Introduction

Surface plasmon polaritons (SPPs) are surface waves formed by the collective oscillation of free electrons on a metal surface under incident light. They can break through the optical diffraction limit and localize light field energy at the nanoscale, enabling deep subwavelength nanofocusing1,2. At the same time, they have the characteristic of localized electric field enhancement, which is essential for enhancing the interaction between light and matter3,4,5. This property makes them highly promising for applications such as near-field super-resolution imaging6,7,8, biomedical detection9,10, materials spectroscopy analysis11,12, and surface localized electric field enhancement13,14,15. Optical probes, as the core carriers for exciting and regulating SPPs, directly determine the resolution and signal-to-noise ratio of measurements. External laser irradiation probes can achieve nanofocusing via metal coatings or grating structures6,16,17. However, they depend on complex optical systems and high-order signal demodulation modules18, leading to increased system complexity and noise interference. Optical probes based on conical fiber waveguides convert the optical mode inside the fiber into SPP mode by covering the surface of the conical fiber with a metal film, ultimately achieving nanofocusing at the probe tip. They have the advantages of low background noise and high signal-to-noise ratio19,20,21.

The combination of conical fiber waveguides with SPPs enables fiber probes to exhibit outstanding performance in scanning near-field optical microscopy (SNOM)19,20,22,23,24,25. In near-field super-resolution imaging measurements, imaging resolution primarily depends on two factors: the curvature radius of the probe tip and the intensity of the focused light spot26,27. Although smaller tip sizes enable higher spatial resolution28, traditional preparation methods like fiber metallization or electrochemical etching struggle to produce stable, ultra-small tips consistently, ultimately limiting probe performance. Focused Ion Beam (FIB), as a mask-free processing technique, can directly pattern materials and offers unique advantages in preparing complex nanostructures29,30. At the same time, high-intensity focused light spots help improve the signal-to-noise ratio and reduce background noise interference. However, scattering during the propagation of SPPs can cause propagation loss4,31,32,33. This phenomenon not only reduces the electric field strength of the light spot formed at the probe tip, but also increases background noise, resulting in a low signal-to-noise ratio for imaging. Therefore, near-field optical imaging systems usually require the use of highly sensitive photodetectors to capture the weak near-field signals34,35. In existing studies, the probe surface has been designed with a grating structure and a platform-based structure, which greatly improves the electric field strength at the probe tip36,37. However, the excitation of SPPs depends on radially polarized light (RPL), which have a complex generation process. In practical applications, even slight deviations between the center of the incident light beam and the probe axis can cause a significant weakening of the local electric field. Existing studies have designed probes with asymmetric geometric structures via 3D printing to reduce polarization dependence19, but insufficient processing precision reduces the electric field enhancement effect at the probe tip. Additionally, researchers have processed asymmetric semi-circular slit structures on the probe surface using FIB, achieving energy focusing at the probe tip using linearly polarized light (LPL)38. However, due to limitations in the energy focusing mechanism and severe short-wavelength energy loss, the electric field enhancement at the probe tip is limited. Currently, research is limited to the focusing mechanism of plasmonic fiber probes, and no near-field imaging-related tests have been conducted, resulting in a lack of sufficient experimental verification.

To overcome these limitations, this study proposes a double-slit plasmonic platform-based fiber probe (DSPP). By combining the interference enhancement of the platform-based structure with the polarization control function of the half-ring slit, efficient nanofocusing was achieved under conventional LPL, solving the problems of light source dependence and energy loss in traditional solutions. At the preparation process level, this study innovatively proposed a sleeve ring etching technique, which uses FIB to perform multiple ring-shaped etching on conical optical fibers to precisely control the structure of the front cone and the curvature radius of the probe tip. Then, combined with the rotation cutting technique, a platform toroidal plane with high flatness is prepared as the reflective surface. Compared with traditional preparation methods, this process achieves controllable preparation of probe morphology parameters, laying the foundation for the standardized manufacture of nano-optical probes. Simulation and experimental verification show that the probe tip electric field strength of the DSPP is 6 times higher than that of an asymmetric double-slit probe (DSP) when LPL with a wavelength of 633 nm is incident, and it maintains stable focusing in a broadband range, especially showing significant advantages in the short wavelength region. In addition, the DSPP achieves an optical resolution of 28.6 nm. In summary, this probe combines conventional incident light polarization, broadband stability, and extremely strong electric field enhancement characteristics, providing a new tool for biological cell detection, subwavelength lithography, and optical chip defect detection.

Results and discussion

The design of DSPP

The DSPP integrates two key mechanisms: the Fabry–Pérot interference enhancement from the platform-based structure (see Section S1 in the Supplementary Information 1 for the focusing enhancement effect of FPFP)39 and the polarization control from the asymmetric half-ring slit. This design eliminates the need for complex RPL and enhances the local electric field at the probe tip.

The three-dimensional structure of DSPP and the nanofocusing process are shown in Fig. 1a. The surface of the optical fiber probe is covered with a layer of Au film, and two semicircular slit structures are distributed on both sides of the probe axis with a specific height difference h (see Fig. S4 in the Supplementary Information 1 for the two-dimensional structure of DSPP). As shown in Fig. 1b, when LPL couples into the DSPP, it initially propagates forward in the linearly polarized mode (\({\text{HE}}_{11}\)). The two asymmetric half-ring slit structures on either side of the probe axis excite \({\text{HE}}_{11}-\text{SPP}\), the asymmetry h introduces precise phase delay, thereby introducing a phase difference ΔΦ between any two symmetrical points along the probe’s circumference. This compensates for the inherent phase difference of \({\text{HE}}_{11}-\text{SPP}\), ultimately converting \({\text{HE}}_{11}-\text{SPP}\) with an antisymmetric phase distribution into \({\text{TM}}_{01}-\text{SPP}\) with a symmetric phase distribution. To achieve polarization reversal and circumferential interference, the following equation must be satisfied:

$$\Delta {\Phi }_{\text{L}}-\Delta {\Phi }_{\text{R}}=(2{\text{N}}-1){\rm{\pi }}$$
(1)
$$\Delta {\Phi }_{\text{L}}=2{\rm{\pi }}{\rm{\cdot }}{\text{h}}/{{\rm{\lambda }}}_{\text{OM}}$$
(2)
$${\Phi }_{\text{R}}=2{\rm{\pi }}\cdot \,\left[{\text{h}}/\cos ({\rm{\theta }}/2)\right]/{{\rm{\lambda }}}_{\text{SPP}}$$
(3)
$${{\rm{\lambda }}}_{\text{OM}}={{\rm{\lambda }}}_{0}/{\text{n}}_{\text{OM}}$$
(4)
$${{\rm{\lambda }}}_{\text{SPP}}={{\rm{\lambda }}}_{0}\sqrt{({{\rm{\varepsilon }}}_{\text{air}}+{{\rm{\varepsilon }}}_{\text{m}}^{\mbox{'}})/{{\rm{\varepsilon }}}_{\text{air}}{{\rm{\varepsilon }}}_{\text{m}}^{\mbox{'}}}$$
(5)
Fig. 1: Structure of DSPP and Principle of Nanofocusing.
Fig. 1: Structure of DSPP and Principle of Nanofocusing.
Full size image

a Schematic diagram of the nanofocusing process of DSPP. Two half-ring slits introduce phase delay, converting \({\text{HE}}_{11}-\text{SPP}\) to \({\text{TM}}_{01}-\text{SPP}\), and then the platform-based structure is used to achieve multiple superpositions of energy at the probe tip. b Electric field distribution of DSPP in the XZ plane. SPPs form standing waves on the surface of the front cone of the probe and achieve nanofocusing at the tip of the probe. c Electric field distribution in the XY plane at different distances from the tip of the probe. \({\text{HE}}_{11}-\text{SPP}\) gradually converts to \({\text{TM}}_{01}-\text{SPP}\) and forms a nanofocusing light spot with high electric field strength at the tip of the probe. d Side view of the prepared DSPP. The probe tip size is 15 nm, the front cone has good surface quality, and the platform toroidal plane is very flat. e Top view of the prepared DSPP

In the formula, \(\Delta {\Phi }_{\text{L}}\) represents the phase delay in the optical fiber, \(\Delta {\Phi }_{\text{R}}\) represents the phase delay of the SPPs mode outside the metal film, N is a positive integer, \({{\rm{\lambda }}}_{\text{SPP}}\) represents the wavelength of the SPPs at the air/gold interface, \({\rm{\theta }}\) is the angle of the back cone, \({{\rm{\lambda }}}_{0}\) represents the wavelength of the incident laser, \({\text{n}}_{\text{OM}}\) and \({{\rm{\lambda }}}_{\text{OM}}\) represent the effective refractive index and effective wavelength of the internal optical waveguide mode, respectively, and \({{\rm{\varepsilon }}}_{\text{m}}\) and \({{\rm{\varepsilon }}}_{\text{air}}\) represent the dielectric constants of the gold film material and the surrounding air, respectively.

After the converted \({\text{TM}}_{01}-\text{SPP}\) propagates to the probe tip and focuses, a portion of the energy continues along the opposite side of the front cone40. Here, the platform-based structure acts as a reflective surface, redirecting these SPPs back towards the tip. This creates a Fabry–Pérot interference condition with incoming \({\text{TM}}_{01}-\text{SPP}\)41, thereby achieving multiple energy superpositions and significantly enhancing the nanofocusing intensity at the probe tip. According to the phase-length interference enhancement mechanism, the phase expression of the SPPs on the front cone of the probe can be obtained as follows:

$${\Phi }_{1}=\frac{4{\rm{\pi }}\text{H}}{{{\rm{\lambda }}}_{0}\sqrt{({{\rm{\varepsilon }}}_{\text{air}}+{{\rm{\varepsilon }}}_{\text{m}}^{\mbox{'}})/{{\rm{\varepsilon }}}_{\text{air}}{{\rm{\varepsilon }}}_{\text{m}}^{\mbox{'}}}{\rm{\cdot }}\cos ({\rm{\alpha }}/2)}+{\Phi }_{0}$$
(6)

In the formula, \({\Phi }_{0}\) is the phase of SPPs at the probe tip, \({\Phi }_{1}\) is the phase when SPPs propagate to the platform toroidal plane and reach the probe tip after reflection, H is the height of the front cone, and \({\rm{\alpha }}\) is the angle of the front cone. Therefore, when \({\Phi }_{1}\)-\({\Phi }_{0}\) = \(2\text{N}{\rm{\pi }}\) (N is a positive integer), the electric field strength at the probe tip can reach its maximum value.

On the front cone surface of the DSPP, as the SPPs on the Au film continuously propagate toward the probe tip, the electric field strength also continuously increases, and ultimately, nanofocusing is achieved at the probe tip of the DSPP (Fig. 1b, c). In addition, it can be seen that standing waves are formed on the surface of the front cone of the probe, and the closer the position is to the probe tip, the shorter the length of the standing waves, indicating that there is stable interference on the surface of the front cone40. The experimentally fabricated DSPP is shown in Fig. 1d, e, with a probe tip size of 15 nm (see the Supplementary Information 2 for the detailed process of preparing DSPP). This nanoscale tip, consistently achievable via our FIB-based sleeve-ring etching technique, ensures the geometric precision required for high-resolution imaging.

Nanofocusing mechanism of probe tip

This paper employs the finite-difference time-domain (FDTD) method to simulate the parameters of the DSPP. The electric field intensity at the probe tip corresponding to different cone heights H is shown in Fig. 2a. It can be observed that the DSPP consistently achieves nanofocusing at the probe tip, with the electric field intensity always exceeding that of the DSP probe tip. In addition, the electric field strength at the probe tip of the DSPP alternates from low to high and then from high to low as the height of the front cone increases. This is because, as the height of the front cone changes, the length of the SPP propagation path from the tip to the platform toroidal plane and then back to the probe tip also changes continuously. Therefore, the phase of the SPP when it reaches the probe tip for the second time will also be different. When the phase difference between the two SPPs at the probe tip continuously changes, the electric field strength at the probe tip will also alternate between strong and weak40.

Fig. 2: Comparison of nanofocusing effects of FPFP, DSP, and DSPP at an incident light wavelength of 633 nm.
Fig. 2: Comparison of nanofocusing effects of FPFP, DSP, and DSPP at an incident light wavelength of 633 nm.
Full size image

a When the front cone angle is 32°, the absolute values of the electric field strength at 2 nm below the probe tip correspond to different cone heights H. The blue line represents the absolute value of the electric field strength of DSP. b Absolute values of electric field strength at 2 nm below the probe tip for DSPP corresponding to different front cone angles \({\rm{\alpha }}\) when the cone height is H = 500 nm and H = 1000 nm, which are the two cone heights with the largest enhancement effect in the simulation results in (a). (c–k) show the electric field and phase distributions under \({\text{HE}}_{11}\). (c–e) represent the electric field distribution of \({\text{HE}}_{11}-\text{SPP}\) excited on the FPFP surface in the cylindrical waveguide in the XY plane, the electric field distribution at the probe tip in the XZ plane, and the \({\text{E}}_{\text{z}}\) phase distribution corresponding to the XZ plane. (f–h) and (i–k) represent the electric field distribution of SPPs in the cylindrical waveguide in the XY plane, the electric field distribution at the tip of the probe in the XZ plane, and the corresponding \({\text{E}}_{\text{z}}\) phase distribution in the XZ plane during DSP and DSPP nanofocusing, respectively

Figure 2b compares the electric field strengths of DSPP and DSP tips by varying the front cone angle. Simulation results show that when the cone height is 500 nm and the angle is 42°, the electric field strength at the DSPP tip increases by a factor of 6 compared to that at the DSP tip. This further highlights the importance of the platform-based structure for DSPP.

In order to more clearly compare the nanofocusing mechanisms of FPFP, DSP, and DSPP, this paper studies the electric field distribution and phase distribution in the SPPs propagation of the three types of probes. As shown in Fig. 2c-e, SPPs are excited in FPFP gold films using \({\text{HE}}_{11}\). Since FPFP is an axially symmetric metal-coated fiber probe, \({\text{HE}}_{11}\) excites \({\text{HE}}_{11}-\text{SPP}\)42,43, and \({\text{HE}}_{11}-\text{SPP}\) exhibits the same antisymmetric polarization distribution at positions symmetrical along the probe axis as \({\text{HE}}_{11}\).

Figure 2f–h shows the SPPs excited by \({\text{HE}}_{11}\) in the DSP gold film. When incident light passes through two asymmetrical half-ring slits, plasmonic modes are excited on both sides of the probe. The DSP utilizes the height difference between the slits to achieve polarization reversal of the plasmonic modes on both sides of the probe. In the cylindrical waveguide at a distance of 30 nm from the probe tip, although a large part of the electric field distribution of the waveguide mode is tightly confined around the metal cylindrical waveguide, a small part has diffused. Therefore, when the SPP continues to the probe tip, the local electric field enhancement at the probe tip is small. Furthermore, although the phases are consistent near the probe tip, the phases on the opposite side of the probe cone are still opposite. This shows that although the DSP can achieve polarization reversal and nanofocusing under LPL using two asymmetrical semicircular slit structures, the efficiency of plasmonic polarization reversal remains relatively low.

Figure 2i–k shows the SPPs excited by \({\text{HE}}_{11}\) in the DSPP gold film. As can be clearly seen from the figures, the plasmonic modes are uniformly distributed around the cylindrical waveguide at a distance of 30 nm from the probe tip, and the electric field intensity at the probe tip is significantly enhanced, while the phase is consistent at positions symmetrical along the probe axis. This is because the two asymmetric slit structures convert the \({\text{HE}}_{11}-\text{SPP}\) with an antisymmetric phase distribution into the \({\text{TM}}_{01}-\text{SPP}\) with a symmetric phase distribution, and the \({\text{TM}}_{01}-\text{SPP}\) does not undergo mode cutoff3,36,42,43,44,45. It can continue to propagate along the probe surface to the probe tip. When it propagates to the probe tip and focuses, part of the energy will continue to propagate along the front cone surface on the opposite side to the platform toroidal plane. The platform toroidal plane acts as a reflective surface, regulating the SPPs propagating to the platform to propagate back to the probe tip, achieving multiple energy superposition37,40.

Nanofocusing experiment verification

The experimental setup for far-field polarization analysis technique is shown in Fig. 3a. In order to study the main electric field component of the DSPP far-field scattered light spot, only the angle of \({\text{LP}}_{2}\) was adjusted during the experiment, while the angles of all other optical components remained unchanged. The focused light intensity at the tip of the probe was measured in the direction of probe scattering. When LPL is coupled into the fiber for transmission, SPPs can be excited on the outer surface of the gold film. Therefore, the probe tips of DSPP and DSP have obvious nanofocusing scattered light spots. Furthermore, due to the reflection of the platform toroidal plane, the brightness of the DSPP probe tip focused light spot is significantly brighter than that of the DSP (Fig. 3b, c). Simultaneously, the axial component of the focused light field at both probe tips is significantly larger than the transverse component, indicating that SPPs have achieved nanofocusing at the tips of both probes. Figure 3e shows the comparison of the focused spots from the two probe tips, we extracted the gray values of the vertical lines at the maximum brightness of the light spots to compare the experiments. From the experimental results, it can be seen that the focused spot size of DSPP is smaller than that of DSP, and the brightness of the focused spot is greater than that of DSP. The experimental results also verify the simulation results mentioned above and further prove the importance of the platform-based structure for DSPP.

Fig. 3: Plasmonic fiber probe focused light field experiment.
Fig. 3: Plasmonic fiber probe focused light field experiment.
Full size image

a Experimental setup for polarization analysis of the probe focused light field. b and c are the scattered light field images of the probe tip under LPL for DSPP and DSP, respectively. The red arrows indicate the polarization output direction of \({\text{LP}}_{2}\). d Far-field scattered light field image of FPFP under LPL. e Comparison of the size and brightness of the focused spots for DSPP and DSP. The regions where the focused spots are located were selected from the axial component images of the focused light fields for the two types of probes in (b) and (c), respectively. These regions are marked with blue rectangular frames in (b) and red rectangular frames in (c). f Relationship between the intensity of the focused light spot at the probe tip and the incident light polarization direction. The figure shows the scattered and transmitted light field images for incident polarization angles of 0°, 90°, 180°, and 270°, along with simulation results at the probe tip. It also illustrates the relationship between the incident light polarization direction and the probe structure for some angles. The blue arrows indicate the polarization direction of LPL. and the red lines indicate the double slit and platform-based structure of the probe. A attenuator, LP linear polarizer, HWP half waveplate, MO micro-objective, FOP fiber optic probe

In order to more clearly illustrate the nanofocusing characteristics of DSPP, the far-field scattered light field images of FPFP under LPL were measured at the same time, as shown in Fig. 3d. In the figure, there are also far-field scattered light spots at the tip of FPFP, but the light spots are completely dispersed, and there are no obvious probe tip focusing light spots, so the specific location of the tip cannot be found. This shows that SPPs can be excited and propagated on the probe surface, but \({\text{HE}}_{11}-\text{SPP}\) has a mode cutoff and is converted into background light at the tip of the probe. Therefore, for FPFP, the axial component of the tip-focused light field is much smaller than the transverse component. The different far-field scattered light field component characteristics of the three probes prove that DSPP excited \({\text{TM}}_{01}-\text{SPP}\) and achieved a larger nanofocusing electric field strength at the tip of the probe.

As shown in Fig. 3f, both the transmitted and scattered light intensities at the probe tip and the simulation results exhibit dipole-like intensity variations in polar coordinates. A clear and bright focal spot is only observable when the incident light polarization is perpendicular to the interface between the double slits. When the linear polarization direction is parallel to this interface, no \({\rm{\pi }}\) phase delay can be introduced due to the absence of asymmetry on both sides of the probe. Therefore, \({\text{HE}}_{11}-\text{SPP}\) cutoff occurs, and only weak background light exists at the probe tip. The dependence of the transmitted and scattered light field intensities at the probe tip on the polarization of the incident light indicates that the focusing at the probe tip arises from the interaction between the linearly polarized incident mode in the optical fiber and the double slit structure.

Research on broadband nanofocusing characteristics

In order to study the broadband nanofocusing characteristics of DSPP, FDTD was also used for simulation in this paper. Based on the simulations in the section “Nanofocusing Mechanism of Probe Tip,” a broadband light source with a wavelength range of 580 nm to 800 nm was used, and the height H of the front cone was varied. The electric field enhancement at 2 nm below the DSPP probe tip under broadband light varies with the height of the front cone, as shown in Fig. 4a. It can be seen that with the change in the height of the front cone, the phenomenon of electric field enhancement always exists at the tip of the probe, proving that DSPP can achieve nanofocusing under broadband (also see Fig. S8 in the Supplementary Information 1). Although DSP can also achieve electric field enhancement at the tip in a broadband range, the effect of electric field enhancement at the tip is significantly weaker for shorter wavelengths, especially in the range of 580–720 nm. This is because SPPs suffer greater losses when propagating at short wavelengths on metal surfaces. However, for DSPP, the probe can use the platform toroidal plane as a reflective surface to regulate the SPPs propagating from the probe tip to the platform toroidal plane, causing them to propagate back to the probe tip, thereby achieving multiple energy superpositions and significantly enhancing the intensity of the nanofocusing energy at the tip.

Fig. 4: Nanofocusing characteristics of DSPP under broadband.
Fig. 4: Nanofocusing characteristics of DSPP under broadband.
Full size image

a Changes in electric field enhancement under incident wavelengths of 580 nm to 800 nm at 2 nm below the DSPP and DSP probe tips, where the lines marked with circles represent changes in the probe tip electric field strength of DSPP at different cone heights, and the line marked with stars represents the electric field enhancement effect of DSP under broadband. b Electric field strength at 2 nm below the probe tip corresponding to different front cone angles at incident wavelengths of 580–720 nm when the cone height H = 500 nm. c Focused light field images of the DSPP during the process where the incident light transitions from yellow (580–590 nm) to red and gradually darkens (730–740 nm). During this experiment, the incident light wavelength increases from short to long, with sampling every 10 nm. The first row of images in the figure shows the transmitted light field images of DSPP in a broadband range, with the bright spot in the center being the transmitted focused light spot of the probe. The second row of images shows the scattered light field images of the probe, and the third row shows the axial component of the scattered light field of the probe. The white arrows in the figure represent the output polarization of the linear polarizer placed between DSPP and CMOS, which is parallel to the probe axis

Following the identification of an optimal cone height through simulation, we further investigated the influence of the front cone angle on the electric field strength at the DSPP tip across the broadband spectrum. The simulation results are shown in Fig. 4b. As the front cone angle varies, the probe tip achieves greater electric field enhancement in a broadband range (also see Fig. S9 in the Supplementary Information 1). This demonstrates that the energy recycling mechanism of this probe can effectively solve the propagation loss of SPPs at shorter wavelengths. Furthermore, despite significant changes in the front cone angle, the electric field enhancement at the tip remains robust without abrupt changes. Consequently, this also relaxes the fabrication requirements for the probe.

Furthermore, this paper experimentally analyzed the broadband nanofocusing characteristics of DSPP7,40,46,47. Figure 4c shows the focused light field image of DSPP under broadband conditions. It can be seen from the transmitted and scattered light fields of DSPP that there is a clear focused light spot at the tip of the probe, and in the axial component image of the scattered light field, the focused light spot at the tip of the probe still exists. The experimental results are consistent with the simulation results, indicating that DSPP can achieve nanofocusing at the tip under broadband conditions through polarization reversal of the double slit and interference modulation of the platform-based structure.

Optical nanoscale imaging using DSPP

To validate the super-resolution measurement application of DSPP, this study selected samples with small-sized features and compared them with commonly used characterization measurement techniques for verification. Figure 5a shows the scanning electron microscope (SEM) image of the measured sample. The sample consists of a \({\text{SiO}}_{2}\) substrate with an Au film deposited on its surface. The structure comprises two opposing triangles and extremely narrow slits between them. To more accurately characterize the dimensions of the slit structure, we used atomic force microscopy (AFM) to measure the sample dimensions, with the results shown in Fig. 5b. By extracting the profile curve at the slit location, the slit width was determined to be 28.2 nm, with the fitted curve shown in Fig. 5d. The optical morphology test results of the sample using DSPP are shown in Fig. 5c. Since the experimental setup uses a transmission signal collection mode, the light field intensity should be higher in areas not covered by the Au film. From the test results, it can be observed that the areas with higher light field intensity are located at the two opposing triangles and the central slit, and the slit and the two opposing triangular structures can be clearly identified. It is worth noting that the higher noise level compared to SEM/AFM images primarily stems from the weak evanescent-field signal and environmental perturbations inherent to near-field optical microscopy. Then, the profile curve at the slit position was extracted, yielding a slit width of 28.6 nm as measured by DSPP. The fitted curve is shown in Fig. 5e, which is consistent with the morphological structure measured by SEM and the slit width measured by AFM. This also demonstrates that the optical image resolution of DSPP measurements can reach 28.6 nm, enabling deep subwavelength resolution imaging of samples under ambient conditions, while also validating that DSPP can detect extremely weak signals.

Fig. 5: Optical nanoscale imaging using DSPP.
Fig. 5: Optical nanoscale imaging using DSPP.
Full size image

a SEM image of the measured sample, which consists of two opposing triangles and a narrow slit between them. The black dashed rectangular box indicates the measurement region for DSPP. b AFM characterization image of the sample. c Optical morphology test results of the sample obtained using DSPP. d Measurement results of the blue line in b as a function of position, with the measured slit width being 28.2 nm. The black dots represent experimental measurement results, and the blue solid line represents the results obtained using Gaussian fitting, where FWHM denotes full width at half maximum. e Near-field intensity of the red line in c as a function of position, with FWHM being 28.6 nm. f Optical measurement results of the sample using confocal microscopy. g Comparison of slit size measurements obtained using AFM, confocal microscopy, and DSPP. The blue dashed line represents the measurement results of the blue line marked in (b), the red solid line represents the near-field intensity results of the red line marked in (c), and the green solid line represents the measurement results of the green line marked in (f)

Furthermore, we performed optical measurement imaging of the slit structure in this sample using confocal microscopy, as shown in Fig. 5f. The results revealed only a blurred outline of the measured structure, failing to clearly image the nanoscale slit. This apparent broadening and the resultant overestimation of the feature size are direct consequences of the diffraction limit inherent to far-field optical microscopy. Figure 5g displays the profile curve distributions at the slit for the three measurement methods. It is evident that both DSPP and atomic force microscopy exhibit significantly higher resolution than confocal microscopy. In addition, since DSPP can simultaneously provide the optical properties and morphological information of the sample, it has more advantages for research in materials science and biological science.

Conclusions

The DSPP achieves efficient nanofocusing under LPL by integrating the Fabry–Pérot interference enhancement mechanism of a platform-based structure with the polarization control function of an asymmetric half-ring slit, thereby eliminating the dependence of traditional axisymmetric probes on complex RPL. Simulations show that the height difference between the two asymmetric slits introduces a phase delay, converting \({\text{HE}}_{11}-\text{SPP}\) into \({\text{TM}}_{01}-\text{SPP}\). The platform-based structure further recycles and superposes energy via reflection, enhancing the electric field at the probe tip sixfold compared to the DSP. Experimental measurements confirm that the brightness of the nanofocused spot at the DSPP tip is significantly greater than that of the DSP, which is consistent with the simulation results. For the fabrication of the DSPP, a novel sleeve ring etching technique was employed. The front cone structure was accurately processed through multiple cyclic etching steps, which improved the tip curvature radius by more than an order of magnitude. Compared to conventional fabrication methods, the proposed approach provides an efficient and applicable strategy for producing structured nanoprobes with high controllability and performance. Broadband characterization further confirms that the DSPP enables nanofocusing at the probe tip across a broad spectrum, with a significantly stronger electric field enhancement in the short-wavelength range compared to the DSP, while also exhibiting robust structural parameter tolerance. Furthermore, optical imaging experiments demonstrated a resolution of 28.6 nm, enabling the simultaneous characterization of both morphological and optical properties of deep subwavelength samples under ambient conditions, significantly surpassing the performance of confocal microscopy. In summary, the DSPP combines high electric field enhancement, simple polarization adaptability, and broadband compatibility, providing an innovative tool for nano-optical detection and super-resolution imaging. Therefore, this probe holds great potential for various applications requiring highly localized field enhancement, such as high-sensitivity single-molecule detection and nanolithography.

Methods

Finite-difference time-domain simulation methods

To investigate the nanofocusing mechanism of DSPP, this paper employs the FDTD method for simulation modeling. We established geometric models of DSPP, DSP, and FPFP. To ensure consistency in comparison, the structural dimensions of FPFP are exactly the same as those of DSPP without a double slit structure. In this paper, the radius R of the platform toroidal plane is set to 500 nm. The cone angle of DSP is set to 32°, which is the same as the back cone angle of DSPP. Since FPFP is an axisymmetric metal-coated fiber probe, \({\text{HE}}_{11}\) inside the fiber will excite \({\text{HE}}_{11}-\text{SPP}\) outside the metal film. Therefore, it is necessary to set an appropriate film thickness to reduce the influence of background light noise. In this paper, the film thickness of DSP and the front cone film thickness of DSPP are both set to 180 nm (see Section S3 in the Supplementary Information 1 for the Au film thickness setting). The radii of the two asymmetric semi-circular slits in DSP and DSPP are identical. The height difference between the dual slits in DSP and DSPP is also consistent (see Section S2 in the Supplementary Information 1 for the model structure and parameter settings in the simulation). The incident laser wavelength used in the simulation experiments is 633 nm.

Far-field polarization analysis techniques and spectral measurement

Far-field polarization analysis techniques were employed to investigate the focusing light field characteristics of DSPP and the interaction between incident light and the probe48. The experimental setup is shown in Fig. 3a. A laser with a wavelength of 633 nm was modulated through an attenuator (A), linear polarizer 1 (\({\text{LP}}_{1}\)), and a half waveplate (HWP) to produce LPL with a specific polarization angle. The light was then focused and coupled into a single-mode fiber using a micro-objective lens (MO), and the far-field scattered and transmitted light spot intensity at the probe tip were measured using a research-grade CMOS camera (Thorlabs, CS126CU). During far-field scattering tests, placing linear polarizer 2 (\({\text{LP}}_{2}\)) between the probe tip and the CMOS allows analysis of the far-field scattering components of the focused light spot at the probe tip. In the experiment, the incident light power can be continuously tuned via an attenuator with an adjustment range of 0–100%, and the incident light polarization angle can be continuously varied by rotating a half-wave plate.

In the spectral measurement experiments, a xenon lamp source (ASB-XE-175-BFEX) was used as the incident light, and the signal was captured by a research-grade CMOS camera (Thorlabs, CS126CU). The detection principle benefits from the advances in broadband photodetection technology49,50. In the experiments, the fiber used was a multimode fiber (see Table S7 in the Supplementary Information 1 for fiber specifications), and the xenon light was modulated by a monochromator (CM110 Compact 1/8 m Monochromator) to produce a monochromatic light source with a wavelength accuracy of 0.2 nm.