At 28nm technology node and below, hot spot prediction and process window control across production wafers have become increasingly critical to prevent hotspots from becoming yield-limiting defects. We previously established proof of concept for a systematic approach to identify the most critical pattern locations, i.e. hotspots, in a reticle layout by computational lithography and combining process window characteristics of these patterns with across-wafer process variation data to predict where hotspots may become yield impacting defects [1,2]. The current paper establishes the impact of micro-topography on a 28nm metal layer, and its correlation with hotspot best focus variations across a production chip layout. Detailed topography measurements are obtained from an offline tool, and pattern-dependent best focus (BF) shifts are determined from litho simulations that include mask-3D effects. We also establish hotspot metrology and defect verification by SEM image contour extraction and contour analysis. This enables detection of catastrophic defects as well as quantitative characterization of pattern variability, i.e. local and global CD uniformity, across a wafer to establish hotspot defect and variability maps. Finally, we combine defect prediction and verification capabilities for process monitoring by on-product, guided hotspot metrology, i.e. with sampling locations being determined from the defect prediction model and achieved prediction accuracy (capture rate) around 75%
In the advent of multiple patterning techniques in semiconductor industry, metrology has progressively become a burden. With multiple patterning techniques such as Litho-Etch-Litho-Etch and Sidewall Assisted Double Patterning, the number of processing step have increased significantly and therefore, so as the amount of metrology steps needed for both control and yield monitoring. The amount of metrology needed is increasing in each and every node as more layers needed multiple patterning steps, and more patterning steps per layer. In addition to this, there is that need for guided defect inspection, which in itself requires substantially denser focus, overlay, and CD metrology as before. Metrology efficiency will therefore be cruicial to the next semiconductor nodes. ASML's emulated wafer concept offers a highly efficient method for hybrid metrology for focus, CD, and overlay. In this concept metrology is combined with scanner's sensor data in order to predict the on-product performance. The principle underlying the method is to isolate and estimate individual root-causes which are then combined to compute the on-product performance. The goal is to use all the information available to avoid ever increasing amounts of metrology.
On product wafers, scanner focus is better controlled at the wafer center than at the wafer edge. This is due, in a large part, to edge roll off effects . This paper quantifies the impact of edge roll off on scanner levelling non-correctable errors and correlates this to on-product effects. The main contributors and mitigation methods are also discussed for a NXT:1950 scanner.
Advancement of the next generation technology nodes and emerging memory devices demand tighter lithographic focus control. Although the leveling performance of the latest-generation scanners is state of the art, challenges remain at the wafer edge due to large process variations. There are several customer configurable leveling control options available in ASML scanners, some of which are application specific in their scope of leveling improvement. In this paper, we assess the usability of leveling non-correctable error models to identify yield limiting edge dies. We introduce a novel dies-inspec based holistic methodology for leveling optimization to guide tool users in selecting an optimal configuration of leveling options. Significant focus gain, and consequently yield gain, can be achieved with this integrated approach. The Samsung site in Hwaseong observed an improved edge focus performance in a production of a mid-end memory product layer running on an ASML NXT 1960 system. 50% improvement in focus and a 1.5%p gain in edge yield were measured with the optimized configurations.
The concept of the multi-source focus correlation method was presented in 2015 [1, 2]. A more accurate understanding of real on-product focus can be obtained by gathering information from different sectors: design, scanner short loop monitoring, scanner leveling, on-product focus and topography.<p> </p>This work will show that chip topography can be predicted from reticle density and perimeter density data, including experimental proof. Different pixel sizes are used to perform the correlation in-line with the minimum resolution, correlation length of CMP effects and the spot size of the scanner level sensor. Potential applications of the topography determination will be evaluated, including optimizing scanner leveling by ignoring non-critical parts of the field, and without the need for time-consuming offline topography measurements.
With continuing dimension shrinkage using the TWINSCAN NXT:1950i scanner on the 28nm node and beyond, the imaging depth of focus (DOF) becomes more critical. Focus budget breakdown studies [Ref 2, 5] show that even though the intrafield component stays the same, it becomes a larger relative percentage of the overall DOF. Process induced topography along with reduced Process Window can lead to yield limitations and defectivity issues on the wafer. In a previous paper, the feasibility of anticipating the scanner levelling measurements (Level Sensor, Agile and Topography) has been shown . This model, built using a multiple variable analysis (PLS: Partial Least Square regression) and GDS densities at different layers showed prediction capabilities of the scanner topography readings up to 0.78 Q² (the equivalent of R² for expected prediction). Using this model, care areas can be defined as parts of the field that cannot be seen nor corrected by the scanner, which can lead to local DOF shrinkage and printing issues. This paper will investigate the link between the care areas and the intrafield focus that can be seen at the wafer level, using offline topography measurements as a reference. Some improvements made on the model are also presented.
With continuing dimension shrinkage using the TWINSCAN NXT:1950i scanner on the 28nm node and beyond, the imaging depth of focus (DOF) becomes more critical. Focus budget breakdown studies [Ref 1, 5] show that even though the intrafield component stays the same this becomes a larger relative percentage of the overall DOF. Process induced topography along with reduced Process Window can lead to yield limitations and defectivity issues on the wafer. To improve focus margin, a study has been started to determine if some correlations between scanner levelling performance, product layout and topography can be observed. Both topography and levelling intrafield fingerprints show a large systematic component that seems to be product related. In particular, scanner levelling measurement maps present a lot of similarities with the layout of the product. The present paper investigates the possibility to model the level sensor’s measured height as a function of layer design densities or perimeter data of the product. As one component of the systematics from the level sensor measurements is process induced topography due to previous deposition, etching and CMP, several layer density parameters were extracted from the GDS’s. These were combined through a multiple variable analysis (PLS: Partial Least Square regression) to determine the weighting of each layer and each parameter. Current work shows very promising results using this methodology, with description quality up to 0.8 R<sup>2</sup> and expected prediction quality up to 0.78 Q<sup>2</sup>. Since product layout drives some intrafield focus component it is also important to be able to assess intrafield focus uniformity from post processing. This has been done through a hyper dense focus map experiment which is presented in this paper.