Scanning Probe Photonic Nanojet Lithography - ACS Publications

This is an open access article published under an ACS AuthorChoice License, which permits copying and redistribution of the article or any adaptations for non-commercial purposes.

Research Article www.acsami.org

Scanning Probe Photonic Nanojet Lithography Andrea Jacassi, Francesco Tantussi, Michele Dipalo, Claudio Biagini, Nicolò Maccaferri, Angelo Bozzola, and Francesco De Angelis* Istituto Italiano di Tecnologia, Via Morego 30, Genova, 16163, Italy S Supporting Information *

ABSTRACT: The use of nano/microspheres or beads for optical nanolithography is a consolidated technique for achieving subwavelength structures using a cost-effective approach; this method exploits the capability of the beads to focus electromagnetic waves into subwavelength beams called photonic nanojets, which are used to expose the photoresist on which the beads are placed. However, this technique has only been used to produce regular patterns based on the spatial arrangement of the beads on the substrate, thus considerably limiting the pool of applications. Here, we present a novel microsphere-based optical lithography technique that offers high subwavelength resolution and the possibility of generating any arbitrary pattern. The presented method consists of a single microsphere embedded in an AFM cantilever, which can be controlled using the AFM motors to write arbitrary patterns with subwavelength resolution (down to 290 nm with a 405 nm laser). The performance of the proposed technique can compete with those of commercial high-resolution standard instruments, with the advantage of a one-order-of-magnitude reduction in costs. This approach paves the way for direct integration of cost-effective, high-resolution optical lithography capabilities into several existing AFM systems. KEYWORDS: photonics, optical lithography, nanojet, microbead, nanofabrication

■

INTRODUCTION Conventional optical lithography is the standard and most common technique in the semiconductor industry and the microfabrication research field.1 Although several approaches have recently been developed to increase the resolution of optical lithography,2−9 the half-wavelength limitation is still a critical obstacle to push this technology into the nanoscale.10−12 Most of the ultrahigh-resolution techniques use light sources with wavelengths down to the deep UV spectral region.13 However, the use of light in the visible range would still be advantageous for many applications in which highly energetic photons are deleterious for the materials. Moreover, deep UV lithography approaches are quite expensive, and modern research laboratories require inexpensive and more affordable instruments. The optics and light sources are often the most expensive components in a setup, making their acquisition challenging for most laboratories. Thus, low-cost equipment with highly enhanced subdiffraction limit capabilities could be used for applications in a wide range of scientific and technological fields. In addition to deep UV light, other techniques have been proposed that use alternative energy sources for exposing photoresists with high resolution. For example, nanonozzles emitting plasmajets have been used for producing structures with resolution of few hundreds of nanometers.14−16 An interesting approach that leads to a subwavelength focusing of light with a simple and cost-effective solution makes use of socalled photonic nanojets.17 Photonic nanojets are narrow, highly © 2017 American Chemical Society

intensive electromagnetic beams that propagate in the free-space or in a material from the bottom-surface of a lossless dielectric microsphere illuminated on the top side by a plane wave with a wavelength shorter than its diameter. Nanojets present very appealing features: they are nonevanescent propagating beams with a fwhm with a subwavelength spatial resolution that can reach ∼ λ/3n.18 Moreover, they represent a nonresonant phenomenon and can appear with a wide range of sphere diameters from ∼2λ to ∼40λ, producing high field enhancement within the focal volume.19 Light nanojets by microsphere focusing can perform a low-cost and high-throughput process to generate two-dimensional periodic arrays with subwavelength features.20 This lithography technique is called bead lithography and is a simple and economical technique that exploits a single layer of densely packed spherical nanolenses deposited on the resist surface. The self-assembled planar array of beads is used as optical lenses exploiting the high-intensity enhancement of the jet to expose the photoresist in the spheres’ focus volume.21 Despite the high resolution obtained with this technique, its lithography capabilities are limited to regular patterns generated by the arrangement of the nanospheres on the substrate.22 A suitable Received: July 12, 2017 Accepted: August 30, 2017 Published: August 30, 2017 32386

DOI: 10.1021/acsami.7b10145 ACS Appl. Mater. Interfaces 2017, 9, 32386−32393

Research Article

ACS Applied Materials & Interfaces approach to exploit bead lithography to generate arbitrary patterns is still lacking. Recently, we showed that nanojets produced using microbeads can be exploited for high-resolution microscopy by integrating the microbead into the cantilever of an AFM system.23 This configuration allows for enhancement of the optical performance of microscopes equipped with a low numerical aperture (NA) objective. Microspheres embedded on cantilevers can act as portable high-NA nanolenses that are capable of extending the response of a low-NA objective to higher spatial frequencies.23 In this work, we show that low-cost systems for high-resolution optical lithography can be achieved by combining scanning probe technologies with microbead lithography. Such a configuration in which an AFM cantilever is exploited for micro and nanopattering has been used in the past for controlled deposition of proteins24 and for Dip-pen nanolithography.25 Our approach relies on the integration of a microsphere into a standard tipless cantilever of an AFM system, as shown in Figure 1a. The AFM system can be used to move the cantilever as a brush while the microbead writes on the optical resist, similarly to what was already shown in literature for scanning Raman imaging.26 The fabrication method and the bead/cantilever configuration are similar to those introduced previously for high-resolution microscopy.23 In the present work, we exploit such cost-effective system as a high-resolution maskless lithography tool that offers several advantages with respect to much more expensive alternatives. We show that the proposed method retains the advantages of the bead lithography technique and offers the possibility of producing arbitrary patterns that are already implemented in the AFM software and controllers. This approach provides a high degree of freedom maintaining lower cost than other standard maskless high-resolution lithographic systems. We experimentally demonstrate that our technique can produce regular and arbitrary patterns with a subwavelength resolution. The proposed method has great potential due to its easy implementation using existing AFM systems that are widely diffused and can be quickly adapted in many laboratories.

■

RESULT AND DISCUSSION 1. Simulation Results. The intensity and the geometry of a nanojet are highly dependent on the shape of the microsphere and on the refractive index difference between the bead and its surrounding medium. To maximize the lithographic performance, the ideal nanojet should present a high-intensity field in a very narrow and long nanojet; this allows for obtaining a high resolution with suitable exposure times and exposing resist layers a few microns thick. Thus, beads with varying radius (R) and fixed refractive index (n) or varying n and fixed R were simulated to determine the best configuration for the nanojet size and depth. A plane wave with a wavelength of 405 nm is shined onto a microsphere inserted into a 2.7 μm thick silicon layer; the sphere is in contact with a photoresist with n = 1.5; the surrounding medium is air (n = 1). Figure 2a−c show the nanojet geometry for a microbead with a 2.35 μm radius and three refractive index values (1.26, 1.46, 1.66). When the refractive index of the bead increases, the nanojet becomes narrower, improving the spatial resolution. However, the focus of the nanojet gets closer to the microbead surface, and its length is considerably reduced. The nanojet length, defined as the nanojet size along the Y direction with intensity above the fwhm measured at the maximum of the nanojets field profile along the X direction, ranges from 5 to 0.5 μm. The Figure 2g depicts the nanojet profile along the Y

Figure 1. (a) Schematic illustration of a cantilever with a bead placed inside the circular rim. The bead is in contact with a substrate and shined with a beam. (b) SEM image of a cantilever with a bead inserted inside the rim. The inset is an optical image taken with an inverted microscope during laser illumination of the microbead. The nanojet is clearly visible as a red dot in the middle of the bead. A magnified image, on the right side, shows the nanojet with the optical microscope resolution (600 nm). The white contours in the insets are artificially superimposed on the optical image for clarity.

direction for various refractive indexes of the microbead; the nanojets at the maximum of the field profiles along the X direction are shown in Figure 2h. Figure 2d−f show how the nanojet profile changes when the radius of the microbead is varied (2, 3.5, and 4.5 μm) with a fixed refractive index of 1.46. Both the length and width of the nanojet are reduced when the radius of the bead is decreased. The nanojet profiles along the Y direction for several radii are presented in Figure 2i, and the profiles along the X direction are shown in Figure 2j. When the radius is below 1 μm, the bead does not produce a noticeable nanojet. Our simulations confirm the expected behavior for dielectric microspheres with different n and R.27 32387


Research Article

ACS Applied Materials & Interfaces

Figure 2. Electric field intensity maps for a microbead with R = 2.35 μm and n = (a) 1.66, (b) 1.46, (c) 1.26. (d−f) Electric field intensity maps for microbeads with three different radii (2, 3.5, and 4.5 μm) and a refractive index of 1.46. Colors have been superimposed on the maps to better identify the cantilever (gray color) and the resist (purple color). (g) Intensity profile plotted along the y-axis (R = 2.35 μm). (h) Intensity profile along the x-axis positioned at the field maximum (R = 2.35 μm). (i,j) Intensity profiles plotted along the y- and x-axes, respectively, for several bead radii at a fixed refractive index of 1.46. (k,l) fwhm at the focus spot and Depth of field as a function of the bead refractive index and the bead radius, respectively. The black circles represent the parameters selected for the experiments.

Figure 3. (a) Trend in the pattern width (Y-axis) as a function of the dose (X-axis). The error bars on the sizes are the standard deviation from the average. The error bars on the dose are calculated assuming 5% variability of the incoming power. (b) SEM image of a pattern with the smallest width achieved, corresponding to a dose of 1.26 × 10−2 J m−1. (c) Cross-sectional view produced by Focus Ion Beam milling of a pattern. The resist slope on the edges is clearly visible.

source. Conversely, the full exposure of thick resist layers requires a long depth of field, which is achievable with a low refractive index and a large radius. Here, we focus on a

The nanojet footprint decreases when a bead with a high refractive index and a small radius is used, below the wavelength dimension, offering a good approach as an ultrahigh-resolution 32388


Research Article

ACS Applied Materials & Interfaces compromise between n and R to obtain a high resolution that preserves a suitable depth of field. Because the nanojet footprint of the bead (R = 2.35 μm, n = 1.46) illuminated with a wavelength of 405 nm is approximately 250 nm and increases only by 20% when n is increased to 1.9, we decided to use silica beads with R = 2.35 μm and n = 1.47, which are low-cost and easily available. Smaller beads with radius down to 1 um have been also tried; however, practical difficulties for embedding small beads into the cantilever prevented us obtaining experimental results with such beads. The selected bead is marked with a black circle in Figure 2k,l, in which the fwhm and the depth of field are plotted as functions of n and R, respectively. 2. Direct Lithography. To investigate the minimum achievable opening size in direct lithography, we performed experiments varying the exposure dose and measuring the resulting aperture in the S1813 photoresist after development. These experiments were performed with the configuration described in detail in the Methods section and sketched in Figure 6. Briefly, a glass coverslip is spin-coated with S1813 photoresist and is exposed with a 405 nm laser impinging on the microbead embedded in the AFM cantilever. The cantilever is positioned so as to have the microbead in contact with the substrate surface and the resulting photonic nanojet immersed in the photoresist. The cantilever is then moved over the substrate by the AFM pattern generator to expose the desired pattern. According to the features of the selected photoresist, doses between 1.01 × 10−2 J m−1 and 1.64 × 10−2 J m−1 were used. The pattern used for exposure was an array of 5 stripes with a length L = 7 μm; for each dose, the width of the stripe was measured by SEM imaging. A representative pattern after resist development is shown in Figure 3b; the pitch between the stripes is 1.5 μm. The resist opening size as a function of the dose is shown in Figure 3a. The error bars on the size measurements are the standard deviation from the average and the error bars on the dose amounts are calculated assuming a 5% variability in the incoming power as declared by the laser data sheet. As shown in the graph, the line width−dose relation agrees well with the expected behavior, in which a higher dose leads to larger lines. The smallest achievable structure, obtained with a dose of 1.26 × 10−2 J m−1, was approximately 290 nm wide, well below the laser wavelength of 405 nm used for the exposure. Lower doses did not result in proper development of the photoresist. To compare the achieved result with the state of the art on optical nanojets, we report resolutions of ∼100 nm obtained by Kim et al.20 and of ∼300 nm obtained by Wu et al.,21 using respectively polystyrene and silica beads with 400 and 405 nm wavelength lamps. However, in these works, the beads were organized as self-assembled arrays directly in contact onto the target substrate, acting as lenses to expose subwavelength regions of the underlying photoresist; the resulting pattern of such configuration was a regular arrangement of disks in a hexagonal frame where the disk diameter could be controlled by varying the exposure time, whereas the lattice period could be controlled by different bead diameters. The target application of these approaches is therefore different from that of our method, which is the cost-effective writing of highresolution arbitrary patterns. Once the best exposure dose was determined, it was used to study the smallest achievable pitch between two stripes. Three different pitches, measured center-to-center from two adjacent stripes, were tested: 1.5 μm, 1 and 0.6 μm. The first two pitches have a good edge shape and narrow width. The 0.6 μm pitch pattern presents good lithographic lines spaced 300 nm apart but the shape and the edges of the resist in the gap between the lines

present irregularities that occurred during the developing step. To confirm the general-purpose potential of the proposed methodology, we repeated the experiments using a second commercial photoresist, AZ-5214 diluted 1:4. The obtained results reflect those achieved with S1813 in terms of maximum resolution and exposure time. To evaluate experimentally the depth of field of the nanojet, we did a further experiment in which we varied the bead/substrate distance while performing the exposure of a 15 μm long line with the optimal dose (sketch in top panel in Figure 4). During the

Figure 4. Top panel is a sketch of the cantilever during an out of plane scanning. Bottom panel represents the SEM image of two lines of an outof-plane scan.

line scanning from right to left, we moved the Z position of the cantilever upward with an angle of 8.5°. The exposure starts with the bead in contact with the substrate (right) and ends with a bead/substrate distance of approximately 1.5 μm (left). To ensure contact, the starting point is set by applying a pressure between the bead and the resist. Once the scan starts, the pressure decreases and at a certain point the cantilever loses the contact with the sample. We call “Hard Contact” the zone where the bead touches the resist ((zone A) in Figure 4), “Soft Contact” the zone where the cantilever is not applying any pressure on the sample ((zone B) in Figure 4), and “Detachment Mode” the part where the cantilever starts rising in the out of plane direction. The SEM image at the bottom of Figure 4 shows the result of such exposure after development. The exposed lines are perfectly developed on the right side where the nanojet is entirely immersed in the photoresist. Moving toward the left, we notice that the development quality gradually degrades, indicating that the nanojet is not completely exposing the full thickness of the photoresist. The left edges of the lines are barely visible in the SEM image, meaning that the nanojet produced by the bead is barely impinging on the photoresist top surface when the bead/ substrate distance is approximately 1.5 μm. This result is in agreement with the depth of field dimension obtained by the simulations. A useful parameter to evaluate the performance of our lithography approach is the scan speed efficiency, defined as the ratio between the area exposure speed and the laser density power (V/σP) used. High scan speed efficiency indicates that 32389


Research Article

ACS Applied Materials & Interfaces large substrate areas can be exposed with low laser power densities in a short time. Considering the properties of the resist and the geometry of the microbead system, the dependence of the scan speed upon the laser power is

D · SB V = σP ρ

(1)

where σP is the laser density power, D is the spatial resolution of the jet, SB is the surface of the bead impinged by the laser (34.7 μm2), ρ is the energy density needed to expose the resist in a line with a lateral size of D (D = 290 nm with ρ = 1.26 × 10−2 J m−1). From this equation, it is apparent that the writing speed is directly dependent on the scan velocity and the laser power. For a system that aims for high efficiency with low cost, the most crucial parameter is the laser power because high power sources and related optics in the UV region are quite expensive. As shown above, the ratio D/ρ is constant and represents the slope of the fit in the plot in Figure 3a with m = 67.03 μm2 μJ−1. Thus, the scan speed efficiency can be written as V = m ·SB = 2.32· 103μm 4 μ J−1 σP

Figure 5. (a) SEM image of the pattern with the smallest width achieved after a lift-off process, corresponding to a dose of 1.19 × 10−2 J m−1. (b) SEM image of the Istituto Italiano di Tecnologia (IIT) logo created with free writing lithography.

after the lift-off process using microsphere lithography. The minimum width of the metal line was slightly less than 400 nm, obtained using a dose of 1.19 × 10−2 J m−1. The introduction of the LOR-3B layer led to an increase of the minimum achievable opening size in the photoresist; however, the correlation between the dose and the opening size remained similar to that observed for S1813 alone. Additional optimizations of the bilayer thickness and development phase should further decrease the final line width. 4. Writing of Arbitrary Patterns. One of the strengths of our approach is the simplicity with which it can be implemented on various existing instruments. The generation of an arbitrary pattern using microbead lithography only requires an AFM that provides a scan pattern generator. Because this software tool is already present in several standard types of equipment, we could write arbitrary patterns on photoresists using our microbead lithography approach. The scanning system was set to deliver a precise dose of 1.19 × 10−2 J m−1 based on the results obtained in the lift-off experiments. The spatial resolution achieved with arbitrary patterning is comparable with the spatial resolution of a regular pattern, with a line width less than 400 nm. An example of the arbitrary pattern of the Ti/Au bilayer is depicted in Figure 5b. Due to the limitations of the pattern generator included in the WiTec AFM system, we could generate arbitrary structures only with orthogonal lines. The creation of more articulated designs would require the use of more complex and advanced pattern generators, which are available on the market and could be easily integrated on the AFM system. Thus, we do not forecast practical or technological obstacles for producing more complex patterns with the presented technique. 5. Comparison with a Laser Writer Instrument. To evaluate the potential of the proposed microsphere lithography approach, we compared the results obtained using an established high-resolution standard maskless lithography technique. We performed the comparison with a Laser writer DL66 Heidelberg Instruments. The laser writer is equipped with a 405 nm laser with 50 mW power. The transmitted power can be modified with two different filtering stages and can be reduced to a minimum of 0.75 μW. For the comparison, we used silicon substrate that was spin-coated with S1813 diluted 1:1 to obtain the same thickness used for microbead-assisted direct lithography. We exposed the substrates with a pattern of regular stripes as in the microbead lithography, using a broad range of exposure powers. By optimizing all the parameters, we achieved a clean and defined resist opening with a width as low as 477 nm, which is larger than the laser wavelength. The measured scan speed efficiency, defined as in microbead lithography, is V/σP = 12.8 μm4 μJ−1,

(2)

Based on the characteristics of our setup, with the laser power density used, assumed constant on the entire spot size, and σP = 10−3 μW μm−2, we worked at a velocity V = 2.32 μm2 s−1. The seemingly low writing velocity is due to the limited laser power available in our lab. In addition, this velocity considers the uniform exposure of a full surface, which is the worst-case scenario. For instance, in a practical case in which the filling factor is reduced to 40%, the writing velocity is approximately 2.5 times higher. 3. Lift-Off Process for Metal Deposition. The high resolution of the aperture into the photoresist is a direct and important indication of the potential of an optical lithography method. However, the high-resolution capabilities must be retained for lift-off processes for a promising and cost-effective lithography technique to be an alternative in a broad range of applications. The resist profile is a crucial parameter to achieve the final structure after a lift-off process with a high resolution. Lift-off processes are normally performed with negative-edge photoresists or with a photoresist bilayer stack that includes a sacrificial layer between the substrate and the photoresist. The inclusion of a sacrificial layer in a bilayer configuration is a delicate step that cannot be used for every type of structure: for example, an array of stripes patterned in a closely packed configuration often collapses. For a fair comparison with the best results of direct lithography, we maintained the use of the S1813 resist, which presents slightly positive edges (as presented in Figure 3c as a cross-sectional view produced with the FIB) and cannot be used alone for lift-off processes. Therefore, we tested our microbead lithography approach using the classic lift-off process based on a photoresist bilayer. A sacrificial layer of LOR3B was spin-coated between the substrate and the S1813. The thickness of the LOR-3B layer was approximately 300 nm after 5 min of baking at 180 °C. The S1813 layer on top of the LOR-3B was spin-coated with the same thickness as in the direct lithography. After exposure using microsphere lithography and subsequent development, a titanium−gold bilayer with thicknesses of 5 and 50 nm, respectively, was evaporated on the substrate using an e-beam evaporator. After metal deposition, the lift-off process was completed by removing the LOR-3B and S1813 in Remover PG. Figure 5a shows the best results obtained 32390


Research Article

ACS Applied Materials & Interfaces

path is used for precise positioning of the cantilever on the substrate. The setup is shown in Figure 6.

obtained using the nominal Laser Writer writing speed (V = 4.09 × 104 μm2 s−1) and laser power required to expose the resist (2.5 mW) focused in a 0.78 μm2 area. Thus, the scan speed efficiency is 190 times lower than that obtained for microbead lithography. These quantitative considerations are specific for the comparison with our DL66 laser system. Nevertheless there are general advantages with respect to all lithography systems based on laser writing. For example, the AFM cantilever with the embedded microbead can also be used to control simultaneously the distance from the substrate; additionally, the cantilever can be used directly in contact mode with the substrate without degradation of lithography performance. In terms of initial costs, the microbead lithography method could be implemented on AFM basic models that have an average cost of few tens of thousands euros, requiring only the addition of few thousands euros for a simple 405 nm laser and dedicated pattern generator. For comparison, a standard maskless lithography system with high resolution, such as the DL66 from Heidelberg Instruments, can have an average cost three to ten times higher. Furthermore, an important advantage of microbead lithography is that AFM systems are already present in the majority of the nanotechnology laboratories, thus strongly reducing the initial costs. Considering recurrent costs, it should be considered that, in case of damage due to collision with the substrate, a cantilever with an embedded microbead is extremely cheaper to replace with respect to the objective of a Laser Writer System.

■

CONCLUSIONS

■

METHODS

Figure 6. Diagram of the lithography setup.

Cantilever Preparation. The positioning of the microsphere on the substrate is controlled with a standard tipless cantilever (Budget Sensor All In One-Al-TL 150 μm long, 30 μm wide, 2.7 μm thick, 150 Hz resonant frequency and 7.4 N/m force constant) that was specifically modified for the experiments. A hosting cavity was drilled into the silicon cantilever using a milling process performed with a Focus Ion Beam (FIB−FEI NanoLab 600 dual beam system) in order to place the microsphere inside it. The FEI instrument is also used to acquire the SEM images reported in the text. The cantilever is integrated into a mechanical arm that allows for movement to a specific position on the sample surface. For this purpose, an AFM tip was modified to include the bead inside a specific compartment. A 4.7-μm-diameter silica beads (Bandglas CS019) with a refractive index of 1.46 are diluted in a water solution. The solution is dropped onto a glass slide and, once the glass surface has dried, the glass slide is placed over the scanning stage of AFM-Raman optical microscope (WiTeC).28,29 The microscope is customized to couple an AFM scanning probe micromanipulator with both a coaxial upright and an inverse optical microscope. To follow the position of the microsphere on the substrate, the cantilever is back-illuminated (see Figure 6), and the image is collected using a 40× objective (Olympus Plan Acrhromat 0.65NA, 0.6 mm WD) and acquired with a CCD camera (Imaging Source DFK 72buc02) placed on the same side as the illumination. The perforated cantilever is mounted on the probe micromanipulator. The hole is aligned with a sphere on the glass slide and is lowered until the cantilever reaches the microsphere. The hole has a diameter of 4.5 μm, slightly smaller than the bead diameter, to promote the adhesion. Once the contact between the bead and the cantilever occurs, the tip is retracted and, in most cases, the sphere adheres to the rim. Sample Preparation. A standard 125-μm-thick coverslip is used as sample substrate. The sample preparation included cleaning for 30 s with a dilute solution of Surfactant Hellmanex III, a rinse with Milli-Q water, and a second rinse with isopropanol followed by nitrogen drying. Two different commercial photoresists were used throughout the experiments, S1813 (Shipley) diluted 1:1 in anisole (methoxybenzene) to obtain a thickness of approximately 500 nm with a 4000 rpm spincoating speed for 60 s and AZ-5214 (Microchemicals) diluted 1:4 with PGMEA (propylene glycol monomethyl ether acetate) to obtain a thickness of approximately 150 nm with a 4000 rpm spin-coating speed for 60 s. Simulation Methods. The formation of a nanojet was simulated using the finite-element method (FEM) with Comsol Multiphysics by illuminating a dielectric bead, inserted into a silicon layer and in contact with bulky resist, with a coherent monochromatic plane wave. Because

In this work, we presented a novel high-resolution low-cost lithographic system based on a microbead embedded in a tipless AFM cantilever; this solution can be implemented in a broad range of applications. We envision that our approach can advance the microbead lithography method toward highly versatile and commercial methodologies, offering increased resolution and a higher degree of freedom compared to standard commercial instruments. The presented approach can be easily implemented on existing relatively low-cost AFMs instruments that are already present in many research laboratories to offer high-resolution lithography capabilities combined with cost-effective tools. Our results show that subwavelength features with spatial resolutions less than 300 nm could be obtained for direct opening of straight lines using a 405 nm laser. For lift-off processes, we obtained metallic nanostructures with a resolution of 395 nm; in this case, further improvements of the bilayer stack are necessary to increase the resolution. Exposures of arbitrary patterns both in direct lithography and in lift-off processes maintained the subwavelength resolution. This new technique has great potential to introduce rapid high-resolution microbead lithography as a new standard technique in several laboratory environments due to its low cost compared with modern commercial lithography instruments. Notably, in future development, the method can be used to expose 3D surfaces due to the capability of the cantilever to follow the substrate topography, which is extremely complex and expensive using current lithographic standard systems based on objective focusing.

Experimental Setup. The experimental setup comprises two distinct optical paths, one below and one above the sample. The upper optical path is dedicated to the laser exposure system and includes the AFM cantilever with the embedded microsphere. The other optical 32391


Research Article

ACS Applied Materials & Interfaces of the radial symmetry of the microbead system, we simulated a twodimensional semicircular structure. Substrate Approach. Once the microsphere is fixed inside the hole, white light from the bottom of the setup is sent to a bandpass filter to avoid contamination of the resist due to the UV component of the incident light. A beam splitter partially reflects the light in the direction of a mirror positioned below a 40× objective. The light illuminates the sample from the bottom and is collected back into the same 40× objective. The beam is then focused on a CCD camera. An optical cage is mounted above the sample plane. Lithography Methods. A 405 nm laser (Roithner Lasertechnik) is sent to a collimator before passing through a mechanical homemade shutter (maximum working frequency 100 Hz), then through a beam splitter cube that sends the beam onto the microsphere through a 50μm-diameter pinhole. After exposure, the sample is developed with Microposit MF-139 Developer from Shipley to remove the exposed resist. For all samples, the developing time was maintained at 15 s. The sample is rinsed with Milli-Q water and dried fluxing nitrogen before plasma cleaning for 1 min at 100 W with 100% oxygen. The sample is coated with an evaporation of titanium and gold, 5 nm of Ti and 50 nm of Au. The evaporation is performed in an ultravertical configuration, placing the sample exactly on the crucible axis with the rotational speed of the stage set at 0 rpm. The evaporation rate for both materials was 0.02 nm s−1. Once the sample is coated, it is immersed in a bath of MNP (N-methylpyrrolidine, Remover PG) at 80 °C for 2 h, then sonicated for 1 min with 100 W and 40 Hz and rinsed with acetone.

■

(6) Selimis, A.; Mironov, V.; Farsari, M. Direct Laser Writing: Principles and Materials for Scaffold 3D Printing. Microelectron. Eng. 2015, 132, 83−89. (7) Cataldo, S.; Zhao, J.; Neubrech, F.; Frank, B.; Zhang, C.; Braun, P. V.; Giessen, H. Hole-Mask Colloidal Nanolithography for Large-Area Low-Cost Metamaterials and Antenna-Assisted Surface-Enhanced Infrared Absorption Substrates. ACS Nano 2012, 6 (1), 979−985. (8) Chen, L.; Milor, L. S.; Ouyang, C. H.; Maly, W.; Peng, Y. K. Analysis of the Impact of Proximity Correction Algorithms on Circuit Performance. IEEE Trans. Semicond. Manuf. 1999, 12 (3), 313−322. (9) Huo, F.; Zheng, G.; Liao, X.; Giam, L. R.; Chai, J.; Chen, X.; Shim, W.; Mirkin, C. a. Beam Pen Lithography. Nat. Nanotechnol. 2010, 5 (9), 637−640. (10) Gan, Z.; Cao, Y.; Evans, R. a; Gu, M. Three-Dimensional Deep Sub-Diffraction Optical Beam Lithography with 9 Nm Feature Size. Nat. Commun. 2013, 4 (May), 2061. (11) Smith, B. W.; Kang, H.; Bourov, A.; Cropanese, F.; Fan, Y. Water Immersion Optical Lithography for 45-Nm Node. Proc. SPIE 2003, 5040, 679−689. (12) Lindquist, N. C.; Jose, J.; Cherukulappurath, S.; Chen, X.; Johnson, T. W.; Oh, S. H. Tip-Based Plasmonics: Squeezing Light with Metallic Nanoprobes. Laser Photonics Rev. 2013, 7 (4), 453−477. (13) Stulen, R. H.; Sweeney, D. W. Extreme Ultraviolet Lithography. IEEE J. Quantum Electron. 1999, 35 (5), 694−699. (14) Rangelow, I. W.; Voigt, J.; Edinger, K. NANOJET”: Tool for the Nanofabrication. J. Vac. Sci. Technol., B: Microelectron. Process. Phenom. 2001, 19 (6), 2723−2726. (15) Voigt, J.; Reinker, B.; Rangelow, I. W.; Mariotto, G.; Shvets, I.; Guethner, P.; Lö schner, H. NANOJET: Nanostructuring via a Downstream Plasmajet. J. Vac. Sci. Technol., B: Microelectron. Process. Phenom. 1999, 17 (6), 2764−2767. (16) Voigt, J.; Shi, F.; Edinger, K.; Guthner, P.; Rangelow, I. W. Nanofabrication with Scanning Nanonozzle “Nanojet. Microelectron. Eng. 2001, 57−58, 1035−1042. (17) Heifetz, A.; Kong, S.; Sahakian, A. V.; Taflove, A.; Backman, V. Photonic Nanojets. J. Comput. Theor. Nanosci. 2009, 6 (9), 1979−1992. (18) Li, X.; Chen, Z.; Taflove, A.; Backman, V. Optical Analysis of Nanoparticles via Enhanced Backscattering Facilitated by 3-D Photonic Nanojets. Opt. Express 2005, 13 (2), 526−533. (19) Devilez, A.; Bonod, N.; Wenger, J.; Gérard, D.; Stout, B.; Rigneault, H.; Popov, E. Three-Dimensional Subwavelength Confinement of Light with Dielectric Microspheres. Opt. Express 2009, 17 (4), 2089−2094. (20) Kim, J.; Cho, K.; Kim, I.; Kim, W. M.; Lee, T. S.; Lee, K.-S. Fabrication of Plasmonic Nanodiscs by Photonic Nanojet Lithography. Appl. Phys. Express 2012, 5, 025201. (21) Wu, W.; Katsnelson, A.; Memis, O. G.; Mohseni, H. A Deep SubWavelength Process for the Formation of Highly Uniform Arrays of Nanoholes and Nanopillars. Nanotechnology 2007, 18, 485302. (22) Horrer, A.; Schafer, C.; Broch, K.; Gollmer, D. A.; Rogalski, J.; Fulmes, J.; Zhang, D.; Meixner, A. J.; Schreiber, F.; Kern, D. P.; Fleischer, M. Parallel Fabrication of Plasmonic Nanocone Sensing Arrays. Small 2013, 9 (23), 3987−3992. (23) Duocastella, M.; Tantussi, F.; Haddadpour, A.; Zaccaria, R. P.; Jacassi, A.; Veronis, G.; Diaspro, A.; Angelis, F. De. Combination of Scanning Probe Technology with Photonic Nanojets. Sci. Rep. 2017, 7 (1), 3474. (24) Zhong, J.; Ma, M.; Zhou, J.; Wei, D.; Yan, Z.; He, D. Tip-Induced Micropatterning of Silk Fibroin Protein Using In Situ Solution Atomic Force Microscopy. ACS Appl. Mater. Interfaces 2013, 5, 737−746. (25) Zhong, J.; Sun, G.; He, D. Classic, Liquid, and Matrix-Assisted Dip-Pen Nanolithography for Materials Research. Nanoscale 2014, 6, 12217−12228. (26) Weber-Bargioni, A.; Schwartzberg, A.; Cornaglia, M.; Ismach, A.; Urban, J. J.; Pang, Y.; Gordon, R.; Bokor, J.; Salmeron, M. B.; Ogletree, D. F.; Ashby, P.; Cabrini, S.; Schuck, P. J. Hyperspectral Nanoscale Imaging on Dielectric Substrates with Coaxial Optical Antenna Scan Probes. Nano Lett. 2011, 11, 1201−1207.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsami.7b10145. Lithography process; comparison with laser writer (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Andrea Jacassi: 0000-0003-1663-6452 Michele Dipalo: 0000-0002-1823-8231 Francesco De Angelis: 0000-0001-6053-2488 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The research leading to these results has received funding from the European Research Council under the European Union’s Seventh Framework Programme (FP/2007-2013)/ERC Grant Agreement no. [616213], CoG: Neuro-Plasmonics.

■

REFERENCES

(1) Nishi, Y.; Doering, R. Handbook of Semiconductor Manufacturing Technology, 2nd ed.; CRC Press: Boca Raton, FL, 2007. (2) Weichelt, T.; Vogler, U.; Stuerzebecher, L.; Voelkel, R.; Zeitner, U. D. Resolution Enhancement for Advanced Mask Aligner Lithography Using Phase-Shifting Photomasks. Opt. Express 2014, 22 (13), 16310− 16321. (3) Kachwala, N.; Petersen, J. S.; Mccallum, M.; Sematech, I.; Drive, M.; Advanced, P.; Box, L. P. O. High Transmission Attenuated PSM Benefits and Limitations through a Validation Study of 33%, 20% and 6% Transmission Masks. Proc. SPIE 2000, 4000 (4), 1163−1174. (4) Wolf, A. J.; Hauser, H.; Kübler, V.; Walk, C.; Höhn, O.; Bläsi, B. Origination of Nano- and Microstructures on Large Areas by Interference Lithography. Microelectron. Eng. 2012, 98, 293−296. (5) Shelden, G.; Marmillion, P. Mask Industry Assessment: 2006. Proc. SPIE 2006, 6349, 634902. 32392


Research Article

ACS Applied Materials & Interfaces (27) Chen, Z.; Taflove, A.; Backman, V. Photonic Nanojet Enhancement of Backscattering of Light by Nanoparticles: A Potential Novel Visible-Light Ultramicroscopy Technique. Opt. Express 2004, 12 (7), 1214−1220. (28) Garoli, D.; Zilio, P.; Gorodetski, Y.; Tantussi, F.; De Angelis, F. Beaming of Helical Light from Plasmonic Vortices via Adiabatically Tapered Nanotip. Nano Lett. 2016, 16 (10), 6636−6643. (29) Zilio, P.; Dipalo, M.; Tantussi, F.; Messina, G. C.; De Angelis, F. Hot Electrons in Water: Injection and Ponderomotive Acceleration by Means of Plasmonic Nanoelectrodes. Light: Sci. Appl. 2017, 6, e17002.

32393
