Enhanced fluorescence at low excitation powers with GaP hybrid nanoantennas

Abstract

Nanoantennas, known for their capability to enhance light–matter interactions at the nanoscale, have emerged as essential components in single-photon generation and fluorescence detection. Compared to metallic nanoantennas, all-dielectric nanoantennas are less vulnerable to ohmic losses and non-radiative quenching. Gallium phosphide (GaP), a dielectric material with a high refractive index and multiple electromagnetic resonances, stands out as a particularly compelling candidate for this purpose. In this work, we present a hybrid nanoantenna composed of GaP nanodisks and hydrogen silsesquioxane (HSQ) slotted antennas, enabling nanofocusing of electric fields and enhancing the local density of optical states (LDOS). Thanks to the suppression of absorption losses, the fluorescence spectral intensity is boosted 124 times, and the fluorescence decay rate is nearly doubled, reaching an 18-fold fluorescence enhancement at an ultralow excitation power of 0.5 μW. By leveraging HSQ as a dual-role photoresist and spacer, this metal-free architecture not only ensures biocompatibility, but also enables scalable fabrication of energy-efficient devices for non-destructive bioimaging.

---

Introduction

In the fields of biological imaging and detection, enhancing the fluorescence signal of molecular structures and bio-processes in cells and tissues is crucial for understanding biological functions. Therefore, fluorescent dyes or proteins with specific markers are widely used.¹ However, prolonged exposure to high-intensity excitation light can cause photobleaching, phototoxicity, and thermal damage to biological samples, compromising the reliability of fluorescence measurements. Reducing excitation light intensity may mitigate the problem by minimizing energy deposition to preserve sample viability. Nevertheless, this approach severely degrades the signal-to-noise ratio, creating an irreconcilable trade-off between measurement reliability and experimental safety.

A promising solution to this dilemma is to utilize nanoantennas, which amplify the fluorescence signals through electromagnetic field confinement. However, metallic nanoantennas suffer from inherent ohmic losses that paradoxically reintroduce thermal effects during prolonged cellular monitoring,² with localized temperature increases exceeding several hundred degrees reported under moderate illumination,³ undermining their core advantage. Hybrid dielectric–plasmonic systems, such as Si–Ag anapole–plasmon slotted nanodisks,⁴ Si nanodisk–gold mirror antennas^5,6 or high refractive index nanohybrid combinations,^7,8 combine multi-resonance coupling and directional emission to overcome individual material limitations. However, their complex fabrication processes and inherent absorption losses from metallic components which lead to both heating and fluorescence quenching, especially for high-quantum-yield emitters,⁹ hinder practical deployment.

The advent of all-dielectric nanoantennas offers a paradigm shift by leveraging low-loss Mie resonances for efficient field enhancement.^10,11 Among various dielectric materials, gallium phosphide (GaP)^12,13 has demonstrated extraordinary potential thanks to its high refractive index and nonlinear optical properties.^3,9,14–19 Recent advances, including nanoparticles,^20,21 slotted nanodisks,^19,22 and metasurfaces,^23–25 have enabled significant fluorescence enhancement in high-Q resonators or sub-50 nm gaps,^16,26 as comprehensively reviewed in our recent work.²⁷ However, such extreme performance typically relies on complex nanostructuring, precise spectral alignment, or cryogenic conditions, and often requires excitation powers exceeding 10 mW. Some studies further employ quenching solvents to artificially enhance signal contrast—a condition that does not reflect realistic biological or sensing environments. Hybrid structures offer an alternative route to field confinement. For example, GaP-ITO-GaP nanodisks²⁸ could strongly localize the electric near-field into the low refractive index region for spectral enhancement. Metal–dielectric nanopillar arrays²⁹ and hybrid Bragg resonators³⁰ further demonstrate enhanced photon collection, yet often involve metallic losses or deterministic coupling under cryogenic operation. Nevertheless, achieving optimal performance in these systems frequently demands sub-20 nm gaps,¹⁸ imposing stringent fabrication demands that hinder scalability and biological compatibility. These challenges highlight the need for a simplified, biocompatible platform that eliminates metallic losses while retaining field enhancement capabilities, ideally with broadband operation, low-power excitation, and compatibility with scalable fabrication.

Here, we propose a hybrid HSQ–GaP nanoantenna that synergizes the high refractive index (n ≈ 3.4) of GaP with the low refractive index (n ≈ 1.5) and biocompatibility of hydrogen silsesquioxane (HSQ).^31–33 Crucially, HSQ acts not only as a low refractive index spacer but also as a direct photoresist for structural patterning, eliminating the need for additional lithography steps and significantly simplifying fabrication. Unlike prior hybrid systems requiring nanoscale gaps or multilayer deposition, our design exploits the inherent dielectric contrast between GaP and HSQ to achieve polarization-selective nanofocusing without metallic components, thereby eliminating absorption losses and significantly boosting fluorescence enhancement. We accomplished a 124-fold fluorescence enhancement, with the fluorescence decay rate nearly doubling compared to that for pure glass conditions. The nanofocusing effect of our antenna design enables up to an 18-fold increase in fluorescence spectra under 0.5 µW excitation power, while maintaining the lifetime nearly consistent. This hybrid design provides a promising solution for low-power fluorescence-based biological detection and light-emitting devices, paving the way for more efficient and reliable imaging technologies.

Results and discussion

We investigate a hybrid nanoantenna design that combines GaP nanodisks and HSQ slotted antennas to enhance fluorescence. Fig. 1 provides a comprehensive overview of our hybrid nanoantenna structure and its optical characteristics. Fig. 1a shows a schematic of the hybrid nanodisks arranged on the surface of a GaP film to enhance fluorescence. The diameter of the hybrid nanoantenna ranges from 250 to 850 nm, with the HSQ slot's width fixed at 50 nm. This design allows for precise control of the optical properties of the antenna. Fig. 1b displays an AFM image of the fabricated hybrid nanodisk, with measurements showing a height of 395.4 nm ± 5.5 nm and RMS roughness of 4.4 nm, consistent with the 400 nm thickness of the GaP film. Details of the GaP thin film and hybrid nanodisk array are provided in the SI, Fig. S1 and S2.

Fig. 1 (a) Schematic of an HSQ slotted antenna on a GaP nanodisk to enhance the fluorescence process. (b) AFM image of the nanoantenna. (c) Simulated scattering of the nanoantenna illuminated with an electric field at either perpendicular or parallel polarization relative to the slot's long axis. Experimental scattering spectra are also shown in (c). (d) Emission spectra of the fluorescent dye Atto 550 in a PMMA solution (yellow area). The green vertical line indicates the excitation wavelength for fluorescence measurements (512 nm). (e and f) Simulations of the electric field energy distribution for perpendicular and parallel polarization excitation.

We investigated the far-field spectra and near-field distributions of the hybrid slotted nanoantenna under visible light excitation. Fig. 1c shows the simulated scattering cross-section for a nanoantenna with 550 nm diameter, excited with an electric field polarized perpendicular and parallel to the slot's long axis. Notably, the scattering behaviors for both excitation types are similar, with both showing the same general trend, indicating that the polarization has a minimal effect on the overall scattering characteristics. The scattering cross-section significantly expands in the range beyond 450 nm, with notable scattering dips appearing near 465 nm, 515 nm, 535 nm, and 615 nm. These dips are attributed to the presence of multiple resonant modes in the Mie scattering process, which arise due to the complex interactions of the nanoantenna with the incident light.³⁴ The experimental dark-field scattering spectra are also presented in Fig. 1c, where a first scattering dip at 473 nm is observed, with substantial peaks near 509 nm, 546 nm, and 632 nm, which is advantageous for enhancing the re-emission relevant to fluorescence. The experimental spectra exhibit a similar trend to the simulations but show a redshift, with slight differences in the intensity ratios of the peaks. These discrepancies are attributed to factors such as the structural roughness of the nanoantenna and limitations in the experimental setup. The nanoantenna's response to different polarizations remains consistent in scattering due to the stability of the HSQ material and the symmetry of the GaP nanodisk.

The high refractive index of GaP enables strong Mie-type resonances, leading to significant electromagnetic field confinement near the top interface of the nanodisk (Fig. S3a–c, SI). When integrated with the HSQ slot (Fig. S3d–f, SI), this resonant field extends into the low-index gap, resulting in enhanced electric field intensity at the center of the slot. Crucially, the HSQ gap is not merely a passive spacer but an engineered optical hotspot, where the local density of optical states (LDOS) is amplified through coupling with the GaP resonance. Changes in the excitation light polarization weaken the electric field. Fig. 1e and f show the electric field distribution for both polarizations, demonstrating the antenna's polarization selectivity, with perpendicular polarization leading to a twofold field enhancement under 512 nm excitation. Fluorescence enhancement is achieved using Atto 550 fluorophores (Sigma–Aldrich).^35–37 The dye is dissolved in anisole, mixed with PMMA, and spin-coated onto the nanoantenna surface, resulting in a uniform fluorescent thin film. The dye's emission spectrum, shown in Fig. 1d, exhibits a pronounced peak at 580 nm. Within the physical dimensions of the HSQ slot (length: 150 nm, width: 50 nm, height: 200 nm), the slot holds up to approximately twenty Atto 550 molecules at a concentration of 20 µM, with a total volume of 1.5 attoliters. The nanofocusing effect of the HSQ slotted antenna is further proven through experimental measurements of both spectra and lifetime.

To further investigate the effect of the excitation light polarization on the fluorescence process, we present the experimental results in Fig. 2. Fig. 2a displays the fluorescence spectra from the antenna's surface, where the fluorescence intensity for light polarized perpendicularly to the slot's long axis is twice as high as that under parallel excitation. This corresponds to a 96-fold enhancement relative to that of the fluorescent film coated on a planar glass surface without the antenna. This enhancement arises from the strong localized electric field generated within the slot during excitation, demonstrating the HSQ slots’ apparent polarization selectivity, which aligns with the near-field enhancement observed under various excitation conditions.

Fig. 2 (a) Fluorescence spectra of the PMMA/Atto 550 coated nanoantenna with the excitation polarization perpendicular (red line) and parallel (yellow line) to the slot's long axis, compared to the fluorescence from the glass-coated surface (gray line). (b) Fluorescence emission decay curves of the nanoantenna under the excitation polarization perpendicular (red line) and parallel (yellow line) to the slot's long axis, with references to the glass-coated surface (gray line) and the instrument response function (IRF).

Additionally, Fig. 2b presents the fluorescence decay rate results, showing that the fluorescence lifetime decreases in the presence of the nanoantenna. Specifically, the lifetime of Atto 550 is reduced from 2.9 ns on the reference glass surface to 1.9 ns with the nanoantenna, corresponding to an ensemble-averaged decay-rate enhancement of ∼1.5×. This ensemble-averaged value reflects the spatial average over all dye molecules, only a fraction of which reside within the high-LDOS region of the HSQ slot; molecules outside the hotspot exhibit near-intrinsic decay rates. This reduction is observed under both perpendicular polarized and parallel polarized excitation, with a consistent decay trend. This reduction in fluorescence lifetime is primarily attributed to an increased radiative decay rate (k_rad), driven by the enhanced local density of optical states within the slot.²⁶ This increase in the radiative decay rate is independent of the excitation polarization, as confirmed by consistent lifetime measurements with the excitation polarization perpendicular and parallel to the slot's long axis.

While the excitation polarization influences the localized electric field distribution within the slot, leading to variations in fluorescence intensity (higher intensity for perpendicular polarization), it does not significantly alter the radiative decay rate driven by the LDOS. This indicates that although the polarization can be used to optimize fluorescence enhancement, the underlying mechanism of decay-rate enhancement remains robust against polarization changes. These findings underscore the potential for designing nanoantennas that are optimized for efficient fluorescence enhancement across various applications and excitation conditions.

We investigated the optical scattering response and excitation enhancement of GaP hybrid nanoantennas with varying diameters. Fig. 3a displays the simulated scattering cross-sections, where an increase in diameter leads to a redshift in the scattering responses, along with an increase in both the scattering cross-section and the number of scattering peaks. The near-field distribution at the excitation wavelength (Fig. S7 and S8, SI) reveals a multi-lobed electric field distribution for larger antennas, with particularly high electric field energy localized inside the slot at a diameter of 550 nm. This enhanced field localization suggests potential to improve fluorescence excitation. Fig. 3b presents the corresponding experimental results, which show similar trends: as the diameter increases, the scattering intensity also exhibits a redshift. When the diameter exceeds 450 nm, multiple scattering peaks emerge, overlapping with the radiation wavelength of the dye, leading to a significant increase in the scattering intensity. These observations validate our modelling approach and demonstrate that optimizing the antenna size can enhance fluorescence excitation and broaden the effective wavelength range for efficient light–matter interactions.

Fig. 3 (a) Simulated and (b) experimental scattering spectra of antennas with diameters ranging from 250 nm to 850 nm. The black dashed line indicates the excitation wavelength (512 nm) and the red dashed line indicates the emission wavelength (580 nm). (c) SEM of the hybrid antenna array after ICP etching. The scale bar is 5 µm.

In exploring the influence of antenna size on fluorescence characteristics, Fig. 4a initially reveals pronounced differences in fluorescence spectra between molecules on a glass substrate and those coupled to GaP nanoantennas of varying diameters. Experimental results demonstrate that all tested antenna sizes significantly enhance the fluorescence emission compared to the glass surface, highlighting their effectiveness in boosting light–matter interactions. For instance, antennas with diameters of 550 nm, 650 nm, and 850 nm exhibit particularly strong fluorescence intensities, underscoring their potential for optimizing fluorescence signals. These findings illustrate how fluorescence enhancement can be effectively tailored by adjusting antenna dimensions. Additionally, as shown in Fig. S5d, a slight redshift of approximately 10 nm in the emission peak is observed across different antenna sizes. While this minor spectral shift has limited impact on practical applications, it provides valuable insights into how the antenna structure influences the optical behavior of emitters. Fig. 4c further illustrates the dependence of the fluorescence enhancement factor (factor calculations in the Methods section) on antenna diameter. Antennas with diameters of 550 nm, 650 nm, and 850 nm achieve the highest enhancement, reaching up to 124-fold enhancement relative to the glass substrate. Even smaller antennas (250 nm in diameter) show appreciable enhancement, with a ∼25-fold increase. The right axis of Fig. 4c displays the simulated fluorescence enhancement, which combines two key contributions: excitation enhancement due to the increased electronic field strength around the slots and the quantum yield improvement resulting from the modified LDOS. The simulation predicts that the optimal antenna sizes correspond closely to the resonance conditions that maximize both effects, thereby achieving the strongest fluorescence enhancement. To enable a meaningful comparison of fluorescence enhancement across different antenna sizes, the simulated enhancement values in the right axis of Fig. 4c were normalized with the spatially integrated electric field energy within the central region of each slot structure. This approach accounts for variations in field confinement and spatial overlap with the uniformly spin-coated fluorophores on the surface. As the excitation enhancement is the dominant contribution in our system, due to the nano-focusing effect in the slot region, this normalization ensures a fair comparison between antennas of varying dimensions. The consistency between experimental and simulated trends further supports the validity of this method. When comparing the experimental and simulated data, we observe a good agreement, confirming the validity of our theoretical model and its ability to accurately predict fluorescence enhancement trends across different antenna sizes. Fig. 4b presents the normalized fluorescence decay curves, demonstrating the effect of antenna size on emission dynamics. Antenna structures of all sizes exhibit faster decay rates than the glass substrate, indicating enhanced radiative processes induced by the nanostructures.

Fig. 4 (a) Fluorescence spectra and (b) emission decay curves for antennas of varying sizes (diameters ranging from 250 nm to 850 nm). (c, d) Scatter plots showing the fluorescence enhancement and lifetime as a function of the nanoantenna diameter (red points). The dotted lines represent the total fluorescence enhancement (c) from the simulation, as well as the inverse of the radiative decay rate (d).

Fig. 4d summarizes the impact of antenna diameter on the experimentally measured fluorescence lifetime (left axis) and the simulated radiative decay rate (right axis). For example, antennas with a diameter of 350 nm reduce the average fluorescence lifetime to 1.8 ns, while other sizes maintain lifetimes up to 2.2 ns. This represents a nearly two-fold reduction compared to the reference lifetimes of 2.9 ns on a glass substrate (Fig. S3b, SI). This trend reflects the ability of resonant GaP antennas to modify the LDOS of fluorescent molecules, thereby enhancing the radiative decay rate via the Purcell effect.³⁵ As illustrated in Fig. S10, electromagnetic simulations reveal that the radiative decay rate (K_r) increases with antenna size, while the non-radiative decay rate (K_nr) remains significantly lower across the visible spectrum (Fig. S10b). The simulated radiative lifetime (τ ∝ 1/K_r) shows a strong correlation with the experimentally measured fluorescence lifetimes (Fig. S10a, right axis), confirming that the observed lifetime reduction is primarily driven by accelerated radiative decay. Furthermore, the quantum yield (QY/QY₀) is enhanced at the resonance wavelength (Fig. S10c), demonstrating an overall improvement in emission efficiency. These results collectively support the interpretation that the fluorescence enhancement is dominated by radiative rate modification, rather than non-radiative quenching.^38,39 Compared to the simple glass structure (Fig. S11), there is a notable shift in the distribution of radiation energy, with the far-field radiation now being confined to a narrower angle range of 10° to 30°. The agreement between theoretical predictions and empirical results underscores the reliability of using GaP nanoantennas to modulate fluorescence properties through dimensional control.

Finally, we examined the influence of excitation power on fluorescence enhancement as shown in Fig. 5. Fig. 5a shows the fluorescence spectra of antennas (550 nm in diameter) under excitation powers ranging from 0.5 µW to 3.5 µW. As the excitation power increases, the fluorescence spectrum grows proportionally, as evidenced by the progressively increasing spectra. According to the peak statistics in Fig. 5c, fluorescence enhancement increases linearly with excitation power, with no observed saturation effect. This linear dependence is consistent with previous studies where the system remains well below the saturation regime.⁴⁰ This indicates that the nanoantenna effectively enhances fluorescence across a wide range of powers. At an excitation power of 3.5 µW, the enhancement reaches over 105-fold, while an 18-fold increase is achieved even at the low excitation power of 0.5 µW. This linear response across a wide power range, with no degradation in lifetime (Fig. 5d), indicates robust performance under low-excitation conditions, making the antenna suitable for applications requiring minimal photodamage and high signal fidelity, such as biosensing and single-molecule tracking. While the system remains linear up to 3.5 µW, nonlinear effects such as fluorophore saturation or thermal response may arise at higher powers, which could be explored in future studies.

Fig. 5 (a) Fluorescence spectra and (b) emission decay curves under different excitation powers (from 0.5 µW to 3.5 µW). (c and d) Scatter plots showing the enhanced fluorescence and fluorescence lifetime as a function of excitation power.

Fig. 5b presents the fluorescence decay curves under varying excitation power levels. Remarkably, the fluorescence lifetime does not decrease with increasing power, demonstrating that the nanoantenna maintains a stable balance between radiative and non-radiative rates across different power conditions. This stability is further quantified as shown in Fig. 5d, where the fluorescence lifetimes remain consistently around 2.2 ns, regardless of the excitation power. These findings suggest that the excitation power can be effectively adjusted to control fluorescence intensity without compromising the fluorescence lifetime. For comparison, Table S1 summarizes several representative plasmonic and dielectric antenna structures reported in the literature, along with their fluorescence enhancement performance under various experimental conditions. Notably, many plasmonic systems achieve higher enhancement factors but typically rely on high excitation power or low-quantum-yield emitters (e.g., QY ≈ 2.5%) to amplify the apparent enhancement, and suffer from metal-induced quenching, optical losses and thermal instability. In contrast, our HSQ/GaP slotted antenna achieves a 124-fold enhancement using a high-quantum-yield dye (Atto 550, QY ≈ 80%), with no measurable fluorescence quenching—evidenced by a nearly doubled radiative decay rate and a stable fluorescence lifetime. Furthermore, this enhancement is realized under ultra-low excitation power (0.5 µW), highlighting the efficiency of our design. These characteristics—low photodamage, high stability, and compatibility with standard fabrication—make our platform particularly well-suited for long-term applications in biological imaging and optical sensing, where sustained signal intensity and emitter integrity are critical.

Conclusions

In this study, we demonstrate that fluorescence can be significantly enhanced by combining HSQ slotted antennas with GaP nanodisks. Under excitation polarization perpendicular to the long axis of the slot, more electric field localization occurs within the HSQ slot, leading to over 2-fold fluorescence spectral enhancement compared to parallel polarization excitation. By optimizing the antenna size, we achieve a 124-fold fluorescence enhancement, with the fluorescence decay rate nearly doubling compared to that for pure glass conditions. The nanofocusing effect of our antenna design enables up to an 18-fold increase in fluorescence spectra under 0.5 µW excitation power, while maintaining the lifetime nearly consistent. This design provides a promising platform for single-molecule fluorescence detection and enhancement through emitter interaction, with potential applications in bio-fluorescence and single-photon source devices.

Methods

Numerical simulations

The electromagnetic responses were calculated using the finite difference time domain (FDTD) method. The total-field scattered-field (TFSF) source was used to investigate the scattering and absorption performance of the nanoantennas while the refractive index of the background stayed at 1.4. The perfectly matched layers (PML) were applied on boundaries of the simulation region with a mesh size of 5 nm around the antennas. The refractive index of the GaP nanodisk with 200 nm height was consistent with ref. 19 The height of the HSQ slotted antenna was set as 200 nm, whereas the typical lengths of slotted long and short axes in nanoantennas were 150 nm and 50 nm. The refractive index of the HSQ was set at 1.5. For the emission calculations, a dipole source was placed at the center of the HSQ slot. The emission emitted by the source was collected by detectors positioned around the antenna. The radiative and non-radiative decay rates were determined by calculating the ratio of the power emitted through radiation to the total power, which includes both radiative and non-radiative components.

Nanofabrication

The fabrication process involves transferring a GaP thin film onto a glass substrate, followed by patterning with HSQ electron beam lithography (EBL) and inductively coupled plasma (ICP) etching. The GaP film originally grown on a gallium arsenide (GaAs) substrate was transferred onto glass substrates via bonding technology. Both the GaAs substrate and the 200 nm thick AlGaInP etch-stop layer were subsequently removed by selective wet etching. The original thickness of the GaP film is 600 ± 20 nm, with a crystal orientation of 〈100〉, deviating towards the 〈011〉 direction by 15°. The hybrid nanoantenna consists of a circular GaP disk featuring a surface-embedded dielectric slot (∼ 50 nm wide) and a stripe of hydrogen silsesquioxane (HSQ) resist, patterned via EBL. This HSQ structure is retained after development and remains intact during ICP etching, resulting in a continuous GaP structure. The designed diameter of the hybrid nanoantenna varies from 250 to 850 nm.

Sample preparation

To prepare the fluorescent thin film layer, Atto 550 fluorophores (Sigma–Aldrich) were first dissolved in anisole. This Atto 550 solution was then mixed with a 3% (w/v) PMMA (Alfa Aesar Chemicals Co. Ltd) in anisole solution to achieve a final concentration of 20 µM. The prepared solution was spin-coated onto the antenna surface at 6000 rpm. Following the spin-coating process, the sample was baked at 100 °C for 2 minutes to form a uniform and stable fluorescent thin film. The film thickness was determined to be ∼ 58 nm using a probing profilometer (Dektak Pro, BRUKER).

Optical microscopy experiments

The fluorescence collection system was employed with an inverted microscopy setup (Olympus IX 73). A picosecond free-space pulsed laser (512 nm, 1 MHz) was used as the excitation source. Utilizing a piezoelectric stage (Mad City Lab) on a 100–200 mm² substrate, the sample is raster-scanned to capture the molecules. The excitation light was precisely focused on the substrate through a 60× objective lens with a numerical aperture (NA) of 0.9. Fluorescence was selectively filtered from the excitation light through a dichroic mirror and a band-pass filter specifically designed for the fluorescence band of organic molecules. Fluorescence spectroscopy was conducted using an Andor grating spectrometer, model Shamrock 193i, and an Andor camera, model iXon 888, to facilitate precise spectral analysis. Dark field scattering was generated by focusing white light on the sample surface through a scattering ring, and the scattered light was collected using a 20× objective lens with a NA of 0.25. Additionally, stray and defocused light was meticulously filtered out at the collection terminal through spatial filtering and a single-mode fiber. For the detection of emitted photons’ lifetime, the system was employed with an avalanche photodiode detector (APD), ensuring sensitive photon capture. Simultaneously, a time-correlated single-photon counting system (TCSPC) was used to accurately record individual photon events. This setup allowed for high-precision fluorescence imaging and spectral analysis, essential for studying molecules immobilized on substrates and their interactions. The spectral and decay-rate acquisition times were 20 s and 120 s. Real time power detection was performed using a power meter (Thorlabs PM130D), with a silicon detector for measuring wavelengths in the range of 400–1100 nm.

Spectral and lifetime analysis

Fig. S4 demonstrates that the tested substrate materials—glass, glass/HSQ, glass/GaP/HSQ, and glass/GaP/HSQ/PMMA—exhibit negligible background fluorescence, confirming that they do not generate intrinsic fluorescence signals and thus will not interfere with subsequent fluorescence measurements. The experimentally obtained fluorescence spectra were processed using smoothing and denoising techniques, with the results before and after processing shown in Fig. S5a, SI. The peak intensity of the spectrum was taken as the fluorescence intensity. The wavelength-dependent fluorescence enhancement factor was calculated as the ratio of the fluorescence intensity from the nanoantenna structure to that from a bare glass substrate:

Enhancement factor (λ) = S_GaP (λ)/S_glass (λ)

The enhancement factor at the emission peak wavelength (580 nm) is reported as the overall enhancement (Fig. S5c, SI). Fig. S5d shows the spectral peak position as a function of antenna diameter. The collected fluorescence decay curves were fitted using a bi-exponential decay model, and the average fluorescence lifetime and lifetime error were calculated based on the correlation coefficients. The results before and after fitting the fluorescence decay curves are presented in SI Fig. S5b. Fig. S12 and S13 show the original spectral data points and the corresponding fitting curves.

Author contributions

Y. Y. and G. M. contributed equally to this work. X. W. S. and Y. L. supervised the project. Y. L., Y. Y., and G. M. initiated the concept. G. M. and Y. Y. finished the optical experiments. Y. Y. performed the simulations and nanofabrication with suggestions from Y. L. and Q. Z. The manuscript was written through the contributions of all authors. All authors have given approval to the final version of the manuscript.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data that support the findings of this study are available within the article and its supplementary information (SI). Supplementary information: additional details of the simulations, fabrication of the samples, and experimental characterization. Fig. S1 and S2 display the testing of the transferred GaP film and nanofabrication details. Fig. S3 illustrates the focusing effect and polarization selectivity of the HSQ slot in the hybrid nanoantenna structure. Fig. S4 shows the comparison of the spectral background signals from different materials without fluorophores: glass, glass/HSQ, glass/GaP/HSQ and glass/GaP/HSQ/PMMA. Fig. S5 shows the comparison of the fluorescence properties of Atto 550 (20 µM) on two different substrates: glass/PMMA and GaP film/PMMA. Fig. S6 shows the fluorescence enhancement spectrum of Atto 550 on GaP nanoantennas with varying disk diameters. Fig. S7–S9 show the electric field energy distributions for varying nanoantenna sizes. Fig. S10–S11 show the emission simulation for a dipole source placed in the HSQ slot center. Fig. S12 shows the original spectral data points and the corresponding fitting curves for antennas of varying sizes. Fig. S13 shows the original spectral data points and the corresponding fitting curves for different excitation powers. Table S1 shows the comparison of the experimental parameters between this work and typical nanophotonic antenna structures. See DOI: https://doi.org/10.1039/d5nr03614c.

Additional data related to this study are available from the corresponding author upon reasonable request.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (No. 62171211),  National Key Research and Development Program of China (No. 2022YFF1203400 and 2022YFB3602903), Guangdong Basic and Applied Basic Foundation (No. 2019A1515111136, 2021B1515120074, and 2023A1515012141), Science and Technology Innovation Commission of Shenzhen (No. JCYJ20220818100218039, JCYJ20220818100411025), Shenzhen Key Laboratory for Advanced Quantum Dot Displays and Lighting (No. ZDSYS201707281632549), and High level of special funds of SUSTech (No. G03034K002).

References

1. D. S. White, M. A. Smith, B. Chanda and R. H. Goldsmith, ACS Meas. Sci. Au, 2023, 3, 239–257.
2. H. Yu, Y. Peng, Y. Yang and Z.-Y. Li, npj Comput. Mater., 2019, 5, 45.
3. P. Albella, R. Alcaraz de la Osa, F. Moreno and S. A. Maier, ACS Photonics, 2014, 1, 524–529.
4. K. As'ham, I. Al-Ani, L. Huang, A. E. Miroshnichenko and H. T. Hattori, ACS Photonics, 2021, 8, 489–496.
5. C. Hong, I. Hong, Y. Jiang and J. C. Ndukaife, Adv. Opt. Mater., 2024, 12, 2302603.
6. P. A. Dmitriev, E. Lassalle, L. Ding, Z. Pan, D. C. J. Neo, V. Valuckas, R. Paniagua-Dominguez, J. K. W. Yang, H. V. Demir and A. I. Kuznetsov, ACS Photonics, 2023, 10, 582–594.
7. S. Bhaskar, A. K. Singh, P. Das, P. Jana, S. Kanvah, S. B. N. Bhaktha and S. S. Ramamurthy, ACS Appl. Mater. Interfaces, 2020, 12, 34323–34336.
8. A. Rai, S. Bhaskar, K. M. Ganesh and S. S. Ramamurthy, ACS Appl. Nano Mater., 2022, 5, 12245–12264.
9. J. Cambiasso, M. König, E. Cortés, S. Schlücker and S. A. Maier, ACS Photonics, 2018, 5, 1546–1557.
10. R. Regmi, J. Berthelot, P. M. Winkler, M. Mivelle, J. Proust, F. Bedu, I. Ozerov, T. Begou, J. Lumeau, H. Rigneault, M. F. García-Parajó, S. Bidault, J. Wenger and N. Bonod, Nano Lett., 2016, 16, 5143–5151.
11. E. A. Gurvitz, K. S. Ladutenko, P. A. Dergachev, A. B. Evlyukhin, A. E. Miroshnichenko and A. S. Shalin, Laser Photonics Rev., 2019, 13, 13.
12. R. Stockill, M. Forsch, F. Hijazi, G. Beaudoin, K. Pantzas, I. Sagnes, R. Braive and S. Gröblacher, Nat. Commun., 2022, 13, 6583.
13. M. Melli, M. West, S. Hickman, S. Dhuey, D. Lin, M. Khorasaninejad, C. Chang, S. Jolly, H. Tae, E. Poliakov, P. St. Hilaire, S. Cabrini, C. Peroz and M. Klug, Sci. Rep., 2020, 10, 1–7.
14. J. Lv, H. Zhang, C. Liu, Z. Yi, F. Wang, H. Mu, X. Li, T. Sun and P. K. Chu, Nanomaterials, 2021, 11, 1490.
15. J. Lv, Y. Ren, D. Wang, X. Xu, W. Liu, J. Wang, C. Liu and P. K. Chu, J. Opt. Soc. Am. A, 2023, 40, 2034.
16. J. Cambiasso, G. Grinblat, Y. Li, A. Rakovich, E. Cortés and S. A. Maier, Nano Lett., 2017, 17, 1219–1225.
17. L. Hüttenhofer, A. Tittl, L. Kühner, E. Cortés and S. A. Maier, ACS Photonics, 2021, 8, 1469–1476.
18. C. Vidal, B. Tilmann, S. Tiwari, T. V. Raziman, S. A. Maier, J. Wenger and R. Sapienza, Nano Lett., 2024, 24, 2437–2443.
19. Y. Yang, Y. Wang, Y. Yan, W. Cheng, Q. Zhao and Y. Li, Adv. Opt. Mater., 2024, 12, 2301444.
20. V. I. Zakomirnyi, A. Moroz, R. Bhargava and I. L. Rasskazov, ACS Nano, 2024, 18, 1621–1628.
21. X. Wu, S. Cheng, G. Huang, S. Zhan, G. Nie, X. Su, D. Cheng and Y. Liu, Ceram. Int., 2024, 50, 3843–3851.
22. Y. Yang, Y. Yan, Y. Wang, W. Cheng, Q. Zhao and Y. Li, in Nanophotonics and Micro/Nano Optics IX, ed. Z. Zhou, K. Wada and L. Tong, SPIE, 2023, p. 41.
23. M. Iwanaga, ACS Nano, 2020, 14, 17458–17467.
24. G. Q. Moretti, E. Cortés, S. A. Maier, A. V. Bragas and G. Grinblat, Nanophotonics, 2021, 10, 4261–4271.
25. X. Luo, Q. Yuan, R. Shi, S. Lin, Y. Liu and G. Li, Chem. Biomed. Imaging, 2025 DOI:10.1021/cbmi.5c00123.
26. R. M. Córdova-Castro, B. van Dam, A. Lauri, S. A. Maier, R. Sapienza, Y. De Wilde, I. Izeddin and V. Krachmalnicoff, Light: Sci. Appl., 2024, 13, 7.
27. Y. Wang, Z. Pan, Y. Yan, Y. Yang, W. Zhao, N. Ding, X. Tang, P. Wu, Q. Zhao and Y. Li, Nanophotonics, 2024, 13, 3207–3252.
28. R. Tirole, B. Tilmann, L. de S. Menezes, S. Vezzoli, R. Sapienza and S. A. Maier, Adv. Opt. Mater., 2024, 12, 1–8.
29. M. Wang, L. Jiang, R. Mei, H. Shi, Y. Xu, Z. Su, G. Ding and F. Zang, Adv. Opt. Mater., 2025, 01591, 1–9.
30. S.-W. Xu, Y.-M. Wei, R.-B. Su, X.-S. Li, P.-N. Huang, S.-F. Liu, X.-Y. Huang, Y. Yu, J. Liu and X.-H. Wang, Photonics Res., 2022, 10, B1.
31. T. Santiago-Cruz, S. D. Gennaro, O. Mitrofanov, S. Addamane, J. Reno, I. Brener and M. V. Chekhova, Science, 2022, 377, 991–995.
32. X. Liu, A. Bruch and H. Tang, Adv. Opt. Photonics, 2023, 15, 236–317.
33. C. Kim, C. Ye, Y. Zheng, E. Semenova, K. Yvind and M. Pu, IEEE J. Sel. Top. Quantum Electron., 2023, 29, 1–13.
34. K. V. Baryshnikova, D. A. Smirnova, B. S. Luk'yanchuk and Y. S. Kivshar, Adv. Opt. Mater., 2019, 7, 13.
35. A. Al Masud, W. E. Martin, F. H. Moonschi, S. M. Park, B. R. Srijanto, K. R. Graham, C. P. Collier and C. I. Richards, Nanoscale Adv., 2020, 2, 1894–1903.
36. M. Baibakov, S. Patra, J.-B. Claude and J. Wenger, ACS Omega, 2020, 5, 6947–6955.
37. E. Torres-García, R. Pinto-Cámara, A. Linares, D. Martínez, V. Abonza, E. Brito-Alarcón, C. Calcines-Cruz, G. Valdés-Galindo, D. Torres, M. Jabloñski, H. H. Torres-Martínez, J. L. Martínez, H. O. Hernández, J. P. Ocelotl-Oviedo, Y. Garcés, M. Barchi, R. D'Antuono, A. Bošković, J. G. Dubrovsky, A. Darszon, M. G. Buffone, R. R. Morales, J. M. Rendon-Mancha, C. D. Wood, A. Hernández-García, D. Krapf, Á. H. Crevenna and A. Guerrero, Nat. Commun., 2022, 13, 7452.
38. W. T. Yip, D. Hu, J. Yu, D. A. Vanden Bout and P. F. Barbara, J. Phys. Chem. A, 1998, 102, 7564–7575.
39. P. Holzmeister, E. Pibiri, J. J. Schmied, T. Sen, G. P. Acuna and P. Tinnefeld, Nat. Commun., 2014, 5, 5356.
40. J. Wenger, D. Gérard, J. Dintinger, O. Mahboub, N. Bonod, E. Popov, T. W. Ebbesen and H. Rigneault, Opt. Express, 2008, 16, 3008.

---