Open Access
Add to BibSonomy
Abstract
We present a procedure for the design of multilevel metalenses and their fabrication with multiphoton-based direct laser writing. This work pushes this fast and versatile fabrication technique to its limits in terms of achievable feature size dimensions for the creation of compact high-numerical aperture metalenses on flat substrates and optical fiber tips. We demonstrate the design of metalenses with various numerical apertures up to 0.96, and optimize the fabrication process towards nanostructure shape reproducibility. We perform optical characterization of the metalenses towards spot size, focusing efficiency, and optical functionality with a fiber beam collimation design, and compare their performance with refractive and diffractive counterparts fabricated with the same technology.
© 2024 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Ultra-thin metamaterials, called metasurfaces, are artificially created surfaces designed to have specific and unique electromagnetic properties [1,2]. By accurate design of the sub-wavelength meta-atoms constituting the metasurface, one is able to realize electromagnetic field modulation leading to the unprecedented control of light. These meta-atoms often come in the form of a periodic or aperiodic lattice of unit cells, typically consisting of nanostructures in the case of optical metasurfaces, to generate the desired electromagnetic response. We can identify multiple application scenarios that have a large potential for the integration of these compact and highly customizable optics, such as in telecommunication, sensing, imaging, and biomedicine. Examples of such metasurfaces that have already been demonstrated include achromatic lenses [3], hologram generators [4], anti-reflection coatings [5], and many other advanced optical elements [6]. One of the interesting properties of light that can be modulated by a metasurface is its phase, thus allowing to manipulate the wave’s propagation direction, polarization, and intensity profile. As such, so-called flat metalenses can be designed to effectively focus the wavefront [7,8]. By accurate engineering of the constituting nanostructures, high-performance metalenses can be achieved with wide control of the numerical aperture (NA), aberrations, and chromaticity [9]. Light focusing can be achieved in different ways, such as with local or extended resonant effects, or by locally changing the rotational orientation (i.e., Pancharatnam-Berry phase control for polarized light) or effective refractive index of the unit cell structures [10].
Furthermore, metasurfaces are often implemented in lab-on-fiber technologies, where the integration of high-performance flat optics directly on optical fibers can play a crucial role [11,12]. Currently, most metasurfaces are being fabricated through high-resolution photolithography, focused ion beam milling, or electron beam lithography, and are often accompanied with nano-transfer techniques to pattern for example the end-facet of an optical fiber. Recently, researchers in different fields are also exploring the attractive technology of multiphoton polymerization [13], also often called 2-photon polymerization (2PP) or direct laser writing, to create metasurfaces on flat substrates and on fiber tips. The technology typically uses an infrared femtosecond pulsed laser beam, together with a high-NA microscope objective to initiate the nonlinear process of 2-photon absorption within a localized focal volume called the voxel. Scanning this voxel within a droplet of liquid photoresin material allows to build 3-dimensional (3D) structures with sub-micrometer resolution. In previous work on 2PP-printed metasurfaces, grating-like [14] and Fresnel lens-like [15] metalenses have been fabricated on single-mode optical fiber (SMF) tips in view of creating high-NA lensed fibers (which are inherently polarization insensitive, and present NA values close to 0.9 and a focusing efficiency of 73 %). Balli et al. explored the 2PP technique to create nanopillar- [16] and nanohole-based [17] hybrid metalenses in combination with phase plates achieving moderate NA values up to 0.27. Later, Pancharatnam-Berry phase modulation was explored with 2PP-printed birefringent nanostructures of cuboid shapes [18,19], whereas cylindrical multiheight metastructures were investigated to modulate the surface’s effective index [20]. In most of these cases, the nanostructures and their periodicity often had relatively large sizes compared to the operational wavelength in order to reach good shape agreement during the fabrication, but this generally also limits the achievable NA values and might lead to additional diffraction orders.
In this work, we investigate multilevel nanopillars, based on the natural ellipsoidal voxel shape obtained by moving the 2PP voxel along the vertical direction, in order to locally modulate a light beam’s phase profile across the metasurface. Since an increase in metasurface performance and modulation resolution is related to achieving smaller unit cell sizes, which often requires time-consuming and cumbersome electron or ion beam lithography and transfer processes, our goal is to explore and push the fast and flexible 2PP technology to its limits in terms of achievable sub-micrometer unit cell dimensions. As opposed to fixed-height metalenses fabricated with conventional lithography techniques, we exploit the 3D fabrication freedom of the 2PP technology to design multilevel metalenses with varying nanopillar heights. In this study, we develop a procedure to design metalenses, based on the effective index theory by modulating the height of the constituting nanostructures, and we print them using 2PP. We show the versatility of this methodology by creating metalenses with 3 different focal lengths and numerical apertures with a maximum value of 0.96. Moreover, we compare our multilevel metalens performance with that of its refractive and diffractive counterparts, all fabricated with the same 2PP technology. Optical characterization is performed by measuring the focal spot sizes and focusing efficiencies of the different lenses.
To validate our metalens concept with a practical demonstrator, we also investigated its use for fiber beam collimation, which is an interesting functionality in many optical interconnect and sensing applications. Researchers have already developed silicon-based metasurfaces to collimate the light beam coming from semiconductor lasers [21], silicon photonic integrated circuits [22], and optical fibers [23]. The latter work consisted of a polarization sensitive metasurface made of high-refractive index silicon nanobricks to enable vortex generation for the transverse electric polarization and beam collimation for the transverse magnetic polarization. Ye et al. theoretically demonstrated a polarization insensitive metasurface profile for fiber beam collimation, where they made use of low-refractive index nanoposts to minimize back-reflections at the fiber end-facet [24]. In this paper we show the practical realization of a low-index fiber beam collimating metalens and evaluate its performance by measuring the divergence angle of the emitted beam.
2. Creation of the metalens phase profile via unit cell simulations
One of the main parameters in the design of a metalens is the unit cell size $U$, which should comply with several considerations, namely the sub-wavelength dimension, sampling, and fabrication tolerance criteria. A first consideration is that the unit cell should be smaller than the operational wavelength in the lens substrate, i.e., $U < \lambda \, / n$, to ensure only zero-order diffraction for normal incidence. In this work, we will use the telecommunication wavelength of $\lambda$ = 1550 nm and print the metalenses on a fused silica substrate with refractive index $n$ = 1.444. Therefore, we should ensure that $U$ < 1.07 $\mu$m. In addition, $U$ should fulfil the Nyquist sampling criterion $U < \lambda \,/\,2\,NA$, with $NA$ the numerical aperture of the lens, to ensure diffraction-limited resolution [10]. In order to be able to create high-NA metalenses (e.g., NA > 0.9) this sampling criterion dictates that $U < 0.86\,\mu m$ at 1550 nm. To comply with the above considerations, we choose to work with a square unit cell with a width of 0.8 $\mu$m.
On the other hand, our goal is to explore the fabrication of these metalenses with the 2PP fabrication technology. The unit cell and nanostructure size should therefore fit within the 2PP dimensional tolerances. Typical minimal feature sizes of commercial 2PP printers are in the order of about 200 nm [25] at low exposure dose of the laser source. In order to push our metalens unit cell towards these limiting feature sizes, the nanostructures are chosen to be cylindrical nanopillars of varying heights consisting of vertically stacked voxels, where their diameter is chosen based on the effective refractive index of the unit cell’s cross-section. Figure 1 (a) shows the basic geometry of the unit cell, where the nanopillar includes an ellipsoidal-shaped tip originating from the fundamental voxel shape that is elongated along the vertical direction, i.e., following the optical axis of the direct laser writing system. Figure 1 (b) shows the electric field intensity distribution for horizontally polarized light incident on a periodic structure with square unit cells having a width of 0.8 $\mu$m and containing a pillar with a diameter of 0.6 $\mu$m, simulated with the Lumerical MODE (Ansys) software. Note that both the square unit cell and the circular nanopillar shape entail a polarization insensitive design.
Fig. 1. Unit cell simulations. (a) Isometric and top-view of the unit cell geometry consisting of a fused silica substrate (blue) and a polymer nanopillar (orange). (b) Electric field intensity distribution for horizontally polarized light on a periodic nanopillar structure for a unit cell width $U$ of 0.8 $\mu$m and a pillar diameter $D$ of 0.6 $\mu$m. (c) Simulation results for the effective refractive index as a function of the nanopillar diameter $D$ and unit cell width $U$. (d) Fitted data of the simulated phase change as a function of the nanopillar height for a periodic nanopillar structure of different fill factors (FF). All simulations are performed at 1550 nm.
As introduced earlier, we design metasurfaces that introduce phase changes based on local variations in effective refractive index. The local accumulated phase difference of a given unit cell geometry can be expressed by
where $n_{\mathit {eff}}$ is the effective refractive index that the incident light beam sees, and H is the height of the nanostructure. It is clear from this equation that the effective index plays an important role in the achievable values of $\phi _{U}$. The parameter $n_{\mathit {eff}}$ comes into play since the refractive index of the nanopillar is larger than that of the surrounding medium (i.e., air), thus the propagating field experiences a certain effective index depending on the nanopillar geometry. In essence, for a fixed nanopillar material (i.e., in our case the commercially available 2PP photoresin IP-Dip from Nanoscribe GmbH [26] with $n$ = 1.53 in its polymerized state [27,28]), we can tune $n_{\mathit {eff}}$ by accurately choosing the fill factor FF (i.e., the ratio of the nanopillar’s areal cross-section over the unit cell area) within the unit cell. Accordingly, we used our unit cell model to simulate the effective index of a periodic lattice of nanopillars with varying unit cell sizes and pillar diameters, as shown in Fig. 1 (c). These results indicate that the highest $n_{eff}$ of 1.53 can be obtained for an FF equal to 1, as expected. A decrease in FF will reduce the $n_{\mathit {eff}}$, which influences the achievable $\phi _{U}$ for a specific nanopillar height, as evident from Eq. (1). Finally, we use finite-difference time-domain (FDTD) simulations (Lumerical FDTD, Ansys) of the unit cell with periodic boundary conditions to numerically verify the phase change that corresponds with different FF values, for a sweep in nanopillar height, as shown in the fitted data in Fig. 1 (d). It is clear that smaller nanopillar heights can be used to cover the full 2$\pi$ phase range when one chooses a larger FF with a correspondingly larger $n_{\mathit {eff}}$. In the end we choose a nanopillar diameter of 0.6 $\mu$m leading to a FF = 0.44 and $n_{eff}$ = 1.23, in order to have sufficient space between the adjacent nanopillars for reliable 2PP fabrication, avoiding problems that can arise due to proximity effects in the printing process [29]. This numerically obtained nanopillar height versus phase change relation is then used to further design the metalenses.During the characterization of our 2PP metalenses we will use a monochromatic plane wave source at normal incidence that will be focused to a single spot by the metalens. Hence, in the design we can use the following geometric phase function
which describes the local phase difference $\phi _{x,y}$ needed at each cross-sectional position (x,y) along the metalens surface in order to reach the desired focal length $f$. In this formula, the plus sign refers to a diverging lens, whereas the minus sign corresponds to a converging lens. Only the latter is of interest in our study. The unit cell design is then used to establish a set of unit cells sampling the phase profile in Eq. (2). To show the applicability of our 2PP metalens design and fabrication method, we create metalenses with 3 different focal lengths and numerical apertures.2.1 Fiber beam collimation metalens with NA of 0.15
First, we create a metalens profile to be used for collimation of the beam emitted by a G.652 standard telecom SMF [30]. In previous work [31,32], we already developed a fiber beam expansion and collimation micro-lens component of which the refractive lens profile was optimized using beam propagation simulations (Zemax OpticStudio, Ansys). In this work, we design an equivalent metalensed fiber beam collimator by using the appropriate focal length of 406.1 $\mu$m and an aperture of 120 $\mu$m, resulting in a metalens with NA = 0.15, where we use the traditional approximation that $NA = D\,/\,2\,f$, in which $D$ is the lens aperture.
2.2 Refractive, diffractive, and metalens with NA of 0.5
For the second metalens design, we aim for an NA of 0.5. For a lens aperture of 100 $\mu$m (chosen in consideration of our collimated light source which has a beam diameter of about 90 $\mu$m) this entails using a focal length of 100 $\mu$m. In the next sections we show the optical characterization of this metalens towards its focusing efficiency, spot size, and focal length. In addition, we design and fabricate diffractive and refractive counterparts of this metalens with the same 2PP technology to enable a direct comparison of their optical performance. As such, a diffractive kinoform lens was designed with the same phase function as given in Eq. (2), using a quasi-continuous surface profile (i.e., 256 height levels), a pixelsize of 0.8 $\mu$m (i.e., equal to the metalens unit cell size), and an aperture of 100 $\mu$m. For the refractive counterpart we targeted a plano-convex spherical lens with a radius of curvature of 53 $\mu$m (based on the well-known lens maker’s equation $1/f = (n-1)\,(1/R_{1} - 1/R_{2})$, to keep its design as simple as possible) and an aperture of 100 $\mu$m. It has to be mentioned that the focal spot of such a spherical lens can be impaired by spherical aberrations for normal incidence of light, which is theoretically mitigated by the geometric phase function in Eq. (2) for the case of our metalens and diffractive lens.
2.3 Metalens with NA of 0.96
Finally, the third metalens was designed to reach the largest NA as allowed by the Nyquist sampling criterion, which is 0.96 for the current unit cell size of 0.8 $\mu$m, leading to a focal length of 52 $\mu$m for a lens aperture of 100 $\mu$m. Figure 2 shows the general design process flow where Eq. (2) is used to generate the desired phase map (Fig. 2 (a)). The phase profile is then wrapped over the 2$\pi$ phase range (Fig. 2 (b)), after which the corresponding height values are linked to the (x,y)-surface positions via the data obtained in Fig. 1 (d).
Fig. 2. Metalens design methodology. (a) Phase profile of the NA 0.96 converging metalens. (b) Wrapped phase profile, also showing the corresponding pillar height values. (c) Scan along the X axis through the center of the continuous phase profile, where the sampling points are indicated with dots. (d) 3D rendering of the final metalens design.
In the next section, we elaborate on the optimization of the fabrication process towards the chosen unit cell size and nanopillar diameter and height, and we show the final lenses fabricated on glass substrates and on the end-facet of a standard telecom SMF.
3. Optimization of the 2PP fabrication process
After designing the optimal metalens profiles, the profile data was transformed into the general writing language of the commercial 2PP printer that we use in this study (Photonic Professional GT+, Nanoscribe GmbH [26]). The generated code describes the start and end point coordinates of each nanopillar. Since the nanopillars are based on vertically-stacked single-voxels and high reproducibility is desired, we decided to use the highly accurate 3D piezo stage of the system to successively print the nanopillars one by one with the high-resolution (63$\times$, NA 1.4) microscope objective immersed into the high-performance IP-Dip photoresin material. In this configuration, we are not using the high-speed galvanometric scanner of the system, therefore exchanging speed for accuracy [33].
During the optimization of the fabrication process, we identified several printing parameters that are of high relevance to reach the desired print quality and reproducibility. First, the system allows to continuously open the laser shutter such that the femtosecond pulsed laser beam irradiates the photoresin in a continuous way, or to modulate the shutter to deliver pulsed shots of laser irradiation. We found that the latter option gives the best control in delivering a fixed exposure dose and thus achieves the most accurate and reproducible nanopillar geometries. This is because the exposure dose determines the voxel’s size, the beam quality of the non-ideal laser source, and the printed structure’s mechanical strength due to its effect on the crosslinking density [34,35]. Indeed, the pulsed exposure mode renders the nanostructures less prone to all these undesired side-effects, and next we have to optimize the applied laser power, the exposure time, and the point distance for the interpolated coordinates between the starting and ending points of the nanopillars. Figure 3 provides an overview of the exploration of the pulsed printing mode, where we show the effect of the applied laser power on the nanopillar width and an evaluation of the multilevel printing of various nanopillar heights. Figure 3 (b) shows an array of nanopillars with a fixed diameter of 0.6 $\mu$m and with an increasing height from 1.3 $\mu$m up to 7 $\mu$m, therefore reaching a maximum aspect ratio of more than 11. This image shows that we achieve good nanopillar shape fidelity and reproducibility when a relatively large spacing between the pillars is used (in this image for example 3 $\mu$m). However, we start to observe deformations when tall nanopillars are positioned closer to each other. That is to say, we find that such high-aspect ratio nanopillars tend to bend towards each other as displayed in Figs. 3 (c)-(f), showing the results of a matrix of nanopillars with different heights and a pitch of 0.8 $\mu$m. This clustering effect is believed to be caused by the last step of the fabrication process, namely the development. Here, all unexposed and hence not polymerized resin is removed by a liquid developer (i.e., propylene glycol methyl ether acetate or PGMEA), after which capillary forces and surface tension impact the thin and long nanopillars during evaporation of the remaining liquids [36,37]. In view of minimizing this effect we submerge the sample into a bath of Novec 7100 Engineered fluid (3M), which has a surface tension of about half of that of PGMEA and traditional isopropanol, before leaving it to dry in air. This way, the bending effect is less pronounced, but still present. Nevertheless, we achieve a very good overall height accuracy and reproducibility as evidenced by comparing the designed nanopillar matrix in Fig. 3 (c) and the measured height map after fabrication in Fig. 3 (e). Moreover, although this bending creates localized small disturbances of the effective index on the metasurface, we show in the next section that the fabricated metalenses indeed do achieve good optical performance. Table 1 lists the optimized fabrication parameters that are finally used for the creation of the metalens components in this work.
Table 1. Overview of the two-photon polymerization-based multilevel nanopillar fabrication parameters for metalens fabrication.
Fig. 3. Two-photon polymerization nanopillar fabrication optimization. (a) Scanning electron microscope image showing the influence of the exposure laser power on the pillar width. (b) Tilted-view image of multilevel printed nanopillars of increasing height and a diameter of 0.6 $\mu$m (c) Design matrix for the evaluation of nanopillars with different aspect ratios. The diameter of the nanopillars is fixed at 0.6 $\mu$m, the pitch equals the unit cell width of 0.8 $\mu$m, and the height is changed over the matrix. (d) Scanning electron microscope image of the resulting matrix after fabrication (top-view). (e) Height map of the fabricated matrix, measured with a confocal laser scanning microscope. (f) Tilted-view of the fabricated matrix of nanopillars.
We continue by printing our 3 different metalenses with the optimized fabrication parameters, together with the refractive and diffractive counterparts discussed in the previous section. As mentioned earlier, we fabricate the NA 0.15 metalens on the tip of a cleaved G.652 standard telecom SMF (SMF-28, Corning). As shown in the scanning electron microscope image in Fig. 4 (a), the metalens design consists of a 606 $\mu$m long base to allow the fundamental mode of the SMF to expand, before collimating the beam with the metalens. Figures 4 (b)-(d) show the 3 fabricated NA 0.5 lenses, respectively the refractive lens, the diffractive lens, and the metalens. Both the refractive and diffractive lenses are printed with the high-speed galvanometric scanner at a print speed of 40 mm/s, leading to print times of 16 minutes and 87 minutes, respectively. The metalens on the other hand is fully printed with the 3D piezo stage, resulting in a total time of about 8 hours to print the more than 12,200 individual nanopillars. Finally, Fig. 4 (e) zooms in on the NA 0.96 metalens with its high spatial frequencies, where the clustering of the high aspect ratio nanopillars is clearly visible.
Fig. 4. Scanning electron microscope images of the fabricated lenses by two-photon polymerization. (a) NA 0.15 fiber collimation metalens printed on the tip of a G.652 standard telecom single-mode optical fiber. (b) Refractive lens, (c) diffractive lens, and (d) metalens with an NA of 0.5. (e) The metalens with an NA of 0.96.
4. Optical characterization of the metalenses
4.1 Fiber beam collimation characterization
After fabrication of the microlenses, we characterize them in terms of optical performance. The collimation characteristics of the metalensed fiber at a wavelength of 1550 nm are verified with a short-wave infrared camera (Bobcat, Xenics) and a 20$\times$ NA 0.4 microscope objective (Seiwa), and compared to those of a cleaved SMF-28 and a refractive lensed fiber. Figure 5 shows the measured spot radii at different distances from the end-facet for the 3 types of fibers. From these measurements we are able to extract the numerical aperture and the divergence angle (as defined by the 1/e$^{2}$ intensity points). The results are given in Table 2, where a clear reduction of the half divergence angle from 4.3$^{\circ }$ to 0.7$^{\circ }$ is observed for the lensed fibers compared to the cleaved SMF-28 (without lens), corresponding to an effective collimation of the output light beam. Moreover, we find that our metalensed fiber performs equally well as its refractive counterpart for beam collimation, where a non-zero divergence is indeed still expected for the free-space propagation of Gaussian-like light beams [21,23].
Table 2. Overview of the fiber experimental divergence results for the SMF-28, refractive lensed fiber, and metalensed fiber, at a wavelength of 1550 nm
Fig. 5. Comparison of the output beam propagation for the case of a cleaved SMF-28, a refractive lensed fiber, and a metalensed fiber. All images were captured at a wavelength of 1550 nm.
4.2 Comparison of the optical performance of the refractive, diffractive, and metalens
Next, we characterize the NA 0.5 microlenses in terms of the achieved spot size, focal length, and focusing efficiency. Figure 6 (a) illustrates the measurement setup, where we use our refractive lensed fiber as a source which approximately delivers a plane wave at 1550 nm with a mode-field diameter of about 90 $\mu$m at the microlens’ substrate plane. The collimated light beam was then focused by the printed microlenses, and the images were captured by a 100$\times$ NA 0.9 microscope objective (Zeiss) and the above-mentioned short-wave infrared camera. The resulting images for the refractive lens, diffractive lens, and metalens are shown in Fig. 6 (b), (c), and (d) respectively. We also placed a non-polarizing cube beamsplitter in the imaging path to split part of the light beam towards an iris connected to a photodetector to be able to measure the focusing efficiency. Here, we choose the iris opening size to be 3 times the full width at half maximum (FWHM) focal spot size [38]. The focusing efficiency is then defined as the ratio of the optical power captured by the photodetector to the total incident power coming from the lensed fiber using the same measurement setup, going through a blank fused silica substrate and without using the iris. The focal length of the microlenses was estimated by finding the appropriate distance from the microlens substrate towards the focal point with a manual translation stage, leading to a measurement uncertainty of $\pm$ 10 $\mu$m. Table 3 summarizes all the results. We found that the spherical refractive lens creates a slightly larger spot size (3.9 $\mu$m) than its diffractive and metalens counterparts (3.5 $\mu$m and 3.4 $\mu$m respectively), but all values are very close to the theoretically expected Airy disk diameter of 3.8 $\mu$m for an NA of 0.5, meaning that the microlenses are indeed diffraction-limited. Moreover, the refractive focal spot shows a clear Airy ring pattern that is much more expressed than for the other microlenses. Next to the 1/e$^{2}$ spot size values, we also provide the FWHM dimensions, as it is a typically used metric in the specification of microlenses. The focal length of the diffractive lens was found to be closest to the designed value of 100 $\mu$m, whereas the focal length of the refractive lens has the largest offset. For the focusing efficiency, we found that the refractive lens performed best (48.3 ${\%}$), compared to the diffractive lens (38.9 ${\%}$) and the metalens (31.8 ${\%}$). For comparison, the 2PP-printed imaging metalenses developed by Wang et al. reached a maximum focusing efficiency of 9.4 ${\%}$ for an NA of 0.6 at 1030 nm [19], whereas Ren et al. created a 2PP-printed achromatic metalens with an efficiency of 30.9 ${\%}$ and NA of 0.2 at 1550 nm [18].
Fig. 6. (a) Illustration of the optical characterization measurement setup. Experimentally visualized focal spots for the (b) NA 0.5 refractive lens, (c) NA 0.5 diffractive lens, and (d) NA 0.5 metalens. (e) Captured focal spot of the NA 0.96 metalens. (f) Comparison of the focal spot cross-sections for the 3 types of NA 0.5 microlenses and the NA 0.96 metalens. All images were captured at a wavelength of 1550 nm. Scale bars = 2 $\mu$m.
4.3 Characterization of the high-NA metalens
Finally, we image the focal spot profile of the NA 0.96 metalens and the results are added to Table 3. We measure a spot size of 2.2 $\mu$m, equivalent to a FWHM = 0.84 $\lambda$, and in line with the theoretically expected diffraction-limited Airy disk diameter of 2.0 $\mu$m for an NA of 0.96. The measured focal length of 53 $\mu$m is found to be extremely close to the designed value of 52 $\mu$m, strengthening the claim that the microlens can be considered as a very high-NA metalens. To the best of our knowledge, an NA of 0.96 is the highest value obtained so far for 2PP-printed metalenses. Since the NA of the metalens is larger than that of the available microscope objectives in our measurement setup (going up to NA 0.9), we are not able to measure its focusing efficiency experimentally, as part of the light would not be captured by the objective lens.
Table 3. Overview of the optical characterization results for the 3 NA 0.5 microlenses and the NA 0.96 metalens. All measurements were performed at a wavelength of 1550 nm
5. Conclusion
We believe that the 2PP technique has become a reliable and valuable technology, complementary to traditional lithographic techniques, enabling in-situ printing of high-quality 3D nanostructures, and allowing the fabrication of high-performance metasurfaces. We showed the design of multilevel metalenses and their fabrication by multiphoton lithography, where we leveraged the 3D capabilities of the 2PP technology as compared to traditional 2D lithography techniques. Moreover, we pushed the technology to its limits in terms of dimensions in order to have better control on the excitation of only one diffraction order and to better comply with the Nyquist sampling criterion for the creation of high-NA metalenses. Although the current nanostructures are circularly symmetric and therefore polarization insensitive, minor modifications to the unit cell and nanopillar geometry could allow for polarization control in the near future. We have demonstrated that our process is capable of creating 2PP metalenses with a record high NA of 0.96, effectively focusing light to a sub-wavelength spot size of 0.84 $\lambda$ (FWHM) at the telecom wavelength of 1550 nm. We have validated that our metalenses reach good optical performance compared with their refractive and diffractive counterparts, reaching a focusing efficiency of 31.8 ${\%}$. Finally, we have demonstrated the versatility of the 2PP technology for the fabrication of metalenses directly on the tip of an optical fiber, with an exemplary use case of fiber beam collimation. Next to their use in optical interconnects and imaging, other applications such as sensing will be investigated in the near future where the sub-wavelength features of metasurfaces can be very important for enhanced light-matter interaction mechanisms.
Funding
Fonds Wetenschappelijk Onderzoek (12E2923N); Interreg Vlaanderen-Nederland; Interreg North-West Europe; Flanders Make; Belgian Federal Science Policy Office; Vrije Universiteit Brussel.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. P. Genevet, F. Capasso, F. Aieta, et al., “Recent advances in planar optics: from plasmonic to dielectric metasurfaces,” Optica 4(1), 139–152 (2017). [CrossRef]
2. D. Neshev, “Optical metasurfaces: new generation building blocks for multi-functional optics,” Light: Sci. Appl. 7(1), 58 (2018). [CrossRef]
3. F. Aieta, P. Kanhaiya, M. Kats, et al., “Achromatic metasurface lens at telecommunication wavelengths,” Nano Lett. 15, 5358–5362 (2015). [CrossRef]
4. L. Huang, S. Zhang, and T. Zentgraf, “Metasurface holography: from fundamentals to applications,” Nanophotonics 7(6), 1169–1190 (2018). [CrossRef]
5. J. Jeon, J. Hwang, K. Bhattarai, et al., “Robust metamaterial-based antireflection coating for surface plasmon polariton resonance,” Opt. Mater. Express 9(3), 1290–1297 (2019). [CrossRef]
6. K. Ou, H. Wan, G. Wang, et al., “Advances in meta-optics and metasurfaces: Fundamentals and applications,” Nanomaterials 13(7), 1235 (2023). [CrossRef]
7. M. Khorasaninejad and F. Capasso, “Metalenses: Versatile multifunctional photonic components,” Science 358(6367), eaam8100 (2017). [CrossRef]
8. I. Moreno, “Optics of the metalens,” Eur. J. Phys. 43(6), 065302 (2022). [CrossRef]
9. A. F. Amir Arbabi, “Advances in optical metalenses,” Nat. Photonics 17, 16–25 (2022). [CrossRef]
10. H. Liang, A. Martins, B.-H. V. Borges, et al., “High performance metalenses: numerical aperture, aberrations, chromaticity, and trade-offs,” Optica 6(12), 1461–1470 (2019). [CrossRef]
11. Q. Zhao, W. Yuan, J. Qu, et al., “Optical fiber-integrated metasurfaces: An emerging platform for multiple optical applications,” Nanomaterials 12(5), 793 (2022). [CrossRef]
12. F. Piccirillo, M. Giaquinto, A. Ricciardi, et al., “(Invited) Miniaturized lenses integrated on optical fibers: Towards a new milestone along the lab-on-fiber technology roadmap,” Results in Optics 6, 100203 (2022). [CrossRef]
13. D. Gonzalez-Hernandez, S. Varapnickas, A. Bertoncini, et al., “Micro-optics 3d printed via multi-photon laser lithography,” Adv. Opt. Mater. 11(1), 2201701 (2023). [CrossRef]
14. W. Hadibrata, H. Wei, S. Krishnaswamy, et al., “Inverse design and 3d printing of a metalens on an optical fiber tip for direct laser lithography,” Nano Lett. 21(6), 2422–2428 (2021). [CrossRef]
15. M. Plidschun, H. Ren, J. Kim, et al., “Ultrahigh numerical aperture meta-fibre for flexible optical trapping,” Light: Sci. Appl. 10(1), 57 (2021). [CrossRef]
16. F. Balli, M. Sultan, and S. K. Lami, “A hybrid achromatic metalens,” Nat. Commun. 11(1), 3892 (2020). [CrossRef]
17. F. Balli, M. A. Sultan, A. Ozdemir, et al., “An ultrabroadband 3d achromatic metalens,” Nanophotonics 10(4), 1259–1264 (2021). [CrossRef]
18. H. Ren, J. Jang, C. Li, et al., “An achromatic metafiber for focusing and imaging across the entire telecommunication range,” Nat. Commun. 13, 4183 (2022). [CrossRef]
19. Z. Wang, Y. Wu, D. Qi, et al., “Two-photon polymerization for fabrication of metalenses for diffraction-limited focusing and high-resolution imaging,” Opt. Laser Technol. 169, 110128 (2024). [CrossRef]
20. F. Ye, M. Pivnenko, H. Huang, et al., “Multi-height metasurface for wavefront manipulation fabricated by direct laser writing lithography,” Nanophotonics 12(17), 3435–3442 (2023). [CrossRef]
21. A. Arbabi, R. M. Briggs, Y. Horie, et al., “Efficient dielectric metasurface collimating lenses for mid-infrared quantum cascade lasers,” Opt. Express 23(26), 33310–33317 (2015). [CrossRef]
22. R. Singh, Y. Nie, M. Gao, et al., “Inverse design of photonic meta-structure for beam collimation in on-chip sensing,” Sci. Rep. 11, 5343 (2021). [CrossRef]
23. C. Zhou, W.-B. Lee, S. Gao, et al., “All-dielectric fiber meta-tip enabling vortex generation and beam collimation for optical interconnect,” Laser Photonics Rev. 15(5), 2000581 (2021). [CrossRef]
24. H. Ye, Q. Sun, Z. Guo, et al., “Theoretical realization of single-mode fiber integrated metalens for beam collimating,” Opt. Express 29(17), 27521–27529 (2021). [CrossRef]
25. F. Cantoni, D. Maher, E. Bosler, et al., “Round-robin testing of commercial two-photon polymerization 3d printers,” Addit. Manuf. 76, 103761 (2023). [CrossRef]
26. “Nanoscribe GmbH: urihttps://www.nanoscribe.de/en/,”.
27. K. Vanmol, S. Tuccio, V. Panapakkam, et al., “Two-photon direct laser writing of beam expansion tapers on single-mode optical fibers,” Opt. Laser Technol. 112, 292–298 (2019). [CrossRef]
28. M. Schmid, D. Ludescher, and H. Giessen, “Optical properties of photoresists for femtosecond 3d printing: refractive index, extinction, luminescence-dose dependence, aging, heat treatment and comparison between 1-photon and 2-photon exposure,” Opt. Mater. Express 9(12), 4564–4577 (2019). [CrossRef]
29. E. H. Waller and G. Von Freymann, “Spatio-temporal proximity characteristics in 3d µ-printing via multi-photon absorption,” Polymers 8(8), 297 (2016). [CrossRef]
30. “International Telecommunication Union 2009 Recommendation G.652 - Characteristics of a single-mode optical fibre and cable,” (2009).
31. K. Vanmol, A. Kandeel, G. Y. Belay, et al., “Two-photon polymerization-based direct laser writing and characterization of micro-lenses for optical interconnect applications,” in 2021 26th Microoptics Conference (MOC), (2021 pp.1–2).
32. K. Vanmol, A. Kyriazis, H. Thienpont, et al., “Design and 3d direct laser writing of micro-optical components to functionalize optical fiber tips,” in Optica Design and Fabrication Congress 2023 (IODC, OFT), (Optica Publishing Group, 2023 p.OM3B.2).
33. X. Zhou, Y. Hou, and J. Lin, “A review on the processing accuracy of two-photon polymerization,” AIP Adv. 5(3), 030701 (2015). [CrossRef]
34. L. J. Jiang, J. H. Campbell, Y. F. Lu, et al., “Direct writing target structures by two-photon polymerization,” Fusion Sci. Technol. 70(2), 295–309 (2016). [CrossRef]
35. S. Schweiger, T. Schulze, S. Schlipf, et al., “Characterization of two-photon-polymerization lithography structures via Raman spectroscopy and nanoindentation,” J. Opt. Microsystems 2(03), 033501 (2022). [CrossRef]
36. T.-H. Kim, J. Kim, and H.-Y. Kim, “Evaporation-driven clustering of microscale pillars and lamellae,” Phys. Fluids 28(2), 022003 (2016). [CrossRef]
37. J. Purtov, A. Verch, P. Rogin, et al., “Improved development procedure to enhance the stability of microstructures created by two-photon polymerization,” Microelectron. Eng. 194, 45–50 (2018). [CrossRef]
38. S. Banerji, M. Meem, A. Majumder, et al., “Imaging with flat optics: metalenses or diffractive lenses?” Optica 6(6), 805–810 (2019). [CrossRef]
