Criticality of Photo Track Monitoring for Lithography Defect Control

Precise control over the lithography process is vital to high volume manufacturing in the semiconductor industry. As integrated circuit design continues to move to smaller and smaller nodes, with increasingly intricate architectures, the number of lithography steps and their importance to the overall process grows. Current advances in Self-Aligned Double Patterning (SADP) and Self-Aligned Quadruple Patterning (SAQP) are driving these increases. As the number of lithography steps increases it becomes critical to have effective monitoring of both the lithography process and tool health. In this paper we present the current methodology for lithography Photo Track Monitoring (PTM) using Inspection, Review, and Classification to allow for excursion and tool health monitoring. This PTM qualification process is advantageous over other methodologies for evaluating lithography performance due to its similarity to production processing. PTM adds a new line of defense and captures a new signal that can be directly correlated to on-product defectivity more effectively than previous dry or coatonly qualification processes. The data generated provides feedback on material and tool issues lending it to be more useful for determining root cause of process, hardware or photo chemical concerns. PTM gives the opportunity to evaluate and determine level of risk without utilizing production wafers. After processing through the lithography track and scanner, PTM wafers are inspected on a high Numerical Aperture, normal illumination, Deep Ultraviolet laser-based inspection platform. A sampling of defect locations, per wafer, are reviewed on Defect Review Scanning Electron Microscope (DR SEM) and classified using Automatic Defect Classification (ADC). Control limits are set for the process based on statistical data trends over time allowing for Statistical Process Control (SPC) charts to be generated (Figure 1). The excursion wafers and, by correlation, lithography excursions are identified based on the SPC methodologies. Inaccurate inspection data, labeled as inspection tool excursions, can cause true lithography-related excursions to be missed. Therefore, stable and reliable inspection data is crucial. Through recipe stabilization we have been able to achieve long term stability. The effectiveness of this PTM inspection flow is highlighted in the case of a stepper striping defect caused by a fiber on the immersion hood assembly. Other lithography monitoring methodologies, track monitors and scanner particle checks, did not show this defect, emphasizing the usefulness of the PTM detection method. The PTM failing on the SPC chart, thus flagged as an excursion and prompting investigation, helped to reduce exposure of product wafers. The product wafers that did process prior to the PTM failure showed an identical striping signature proving a direct correlation of PTM to product. This correlation allowed for PTM wafers to be used to run partitions within the tool to identify root-source of defect and to verify the fiber was successfully removed instead of risking production for requalification. The importance of lithography in current semiconductor processing necessitates the ability to monitor and control the lithography tools and process. As demonstrated over three years, inspection of PTM provides stable reliable defect maps and SEM classified images that are utilized for lithography excursion monitoring.