Mask Error Enhancement Factor Research Articles

Generation of inverse assist features using generative adversarial networks

Adding the assist feature in the optical proximity correction (OPC) procedure is crucial for improving the process window. The rule-based assist feature placement is very fast; however, the resultant process window still requires improvement. The model-based assist feature placement presents a better process window, but its iterative nature makes it extremely time-consuming. The assist features (AFs) that are optimized by using the inverse lithography technique perform best in terms of lithographic metrics such as the Depth of Focus, Mask Error Enhancement Factor, and Normalized Image Log Slope; these features can be added to the objective function. However, the inverse lithography technique is a pixel-level correction method, which is even more time-consuming. Consequently, it is typically used to provide guidance for rule-based AF placement. In this study, we design novel generative adversarial networks (GANs) that can generate AFs to increase the speed of assist feature placement while maintaining relatively high lithographic performance using inverse lithography techniques. The results demonstrate that the proposed generative adversarial network methods are three orders of magnitude faster than the inverse lithography techniques. Additionally, they exhibit lower edge placement errors and process variation bands when compared to the inverse lithography techniques.

Curvilinear mask optimization with refined generative adversarial nets

Inverse lithography technology (ILT) can optimize the mask to gain the best process window and image quality when the design dimension shrinks. However, as a pixel level correction method, ILT is very time-consuming. In order to make the ILT method useful in real mask fabrication, the runtime of ILT-based optical proximity correction mask must evidently decrease while keeping the good lithographic metric performance. Our study proposes a framework to obtain the curvilinear ILT mask with generative adversarial network (GAN). It is subsequently refined with the traditional ILT to exclude unexpected outliers generated by the GAN method. We design conditional GAN, reverse GAN (RGAN), and high discretion GAN (HDGAN) to generate curvilinear ILT mask. Their runtime and the performance are compared. Compared with the CILT method, the speed of GAN type methods with the afterward refinement is increased by an order of magnitude. The RGAN has a better performance in edge placement error and process variation band evaluation, and HDGAN has a better performance in the mask error enhancement factor evaluation. The designed RGAN and HDGAN are promising in actual application to generate the curvilinear mask. They can evidently decrease the runtime and have better lithographic metric performance.

Aberration analysis and compensate method of a BP neural network and sparrow search algorithm in deep ultraviolet lithography.

Mass production can be planned by utilizing the multiple patterning technology of 193 nm immersion scanners at the 7 nm technology node. In deep ultraviolet lithography, imaging performance is significantly affected by distortions of projection optics. For 7 nm immersion lithography layer patterns, distortions of the projection optics must be tightly controlled. This paper proposes an optimization method to determine the distribution of Zernike aberration coefficients. First, we build aberration prediction models using the backpropagation (BP) neural network. Then, we propose an aberration optimization method based on the sparrow search algorithm (SSA), using the common indicators of the lithography process window, depth of focus, mask error enhancement factor, and image log slope as the objective function. Some sets of optimized aberration distributions are obtained using the SSA optimization method. Finally, we compare the results of the SSA optimization algorithm with those obtained by rigorous computational simulations. The aberration combination distribution optimized by the SSA method is much more significant than the value under the zero aberration (ideal conditions), a nonoptimal distribution in deep ultraviolet lithography image simulation. Furthermore, the results indicate that the aberration optimization method has a high prediction accuracy.

Aberration analysis and compensate method of fully connected neural network in deep ultraviolet lithography

Thermal aberration of a projection lens is caused by lens heating and leads to degradation of image quality. Two tasks have been undertaken in this article: first, analysis of the impact of single aberration on imaging quality, and second, use of machine learning to calculate the combination of Zernike aberration coefficients to reduce the influence of aberration on the process window. The impact of aberration on deep ultraviolet (DUV) lithography imaging has been analyzed separately for 16 Zernike terms. The influence of single Zernike coefficient on depth of focus, mask error enhancement factor, and image log slope is comprehensively analyzed by source mask optimization. However, more than one Zernike term is present in the aberration of a real DUV optics. A fully connected neural network (FCNN) model based on machine learning is widely used in the screening of multivariate experiments and can solve the problem of the large number of input variables and the random values of the input data. Its network structure is simple to debug, with fast screening speed and high precision. To make reasonable budget for aberration, an FCNN model is used to find some combinations of Zernike coefficients for different lithographic performance indicators. This implies that the FCNN model has the ability to select a better Zernike combination while each individual Zernike term value falls in the range of −30 to 30 mλ, which plays a guiding role in DUV optics design.

P‐108: Development of Photomask Quality‐Assurance System Using Aerial Image Analysis

According as phase‐shift layer and OPC techniques are applied to photomasks for novel display with a resolution of higher than 1000ppi, it is indispensable to qualify photomask printability of area of interest‐ especially the area where the photomask repair processes were performed ‐ before the photomask is used in panel manufacturing line[1]. For this photomask qualification, new imaging system to emulate photo exposure process was developed; which enables to assess the printability of photomask by analyzing the aerial image. To verify the results from the imaging system, simulation and analytical solution were performed with various optical parameters such as NA and coherence. As an application of the imaging system, how to quantify the pattern abnormality using process window analysis and to measure in maskshop the mask error enhancement factor in aerial image was introduced.

Effects of mask pattern transmission on ArF lithographic performance in contact hole patterning

Even with the increase in need for next-generation lithography, immersion ArF lithography is still applied to the majority of critical layers. However, as circuit designs shrink, conventional 6% phase-shift mask (PSM) will become difficult to meet the lithography requirements for dense dot pattern compared to dense line pattern. To enhance immersion ArF lithographic performances for dot pattern, high-transmission PSM (High-T PSM) is attracting attention because the transmission of PSM has a significant impact on lithographic performances. From results of transmission dependency evaluated by mask three-dimensional (3D) simulation, it was found that 30% transmission has the best lithographic performances for dense dot. Based on these results, mask blank and mask making process for the new 30% PSM were developed. The results showed good cross-section profile, mask pattern resolution, and defect repairability. In addition, the durability against chemical cleaning and ArF irradiation were also improved. Wafer printability test using negative tone development demonstrated that new PSM has advantages in process window and mask error enhancement factor for dense dots (holes on wafer). Finally, the potential for further application of new PSM was investigated by mask 3D simulation. The results showed that new PSM has lithographic benefits not only for dense dots but also other patterns.

Gradient-based source mask and polarization optimization with the hybrid Hopkins–Abbe model

Source mask and polarization optimization (SMPO) is a promising extension of the widely used resolution enhancement technology, source mask optimization (SMO), to further enhance chip manufacturability beyond 28-nm node. Our work is aimed to develop an efficient gradient-based SMPO method by employing the hybrid Hopkins–Abbe imaging model to fulfill the goal. In addition to source and mask variables, the model is adapted to also include polarization variables to realize the optimization. Compact formulas for forward and backward model application are derived. The computation benefits from precomputed transmission cross coefficients and features high efficiency. Validity of the method is confirmed by case studies. For dense array pattern case, the optimal source and polarization can be found analytically. SMPO optimized results match well with the theoretical expectations. In addition, process window, mask error enhancement factor, and normalized image log-slope for the studied cases all get improved over the counterpart SMO results, which employ commonly used polarization. Runtime analysis shows the method is computationally efficient. Our work provides a valid way to optimize polarization together with source and mask.

Ant Colony Optimization-Based Freeform Sources for Enhancing Nanolithographic Imaging Performance

As the complexity of mask patterns increases, the standard off-axis illumination in 193-nm immersion systems is increasingly becoming inadequate for high-quality lithography. In this study, a source optimization (SO) technique was employed for improving lithographic resolution. A pixel-based optimization process was developed for constructing freeform sources on a PC platform to enhance contrast and improve process window (PW). In the proposed SO algorithm, the metaheuristic binary ant colony optimization algorithm was applied for constructing the freeform sources. Moreover, the proposed algorithm converged to the optimal solution between the 300th and 400th iterations. The performance of this algorithm was assessed using the half pitch (hp) 45-nm line/space (L/S) and 55-nm SRAM contact hole (CH) patterns. The fitness function was quality of aerial image, exposure latitude (EL), and depth of focus (DOF); the constructed freeform sources enhanced the contrast and extended the PW for the patterns. The DOF increased from 0.91 to 1.1 μm and 0.042 to 0.061 μm for the L/S and SRAM CH patterns, respectively, at an EL of 5% when using the freeform source. Moreover, the freeform source improved the mask error enhancement factor by 22% for the SRAM pattern.

Model-based breakdown of resist and mask contributions to local CDU for sub-30-nm contact holes in EUV lithography

The resist and mask contributions to the local critical dimension uniformity (L.CDU) of 30-nm contact holes (CHs) printed using extreme ultra-violet lithography (EUVL) are decomposed quantitatively by three independent budget breakdown methods. An asymptotic behavior model is proposed for the prediction of the resist-only contribution via a shifted and averaged multiple exposure (SAME) method that statistically averages out the collective mask local contributions. The other two budget breakdown methods of assessing, the variance of site-to-site difference for the derivation of resist-only contribution and the global mask error enhancement factor (G.MEEF) for the mask-only contribution, are performed and compared to the resist and mask contributions from the SAME method. The three budget breakdown methods consistently indicate that approximately 75–80% of L.CDU is attributed to the resist and process regardless of the process conditions, such as post exposure bake (PEB) temperature, PEB time, and exposure dose. L.CDU exhibits a linear correlation with 1/sqrt(dose), which is a potential indication that photon shot noise (PSN) could be a key contributor to the L.CDU on a wafer. Although a significant reduction of the total L.CDU is observed in the low PEB temperature process, i.e., high exposure dose, a relatively small change in the ratio of the resist-to-mask contribution may be observed because the chemical contrast enhancement by acid diffusion reduction is relatively small compared to PSN reduction.

Mask characterization for critical dimension uniformity budget breakdown in advanced extreme ultraviolet lithography

As the International Technology Roadmap for Semiconductors critical dimension uniformity (CDU) specification shrinks, semiconductor companies need to maintain a high yield of good wafers per day and high performance (and hence market value) of finished products. This cannot be achieved without continuous analysis and improvement of on-product CDU as one of the main drivers for process control and optimization with better understanding of main contributors from the litho cluster: mask, process, metrology and scanner. We will demonstrate a study of mask CDU characterization and its impact on CDU Budget Breakdown (CDU BB) performed for advanced extreme ultraviolet (EUV) lithography with 1D (dense lines) and 2D (dense contacts) feature cases. We will show that this CDU contributor is one of the main differentiators between well-known ArFi and new EUV CDU budgeting principles. We found that reticle contribution to intrafield CDU should be characterized in a specific way: mask absorber thickness fingerprints play a role comparable with reticle CDU in the total reticle part of the CDU budget. Wafer CD fingerprints, introduced by this contributor, may or may not compensate variations of mask CDs and hence influence on total mask impact on intrafield CDU at the wafer level. This will be shown on 1D and 2D feature examples. Mask stack reflectivity variations should also be taken into account: these fingerprints have visible impact on intrafield CDs at the wafer level and should be considered as another contributor to the reticle part of EUV CDU budget. We also observed mask error enhancement factor (MEEF) through field fingerprints in the studied EUV cases. Variations of MEEF may play a role towards the total intrafield CDU and may need to be taken into account for EUV lithography. We characterized MEEF-through-field for the reviewed features, with results herein, but further analysis of this phenomenon is required. This comprehensive approach to quantifying the mask part of the overall EUV CDU contribution helps deliver an accurate and integral CDU BB per product/process and litho tool. The better understanding of the entire CDU budget for advanced EUVL nodes achieved by Samsung and ASML helps extend the limits of Moore’s Law and to deliver successful implementation of smaller, faster and smarter chips in semiconductor industry.

Improved imaging properties of thin attenuated phase shift masks for extreme ultraviolet lithography

The mask shadowing effect is a unique phenomenon caused by a mirror-based mask structure and an oblique incident angle of light in the extreme ultraviolet lithography process. This results in a horizontal–vertical (H–V) bias, an ellipticity in the contact hole pattern and, eventually, a patterning limit and critical dimension (CD) nonuniformity. Reducing the absorber thickness is the most effective method to minimize the mask shadowing effect, but this can deteriorate the mask image contrast. In this paper, an attenuated phase shift mask (PSM) is proposed as a potential method for extending the patterning limit to below 16 nm. By applying Mo as a phase shift layer, which has a refractive index (n) similar to that of the TaN absorber but with a lower absorption value, thin attenuated PSMs with various reflectivity values were obtained by controlling the Mo thickness. The proposed PSM consists of a 16.5-nm-thick TaN absorber layer and a 24-nm-thick Mo phase shifter on 2-nm-thick Ru-capped Mo/Si multilayers. This attenuated PSM results in 17.98% and 27.52% reductions in the mask error enhancement factor as well as 3.65 and 1.84 nm reductions in the H–V CD bias compared with the conventional binary intensity mask with a 70-nm-thick TaN absorber for the 22 nm line and space 1:1 vertical dense pattern under 0.25 and 0.33 numerical aperture illumination conditions, respectively. Moreover, a 10%–12% improved image contrast was obtained with 11%–17% reflectivity on the absorber stack, which corresponds to a 24–29 nm Mo thickness under 0.25 NA illumination conditions.

Evaluation of lithographic performance of extreme ultra violet mask using coherent scattering microscope

The authors report the application of a coherent scattering microscope (CSM) equipped with a femtosecond laser based extreme ultra violet (EUV) source to analyze aerial images for an EUV lithography mask. The CSM employs an EUV beam with a 13.5 nm wavelength and an incident angle of 6° for the wafer scanner emulation. The CSM consists of a condenser optics that delivers the EUV beam to the mask and detector that measures the field spectrum diffracted by the mask pattern. Using a pair of spherical and plane EUV mirrors, coherent EUV light is relayed to the mask and the light diffracted by the mask pattern is measured by a 1 in. size charge-coupled device detector. The integration time of the EUV beam in the detector is controlled by a high speed shutter. This device experimentally records the coherent diffraction pattern from the EUV mask and reconstructs its aerial image using a phase retrieval algorithm. These features made it possible to investigate the projected image properties, such as depth-of-focus, exposure latitude, mask error enhancement factor, horizontal–vertical critical dimension (H–V CD) bias, normalized image log slope (NILS), and so on. The device also provides a mask CD from a rigorous coupled-wave analysis algorithm modeling based on the Maxwell equation through measured diffraction patterns. NILS according to half-pitch design node was measured by CSM and simulated using a Solid-EUV simulator. Illumination conditions of NA and sigma (σ) values are 0.25 and 0.8, respectively, which are the same as those of the EUV exposure tool. The NILS measured by CSM and that simulated using the Solid-EUV have good consistency.

Performance Comparison between Attenuated PSM and Opaque MoSi on Glass (OMOG) Mask in Sub-32nm Litho Process

During the development of optical lithography that has been extending to 32 nm and beyond, it is found that the selection of mask blank types has a very strong impact on minimum feature process window, or critical dimension performance that can be achieved, which will affect ultimate wafer lithographic performance. Although the industry-standard 6% attenuated phase shifting mask (PSM) with 67nm MoSi has been widely used, as the manufacturing design rule continues to shrink, it will introduce quite a few challenges, such as, defect inspections, imaging contrast, or NILS, and mask error enhancement factor (MEEF) [1]. Meanwhile, the standard chrome-on-glass mask, compared with 6% Att. PSM, is unable to support the minimum feature required in lithographic process windows and through pitch requirements. To alleviate this problem, a new type of binary mask with Opaque MoSi On glass (OMOG) is gradually accepted and used in 32 nm. In this paper, both 6% attenuated PSM and OMOG mask have been evaluated with 28 nm patterns. PSM and OMOG mask's critical dimension uniformities (CDU) are compared. Further more; we have also compared wafer line width roughness (LWR), After Development Inspection (ADI) CDU, MEEF, through pitch process window including Depth of Focus (DOF) and Exposure Latitude(EL). Our result shows that, for bright field and dark field mask, OMOG mask show different behavior than PSM mask in lithographic process performance. As above, whether OMOG can replace 6% attenuated PSM as 32 nm/28 nm main stream mask blank will depend on mask performance, litho performance, and device requirements.

Influence of MEEF change on the mask shadowing effect in extreme ultraviolet lithography

The influence of mask error enhancement factor (MEEF) on the mask shadowing effect was investigated for extreme ultraviolet lithography. Imaging properties including horizontal–vertical (H–V) CD (critical dimension) bias and MEEF change through the pitch according to absorber thickness and process condition were studied using aerial image simulation. The mask structure used in this study consisted of tantalum nitride (TaN) absorber and 2 nm ruthenium (Ru) capping layer on the 40 pair of Mo–Si multilayer. As the absorber thickness increased and the pattern pitch decreased, both H–V CD bias and MEEF increased. At the illumination condition of 0.32 numerical aperture (NA), the H–V CD bias variation through the pitch was negligible and slightly increased at 1:1 pitch, while it steeply increased at 1:1.2 and 1:1 pitch for NA of 0.25. The MEEF value was below 1.5 for all calculated absorber thicknesses when the pitch was from 1:1.2 to 1:5, whereas it was 3 with 64 nm thick TaN for 1:1 pitch at vertical pattern. With the increment of absorber thickness, the MEEF difference between the horizontal and vertical pattern increased. We also calculated the H–V overlapping process window (PW) according to TaN thickness using 22 nm 1:1 line and space (L/S) pattern. As absorber thickness decreased, the overlapping zone in the EL of the focus–exposure plots between the horizontal and vertical features increased. Enough image contrast and H–V overlapping PW could be achieved by applying 38 nm thick TaN.

Lithography Process Transfer from ArF to High-NA KrF Tool

In this paper, we present the capabilities of lithographic process transfer by use of high NA KrF tool (TWINSCAN XT:1000H, Maximal NA 0.93). The Metal 1 (100 nm half-pitch)and Via (110 nm half-pitch) layers of 65 nm technology have been chosen for this research. The layers belong to customer product that is currently manufactured on ArF tool (TWINSCAN XT:1400). By optimizing illumination setting and applying advanced OPC solution, we have obtained the overlapping process window of critical features for the KrF process that is comparable with parameters of the initial ArF process. A good performance of Metal layer with respect to MEEF (Mask Error Enhancement Factor) also allows for stable process characteristics within wide line through pitch range. However, to ensure CD uniformity for Via layer an aggressive illumination solution such as soft-QUASAR or customized illumination may be required.

Analytical optimization of high-transmission attenuated phase-shifting reticles

For sub-50-nm semiconductor patterning processes, extremely coherent illumination and high-transmission (Hi-T) attenuated background phase-shifting reticles are becoming increasingly important in delivering high-quality aerial images. Here, a Hi-T reticle is defined as any attenuated phase-shifting mask with a background transmission intensity of 12% or higher of incident light. While numerical simulation has served as a powerful tool in reticle design, the analytical relation of image log slope (ILS) and mask error enhancement factor (MEEF) are calculable in concise, exact form to reveal additional insight into reticle optimization. The properties of ILS and MEEF can be derived analytically in predicting the optimal range of mask bias in terms of image contrast and sensitivity for improved critical dimension (CD) control. These derived relations also demonstrate the capability for analysis beyond extensive numerical simulation and can be easily scaled and applied to various pitches and mask CDs. These relations are also verified in this paper with simulation of realistic process parameters and thus prove the applicability in optimization of Hi-T reticles.

Accurate Systematic Hot-Spot Scoring Method and Score-Based Fixing Guidance Generation

The researches on predicting and removing of lithographic hot-spots have been prevalent in recent semiconductor industries, and known to be one of the most difficult challenges to achieve high quality detection coverage. To provide physical design implementation with designer's favors on fixing hot-spots, in this paper, we present a noble and accurate hot-spot detection method, so-called “leveling and scoring” algorithm based on weighted combination of image quality parameters (i.e., normalized image log-slope (NILS), mask error enhancement factor (MEEF), and depth of focus (DOF)) from lithography simulation. In our algorithm, firstly, hot-spot scoring function considering severity level is calibrated with process window qualification, and then least-square regression method is used to calibrate weighting coefficients for each image quality parameter. In this way, after we obtain the scoring function with wafer results, our method can be applied to future designs of using the same process. Using this calibrated scoring function, we can successfully generate fixing guidance and rule to detect hot-spot area by locating edge bias value which leads to a hot-spot-free score level. Finally, we integrate the hot-spot fixing guidance information into layout editor to facilitate the user-favorable design environment. Applying our method to memory devices of 60nm node and below, we could successfully attain sufficient process window margin to yield high mass production.

Hot-Spot Detection and Correction Using Full-Chip-Based Process Window Analysis

In semiconductor industry, lithography technology has advanced with higher numerical aperture (NA) and/or lower-wavelength scanner on a yearly basis, which made it possible to keep shrinking the pattern size of semiconductor devices. Unfortunately, since chip size shrinking is currently limited to 193 nm ArF immersion scanners, lithographers have no further help from the equipment side for technology development. At this time, obtaining a compact design of a semiconductor device having a wide process window will be a key solution to realize the semiconductor industry's primary goal of fabricating high-density products with high yield. For a given particular design layout, the capability to analyze how difficult the process will be, which regions need to be modified to enlarge the process margin, and to provide guidelines to obtain a high-quality design are some of the key issues for the design for manufacturability (DfM). Many electronic design automation (EDA) companies are aggressively developing the method for DfM to meet the requirement for improved designs. In this paper, we discuss a method of full-chip-level lithography simulation using the optical proximity correction (OPC) engine and the calculation of process-limiting index parameters: mask error enhancement factor (MEEF), image log slope (ILS), and focus sensitivity (FS). Furthermore, this work includes a method of the quantitative rating of each design.

Observation of negative mask error enhancement factor due to mask transmission resonance

We report on the observation of mask transmission resonances in line/space gratings on bilayer mask stacks, both experimentally and in rigorous electromagnetic field (EMF) lithographic simulations. The mask transmission resonances manifest themselves as a local enhancement in the dose-to-size curve through mask linewidth for a given pitch targeted to a fixed wafer CD. We relate this finding to a local transmission loss of the propagating diffraction orders. We investigate both by experiments and by simulations the dependency of this mask transmission resonance on various lithographic conditions such as mask material, grating pitch, and incidence angle and polarization state of the incoming light. The resonant-type anomaly has a large impact on the mask error enhancement factor (MEEF) behavior of these bilayer mask stacks. The dependence of the MEEF as a function of mask linewidth shows strong variations and even negative MEEF values around the position of this resonance, which stem from the local dose-to-size enhancement. More precisely, the behavior of the MEEF can be well predicted by the differential of the dose-to-size curve.

Experimental demonstration of dark field illumination using contact hole features

The use of off-axis illumination in present lithography processes has enabled patterning to push below k1<0.40. This continual drive toward printing smaller feature sizes at the 193nm wavelength has led to the investigation and successful implementation of a number of new resolution enhancement technique (RET) approaches. In this work, the application of dark field imaging as a RET for low-k1 contact hole features was examined using both simulation and experiment on a standard lithography tool set. Here dark field imaging is defined as imaging where a portion of the illumination source is placed outside of the projection optics. This approach is similar to standard off-axis illumination approaches using the familiar two-beam imaging but at greater incident angle where the zeroth diffraction order is not captured and image modulation is created by higher diffraction orders. This approach creates a number of “unexpected” effects and the authors show these for contact hole patterns using ASML Twinscan lithography tools operationally modified to operate in dark field mode. They demonstrate the following new dark field effects for contact holes: reduced and negative mask error enhancement factor, inverted develop critical dimension as a function of pitch, and image tone reversal at phase edges. Additionally, experimentally relevant items such as stray light and dark field exposure energy requirements are discussed.