imaginary parts of 1/α iso and S(denoted as Im(1/α iso )andS i , respectively) should be sufficiently small to ensure low damping relaxation rates and to enable spectrally sharp |α p |for narrow resonances6. Considering corrections for the modified longwavelength approximation34,37,α iso of isolated PEDOT nanodisk antennas is evaluated through34: αiso =αs1k2 aαs2 3ik3αs "# 1 ð3Þ where α s is the static non-corrected polarizability obtained by approximating the nanodisks as oblate spheroids via: αs=a2b 3 εmεs εs+LðεmεsÞð4Þ where aand bdenote the long and short semiaxes of the oblate spheroid, respectively. Lis a geometric factor, and ε m and ε s are the relative permittivity of PEDOT (Supplementary Fig. 3) and the surrounding medium, respectively. Considering the inhomogeneous environment consisting of dielectric substrate (n s =1.5) and air, we applied ε s = 1.3 in the calculations of α iso . Figure 2a presents the real and imaginary components of S(full lines) for four periodicities together with 1/α iso (dashed lines) for an isolated PEDOT nanoantenna (diameter [d]=0.52μmandheight=0.2 μm). To ensure efficient dipolar coupling interactions from sufficient nanoantenna units, we set N= 1000 in the calculation of S(Supplementary Fig. 4). The conditions that satisfy both Re(1/α iso )=S r and minor |Im(1/α iso )-S i | values are marked by purple dots, which represent the resonance wavelengths λ r for respective periodic distance (r). By increasing rfrom 1.0 to 1.6 μm, S r increases from -6 to 33 μm−3at respective λ r (Figs. 2a, b), leading to red-shifted resonances due to dipolar coupling interactions35,39. Importantly, increasing rmoves λ r closer to the spectral position of the peak wavelength of S r for the respective period (λ≈n s r), indicating the enhancement of coupling to in-plane diffraction orders from the lattice effect37. We further note that the value of |Im(1/α iso )−S i | at respective λ r also gradually reduces with increasing periodicity, from 41 (r=1.0μm) to 17 μm−3(r=1.6μm). Corresponding decreased damping relaxation rate should aid resonance strength and narrowing of linewidths for larger r34,35.Figure2c confirmsthisbehavior,whichpresents|α p | for the different PEDOT arrays along with |α iso | for the isolated plasmonic PEDOT nanoantenna (also see Supplementary Fig. 6). While most arrays show |α p | peaks that are sharper and more intense compared with the isolated nanoantenna, the peak gradually increases in intensity and narrows for rup to 1.6 μm. For yet larger r(r≥1.8 μm), the intensity of |α p | sharply reduces at λ r ≈n s r, because the peak intensity of S r becomes too low to satisfy Re(1/α iso )=S r (Supplementary Fig. 7 and 8). Combined, the above results suggest that the optical extinction of CLRs based on these PEDOT nanoantennas can be sharper and more intense under optimized periodic condition r=1.4–1.6 μm. The above strategy to optimize CLRs also work for other nanoantenna dimensions (Supplementary Fig. 9-16) and for other types of PEDOT1(Supplementary Figs. 17–19), demonstrating good generality. Based on the prerequisite of Re(1/α iso )=S r ,wededucedtherelationship between Sand normalized detuning wavelength (Δ)atλ r for square arrays with varying periodicity (mathematic processing shown in Supplementary Note 2). Δis used to evaluate the wavelength difference between the surface plasmon resonance (λ r ) and latticeinduced diffraction orders (λ=n s r) via the equation32 Δ=(λ r -n s r)/n s r. The smaller wavelength difference (Δ→0) reflects the more significant contribution of nonlocal diffraction orders to the surface plasmon resonance, which serves as prerequisite for CLRs. S(r,Δ) mappings are shown in Fig. 2d (real part) and 2e (imaginary part). Under various periodic distances r=0.9~1.7 μm, S r increases for reducing Δfrom Δ>0.1toΔ< 0.001 (Fig. 2d), according to the scaling law S r ~(1+Δ)−2 (extracted from Supplementary Equation [10]). These results confirm that the intense dipolar interactions are closely related to efficient coupling to nonlocal diffraction orders from the lattice32.InFig.2e, diminishing Δresults in more negative S i according to the scaling law S i ~-(1 + Δ)−1(extracted from Supplementary Equation [11]), and achieves lower |Im(1/α iso )-S i | , which is correlated to increasing damping relaxation lifetimes and reducing resonance linewidths. By fixing nanoantenna dimensions (with unchangeable 1/α iso ), Δis lower for increasing periodicity (Fig. 2d), which fulfills the scaling law r~(1+ Δ)−2/3 (extracted from Supplementary Equation [10]). Therefore, suitably large periodic distances (r=1.4–1.6 μm) should be designed to access low-Δnarrow resonances. We also evaluate hexagonal arrays of PEDOT nanoantennas to theoretically test the generality of CLR actions (Supplementary Note 2). Compared with square arrays under the same r,hexagonal arrays increase Sby ~1.5 times, owing to enhanced dipolar coupling interactions from more adjacent nanoantennas. This feature allows hexagonal arrays to fulfill Re(1/α iso )=S r under more extensive periodic distances (optimized rup to 1.8 μm, Fig. 2b and Supplementary Fig. 2023). Meanwhile, larger-magnitude negative S i of hexagonal arrays is more effective to cancel out the optical loss from damping relaxation, confirmed by lower values of |Im(1/α iso )-S i | = 12-14 μm−3than those in square arrays (|Im(1/α iso )-S i |≥17 μm−3) under the individually optimized periodic conditions (Fig. 2b). In addition, hexagonal arrays with low Δcan be reached at r= 1.6-1.8 μm (Supplementary Fig. 20 and 22). Article https://doi.org/10.1038/s41467-025-59764-5 Nature Communications | (2025) 16:4484 3 These results confirm the possibility of high-QCLRs also in PEDOTbased hexagonal arrays. Experimental demonstration of spectrally narrow conducting polymer CLRs To put the above analysis to experimental scrutiny, we prepared PEDOT nanodisk arrays (see scanning electron microscopy [SEM] images in Supplementary Fig. 24-25) on glass substrates (n s ≈1.5) by electron beam lithography and compare experimental extinction spectra with results from finite-difference time-domain (FDTD) simulations. Beginning with square arrays, we note that all systems exhibited clear extinction peaks that redshifted (λ r = 1.80 to 2.32 μm) when increasing rfrom 1.0 to 1.5 μm(Fig.3a). The simulations show similar results (λ r =1.80–2.41 μm, for r=1.0~1.6 μm), consistent with more significant S r originating from dipolar coupling35. Importantly, the redshift is accompanied with significant narrowing of the peak linewidth from 1.0 μm(r=1.0 μm) to 0.39 μm(r=1.5 μm), corresponding to improvement of Qfrom 1.8 to 6.0. FDTD simulations confirm the reduction in linewidth (from 0.9 ~ 1.0 μmatr=1.0μmto0.25μmat r=1.5–1.6 μm) and show Q-factors enhanced to 10. We note that λ r for these narrow resonance peaks approaches the wavelength λ≈n s r,in agreement with enhancement of nonlocal coupling to lattice modes. In addition, the narrowing of resonance linewidths for redshifted λ r can be connected to lower radiation loss33,35 from damping relaxation, confirmed by smaller values of |Im(1/α iso )-S i | at larger r(Fig. 2b). Further increasing r≥1.8 μm sharply diminishes the optical extinction of CLR at λ r ≈n s r(Supplementary Fig. 7), consistent with lowering of | α p | (Supplementary Fig. 8) around λ r . If the periodicity is r=3.0 μm, the linewidth of the surface plasmon resonance at λ r ≈2.1 μmbecomes much broader (FWHM = 1.1 μmandQ≈1.9), which is similar to the LSPR of single nanoantennas (Q≈2) observed by FDTD simulations (Supplementary Fig. 27). PEDOT-based hexagonal arrays confirm the generality of forming narrow CLRs with conducting polymers (SEM images in Supplementary Figs. 28, 29). The experimental results (Fig. 3b) show clear CLRs with λ r redshifted from 1.78 (r=1.2μm) to 2.25 μm(r=1.7μm), and the resonance linewidth is reduced to 0.36 μm, corresponding to enhancement in Q-factor up to 6.5. These results are corroborated by the FDTD simulations, which suggests Q-factors as high as 10. The reduction in resonance linewidth can be directly correlated with decreased Δthrough the scaling law FWHM ~ Δ0.5, which shows good agreement when jointly plotting results for both square arrays and hexagonal arrays (Fig. 3c). These results reflect that the surface plasmon resonance from PEDOT nanoantennas is transitioned from local (Δ→1) to nonlocal modes (Δ→0) by increasing the periodicity. Considering Q d ~Δ−0.5,whereQ d is the dissipation quality factor32,the nonlocal resonance mode diminishes the dissipation loss in the PEDOT nanoantenna arrays. In addition, minor Δcorrelates with more negative S i (Fig. 2e), which suppresses the damping relaxation in LSPR and reduces the radiative loss. Therefore, low Δfavorably lowers optical losses and aids the narrow conducting polymer plasmon resonances in both square and hexagonal arrays. So far, the study has focused on CLRs based on one particular PEDOT nanodisk diameter (d). We will now show that the narrow CLRs can be tailored to different spectral positions across the midinfrared wavelength region by also allowing variation of d.Figure3d shows experimental results for square (top) and hexagonal (bottom) arrays with jointly optimized dand r(on calcium fluoride substrate). For the square arrays, increasing rfrom 2.4, 2.7, 2.9, and 3.0 μmand optimizing nanodisk diameters for each period (d= 0.9, 1.08, 1.07, and 1.1 um, respectively, Supplementary Fig. 31), polymer CLRs with FWHM = 0.4–0.5 μm were obtained at λ r from 3.48, 3.88, 4.15 to 4.40 1.2 1.6 2.0 2.4 -80 -40 0 40 0 100 200 300 S r or Re(1/ D iso )(μm -3) Wavelength (μm) Im(1/ D iso ) S i or Im(1/ D iso )(μm -3) r=1.0μm r=1.2μm r=1.4μm r=1.6μm Re(1/ D iso ) 248 0 Sr 0 -87 Si ab c d e 1.0 1.2 1.4 1.6 1.8 -30 -15 0 15 30 S r at O r( μm-3) r(μm) Square array Hexagonal array 10 20 30 40 50 60 |Im 1/ D iso -S i |at O r( μm-3) Square array Hexagonal array 10 -2 10 -1 1.0 1.2 1.4 1.6 Δ r(μm) 10-2 10-1 1.0 1.2 1.4 1.6 Δ r(μm) 10 -3 1.5 1.8 2.1 2.4 2.7 0.00 0.02 0.04 0.06 0.08 r=1.0μm r=1.2μm r=1.4μm r=1.6μm | D iso | ×2 | D p| ( μm3 ) Wavelength (μm) Fig. 2 | Calculations of array factors for PEDOT nanoantenna arrays with different periodic distances. a Analysis of the real and imaginary part of array factors (S r and S i , respectively), together with 1/α iso , under the periodic distance r= 1.0, 1.2, 1.4, and, 1.6 μm. 1/α iso also has a complex form, where Re(1/α iso ) and Im(1/α iso )are the real and imaginary part of the reverse isolated polarizability, respectively. The length of the region with 0% transparency for each dashed line indicates the magnitude of |Im(1/α iso )-S i |. bS r and |Im(1/α iso )-S i |atλ r , for both square and hexagonal arrays. cMagnitude of periodic polarizability (|α p |) for PEDOT nanoantenna square arrays, along with |α iso | from LSPR of an isolated nanoantenna. PEDOT nanodisks are set with d=0.52μm and a height of 0.2 μm. d,eS r and S i as a function of normalized detuning wavelength (Δ)andr,atλ r . Article https://doi.org/10.1038/s41467-025-59764-5 Nature Communications | (2025) 16:4484 4 μm (approximate to λ r ≈n s rfor the square arrays). The Q-factors for these optimized CLRs ranges from 9 to 11. Similarly, increasing rfor hexagonal PEDOT nanoantenna arrays (2.7, 3.0, and 3.2 μm, respectively, Supplementary Fig. 33) and optimizing d(1.10, 1.16, and 1.21 um, respectively) enabled λ r modulation from 3.47, 3.75 to 4.03 μm. These peak wavelengths are consistent with λ r ≈0.874n s rfor hexagonal arrays. The resonance linewidths for these hexagonal CLR arrays maintained below 0.4 μm, corresponding to experimental Qfactors up to 12. As illustrated in Fig. 3e, the Q-factors reported in this study are 4-10 times higher than those in previous works of conducting polymer nanoantennas (FWHM ≥2-3 μmandQ≈1–2) in the mid-infrared wavelength region, the latter of which are dominated by LSPRs1–3,8–12. In the shorter wavelength regions (λ r < 2.5 μm), the Qfactors could be enhanced by 1–4 times by forming CLRs (versus LSPR with FWHM ≈0.8–1.0 μm). The concept is not restricted to the type of PEDOT that we focus on in this work, but forms a general strategy to obtain narrow resonances with optically metallic polymers (see simulation results in Supplementary Fig. 35). Understanding conducting polymer CLRs via angle-dependence To determine the mechanisms for nonlocal coupling in PEDOT nanoantenna arrays, we study the effect of varying the incident angle (θ i ). The incident light was polarized in transverse-magnetic (TM) mode (p-polarization, see Fig. 4a, b). As illustrated in Fig. 4b, the coupling with (±1, 0) Rayleigh anomaly (RA) can be assigned to the diffraction grating effect that obeys the equation31 λ r =r(n s ±sinθ i ) along the X-axis. Along the Y-axis, the coupling with (0, ±1) RA that follows the equation λ r 2=r2(n s 2–sin2θ i ) can be assigned to the dipole ab d e c 234567 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 Normalized Extinction Wavelength (μm) r = 2.7 μm (d = 1.1 μm) r = 3.0 μm (d = 1.16 μm) r = 3.2 μm (d = 1.21 μm) Normalized Extinction r = 2.4 μm (d = 0.9 μm) r = 2.7 μm (d = 1.08 μm ) r = 3.0 μm (d = 1.07 μm) r = 3.2 μm (d = 1.1 μm) Square arrays Hexagonal arrays 12345 2 4 6 8 10 12 14 [1] [2] [3] [8] [9] [10] [11] [12] This work Q-factor Wavelength (μm) Previous works r r 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.2 0.4 0.6 0.8 1.0 1.2 Square array (simulation results) Hexagonal array (simulation plots) Square array (experimental results) Hexagonal array (experimental results) Fitting FWHM (μm) '0.5 FWHM ~ '  0.0 0.2 0.4 0.6 0.8 1.0 1.0 1.5 2.0 2.5 3.0 0.2 0.4 0.6 0.8 1.0 Normalized Extinction 1.0 μm 1.1 μm 1.2 μm 1.3 μm 1.4 μm 1.5 μm r Normalized Extinction Wavelength (μm) 1.0 μm 1.1 μm 1.2 μm 1.3 μm 1.4 μm 1.5 μm 1.6 μm r Experiments Simulations 0.0 0.2 0.4 0.6 0.8 1.0 1.0 1.5 2.0 2.5 3.0 0.2 0.4 0.6 0.8 1.0 Normalized Extinction 1.2 μm 1.3 μm 1.4 μm 1.5 μm 1.6 μm 1.7 μm Normalized Extinction Wavelength (μm) 1.0 μm 1.1 μm 1.2 μm 1.3 μm 1.4 μm 1.5 μm 1.6 μm 1.7 μm 1.8 μm r r Experiments Simulations Fig. 3 | Extinction spectra and corresponding performance parameters for PEDOT nanoantenna arrays under normal incidence. a Experimental results and FDTD simulations of r-dependent extinction spectra for square arrays. The pattern of square array is shown with a periodic boundary (marked by the orange dashed line). bExperimental results and FDTD simulations of r-dependent extinction spectra for hexagonal arrays. The pattern of the hexagonal array is shown with a periodic boundary (marked by the orange dashed line). For (a)and(b), the nanoantennas’diameter (d)wasfixed as 0.52 μm in the FDTD simulations, while the measured diameter ranged from around 0.40 to 0.50 μm for the experimental results (along with the height of ~0.2 μm). cLinewidths (represented by FWHM) for different experimental and simulated arrays plotted as a function of normalized detuning wavelength (Δ). dExtinction spectra of PEDOT nanoantenna arrays optimized for CLRs at longer peak wavelengths, with rincreased to 2.4–3.2 μmand dincreased to 0.8–1.3 μm. All PEDOT materials are based on acid-treated PEDOT:Tos. eExperimental Q-factors according to this work and previous works on conducting polymer nanoantennas1–3,8–12. Article https://doi.org/10.1038/s41467-025-59764-5 Nature Communications | (2025) 16:4484 5 radiation effect33, as the long-range coupling interaction perpendicular to the electric polarization. Figure 4c-h show simulated (Fig. 4c, e, g) and experimental (Fig. 4d, f, h) angle-dependent extinction spectra for PEDOT nanoantenna arrays with constant d(0.52 μm for simulations and ~0.5 μm also for the experiments, Supplementary Figs. 24, 25) and increasing ras detailed below. In Fig. 4c, d, the smallest array period (r=1.0μm) shows the least clear CLR at normal incidence among the studied systems. Under small incident angles (θ i ≤10°), the array structure exhibits an angle-independent broad feature. Such nondispersive mode reflects low coupling interactions from lattice effects, similar to LSPR31, consistent with low contribution from S r and high Δ. Further increasing sin θ i enables the introduction of (−1, 0) RA coupling (on glass substrate n s = 1.5) from diffraction grating effects, which facilitates the transformation from LSPR into a clear but weak CLR. However, (0, ±1) RA coupling from dipole radiation is not observed even for larger angles, in agreement with a large |Im(1/α iso )−S i |that indicates large damping relaxation. These features without efficient radiative coupling are also observed in periodic arrays with r=1.1μm (Supplementary Fig. 36). When increasing rto 1.3 μm (simulation results in Fig. 4eand experimental results in Fig. 4f), the resonance peak is split into two dispersive modes upon increasing the incidence angle, where one mode has the coupling with (−1, 0) RA (n s = 1.5) from diffraction grating effect and the other mode exhibits coupling with (0, ±1) RA from dipole radiation. At even larger sin θ i =0.4–0.5, such radiative (0, ±1) RA coupling interacts with the (−1, 0) RA (n s = 1.0, in the air), which generates an additional sharp resonance. These results verify the enhancement of nonlocal diffraction orders in plasmonic resonances when the CLR matching conditions are combined with moderate S r =12μm−3and Δ= 0.04. As expected, the above coupling features are 0.0 0.1 0.2 0.3 0.4 0.5 1.4 1.6 1.8 2.0 2.2 sin θ i Wavelength (μm) (-1,0) RA with n s = 1.5 (-1,0) RA with n s = 1.0 0.0 0.1 0.2 0.3 0.4 0.5 1.6 1.8 2.0 2.2 2.4 sin θ i Wavelength (μm) (-1,0) RA with n s = 1.5 (-1,0) RA with n s = 1 (0,±1) RA 0.00.10.20.30.40.5 1.8 2.0 2.2 2.4 2.6 sin θ i Wavelength (μm) (-1,0) RA with n s = 1.5 (0,±1) RA (-1,0) RA with n s = 1.0 0.5 1 1 0.5 1 0 1 0.5 1 0 1 X θi E(TM mode) k 0.82 4.23 |E| a b cd ef gh i 0 0.0 0.1 0.2 0.3 0.4 0.5 1.8 2.0 2.2 2.4 sin θ i Wavelength (μm) (0,±1) RA (-1,0) RA with n s = 1 (-1,0) RA with n s = 1.5 0.0 0.1 0.2 0.3 0.4 0.5 1.6 1.8 2.0 2.2 sin θ i Wavelength (μm) (-1,0) RA with n s = 1.5 -1.2 -0.6 0.0 0.6 1.2 -1.2 -0.6 0.0 0.6 1.2 X (μm) Y (μm) (0,±1)RA E (±1,0)RA r = 1.0 μm r = 1.3 μm r = 1.5 μm r = 1.0 μm r = 1.3 μm r = 1.5 μm Simulations Experiments -0.9 -0.6 -0.3 0.0 0.3 0.6 0.9 0 1 2 3 4 5 1.0 μm 1.1 μm 1.2 μm 1.3 μm 1.4 μm 1.5 μm 1.6 μm |E| X (μm) r kx 0.0 0.1 0.2 0.3 0.4 0.5 1.8 2.0 2.2 2.4 2.6 sin θ i Wavelength (μm) (0,±1) RA (-1,0) RA with n s = 1 (-1,0) RA with n s = 1.5 Fig. 4 | Angle-dependent optical response of PEDOT nanoantenna arrays. aIllustration of the experimental conditions based on the transverse magnetic (TM) mode, where the orientation of the electrical field (E) and wave vectors (k)and incident angle (θ i )arealsodefined. bIllustrations of the in-plane components ofthe electric field polarization and wave vectors(along the X-axis, marked ask x )fromthe incident light, and the diffraction orders of Rayleigh anomaly (RA) based on the electrical field distribution of square arrays, as calculated by FDTD simulations. c,e,andgSimulated angle-dependent extinction spectra (normalized) for PEDOT nanoantenna square arrays with r=1.0,1.3,and1.5μm, respectively. The experimental results (normalized extinction) with r=1.0,1.3,and1.5μm, are shown in (d,f,h), respectively. The red dashed line represents the RA originating from dipolar radiation, and the yellow dashed lines are the RAs derived from diffraction grating effect based on refractive index of medium (n s ). iElectric field distribution (denoted by the magnitude |E|) of square arrays with different r=1.0–1.6 μmatλ r , where the nanoantenna was centered at (X=0,Y= 0). For band i,theZcoordinate for the electric field distribution was set at the interface between the PEDOT nanoantennas and the dielectric substrate. The diameter of the PEDOT nanoantennas was set to 0.52 μm for all panels (≈0.5 μm in the experiments). Article https://doi.org/10.1038/s41467-025-59764-5 Nature Communications | (2025) 16:4484 6 also observed for the PEDOT nanoantenna array with optimized period (r=1.5μm, simulations in Fig. 4g and experimental results in Fig. 4h). The peak associated with diffraction orders from dipole radiation is more intense than that from diffraction grating effects. This enhanced radiative coupling is mainly attributed to slow-decaying radiation fields, which agrees with further decrease of |Im(1/α iso )−S i |when increasing rfrom 1.3 μmto1.5μm. To confirm the contribution of dipole radiations to the CLR actions, we investigate the electric field distributions of the PEDOT nanoantenna arrays (at λ r ) via FDTD simulations. At the hotspot positions of the PEDOT nanoantennas (at X=±0.26-0.28μminFig.4i), increasing rfrom 1.0 to 1.6 μmgradually enhances the magnitude of the electric field |E|, which is also observed in hexagonal arrays (Supplementary Fig. 39). These results reflect the intensified dipole radiations due to the lower |Im(1/α iso )−S i |(Fig.2b) and the enhanced |α p |(Fig.2c) for the arrays with larger r. However, if the periodicity is r=3.0 μm, the broad peak at λ r ≈2.1 μmdoesnot showthetypeofangle-dependentfeaturesthatwouldberequiredfor either diffraction grating effect or lattice dipole radiation coupling (Supplementary Fig. 40). This confirms that such large periodicity for the same nanoantenna dimensions leads to optical response dominated by LSPR, without major contribution from lattice effects. Localized resonance modes were also observed if limiting the lattice size [(N+1)×(N+ 1)] to only allow coupling to the closest neighboring nanoantennas (N= 1, Supplementary Fig. 41). We estimate a lattice size of N≈150 to be sufficient to support efficient nonlocal resonance with intense coupling to dipole radiation (Supplementary Fig. 42). Therefore, effective radiation coupling from suitably periodic long-range lattices is key in narrow conducting polymer CLRs, similar to goldbased systems33. Redox-switching of the CLR matching conditions We further demonstrate switching of the PEDOT-based CLR actions by modulating the redox-state of the polymer. While such tuning is known to affect the response of individual nanoantennas, effects here also involve modulation of theCLR matching conditions. We selected a square array with r=1.3 μm(λ r = 2.03 μm) of PEDOT nanoantennas with d=0.5 μmandaheightof0.2μm. Figure 5a illustrates how to switch the redox state of PEDOT by oxidation with hydrochloric (HCl) acid vapor40 or by reduction with poly(ethylenimine) (PEI) vapor1,40, where the latter diminishes π-electron delocalization and decreases 1.5 1.8 2.1 2.4 2.7 0.00 0.01 0.02 0.03 0.04 0.05 Oxidized state Reduced state Im( D p ) Wavelength (μm) PEI HCl Oxidized state Reduced state a 0.5 b On state Off state + - + - 0.0 0.1 0.2 0.3 0.4 0.5 1.6 1.8 2.0 2.2 2.4 sin θ i Wavelength (μm) (-1,0) RA with ns=1.5 (-1,0) RA with ns=1 (0,±1) RA 0.5 1 Normalized Extinction Re-oxidized + e - - e - cd e f Off state On state mismatch match On state Off state 1.0 1.5 2.0 2.5 0.0 0.2 0.4 0.6 0.8 1.0 Oxidized state 1st 2nd 3rd 4th Reduced state 1st 2nd 3rd 4th Normalized Extinction Wavelength (μm) r=1.3μm 0 110 220 330 1.8 2.0 2.2 2.4 -207 -138 -69 0 S r or Re(1/ D iso )(μm -3 ) S r Oxidized Re(1/ D iso ) Reduced Re(1/ D iso ) S i or Im(1/ D iso )(μm -3 ) Wavelength (μm) S i Oxidized Im(1/ D iso ) Reduced Im(1/ D iso ) 1.0 1.5 2.0 2.5 10 20 30 40 50 Oxidized Reduced Re-oxidized Extinction (%) Wavelength (μm) r=1.3μm S O O O O S S O O S O O O O S S O O Fig. 5 | Redox switching of PEDOT nanoantenna CLRs. a Illustration of the redox modulation process of PEDOT nanoantennas. The reduced state is formed by PEI vapor treatment while the oxidized state could be retrieved via hydrochloric (HCl) vapor treatment. bMeasured extinction spectra of a PEDOT-based periodic nanoantenna array in the oxidized state (on), reduced state (off), and re-oxidized state after one full redox-cycle (on). cMeasured extinction spectra (normalized) of the PEDOT nanoantenna array in the oxidized (during the first redox cycle) and reoxidized states (during the 2nd~4th redox cycle), along with the reduced state (during the 1st–4th redox cycle). dMeasured angle-dependent extinction spectra (TM mode) of the PEDOT nanoantenna array in the re-oxidized state during the 4th redox cycle. eAnalysis of (1/α iso )-Srelationships for both oxidized and reduced states of the PEDOT nanoantenna array, including both the real (Re[1/α iso ]andS r ) and imaginary (Im[1/α iso ]andS i ) components. The calculations of oxidized and reduced 1/α iso are based on the permittivity of the PEDOT material in the oxidized and reduced states, respectively. The red shaded point represents the matching conditions for CLRin the on state. fThe imaginary part of the periodic polarizability (Im[α p ]) of the PEDOT-based periodic array in the different redox states, as obtained by calculations through Eq. (1). For both experimental results and theoretical calculations, the PEDOT nanoantenna units are on glass and have a diameter of 0.5 μm and a height of 0.2 μm in an array structure with periodicity r=1.3μm. Article https://doi.org/10.1038/s41467-025-59764-5 Nature Communications | (2025) 16:4484 7 the electrical conductivity by at least three orders of magnitude. In the oxidized high-conducting state of the PEDOT, the nanoantenna array provides a narrow CLR extinction peak at 2 μm, with peak extinction of around 45% and FWHM = 0.43 μm(Fig.5b). When reducing the material by exposure to PEI vapor for 5 min (under nitrogen atmosphere), this intense plasmonic CLR peak is eliminated, which can be denoted as the “off state”of the CLR. Acid treatment of the reduced PEDOT nanoantenna array then enables the regeneration of the CLR to its “on state”. The modulation depth of the extinction is up to 40 percentage points between the oxidized (≈45% extinction) and reduced (≈5% extinction) states, corresponding to around 2 times larger modulation depth than for periodic arrays without CLR in previous works (20–25 percentage points)2,12. Apart from extinction intensities, narrow resonance linewidth is also observed (FWHM ≈0.37–0.45 μm) after reoxidation. Even under 3–4 redox cycles, we still observe the disappearance and re-appearance of the intense and narrow plasmonic CLR (Fig. 5c), with a small (~10 nm) blue-shift after each redox cycle as discussed more below. Angle-dependent measurements confirm that the CLR in the re-oxidized state (during the 4th cycle, Fig. 5d) still shows efficient radiative coupling with (0, ±1) RA and the diffraction grating effect with (−1, 0) RA. The generality of redox-switchable conducting polymer CLRs is shown through another array structure with a longer periodic distance r=1.5μm (Supplementary Fig. 48). To understand the mechanism of the switchable CLR response, we explore changes in the relationship of (1/α iso )−Sbetween the oxidized and reduced states of the PEDOT nanoantenna array. Since the periodic distance is not modulated during redox cycling, S r and S i should be the same in the oxidized (“on state”) and reduced state (“off state”), according to Eq. (2). By contrast, the transformation from the oxidized to the reduced state enables drastic variations in 1/ α iso through permittivity modulation, because reducing the material to a low-conducting state drastically decreases the imaginary part of the permittivity and also switches sign of the real permittivity from negative to positive (Supplementary Fig. 49). According to Eqs. (3) and (4), the significant rise of Re(1/α iso ) for the reduced PEDOT nanoantennas generates a large mismatch gap by |Re(1/α iso )−S r |≈ 210 μm−3at the S r peak (λ=n s r=1.95 μm, in Fig. 5e). This result suggests that the dipolar coupling interaction from S r becomes insufficient to eliminate Re(1/α iso ), making it impossible to fulfill the CLR matching condition (Fig. 5e). Furthermore, the more negative Im(1/α iso ) of the reduced PEDOT nanoantennas enlarges the value of | Im(1/α iso )−S i | by ~3 times at λ=1.95–2.03 μm, indicating more serious damping relaxation. Figure 5f shows that these two changes in 1/ α iso leads to vanishing of the sharp peak of Im(α p ) in the reduced state (in the “off state”). Meanwhile, because the real permittivity is not negative around 1-3 μm, we also do not observe the LSPR signal of the individual nanoantennas in the reduced state. The above analysis is consistent with simulated extinction spectra (Supplementary Fig. 50) based on the permittivity of the reduced state. When returning to the high-conducting oxidized state via acid treatment, Re(1/α iso ) is reduced approximately to the original “on state”so that dipolar coupling interactions can again cancel out Re(1/ α iso )atλ r , which favorably re-generates the intense Im(α p ) peak and reexcites the CLR (“on state”). Moreover, Im(1/α iso )becomeslessnegative to diminish differences from S i , which allows the recovery of lowloss radiative coupling to ensure the narrow linewidth and high intensity of the re-generated CLR. The small changes observed for the “on state”between cycles may be associated with slightly lower conductivity of the re-oxidized state, leading to marginally larger Re(1/α iso ) which slight blue-shifts the resonance peak and lowers Δ, while the CLR matching conditions can be still fulfilled for the “on state”. We conclude that the redox-induced switching of the (1/α iso )-Srelationship allows for large modulation of intense conducting polymer CLRs. Discussion Our study improves the Q-factor of conducting polymer plasmonics from typical values below 2 to values up to above 10, by nonlocal coupling of individual nanoantennas to CLRs. Optimizing geometric parameters including array periodicity and nanoantenna dimensions enable extinction peak linewidths <0.4 μm in the mid-infrared wavelength region of 2–4.5 μm, with experimental Q-factors up to 12. Angledependent extinction spectra show that efficient radiative coupling from diffractive lattice structures dominates the narrow CLR action. Furthermore, redox-modulation can tune the PEDOT nanoantenna arrays between matching and mismatching the CLR conditions, leading to on/off switching with extinction modulation depth as large as 40 percentage points. The work takes important steps for conducting polymer plasmonics by overcoming the major challenges of weak and broad resonances and providing performance values suitable for practical applications. Methods Thin-film deposition The PEDOT nanoantennas were prepared based on the acid-treated PEDOT:Tosylate (PEDOT:Tos) thin films using the procedure described in previously reported literature36,41. Briefly, precleaned glass substrates were spin-coated with the oxidant reagent (the mixture of 2 g of tri-block PEGPPG-PEG co-polymer (from Sigma Aldrich company), 2 g of Clevios C-B 54 V3 (from Heraeus company in German), and 5 g of absolute ethanol) with a spinning speed of 1500 RPM, followed by annealing at 70 °C for 1 min. The oxidant-coated films were then placed in a vacuum chamber for vapor phase polymerization (VPP). In this step, the oxidant-coated films were exposed to EDOT vapor, generated by heating liquid EDOT at 60 °C under vacuum, and then reacted with EDOT vapor for 40 min. Then, films were rinsed with ethanol to remove any unreacted residues and subsequently dried using a gentle flow of nitrogen gas. The resulting VPP films with a thickness of about 0.2 μm were produced. For acid treatment, we soaked these PEDOT films in 3 M sulfuric acid solution for 10 min at room temperature and then washed these samples with deionized (DI) water, followed by nitrogen flow drying and annealed at 140 °C for 10 min, which can improve the electrical conductivity of PEDOT films. If using CaF 2 substrates, the VPP conditions and procedures were the same as those using glass substrates. Still, during the acid treatment, we use 1.5 M sulfuric acid solution containing 50% volume ratio of ethanol, because the 3 M sulfuric acid solution can easily separate PEDOT film from CaF 2 substrates. Periodic array fabrication and characterization The acid-treated PEDOT:Tos films on glass substrate were covered with ZEP520A positive resist (ZEONREX Electronic Chemicals) using spincoatingat2000rpm.Webakedtheresistat130oC for 3 min. The resist thickness was 0.5–0.6 μm. Then the specific regions for periodic arrays were exposed using Raith Voyager 100 electron beam lithographer operating at 50 kV. We used beam current of 1.6–1.7 nA and set an area dose and curved elements dose to 113 μCcm −2.Wealsooptimized pattern design to reach the target size of the nanodisks. We did not deposit any additional conductive layer because the PEDOT:Tos films provide good charge dissipation during e-beam lithography. After competing exposure, we developed the resist for 75–90 s using ZEDN50 developer (ZEONREX Electronic Chemicals) and blow-dried the samples. Then we dry-etched PEDOT:Tos film through ZEP520A mask by reactive oxygen plasma in a RIE Vacutec etcher. We etched at 150 mTorr pressure and 50 W forward power for 155 s to reach slight overetching of the areas around the nanodisks. Finally, we soaked the etched sample into the remover AR 300-72 at 60 °C for 3 min to remove the remaining positive resist from the top of PEDOT:Tos nanoantennas. The final periodic structures were visualized without Article https://doi.org/10.1038/s41467-025-59764-5 Nature Communications | (2025) 16:4484 8 any additional coating using Carl Zeiss Sigma 500 scanning electron microscope operating at 1 kV voltage with SE2 detector. Redox-recycles The chemical redox reactions were referred from previous works1,40.In brief, the reduction reaction was performed via the vapor of branched poly(ethylenimine) (M w ≈800, from Sigma–Aldrich company) in a nitrogen-filled glove box (to ensure the full effectiveness of the reduction reaction with PEDOT). The reduction reaction was carried out under 120°C for 5 min, followed by annealing at 120 °C for 10 min. For the re-oxidized state of PEDOT-based periodic arrays, we performed the oxidation reactions with the HCl vapor (Sigma–Aldrich, 37%) under 20 °C for 15–20 min. It is noted that although 3 M sulfuric acid solution can also induce oxidation reactions via soaking the sample, but such solution can easily destroy the periodic arrays. Extinction spectra characterization The extinction spectra were recorded using UV–Vis–NIR spectrophotometer (Perkin Elmer Lambda 900) in the wavelength regions of 0.9–3μm (when using glass substrates. The angle-dependent spectra were collected through the angle-resolution transmission accessory, and the aperture (a diameter of 5 mm) and polarizer are also used. As the spectra intensity becomes much lower when additionally using an aperture and a polarizer, we performed the smoothing process for extinction spectra to testify the resonance peak. The extinction spectra in the longer-wavelength region of 2−6μm were based on the transmission characterization via FTIR machine (PerkinElmer Spectron 3), when using CaF 2 substrates. During all above extinction characterizations, the sample of PEDOT nanoantenna array was placed in air. Electrical conductivity characterization The sheet resistance R s of PEDOT films was measured through a fourpoint probe machine consisting of a Keithley 2400 part and a Signatone Pro4 S-302 resistivity stand. The surface profiler (Dektak XT Bruker) is used to evaluate the acid-treated PEDOT:Tos film thickness (h). The electrical conductivity was evaluated based on the equation σ=1/(R s h). Ellipsometry characterization Two different ellipsometers, each with different spectra ranges, were used to collect the ellipsometry data of the acid-treated PEDOT:Tos film and the PEI-reacted PEDOT film on silica substrates (with an ultrathin SiO 2 layer of 1 nm on the top of silica layer). All ellipsometry measurements were conducted at 20 °C. For the wavelength range of 0.4–1.69 μm, the UV–vis–NIR measurements were carried out with the J. A. Woollam Co. RC2 spectroscopic ellipsometer using four incident angles (45°, 55°, 65°, and 75°). For the wavelength range within 1.69–10 μm, infrared measurements were based on the J. A. Woollam Co. IRVASE spectroscopic ellipsometer using three incident angles (50°, 60°, and 70°). We used the CompleteEASE (J. A. Woollam Co.) software to fit and analyze the refractive index based on the anisotropic Drude–Lorentz model41, and the parameters in Drude-Lorentz models such as frequencies, damping parts, and amplitudes were further confirmed by Matlab software. Optical numerical simulations FDTD simulations were based on the Ansys Lumerical FDTD software (https://www.ansys.com/products/optics/fdtd)toachievefar-field (extinction spectra) and near-field (electrical field distribution) optical response of PEDOT-based periodic arrays. The PEDOT-based periodic arrays were constructed with periodic boundaries (Bloch type) along the X-axis (parallel to the electric field polarization direction) and the Y-axis. By contrast, PML boundaries were used along Z-axis (as parallel to the direction of wave vector from incident light). A dielectric material with the constant refractive index n s = 1.5 was set as the substrate. The refractive index of the PEDOT nanodisk units (with a height of 0.2 μmandd= 0.5 ~ 0.52 μm) was based on the anisotropic Drude–Lorentz model analysis from the experimentally obtained ellipsometry data. We used a Bloch/periodic type plane wave to illuminate the nanoantenna at normal incidence along the Z-axis from air medium. For angle-dependent simulations, we used a BFAST type plane wave as the light source. Power monitors wereused to record the far-field and near-field distribution profiles.Themeshsizewas given as 20 × 20 × 10 nm3. For the localized model consisting of a nanoantenna (a diameter of 0.52 μm) on the dielectric substrate (n s = 1.5), we used PML boundaries along all directions and selected Total-Field Scattered-Field (TFSF) source as the light source. The cross-section spectra of LSPR were achieved by boxlike power detectors. Data availability All data generated in this study is provided in the Supplementary Information and/or available at Zenodo at https://doi.org/10.5281/ zenodo.15100375. References 1. Chen, S. et al. Conductive polymer nanoantennas for dynamic organic plasmonics. Nat. Nanotechnol. 15,35–40 (2020). 2. Karst, J. et al. Electrically switchable metallic polymer nanoantennas. Science 374,612–616 (2021). 3. Karki, A. et al. Electrical tuning of plasmonic conducting polymer nanoantennas. Adv. Mater. 34, 2107172 (2022). 4. Bubnova, O. et al. Semi-metallic polymers. Nat. Mater. 13, 190–194 (2014). 5. Meinzer, N., Barnes, W. L. & Hooper, I. R. Plasmonic meta-atoms and metasurfaces. Nat. Photonics 8, 889–898 (2014). 6. Kravets, V. G., Kabashin, A. V., Barnes, W. L. & Grigorenko, A. N. Plasmonic surface lattice resonances: a review of properties and applications. Chem. Rev. 118,5912–5951 (2018). 7. Chen, S. & Jonsson, M. P. Dynamic conducting polymer plasmonics and metasurfaces. ACS Photonics 10,571–581 (2023). 8. Karki, A. et al. Doped semiconducting polymer nanoantennas for tunable organic plasmonics. Commun. Mater. 3, 48 (2022). 9. Kazi, S. et al. n-Type redox-tuneable conducting polymer optical nanoantennas. J. Mater. Chem. C. 12, 17469–17474 (2024). 10. Bléteau, P., Gam-Derouich, S., Sun, X. & Lacroix, J.-C. Organic plasmonics: PEDOT nanoparticles are getting closer to the visible range. J. Mater. Chem. C. 12, 11579–11585 (2024). 11. Duan, Y. et al. Tuneable anisotropic plasmonics with shapesymmetric conducting polymer nanoantennas. Adv. Mater. 35, 2303949 (2023). 12. Lee, S. et al. Plasmonic polymer nanoantenna arrays for electrically tunable and electrode-free metasurfaces. J. Mater. Chem. A 11, 21569–21576 (2023). 13. Huang, X. et al. Organic metasurfaces with contrasting conducting polymers. Nano Lett. 25,890–897 (2025). 14. Wang, B. et al. High-Q plasmonic resonances: fundamentals and applications. Adv. Opt. Mater. 9, 2001520 (2021). 15. Lu, Z. et al. Quantifying the ultimate limit of plasmonic near-field enhancement. Nat. Commun. 15, 8803 (2024). 16. Wen, X. & Deng, S. Plasmonic nanostructure lattices for highperformance sensing. Adv. Opt. Mater. 11, 2300401 (2023). 17. Sharma, M., Tal, M., McDonnell, C. & Ellenbogen, T. Electrically and all-optically switchable nonlocal nonlinear metasurfaces. Sci. Adv. 9, eadh2353 (2023). 18. Zhang, J. et al. Spatially entangled photon pairs from lithium niobate nonlocal metasurfaces. Sci. Adv. 8, eabq4240 (2022). 19. Hail, C. U., Michaeli, L. & Atwater, H. A. Third harmonic generation enhancement and wavefront control using a local high-Q metasurface. Nano Lett. 24, 2257–2263 (2024). Article https://doi.org/10.1038/s41467-025-59764-5 Nature Communications | (2025) 16:4484 9