A method for fabricating superlenses on semiconductors using electrochemical nanoimprinting

Abstract

This invention discloses a method for fabricating superlenses by electrochemical nanoimprinting on semiconductors. Based on a target phase distribution, structural design data for the superlens is generated. Using this design data, a metallized nanopillar array mold is fabricated. Then, in an electrolyte environment containing an oxidant, the mold is brought into contact with a semiconductor substrate and pressure is applied. The contact potential generated by the difference in work functions between the two substrates drives a localized electrochemical oxidation-dissolution reaction at the interface, thereby transferring the mold pattern to the surface of the semiconductor substrate with high fidelity, forming a superlens structure composed of a nanocavity array. This invention achieves short-process, reproducible structure transfer while meeting the requirements of superlens structural consistency and nanoscale processing accuracy, providing a feasible solution for the mass production of high-performance planar optical devices for imaging and communication applications.

Description

Technical Field

[0001] This invention belongs to the technical field of micro-nano optical structure fabrication, specifically relating to a method for preparing a superlens by electrochemical nanoimprinting on a semiconductor and the prepared semiconductor superlens structure. Background Technology

[0002] Superlenses are a class of planar optical devices based on metasurfaces. They achieve spatial phase modulation by arranging subwavelength nanostructures on the device surface, thereby enabling focusing, imaging, and wavefront shaping. Compared to traditional refractive lenses, superlenses have the advantages of being thin, lightweight, and easy to integrate, making them suitable for lightweight optical systems and facilitating system-level integration with miniature cameras, AR / VR, on-chip optics, and sensing platforms.

[0003] To achieve good imaging and high-efficiency focusing, precise control of the unit structure size and spatial arrangement of the superlens is essential, posing a significant challenge to the fabrication process. Photolithography, as a mature technology supporting large-scale manufacturing in the modern microelectronics industry, has been widely used in superlens fabrication. Currently, there is considerable research on superlens fabrication based on the most cost-effective deep ultraviolet lithography (such as 248nm KrF lithography), but the feature sizes that can be stably achieved at the device level are typically on the order of hundreds of nanometers, making it difficult to support higher nanometer precision requirements. Extreme ultraviolet (EUV) lithography can improve processing precision to the order of tens of nanometers, but it requires high equipment investment and stringent requirements for process window and yield control, resulting in high manufacturing costs. In contrast, electron beam lithography (EBL) can achieve high-precision processing at the sub-10 nm scale, but its serial writing characteristics lead to low throughput and high cost per unit area. In general, the aforementioned technologies are all patterning processes, which typically require first forming a pattern on photoresist, followed by multiple steps such as pattern transfer and dry etching. This process is lengthy and prone to error accumulation, causing unit structure parameters to deviate from design values, resulting in degraded image quality and reduced focusing efficiency. Therefore, there is an urgent need for a new manufacturing route that can balance nanoscale precision, high consistency, scalability, and cost-effectiveness. Summary of the Invention

[0004] This invention addresses the shortcomings of existing technologies by providing a method for fabricating superlenses by electrochemical nanoimprinting on semiconductors, offering a scalable technical approach for manufacturing large-area, high-fidelity superlenses.

[0005] In a first aspect, the present invention provides a method for fabricating superlenses by electrochemical nanoimprinting on semiconductors, comprising the following steps:

[0006] S1. An optical response database of the superlens unit structure is established through electromagnetic numerical simulation. Based on the target phase distribution calculated with a preset focal length and working wavelength, the corresponding unit structure geometric parameters are matched for each position within the superlens aperture from the optical response database to generate superlens structured design data and layout data.

[0007] S2. Using the layout data as input, a metallized nanopillar array mold is fabricated using micro-nano fabrication technology, wherein the surface of the nanopillars of the metallized nanopillar array mold is provided with a metal catalyst layer;

[0008] S3. The semiconductor substrate and the metallized nanopillar array mold are brought into contact in an electrolyte environment containing an oxidant and an imprint load is applied to cause an electrochemically induced localized oxidation and dissolution reaction at the interface between the metallized nanopillar array and the semiconductor substrate, thereby forming a nanocavity array complementary to the nanopillar array on the surface of the semiconductor substrate, thus forming a superlens structure.

[0009] Optionally, the work function of the metal catalyst layer is not equal to that of the semiconductor substrate, and a contact potential is formed when the metallized nanopillar array mold comes into contact with the semiconductor substrate, thereby inducing a local electrochemical reaction.

[0010] Optionally, in S1, the target phase distribution is calculated according to the preset focal length f and working wavelength λ of the superlens, based on the ideal focusing phase function, and the target phase distribution is subjected to 0~2π phase folding processing.

[0011] Optionally, S1 includes:

[0012] Electromagnetic numerical simulation was used to perform parameter scanning on the geometric parameters of the superlens unit structure, and an optical response database containing the transmission phase and transmittance of unit structures with different geometric parameters at the target wavelength was established.

[0013] Candidate unit structures are screened from the optical response database based on a preset transmittance threshold, and geometric parameters that minimize the phase error between the transmitted phase of the candidate unit structure and the target phase distribution at the corresponding positions are selected.

[0014] Based on the geometric parameters, structured design data and layout data of the superlens are generated, and the nanopillar arrangement of the metallized nanopillar array mold in step S2 is determined by the layout data.

[0015] Furthermore, numerical simulation is used to scan the unit structure parameters of the superlens to obtain the optimal parameter combination that satisfies the phase modulation range. From the optimal parameter combination that satisfies the phase modulation range, the optimal parameters corresponding to the phase distribution map are determined, specifically including:

[0016] A unit periodic model was established using the finite-difference time-domain method: the long axis of the nanostructure was defined as the x-direction, and periodic boundary conditions were applied in the x–y plane to simulate an infinite periodic array; a perfectly matched layer (PML) was set in the z-direction (perpendicular to the metasurface plane) to absorb leaked energy and suppress boundary reflection; a plane wave incident along the +z direction was used as the light source to ensure that the extracted transmission phase corresponds to the anisotropic response of the structure.

[0017] Numerical simulations were performed on the unit structure parameters to select the optimal parameter combination that satisfies the phase modulation range. The parameter combination includes the type of semiconductor material, the depth, period, length, and width of the unit structure, and the wavelength of the incident light. Based on the scanning results, a mapping relationship between the phase and the length L and width W in the optimal parameter combination that satisfies the phase modulation range was constructed to obtain a database of phase and transmittance.

[0018] Optionally, for each discrete unit position (x) within the lens aperture i ,y i Among the candidate cells that meet the transmittance threshold, the phase response ϕ(L,W) and the target phase ϕ are selected. tar (x i ,y i The L / W combination with the smallest difference is used to generate a complete structural size distribution map.

[0019] Optionally, the depth scanning range of the periodic unit structure is 100nm to 2000nm, the period range is 100nm to 4000nm, and the length and width range is 50nm to 5000nm; the wavelength of the incident light is 300nm to 5000nm.

[0020] Optionally, the template preparation error Δ1 is obtained by comparing the design dimensions with the measured dimensions of the mold, and the structure transfer error Δ2 is obtained by comparing the measured dimensions of the imprinted structure with the measured dimensions of the mold. Based on the template preparation error Δ1 and the structure transfer error Δ2, the geometric parameters of the unit structure in the structured design data are pre-compensated and corrected, and the layout data is generated based on the corrected structured design data.

[0021] Optionally, the metallized nanopillar array mold includes a template substrate, a protective layer, an adhesion layer, and a metal catalyst layer arranged in sequence, wherein a mask layer is provided between the top surface of the nanopillars formed by the template substrate and the protective layer; the template substrate is at least one of silicon, glassy carbon, and silicon carbide; the mask layer is at least one of gold, silver, copper, platinum, titanium, aluminum, nickel, and chromium; the protective layer is silicon dioxide; the adhesion layer is at least one of titanium and chromium; and the metal catalyst layer is at least one of platinum, gold, silver, copper, and aluminum.

[0022] Optionally, the fabrication process of the metallized nanopillar array mold is as follows:

[0023] Photoresist is spin-coated on the surface of a template substrate, and the photoresist is patterned at the nanoscale by electron beam exposure according to the unit structure arrangement pattern corresponding to the phase map of the superlens target, so as to obtain a template substrate with patterned photoresist.

[0024] Electron beam evaporation is used to deposit a mask layer metal on its surface, and the metal on the photoresist is removed by a lift-off process, leaving only the mask layer on the template substrate surface;

[0025] Inductively coupled plasma etching is used to etch the area outside the mask layer to form a nanopillar array structure;

[0026] A protective layer is formed on the surface of a nanopillar array by atomic layer deposition, and an adhesion layer and a metal catalyst layer are sequentially deposited on the protective layer by magnetron sputtering to obtain a metallized template.

[0027] Optionally, the electrochemical nanoimprinting process of S3 specifically includes:

[0028] Pretreatment: The semiconductor substrate is cleaned by first soaking in hydrochloric acid to remove the natural oxide layer on the surface, then ultrasonically cleaning with ethanol and rinsing thoroughly with deionized water, and finally drying with nitrogen and setting aside.

[0029] Mold fixing: After heating the electrolytic cell, apply a thin layer of paraffin wax to the center of the bottom, then fix the mold on the paraffin wax and wait for the paraffin wax to solidify;

[0030] Substrate fixing: The pre-treated semiconductor substrate is fixed on the imprint rod of the Z-axis displacement stage with curing adhesive. The stepper motor is driven to gradually bring the substrate closer to the template so that the two come into contact. During the process of curing adhesive gradually curing, the mold and semiconductor substrate can be self-adaptive leveling.

[0031] Etching process: Drive the stepper motor to lift up and add electrolyte to the electrolytic cell. After the solution fully immerses the mold, drive the stepper motor to gradually bring the semiconductor substrate close to the mold so that the two come into contact. Under the set contact pressure, the in-situ corrosion galvanic cell reaction driven by the contact potential is induced at the template-substrate interface formed by the catalyst, so as to achieve local selective etching. After etching, the mold and the substrate are separated so that the substrate surface forms a nanocavity array.

[0032] Optionally, the electrolyte includes an oxidant and a supporting electrolyte; the oxidant is K2S2O8, H2O2, KMnO4, K2Cr2O7, HNO3, or Ag. +At least one of the following, with a concentration of 0.001–5 mol / L; the supporting electrolyte is at least one of hydrochloric acid, sulfuric acid, phosphate, and hydrofluoric acid, with a concentration of 0.01–10 mol / L; the mass ratio of the oxidant to the supporting electrolyte is 1:1 to 1:100.

[0033] Optionally, the imprinting load ranges from 0.1 to 60 N, and the duration is from 5 to 45 min.

[0034] Optionally, the semiconductor substrate material includes gallium phosphide wafers (which may be sulfur-doped gallium phosphide), gallium arsenide wafers, indium phosphide wafers, gallium nitride wafers, or silicon carbide wafers.

[0035] Secondly, the present invention provides a semiconductor superlens optical element prepared by the above-described preparation method.

[0036] The beneficial effects of this invention are as follows:

[0037] (1) Compared with traditional processes, this invention does not require coating, curing and demolding of photoresist or resin, but directly transfers the structure to the surface of functional semiconductor materials, which not only significantly improves processing efficiency, but also effectively reduces the risk of mold damage and structural deformation. It can directly process semiconductors, simplifies the process, and the template can be reused.

[0038] (2) The present invention can meet the requirements of super lens structure consistency and nanoscale processing accuracy. The same mold can be used to imprint multiple times and maintain good stability. It provides a feasible and efficient new technology path for realizing high-throughput and high-precision manufacturing of metasurfaces and mass production of high-performance planar optical devices for imaging and communication applications.

[0039] (3) The superlens fabrication scheme proposed in this invention takes electrochemical nanoimprinting as its core and uses the metal-assisted chemical etching (MACE) mechanism to achieve selective transfer and material removal of nanostructures. The structure fabricated by this method has high sidewall steepness and no burrs at the cavity edges. The fabrication process can be completed under relatively mild process conditions, thereby reducing the potential damage to the integrity of semiconductor lattice and optical properties caused by traditional high-energy ion bombardment or plasma treatment, and achieving high-fidelity replication of nanostructures. Attached Figure Description

[0040] Figure 1 (a) Schematic diagram of the focusing principle of the cavity-type superlens, (b) Schematic diagram of the nano-cavity unit structure, (c) Phase library, (d) Transmittance library, (e) Target phase distribution, (f) Superlens structure generated according to the target phase, (g) Simulated light intensity distribution of the XZ plane of the superlens, (h) Simulated light intensity distribution of the XY plane of the superlens, (i) Simulated PSF curve of the superlens.

[0041] Figure 2A schematic diagram of the unit structure model constructed in FDTD;

[0042] Figure 3 (a) Causes of error, (b) Error compensation for attenuation dimensions, (c) Error between template and design value, (d) Error between template and imprint value, (e) Error between target and imprint.

[0043] Figure 4 (ac) Overall, partial and microscopic images of the superlens template observed under a scanning electron microscope (SEM); (df) Overall, partial and microscopic images of the imprinted structure observed under a scanning electron microscope (SEM).

[0044] Figure 5 A schematic diagram illustrating the principle of electrochemical nanoimprinting.

[0045] Figure 6 (ae) SEM image, XZ plane light intensity distribution, XY plane light intensity distribution, PSF curve and MTF curve obtained from the test of sample 1; (fj) SEM image, XZ plane light intensity distribution, XY plane light intensity distribution, PSF curve and MTF curve obtained from the test of sample 2; (ko) SEM image, XZ plane light intensity distribution, XY plane light intensity distribution, PSF curve and MTF curve obtained from the test of sample 3. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will now be described in further detail with reference to the accompanying drawings.

[0047] In this example, a superlens (200 μm diameter, 1.5 mm focal length, NA≈0.067) with a working wavelength of 650 nm was fabricated on a gallium phosphide (GaP) substrate using electrochemical nanoimprinting (ECNL) technology. Gallium phosphide (GaP) is a III-V group semiconductor that exhibits both high transmittance and high refractive index in the visible light region. The gallium phosphide (GaP) substrate used in this example had a doping concentration of 4.0 × 10⁻⁶. 17 cm -3 The sulfur-doped gallium phosphide wafers were purchased from China Crystal Technologies Co. Prior to processing, the gallium phosphide wafers were rinsed sequentially with hydrochloric acid, ethanol, and deionized water to remove native oxides and surface contaminants. All chemicals used were of analytical grade (China National Pharmaceutical Group), and all aqueous solutions were prepared using deionized water (18.2 MΩ·cm, Milli-Q, Millipore Co.).

[0048] The embodiments employ numerical simulations using Lumerical FDTD Solutions (Finite-Difference Time-Domain Method) for unit structure selection and overall performance verification of the superlens. The dispersive optical parameters of the materials are taken from refractive index data compiled by Edward D. Palik.

[0049] In this embodiment, the imprinted workpiece and mold were characterized using a field emission scanning electron microscope (Gemini SEM 500, ZEISS, Germany). To analyze the cross-sectional morphology of the nanostructure, milling was performed using a focused ion beam (FIB) at a beam current of 20 pA (Orion Nanofab, ZEISS, Germany), followed by imaging using an integrated electron microscopy imaging system. The performance of the imprinted GaP superlens was tested using an optical testing platform including a light source module, a beam expander and collimator system, a linear polarizer, a sample stage, a focal plane receiving screen, and a camera (CCD). A 650 nm laser was used, which, after beam expansion and polarization modulation, was incident perpendicularly on the sample. The focusing effect was recorded using a CCD camera, and the focal length and focusing efficiency were measured.

[0050] In this embodiment, numerical simulation is used to scan the unit structure parameters of the superlens to obtain parameter combinations that can satisfy phase modulation. The optimal parameters for the phase distribution map are then determined from the optimal parameter combinations within the phase modulation range. Specifically, this includes:

[0051] Figure 1 (a) illustrates a focusing schematic of the superlens. To achieve focusing, a spatial phase distribution covering 0–2π needs to be achieved within the device aperture to reconstruct the incident plane wave into a converging wavefront. In this embodiment, the superlens is composed of periodically arranged nanocavity units, such as… Figure 1 As shown in (b), the device consists of periodically arranged nanocavity units, with cavity depth denoted as H, cavity period denoted as P, cavity length denoted as L, and cavity width denoted as W. In this embodiment, the cavity depth H = 600 nm and the period P = 2 μm are fixed. The cavity length L and cavity width W are the main design degrees of freedom. By adjusting different combinations of L / W to change the transmission phase response of the unit, phase modulation in the range of 0 to 2π can be achieved. The design goal is to achieve phase coverage of 0 to 2π at the operating wavelength of 650 nm while maintaining a high transmittance (e.g., not less than a transmittance threshold of 60%).

[0052] Electromagnetic numerical simulations were performed using Lumerical FDTD Solutions, with material dispersion parameters derived from Edward D. Palik's refractive index data. The long axis of the nanostructure was defined as the x-direction, and periodic boundary conditions were applied in the x–y plane to simulate an infinite periodic array. A perfectly matched layer (PML) was placed in the z-direction (perpendicular to the metasurface plane) to absorb leaked energy and suppress boundary reflections. A linearly polarized plane wave (x-polarized) incident along the +z direction was used as the light source to extract the transmission phase and transmittance of the unit structure under a given polarization. A transmission monitor was placed below the structure to obtain the transmission complex amplitude, thereby calculating the transmission phase and transmittance. In this paper, T... x T represents the co-polarized transmittance (or equivalent transmittance index) under x-polarized incident conditions, used for candidate cell screening; if other definitions are used (such as total transmittance / cross-polarized transmittance), T can also be used. x Replace with the corresponding transmittance definition.

[0053] To establish the phase-structure mapping relationship, FDTD was used to perform parameter scanning of L and W in the range of 500–1800 nm (step size set to 20 nm) to obtain the transmission phase and transmittance corresponding to each combination of geometric parameters, forming a phase database and a transmittance database, such as... Figure 1 As shown in (c)-(d), the results demonstrate that scanning within this range can achieve full 2π phase coverage, with transmittance exceeding 60% (and sometimes exceeding 70%) under most parameter combinations. The following criteria were used to select units for subsequent lens stitching: prioritizing structural combinations with continuous phase coverage and minimal phase error, while meeting the transmittance threshold, thus achieving a balance between efficiency and phase approximation accuracy.

[0054] Subsequently, the target phase distribution is calculated using an analytical phase function, and structural matching and layout generation are completed by combining the phase database and the transmittance database, thus encoding the superlens structure onto the metasurface. The target phase of the superlens is calculated using an ideal focusing phase function (corresponding to a focal length f = 1500 μm in this embodiment). For an ideal focusing superlens with a focal length f, its target phase distribution can be expressed as:

[0055]

[0056] Where ϕ(x,y) is the target phase of the superlens at the plane coordinates (x,y); x and y represent the lateral coordinates in the superlens plane, respectively; f is the design focal length, λ is the working wavelength, and (mod 2π) indicates that the phase result is moduloed by a 2π period (folded into a 2π interval), and its target phase distribution is as follows: Figure 1 As shown in (e), a lookup-table "phase-structure size" mapping is used to complete the unit matching: for each discrete unit position (x) within the lens aperture...i ,y i ), search the phase library for values ​​that satisfy the transmittance threshold (e.g., T). x Candidate structures with a phase error greater than 0.6 were selected, with the minimum phase error used as the selection criterion.

[0057]

[0058] Where J(L,W) represents the phase matching error corresponding to the candidate nanocavity size, and L and W are the cavity length and width of the nanocavity, respectively; (x i ,y i ) represents the position of the i-th discrete unit within the lens aperture; ϕ(x) i ,y i ) represents the target phase at that location; ϕ LW The phase response of a unit cell of size (L, W) in the phase library at the operating wavelength is given; `wrap` is used to fold the phase difference to [−π, π]. This establishes a correspondence between the target phase and the nanocavity geometric parameters (L, W), allowing the phase distribution data to be directly converted into the nanocavity size parameter distribution, thus completing the spatial encoding of the superlens unit structure. Based on this mapping relationship, a nanocavity array is arranged with a 2 μm period along the x–y direction on a GaP substrate, resulting in... Figure 1 (f) shows the superlens structure design.

[0059] This embodiment uses the propagation phase mechanism to achieve phase modulation. The phase response of the anisotropic unit is determined by its geometric parameters (L, W). Therefore, by adjusting L and W, a phase coverage of 0~2π is achieved, and the spatial phase encoding of the superlens is completed while ensuring high transmission efficiency, resulting in the final superlens structure layout.

[0060] Considering the relatively long focal length (f=1500μm) in this embodiment, to reduce the computational burden of direct large-volume three-dimensional far-field simulation and improve numerical stability, a near-field monitor is set 1µm above the top surface of the nanostructure to record the transmission complex amplitude field (amplitude and phase). Subsequently, the near field is propagated in free space to the focal plane and the corresponding propagation distance using the angular spectrum propagation method, and key imaging parameters such as the light intensity distribution of the XY plane of the focal plane, the longitudinal focusing process of the XZ plane, the phase distribution, and the actual focal length are reconstructed.

[0061] The focusing performance of the superlens design was verified through numerical simulation. A complete superlens model was established, and full-field simulations were performed using PML boundaries (full PML) in the x, y, and z directions. Combined with the aforementioned near-field and far-field joint analysis, results such as focal spot, PSF, and focal length deviation were obtained, and the focusing performance and image quality of the superlens were evaluated accordingly. Focusing efficiency and other indicators were obtained by normalizing the integrated light intensity within a specific region of the focal plane and the total transmitted light intensity within the lens aperture range. Post-processing of key data, curve statistics, and image analysis were performed using MATLAB.

[0062] Figure 1 (g) gives the light intensity distribution of the XZ cross section. The incident beam gradually converges along the propagation direction and reaches the maximum intensity at z≈1500 μm. The simulated focal length is consistent with the design value f=1500 μm. Figure 1 (h) represents the XY light intensity distribution at the focal plane (z=f). The FWHM of the light intensity distribution on this plane is used to characterize the spot size, and the spot size is 4.4μm. A main peak with concentrated intensity is formed at the focal point. The whole is paraxially symmetric and has low background energy. The weak annular side lobes around the main peak are mainly caused by finite aperture diffraction and residual phase errors introduced by phase discretization and lookup table matching. Figure 1 (i) shows a PSF profile with a clear main lobe and controlled side lobes, with the main lobe width approaching the diffraction limit. Based on the focal plane intensity integral, the focusing efficiency of this design at a wavelength of 650 nm is calculated to be 63%.

[0063] To further quantify and correct potential dimensional deviations introduced during structure transfer, we analyze the sources of error and establish a compensation strategy. We categorize dimensional deviations into two types: template preparation error Δ1 originates from systematic deviations introduced during the EBL exposure, development, and pattern transfer stages; structure transfer error Δ2 arises from the contact potential formed when the metal mold and GaP come into contact in the electrolyte. The interfacial electric field causes hole accumulation in the contact / near-contact region, triggering dimensional shifts caused by nanoscale localized oxidation and dissolution. Figure 3 As shown in (a).

[0064] Therefore, when using layout for mold design, a pre-compensation strategy is introduced to apply an equivalent correction of Δ1 + Δ2 to the dimensions of each unit, as shown in Figure (b). Here, Δ1 is obtained by comparing the designed dimensions with the measured dimensions of the mold, and Δ2 is obtained by comparing the measured dimensions of the embossing structure with the measured dimensions of the mold (statistical methods can be used to calculate L and W separately). Based on Δ1 and Δ2, pre-compensation corrections are applied to the key dimensions of each unit in the layout, for example, using equivalent compensation:

[0065] L′=L−(Δ1+Δ2), W′=W−(Δ1+Δ2)

[0066] Figure 3(c–e) provides the error statistics: Figure 3 (c) shows the fabrication error distribution of the template preparation relative to the design value (i.e., the statistics of Δ1, approximately 140 nm). Figure 3 (d) shows the distribution of processing errors between the template and the imprinted structure (i.e., the statistics of Δ2, approximately 60 nm). Figure 3 (e) shows the residual error distribution of the imprinted structure relative to the design target after Δ1+Δ2 compensation. The results show that pre-compensation can significantly reduce the systematic dimensional offset of the final device, with the residual error fluctuating near zero and without significant bias. Meanwhile, to avoid structural overlap and phase crosstalk at the array boundary, some small periodic units are selectively removed at the edge of the superlens to ensure boundary continuity and local phase consistency.

[0067] The fabrication process of the metallized nanopillar array mold is as follows: First, a polymethyl methacrylate (PMMA) photoresist layer with a thickness of 300 nm is spin-coated onto a silicon substrate. Nanoscale patterning is then performed using an electron beam lithography system (Sigma300 ELPHY Quantum, ZEISS, Germany) to form a patterned photoresist corresponding to the compensated layout. The exposure dose is 370 μC / cm² to precisely define the size and period of the nanopillars. After development, a 30 nm thick chromium (Cr) layer is deposited on the patterned photoresist surface using an electron beam evaporation system (DE400, DETechnology Inc., USA). Subsequently, the PMMA photoresist and the Cr layer above it are removed using an acetone stripping method, forming a complementary Cr mask layer on the silicon substrate for subsequent inductively coupled plasma (ICP) etching. The exposed silicon substrate is then etched using SF6 / CF4 plasma (HAASRODE-E200A, Jiangsu Luwen Instrument Co., Ltd., China) to form the nanopillar array. Finally, a 5 nm thick silicon dioxide (SiO2) protective layer was deposited using an atomic layer deposition system (R-200 Adv Thermal LC, Picosun Oy, Finland), followed by the sequential deposition of a 5 nm thick titanium (Ti) adhesion layer and a 50 nm thick platinum (Pt) catalyst layer using a magnetron sputtering system (Discovery 635, Denton Vacuum, USA).

[0068] like Figure 4 As shown in (a–c), the prepared metallized nanopillar array mold has clear boundaries and steep sidewalls, and its overall morphology is consistent with the design, providing a reliable imprint for subsequent ECNL structure transfer.

[0069] The ECNL process is as follows: The GaP substrate undergoes surface cleaning, including immersion in hydrochloric acid to remove the natural oxide layer, followed by ultrasonic cleaning with ethanol and thorough rinsing with deionized water. Finally, it is dried with nitrogen and set aside. After heating the electrolytic cell, a thin layer of paraffin wax is applied to the center of the bottom. The mold is then fixed onto the paraffin wax and allowed to solidify. The pre-treated GaP substrate is fixed to the imprinting rod end of the z-axis displacement stage using a curing adhesive. A stepper motor is driven to gradually bring the substrate closer to the template until they contact each other. During the gradual curing of the adhesive, adaptive leveling of the metallization template and the GaP substrate is achieved. Subsequently, the stepper motor is driven to lift the substrate, and an electrolyte solution of 1.0 mol / L H2SO4 and 1.6 g / L KMnO4 is added to the electrolytic cell. After the solution fully immerses the mold, the stepper motor is driven to gradually bring the semiconductor substrate closer to the mold until they contact each other. To maintain a constant contact pressure, the system uses a force sensor (LSB200, FUTEK, USA) with integrated feedback control circuitry for real-time monitoring and adjustment. In this example, the imprinting load is 10 N. Figure 5 As shown, when an n-type doped GaP wafer (W f =4.34eV) and Pt mold (W f When the Pt and GaP layers come into contact (electrode work function unequal at 5.36 eV), charge transfer spontaneously occurs at the interface due to the unequal electron work function. When the interface reaches electrostatic equilibrium, the Pt surface becomes negatively charged and the GaP surface becomes positively charged. In an etching solution containing an oxidant, the Pt surface undergoes a reduction reaction, and the GaP wafer undergoes self-etching along the Pt / GaP / electrolyte three-phase interface, generating a three-dimensional micro / nano structure complementary to the metal layer. After maintaining contact and etching for 20 minutes, the mold pattern is faithfully replicated onto the GaP surface. After etching, the mold is separated from the substrate, cleaned, and dried. Ten parallel imprinting operations are then performed to obtain different batches of superlens structures.

[0070] The imprinted structure was characterized using SEM, such as... Figure 4 As shown in (d–f), the imprinted structure has clear boundaries and complete morphology. The nanocavities can maintain the contour features defined by the template well, and no obvious overall collapse or large-area over-etching was observed.

[0071] To evaluate imaging performance, an optical testing platform was constructed, including a light source module, a beam expander and collimator system, a linear polarizer, a sample stage, a focal plane receiving screen, and a camera (CCD). A 650nm laser was used, which, after beam expansion and polarization modulation, was incident perpendicularly on the sample. The focal plane and axial light intensity distribution were recorded by the CCD camera. The axial light intensity was scanned and recorded along the z-direction, and the z-value corresponding to the position of maximum light intensity was taken as the measured focal length. The intensity cutoff line was extracted on the focal plane, and the full width at half maximum (FWHM) was used to characterize the focal spot size. The PSF was obtained from the focal plane light intensity distribution, and the MTF curve was calculated based on it to evaluate the imaging frequency response. The integrated light intensity within a preset area of ​​the focal plane (e.g., a circular area centered on the focal point) was divided by the total transmitted light intensity within the lens aperture range to obtain the focusing efficiency. In this embodiment, the measured focusing efficiencies were 47%, 43%, and 44%, respectively.

[0072] Figure 6 The test results of three superlens samples prepared by different imprinting cycles are presented, where Sample 1, Sample 2, and Sample 3 correspond to devices obtained by 1, 5, and 10 imprinting cycles, respectively. For each sample, its SEM morphology, XZ plane light intensity distribution, XY plane light intensity distribution, PSF curve, and MTF curve are presented sequentially. From the SEM images ( Figure 6 As can be seen from (a), (f), and (k), all three samples maintained a complete annular superlens pattern and clear nanocavity structure boundaries, with no systematic collapse or significant degradation observed due to multiple imprintings. The corresponding XZ plane light intensity distribution is as follows: Figure 6 As shown in (b), (g), and (l), the measured focal lengths of the three groups of samples are approximately 1482 μm, 1450 μm, and 1530 μm, respectively, all close to the simulated design value of 1500 μm (relative deviations of approximately −1.2%, −3.3%, and +2.0%). The corresponding XY plane light intensity distribution ( Figure 6 (c), (h), and (m) show that all three samples formed symmetrical and concentrated focused spots at the designed focal plane position, with sizes of 4.48 μm, 4.86 μm, and 4.85 μm, respectively. This is consistent with simulation results ( Figure 1 Compared to (h), FWHM≈4.40 μm), the experimental spot only showed a slight widening, and the overall result still matched well with the simulated value, indicating that the phase modulation capability and focusing performance of the device did not change significantly after multiple imprintings. Similarly, it can be seen that the experimentally measured PSF (h) Figure 6 (d), (i), (n)) show good consistency with the simulation results in terms of main lobe position and half-width at half-maximum, with no obvious broadening or distortion; the MTF curves calculated accordingly ( Figure 6 (e), (j), and (o) exhibit similar attenuation trends across the entire spatial frequency range.

[0073] In summary, the gallium phosphide (GaP) superlens of this embodiment has a measured focal length that deviates from the design value by less than 3.3%, a focusing efficiency of 47%, a symmetrical focal spot, and a point spread function main lobe width that is close to the diffraction limit. It meets the requirements of superlens structure consistency and nanoscale processing precision while achieving short-process, reproducible structure transfer.

[0074] The above embodiments are only used to further illustrate a method for preparing a superlens by electrochemical nanoimprinting on a semiconductor according to the present invention. However, the present invention is not limited to the embodiments. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention shall fall within the protection scope of the technical solution of the present invention.

Claims

1. A method for fabricating superlenses by electrochemical nanoimprinting on semiconductors, characterized in that, Includes the following steps: S1. An optical response database of the superlens unit structure is established through electromagnetic numerical simulation. Based on the target phase distribution calculated with a preset focal length and working wavelength, the corresponding unit structure geometric parameters are matched for each position within the superlens aperture from the optical response database to generate superlens structured design data and layout data. S2. Using the layout data as input, a metallized nanopillar array mold is fabricated using micro-nano fabrication technology, wherein the surface of the nanopillars of the metallized nanopillar array mold is provided with a metal catalyst layer; S3. The semiconductor substrate and the metallized nanopillar array mold are brought into contact in an electrolyte environment containing an oxidant and an imprint load is applied to cause an electrochemically induced localized oxidation and dissolution reaction at the interface between the metallized nanopillar array and the semiconductor substrate, thereby forming a nanocavity array complementary to the nanopillar array on the surface of the semiconductor substrate, thus forming a superlens structure.

2. The method for fabricating superlenses on semiconductors using electrochemical nanoimprinting according to claim 1, characterized in that: The work function of the metal catalyst layer is not equal to that of the semiconductor substrate. When the metallized nanopillar array mold comes into contact with the semiconductor substrate, a contact potential is formed, thereby inducing a local electrochemical reaction.

3. The method for fabricating superlenses on semiconductors using electrochemical nanoimprinting according to claim 1, characterized in that: In S1, the target phase distribution is calculated according to the preset focal length f and working wavelength λ of the superlens, based on the ideal focusing phase function, and the target phase distribution is subjected to 0~2π phase folding processing.

4. The method for fabricating a superlens on a semiconductor by electrochemical nanoimprinting according to claim 1, characterized in that, S1 includes: Electromagnetic numerical simulation was used to perform parameter scanning on the geometric parameters of the superlens unit structure, and an optical response database containing the transmission phase and transmittance of unit structures with different geometric parameters at the target wavelength was established. Candidate unit structures are screened from the optical response database based on a preset transmittance threshold, and geometric parameters that minimize the phase error between the transmitted phase of the candidate unit structure and the target phase distribution at the corresponding positions are selected. Based on the geometric parameters, structured design data and layout data of the superlens are generated, and the nanopillar arrangement of the metallized nanopillar array mold in step S2 is determined by the layout data.

5. The method for fabricating a superlens on a semiconductor by electrochemical nanoimprinting according to claim 4, characterized in that: The template preparation error Δ1 is obtained by comparing the design dimensions with the measured dimensions of the mold, and the structure transfer error Δ2 is obtained by comparing the measured dimensions of the imprinted structure with the measured dimensions of the mold. Based on the template preparation error Δ1 and the structure transfer error Δ2, the geometric parameters of the unit structure in the structured design data are pre-compensated and corrected, and the layout data is generated based on the corrected structured design data.

6. The method for fabricating a superlens on a semiconductor by electrochemical nanoimprinting according to claim 1, characterized in that, The metallized nanopillar array mold includes a template substrate, a protective layer, an adhesive layer, and a metal catalyst layer arranged in sequence. A mask layer is provided between the top surface of the nanopillars formed by the template substrate and the protective layer. The template substrate is at least one of silicon, glassy carbon, and silicon carbide. The mask layer is at least one of gold, silver, copper, platinum, titanium, aluminum, nickel, and chromium. The protective layer is silicon dioxide. The adhesive layer is at least one of titanium and chromium. The metal catalyst layer is at least one of platinum, gold, silver, copper, and aluminum.

7. The method for preparing semiconductor superlenses by electrochemical nanoimprinting according to claim 1, characterized in that: The semiconductor substrate is a gallium phosphide wafer, a gallium arsenide wafer, an indium phosphide wafer, a gallium nitride wafer, or a silicon carbide wafer.

8. The method for fabricating a superlens on a semiconductor by electrochemical nanoimprinting according to claim 1, characterized in that: In S3, the electrolyte includes an oxidant and a supporting electrolyte; the oxidant is K2S2O8, H2O2, KMnO4, K2Cr2O7, HNO3, or Ag. + At least one of the following, with a concentration of 0.001–5 mol / L; the supporting electrolyte is at least one of hydrochloric acid, sulfuric acid, phosphate, and hydrofluoric acid, with a concentration of 0.01–10 mol / L; the mass ratio of the oxidant to the supporting electrolyte is 1:1 to 1:

100.

9. The method for fabricating a superlens on a semiconductor by electrochemical nanoimprinting according to claim 1, characterized in that: In S3, the imprinting load ranges from 0.1 to 60 N, and the duration is from 5 to 45 min.

10. A semiconductor superlens structure, which is prepared by the method of preparing a semiconductor superlens by electrochemical nanoimprinting according to any one of claims 1-9.

Images