Background: With continuous decrease in the technology node of dynamic random access memories (DRAMs) down to sub-20 nm, the self-aligned double patterning (SADP) is an effective approach to generate two-dimensional (2-D) patterns, particularly contact arrays.

Aim: We demonstrate a patterning scheme using the SADP technique to produce active trim contacts with anisotropic pattern pitches.

Approach: The proposed scheme uses two consecutive spacer-formation processes.

Results: By making the ellipsoidal core pillars and minimizing the spacer thickness, 2-D critical dimensions (CDs) for self-generated contacts match well with those for core contacts. In addition, an interesting cross-dependence of X-CD and Y-CD variations for the core and self-generated contacts is observed.

Conclusion: This patterning approach is useful for forming active trim contacts in sub-20 nm DRAMs using fewer numbers of ArF immersion photolithography steps.

Background: Switchable glasses allow the control of light transmission—an attractive property for applications such as car sunroofs, aircraft windows, building windows, augmented reality, imaging, and displays. Commercialized switchable glasses have severe limitations, such as speed, cost, and operating conditions, among others. Microshutters, a type of switchable glass with very distinctive properties, are reviewed, as they are a technology that could significantly improve some or all of the shortcomings mentioned above.

Aim: We will summarize the various types of microshutters and tentatively identify various critical designs, fabrication schemes, and performance criteria by the many research groups implementing them and investigating their properties.

Approach: We will describe the various approaches used to control light transmission through microelectromechanical systems. It will compare their performances and comment on fabrication and implementation challenges.

Conclusions: Microshutters have performance levels that could make them good candidates for switchable glasses. Many research groups have investigated various approaches to fabricate microshutters and have shown that they can be implemented reliably on a small scale, with fast actuation, low power, and high contrast and are relatively easy to manufacture. Work is needed to demonstrate that they can be scaled-up and still be economical to produce.

In recent decades, the rate of shrinking integrated-circuit components has slowed as challenges accumulate. Yet, in part by virtue of an accelerating rate of cleverness, the end-user value of new semiconductor processes steadily advances. On top of the miniaturization benefits delivered by optical lithography, value is boosted by innovations in wafer processing, mask synthesis, materials and devices, microarchitecture, and circuit design. Focusing on three decades of microprocessor data enables quantification of how innovations from those domains have contributed over time to integrated-circuit “value scaling” in terms of performance, power, and cost. At some point, lateral shrinking will end altogether and the kinds of ingenuity emerging from those domains may provide clues for how very large-scale integration value creation will advance beyond that point.

Background: The resolution capability of EUV lithography has reached parity with e-beam, raising the possibility that maskless EUV could supplant e-beam for mask writing and low-volume wafer patterning.

Aim: We outline a maskless EUV scanner design with a 13.5-nm operating wavelength and numerical aperture of 0.55.

Approach: A microlens array partitions radiation from a commercial laser-produced plasma EUV source into ∼2 million individual beams, which are focused to separate, diffraction-limited focal points on a writing surface, and the surface is raster-scanned across the focal point array as the beams are individually modulated by MEMS microshutters integrated within the microlens array to construct a digitally synthesized raster exposure image.

Results: Compared to state-of-the-art mask-projection EUV lithography, the system would have ∼1000  ×   lower throughput, but its power requirement would also be ∼1000  ×   lower, the exposure dose would be ∼10  ×   higher, scan velocity and acceleration would be ∼1000  ×   lower, and it would have the advantage of maskless operation. In comparison to e-beam mask writers, a maskless EUV scanner could provide higher resolution with at least double the throughput and over 10  ×   higher dose.

Conclusions: Maskless EUV lithography could provide significant cost and performance benefits for both direct-write applications and photomask production for mask-projection lithography.

As a potential solution to next-generation nanolithography, directed self-assembly (DSA) of block copolymers (BCPs) is still restrained in high-volume manufacturing primarily due to its defectivity issue. Though defects possess greater free energies than aligned morphologies and are highly energetically unfavorable, they can be kinetically trapped by the energy barriers and persist for a long time during annealing. Therefore, understanding the kinetics of defect annihilation is crucial in revealing the mechanism of defect formation and in further reducing defectivity in DSA. We focus on two types of predominant defects in DSA—dislocation and bridge. A kinetic model of each defect type is developed through statistical analysis of experimental data, providing insight into possible approaches of further defect reduction. We also investigate the impact of annealing temperature and film thickness on annihilation kinetics and discuss the reasons behind the observed results. By simply optimizing annealing conditions and film thickness, we have successfully reduced the total defect density by 1 order of magnitude. Though these findings are based on polystyrene-b-poly(methyl methacrylate) (PS-b-PMMA), we anticipate they could be readily applied to other BCP platforms as well.

A method of tapering waveguides using fixed-electronic-beam-moving-stage (FBMS) paths is presented. The tapering is achieved by joining two FBMS paths to a common point. Compared to conventional area and FBMS tapering methods, the proposed method offers smooth and alignment-error-free tapering between waveguides of different widths. We experimentally demonstrate a fully functional FBMS patterned photonic circuit with a power splitter, wire-to-slot coupler, slot waveguide, and a slotted ring resonator. The device response with an insertion loss of −1.35  dB is measured around 1550-nm wavelength.

Background: Hybrid inorganic-organic materials have emerged as promising candidates for EUV resists. However, knowledge on their stability when deposited as thin films is essential for their performance in EUV lithography.

Aim: We investigate whether the molecular structure of Zn-based metal oxoclusters is preserved upon thin film deposition and study aging processes of the thin film under different atmospheres, since these chemical changes affect the solubility properties of the material.

Approach: A hybrid cluster that combines the high EUV photon absorption cross-sections of zinc and fluorine with the reactivity of methacrylate organic ligands was synthesized. The structural modifications upon thin film formation and after aging in air, nitrogen, and vacuum were studied using a combination of spectroscopic techniques. Preliminary studies on the lithographic performance of this material were performed by EUV interference lithography.

Results: The Zn-based compound undergoes structural rearrangements upon thin film deposition as compared to the bulk material. The thin films degrade in air over 24 h, yet they are found to be stable for the duration and conditions of the lithography process and show high sensitivity.

Conclusions: The easy dissociation of the ligands might facilitate hydrolysis and rearrangements after spin-coating, which could affect the reproducibility of EUV lithography.

Background: To reduce defocus from leveling errors at the wafer edge, modern exposure tools offer a broad range of advanced leveling controls. These additional degrees of freedom offer better leveling performance, but users hesitate to spend the tool time, wafers, and engineering hours necessary to find and maintain the optimal settings experimentally.

Aim: In order to fully explore the potential of advanced leveling controls, an automated, fast simulation method should be introduced.

Approach: Alternative set-point curves and resulting focus residuals are simulated from existing wafer height maps. Optimizations are carried out to obtain the best edge exclusion settings for several dynamic random access memory and NAND flash memory products, across different layers and exposure tools. The simulated focus errors are compared to the POR settings and verified with electrical results.

Results: An efficient optimization algorithm was demonstrated and significant leveling improvements found for a number of use cases. The resulting settings vary substantially between different products, layers, and exposure tools. The impact of the improved leveling performance is verified using electrical data.

Conclusions: The speed of the presented method proves crucial to help lithographers dial in and maintain numerous settings at optimal values across a typical production line.

An optimized source has the ability to improve the process window during lithography in semiconductor manufacturing. Source optimization is always a key technique to improve printing performance. Conventionally, source optimization relies on mathematical–physical model calibration, which is computationally expensive and extremely time-consuming. Machine learning could learn from existing data, construct a prediction model, and speed up the whole process. We propose the first source optimization process based on autoencoder neural networks. The goal of this autoencoder-based process is to increase the speed of the source optimization process with high-quality imaging results. We also make additional technical efforts to improve the performance of our work, including data augmentation and batch normalization. Experimental results demonstrate that our autoencoder-based source optimization achieves about 10⁵  ×   speed up with 4.67% compromise on depth of focus (DOF), when compared to conventional model-based source optimization method.

Grayscale lithography is a widely known but underutilized microfabrication technique for creating three-dimensional (3-D) microstructures in photoresist. One of the hurdles for its widespread use is that developing the grayscale photolithography masks can be time-consuming and costly since it often requires an iterative process, especially for complex geometries. We discuss the use of PROLITH, a lithography simulation tool, to predict 3-D photoresist profiles from grayscale mask designs. Several examples of optical microsystems and microelectromechanical systems where PROLITH was used to validate the mask design prior to implementation in the microfabrication process are presented. In all examples, PROLITH was able to accurately and quantitatively predict resist profiles, which reduced both design time and the number of trial photomasks, effectively reducing the cost of component fabrication.

Metal containing resists (MCR) are one of the candidates for extreme ultraviolet resists aiming to achieve the resolution, linewidth roughness, and sensitivity requirements of advanced design nodes. MCRs intrinsically have high etch resistance owing to their metal content. Therefore, low resist thickness (∼18  nm) is sufficient to transfer resist patterns into an underlying hard mask. Also, the thin resist reduces susceptibility to pattern collapse during development because of low aspect ratio. However, thus far, little attention has been paid to optical metrology and inspectability (overlay, defect inspection, scatterometry, etc.) of these resists, which is another critical requirement to move MCR toward high-volume manufacturing. We investigate the overlay metrology and overlay correction with MCR. Even though the optical contrast for MCR is slightly lower than for chemically amplified resist (CAR), it seemed sufficient for high-quality overlay metrology. However, the measurement precision for MCR is deteriorated compared to that for CAR, resulting in significantly higher residuals. The root cause of the deteriorated measurement precision was found in grains in the optical image after MCR development. Interestingly, the after etch performance of CAR and MCR is identical. We demonstrate that with sufficient sampling, appropriate correctables can be extracted from the MCR overlay results. Finally, we discuss how the increased image noise can be compensated by the applied sampling scheme.

Background: Monte Carlo simulations of scanning electron microscopy (SEM) images ignore most surface effects, such as surface plasmons. Previous experiments have shown that surface plasmons play an important role in the emission of secondary electrons (SEs).

Aim: We investigate the influence of incorporating surface plasmons into simulations of low-voltage critical dimension SEM (CD-SEM).

Approach: We use a modified inelastic scattering model, derived for infinite flat surfaces, and apply it to nonflat, but smooth, geometries. This simplification captures most qualitative effects, including both surface plasmons and a reduced interaction with bulk plasmons near interfaces.

Results: We find that the SE signal hardly changes when surface interactions are turned on for a perpendicularly incident beam. When the incident beam is perfectly parallel to a surface, the SE signal does significantly increase. However, the beam must be extremely close to the surface for this effect to be appreciable. An SEM is unable to produce a beam that is both narrow and parallel enough to be noticeably affected.

Conclusions: The position of edges may appear shifted under specific circumstances. In realistic situations, it is unlikely to be a large effect.

Background: Charging of insulators is a complex phenomenon to simulate since the accuracy of the simulations is very sensitive to the interaction of electrons with matter and electric fields.

Aim: In this study, we report model improvements for a previously developed Monte-Carlo simulator to more accurately simulate samples that charge.

Approach: The improvements include both modeling of low energy electron scattering by first-principle approaches and charging of insulators by the redistribution of the charge carriers in the material with an electron beam-induced conductivity and a dielectric breakdown model.

Results: The first-principle scattering models provide a more realistic charge distribution cloud in the material and a better match between noncharging simulations and experimental results. The improvements on the charging models, which mainly focus on the redistribution of the charge carriers, lead to a smoother distribution of the charges and better experimental agreement of charging simulations.

Conclusions: Combined with a more accurate tracing of low energy electrons in the electric field, we managed to reproduce the dynamically changing charging contrast due to an induced positive surface potential.

Microsystem technology is well suited to batch fabricate microhemispherical resonator gyroscopes (HRG) to reduce cost and volume. In the processing of micro-HRG, a crucial step is to get a 3-D hemispherical mold with the large-scale, high-symmetry, and smooth surface. Compared with the hemispherical resonator, the toroidal resonator has the smaller frequency split and larger effective resonance mass under the same processing accuracy. A wafer-scale etching method for the toroidal resonator mold was presented, which is based on the deep reactive ion etching and improved HNA isotropic etching. The advantages of this method include low cost, time savings, and easy operation. With this method, toroidal molds with an average diameter over 1900  μm, asymmetry <0.2  %  , and roughness <10  nm were successfully fabricated. The uniformity and surface smoothness of the molds are mainly determined by the parameters of HNA etching. A series of controlled experiments were conducted to optimize isotropic etching parameters that include mask design, bath agitation, HNA composition, and temperature. The influence of these parameters on etching rate and uniformity was discussed. The result shows that the composition of 2.5:7:1 (HF  :  HNO₃  :  CH₃COOH), temperature of 30°C, and bath agitation of 20 revolutions per minute are optimal etching conditions to achieve high-performance molds.

Following the need to improve packaging and contact layers for photovoltaics and other optoelectronic applications, a renewed interest in the fabrication of thin, low-density silicon films has arisen. We demonstrate a reactive sputtering technique utilizing a secondary plasma to crack hydrogen gas during physical vapor deposition of silicon layers. Cracking efficiency of the gas varies heavily with pressure and power from under 10% to nearly 100% conversion to hydrogen radicals. Radicals incorporated into the film produce amorphous silicon films with densities as low as 1.73  g  /  cm³, compared to 2.2  g  /  cm³ in their nonhydrogenated counterparts. Reduced density films likewise have a reduction in index of refraction comparable to other hydrogenated amorphous silicon produced by other techniques with indices close to 2 across the visible portion of the spectrum. Our work represents a useful, scalable advance in the production of amorphous hydrogenated silicon for a variety of applications requiring large areas.

Background: The benefits of microresonator lasers include easy and low-cost fabrication methods and high-quality factor of microresonators, which results in low threshold current of the fabricated laser. However, the lack of tunable fabrication methods with precise size and spacing is a great challenge.

Aim: Based on a microsphere fabrication method that relies on injection and surface tension effects, simultaneous demonstration of an array of cavities with precise spacing and sizes is possible where the size of cavities is widely tunable.

Approach: Using a maskless lithography setup, pillars with base and tip diameters of 250 and 60μm are fabricated, which are used for demonstration of microspheres with tunable size, depending on the applied pump pressure. Microcavities composed of Rhodamine B dissolved in glycerol are injected into the polydimethylsiloxane carrier through a nozzle with a diameter of 70  μm.

Results: An ∼4-μJ  /  mm² lasing threshold has been obtained for a laser with a microsphere diameter of ∼350  μm. A redshift originated from the increment of dye concentration has also been observed.

Conclusions: Fabrication of precise microspheres for a microresonator laser is feasible using the introduced method, which can be further generalized using different active materials.