Enabling scanning electron microscope contour-based optical proximity correction models

As High Numerical Aperture Extreme Ultraviolet Lithography (0.55NA EUVL) gets ready to step in the integrated circuit manufacturing world, advanced materials and patterning becomes very critical in terms of resolution, roughness, defectivity and process window. Among various combinations of mask absorber materials, mask tones and processes, bright field masks with utilization of negative-tone metal oxide resists (MORs) have emerged as one of the most promising candidates for 0.55NA EUVL patterning. Therefore, in this paper, we investigate the optical proximity correction (OPC) model accuracy of the dry resist process using a bright field low-n attenuated phase-shift mask. The use case is conducted on imec N2 (pitch 28nm) metal design, with wafer exposure performed using the imec NXE:3400 scanner. Thousands of OPC-relevant features are selected, and all the SEM images (both after develop inspection and after etch inspection) are acquired using a HITACHI CDSEM. The capability and accuracy of machine learning-based OPC models for resist (N2R) and etch (N2E) are studied.

The concept of focus blur encompasses the effect of laser bandwidth longitudinal chromatic aberration and scanner stage vertical vibration. The finite bandwidth of excimer laser source causes a corresponding finite distribution of focal planes in a range of 100nm or larger for the optical lithography system. Similarly, scanner vertical stage vibration puts the wafer in a finite distribution of focal planes. Both chromatic aberration and vertical stage vibration could introduce significant CD errors, hence can no longer be ignored in current lithography processes development and OPC development that require CD control within a few nanometers. We developed several methodologies to model the laser chromatic aberration and vertical stage vibration in OPC (Optical Proximity Correction) modeling tool. Extensive simulations were done to calculate chromatic aberration and vertical stage vibration focus blur's impact on lithography patterning for a variety of test structures. Chromatic aberration and vertical stage vibration focus blur effect was further included as an regression term in experimental OPC model calibration to capture its impact on litho patterning, and significant benefit to OPC model calibration was observed.

As the design technology node becomes smaller, k1 factor is decreasing below 0.3 and optical proximity correction (OPC) divergence is increasing. The gate critical dimension (CD) control and systematic defect inspection is becoming critical to improving circuit yield. For more accurate OPC verification and systematic defect inspection, design based metrology become increasingly important, because accuracy of simulation based OPC model verification has its limitation. In this paper, we used NGR-2100 as a design based metrology tool to confirm the accuracy of OPC modeling and process window qualification. NGR-2100 uses high energy wide-beam for high speed secondary electron sampling and large field of view. It can measure full chip CD distribution and more accurate process window compared to optical inspection tool. Because of using high energy beam, conducting layer like carbon film should be coated on photo resist patterned sample wafer to prevent local electron charging. However, coated carbon may increase CD variation. By using atomic layer deposition-type TiN layer instead of carbon, CD variation could be reduced.

As the semiconductor industry moves to the 45nm node and beyond, the tolerable lithography process window significantly shrinks due to the combined use of high NA and low k1 factor. This is exacerbated by the fact that the usable depth of focus at 45nm node for critical layer is 200nm or less. Traditional Optical Proximity Correction (OPC) only computes the optimal pattern layout to optimize its lithography patterning at nominal process condition (nominal defocus and nominal exposure dose) according to an OPC model calibrated at this nominal condition, and this may put the post-OPC layout at non-negligible patterning failure risk due to the inevitable process variation (mainly defocus and dose variations). With a little sacrifice at the nominal condition, process variation aware OPC can greatly enhance the robustness of post-OPC layout patterning in the presence of defocus and dose variation. There is also an increasing demand for through process window lithography verification for post-OPC circuit layout. The corner stone for successful process variation aware OPC and lithography verification is an accurately calibrated continuous process window model which is a continuous function of defocus and dose. This calibrated model needs to be able to interpolate and extrapolate in the usable process window. Based on Synopsys' OPC modeling software package-ProGen and ProGenPlus, we developed an automated process window (PW) modeling module, which can build process variation aware process window OPC model with continuously adjustable process parameters: defocus and dose. The calibration of this continuous PW model was performed in a single calibration process using silicon measurement at nominal condition and off-focus-off-dose conditions. Through the example of several process window models for layers at 45nm technology nodes, we demonstrated that this novel continuous PW modeling approach can achieve very good performance both at nominal condition and at interpolated or extrapolated off-focus-off-dose conditions.

The super-sensitivity of wafer critical dimensions (CDs) t o mask CDs at low k 1, known as the Mask Error Enhancement Factor (MEEF) drives the need for increasingly tighter m ask CD control. In addition, the accuracy of the model based optical proximity correction (OPC) used to compensate systematic lithographic errors is partially dependent on a stable mask CD error signature that expands mask CD control requir ements over multiple feature types. This paper presents the need for improved quantification and m onitoring of mask CD signatures that includes CD characteristics relevant to OPC model calibration. It a lso introduces and discusses a new method to characterize, quantify, and control mask signatures in a mask manufacturing environment to limit the impact of mask CD variations on the OPC model validity. Multiple approaches to imple menting this golden curve method are discussed in terms of their advantages and disadvantages.

Model based optical proximity correction (OPC) to enhance image fidelity and process robustness has become one of the most critical components that enable the low k₁ optical lithography. To meet the challenges imposed by the previously unthinkable low k₁ for manufacturing with most stringent dimension control requirements, a capable OPC model to meet such an aggressive lithography challenges has been urgently called upon. In addition to providing better accuracy for the currently implemented process technologies, the new OPC model must work well with Chromeless Phase Lithography (CPL) in which the topography on the mask is rather significant, and Double Dipole Lithography (DDL) in which two masks and two exposures are needed. It must also be able to intelligently take into account the effect from the more aggressive illuminations, usch as customer designed illuminator and experimental measured illuminator profile from the scanners. The physical and mathematical foundation of the model must be well thought of to meet the requirements for the above-mentioned applications. We have extended our Eigen Decomposition Model (EDM) for model OPC treatment into the high NA regime, in which the vector characteristics of light and thin film stack are taken into account. For CPL calibration, it has been found that 3D mask topography effect cannot be ignored in order to achieve satisfactory model accuracy.

Optical proximity correction (OPC) plays a vital role in the lithography process development of current semiconductor devices. OPC is utilized to achieve the ideal pattern shape because of the limitations of optical resolution. However, the lithography process design has become increasingly more complex due to the abundant use of OPC features. Hence, metrology requests for CD-SEM have also become more complex and diverse in order to characterize the critical OPC models. The number of measurement points for OPC model evaluation has increased to several hundred points per layer, and metrology requests for realized pattern shapes on the wafer are no longer simple one-dimensional measurements. Metrology requests include not only the traditional line width measurements, but also edge placement error (EPE) and corner rounding to identify line end shortening. Several researchers have proposed using the design layout as a template instead of the SEM image for the recipe creation of CD-SEM and EPE measurement. However, it is very difficult to achieve good matching results between the design layout and the SEM image in practical processing times. Hitachi High-Technologies has developed a robust and quick matching engine between the design layout and SEM image bitmap. The new system, incorporating this new matching engine, can automatically create a practical recipe from the coordinate information of measurement point and the design layout information, such as GDSII. As a result, the new system can vastly reduce the amount of time and number of operations required to generate a several-hundred point CD-SEM recipe for OPC evaluation. This study demonstrates the capability and presents evaluation results of this new matching engine. This new capability has proven to be a viable solution for OPC evaluation, and its efficiency will allow for quicker information turns between design and manufacturing.

Optical Proximity Correction (OPC) is the most popular technique to handle design shape distortions arising from subwavelength lithography. Existing OPC models are typically very computationally expensive and thus not efficient to be incorporated for layout optimization. In this paper, we present an efficient, yet sufficiently accurate OPC cost model which can predict the optimal location of a wire segment for OPC optimization and give an upper bound of the interference amount, guaranteeing that the interference amount is never underestimated. Based on this cost model, we propose an OPC-aware wire perturbation algorithm for post-layout interconnect optimization. We show that the effects of wire perturbation have the concavity or monotonicity property which can dramatically reduce the search space for finding the optimal location of each wire for OPC optimization. Further, we can incrementally update the OPC cost of a wire by recomputing only the affected wires because of the property of superposition of our model. Experimental results show that our algorithm can efficiently obtain much better OPC results than a state-of-the-art OPC-friendly router, based on a leading commercial OPC tool.

Deviations from paraxial image models are significant at numerical apertures (NA) planned to support immersion lithography. Apodization and rotation of polarization by high numerical optics are well-characterized phenomena. Similar behaviors follow from intrinsic properties of photomask patterns at high spatial frequencies. Diffraction efficiencies differ from values predicted using Kirchhoff boundary conditions, and depend on polarization. Pellicles also apodize diffraction patterns because reflection losses depend on angle of incidence and polarization. These effects are large enough to influence contrast, critical dimension and depth of focus of images thru pitch. These vectorial effects may degrade image quality, or invalidate models for optical proximity correction (OPC) that do not properly comprehend them. What matters for the image is the cumulative contribution of mask pattern, pellicle and optics. The objective in this study is to provide a systematic and unified mapping the optical landscape of mask patterned film and pellicle membrane materials as well as their impacts to vectorial imaging performance. We conclude that to optimize image performance, it is essential to balance the diffraction beam's angle, phase and amplitude. And there are non-unique reticle (mask scatter and pellicle membrane) material solutions to enable high (NA) immersion lithography.

Driven by steadily decreasing critical dimensions in the semiconductor technology, sub-resolution assisting feature (SRAF) has been proposed to enlarge the capability of the exposure technique to smaller pitch requirements exceeding the standard lithography specification by improving the through-focus process window. There are generally two types of SRAF implementation methodologies: the rule-based approach and the model-based approach. In this paper, the hybrid SRAF implementation methodology is proposed to benefit from these two methodologies. Firstly, the candidate rules are selected by using the process window aware optical proximity correction (OPC) model. Secondly, the candidate SRAF rules are taped out. Finally, the optimized candidates among the taped out SRAF rules are verified by wafer data. If the wafer data show that the SRAF rule will not meet the criteria, the SRAF implementation methodology will go back to evaluate other candidates. From the wafer data, it is proved that the hybrid SRAF implementation methodology can achieve an optimized through-focus process window in a short turn around time (TAT) with lower engineering and mask cost.

We have calibrated a physical resist model for extreme ultra-violet (EUV) lithography, and discuss model calibration and validation over a larger set of structures. The study is conducted on an extensive data set, collected at imec, for ShinEtsu resist SEVR-59 exposed on the ASML EUV alpha demo tool (ADT). The data set included more than a thousand measured feature widths (critical dimensions or CD) on wafer and mask. We address practical aspects of the calibration, such as the speed of calibration and selection of calibration input. The model is calibrated by simultaneously fitting 12 process windows of features with different mask CD (32, 36, 40 nm), orientation (horizontal, vertical), and pitch (dense, isolated). The smallest feature size at nominal process condition is a 32 nm CD at a dense pitch of 64 nm. Mask CD metrology was used to fit the model versus actually measured mask CD's. Cross-sectional scanning electron microscopy information was included in the calibration, to tune the simulated resist loss and sidewall angle. The achieved calibration root-mean-squared (RMS) error is ∼1.0 nm. We discuss the elements that were essential to obtain a well calibrated model. We discuss the impact of 3-D mask effects on the Bossung tilt. We demonstrate that a correct representation of the flare level during the calibration is key in order to achieve a high CD predictability at various flare levels. Although the model calibration is performed on a limited subset of the measurement data collected on 12 different patterns (one dimensional structure process windows), its accuracy is validated on a large number of patterns used to calibrate models for optical proximity correction―several hundred different feature types, at nominal dose and focus conditions. These were not included in the calibration; validation RMS results as small as 1 nm can be reached. Furthermore, we study the model's extendibility to two-dimensional end of line structures.

In the integrated circuits manufacturing process, a critical dimension scanning electron microscope machine provides high-resolution images of geometries on semiconductor wafers and measurements of the printed polygons after the photolithography process. These images and measurements are crucial for calibrating lithography process models such as the etch model, optical proximity correction model, and resist model, which are necessary for simulating the photolithography process and hence reducing the probability of defects occurrence. However, The test chips used for model calibration often contain imperfections or anomalies that require accurate detection and classification to be removed from the model calibration process. Traditionally, engineers manually filter semiconductor wafer images and their measurements, which is a time-consuming and error-prone task. Therefore, there have been attempts to use machine learning to automate image filtering. In this research, we introduce a novel approach using an image preprocessing stage followed by a transformer-based object detection model to automate detecting and classifying defects in semiconductor wafer images containing an array of lines or a matrix of contacts. The proposed approach shows improved accuracy compared to traditional methods.

OPC (Optical Proximity Correction) has been employed for over decade to address local lithographic printing effects. With the intensive scaling down of the designs as well as the increasing complexity of layout routing, lithographic process is being pushed to its theoretical limit and it has led to continuously shrinking DoF (Depth of Focus). Complex OPC model components are hence included into optical lithography simulation to ensure tolerable CD (Critical Dimension) variation and sustainable DOF of concerned layouts. For example, very complicated segmentation needs to be applied in mask correction, which comes at the cost of long runtime and requires an effective approach to consolidate the adequacy of model components during the flow of correction parameter tuning. In this paper, an approach is demonstrated to improve the accuracy and efficiency of OPC parameter tuning for mask correction. The approach starts with analyzing the target points in post-OPC database to identify those intolerable variations, followed by a pattern similarity grouping for the above intolerable layouts. Then, a concern index is established based on the CD out-of-tolerance ratio, dissection and pattern type for prioritizing the problematic variations. Then the corrective parameters are accordingly optimized to reduce the variation on highly prioritized patterns. During the iteration flow of OPC parameter optimization, the combination of pattern grouping and concern index greatly reduces required optimization iterations for OPC recipe tuning and enhances OPC convergence.

While optical lithography is being pushed to its limits, there is a general concern as to which metrology tool is more suitable for inspection of new generation devices. Scatterometry is one of the few types of metrology that has true in-situ potential for deep submicron critical dimension and profile analysis. Physical metrology is the key element in maintaining adequate and affordable process latitude in lithography processing. Accurate metrology is needed for characterizing and monitoring the processing states, such as exposure, focus, post-exposure bake (PEB), critical dimension (CD) resolution, and uniformity. In addition, scatterometry is a good candidate tool to obtain data necessary to perform model-based optical proximity correction (OPC). However, it is unknown as to current scatterometry provides necessary sensitivity to yield results acceptable for OPC usage. In this paper, we have utilized scatterometry to measure test patterns used in a model-based OPC and performed OPC on DRAM bitline core and periphery adjoining region then, its results are compared to those model-based OPC performed using data obtained from CD-SEM and V-SEM. In doing so, we have attempted to obtain an ideal model which provides best performance in context of OPC. Furthermore, we have discussed 1-D and 2-D types of test patterns that are acceptable for OPC purpose and provided the verification results for each model using commercially available software.

The evolution of high-NA projection optics design and manufacturing tolerances has been remarkable in recent years. Nevertheless, different instances of identical scanner models can still exhibit unique optical fingerprints which can impart subtle patterning differences for a given mask exposed at nominally identical conditions on different scanners. In some cases, a product can be shown statistically to yield lower when a certain layer is exposed on a particular scanner. Thus it is common to have a certain subset of the total population of tools allowed for certain critical levels, such as gate. Since a single mask is typically shared between the multiple allowed scanners, the optical proximity correction model which is employed in the generation of that mask must represent the average fingerprint of those tools. In practice, CD data may be collected from multiple tools, but more often a single tool is somehow identified as the "golden" tool for the purpose of calibrating the OPC model. Once the mask is generated, however, its printing behavior on multiple scanners can readily be simulated using tool-specific optical models. Such models can be easily generated based upon known optical fingerprint data, such as measured illumination source maps, Jones pupil or Zernike aberration files. This paper investigates the use of tool-specific optical models to elucidate the intersection of design and process variability, which will manifest differently on each scanner, depending upon subtle details of the scanner fingerprint.