Achieve High Hotspot Detection Accuracy by Pattern Scoring

How to automate finding high-risk hotspots with a modularized hotspot-checking rule deck.


In this paper, we combined the hotspot pattern library and the rule-based scoring system into a modularized hotspot-checking rule deck running on an automatic flow. Several DFM (design for manufacture) properties criteria will be defined to build a “score board” for hotspot candidates. When hotspots in the input design are highlighted, the scoring system can identify whether a hotspot is a high-risk hotspot or not, and define the severity of the hotspots by extracted DFM properties. The automatic flow will detect which layers are contained in the design then generate a modular rule deck with several corresponding hotspot check modules. The flow also takes snapshots of the high-risk hotspots according to the score board automatically. After all the essential hotspot data is collected, the flow will automatically create an HTML-format report which has histograms of properties and overview graph that shows the distribution of hotspots. The aforementioned HTML report containing scored DFM properties and snapshots can help result-viewers to identify the high risk hotspots on the design quickly; namely, users can examine hotspots by snapshots without loading the whole design into layout viewer tools. By comparing the hotspot checking result with real defects from wafer data, a true hotspot’s values of DFM properties can be obtained. We believe this is helpful for users to improve their hotspot rules in accuracy.

Read more here.

Authors: Xiang Fang, Shou-Yuan Ma
Mentor Graphics Taiwan Ltd. (Taiwan)
Chien-Nan Lin, Hsu-Tang Liu, Chuan-Chun Lee, Chuen-Huei Yang, I-Y. Chang, Erwin Deng
United Microelectronics Corp. (Taiwan)
Ming-Feng Shen, Min-Ying Lu, Ling-Chieh Lin, Hung-Yueh Liao
Mentor Graphics Corp (Taiwan)
Elven Huang, Jonathan Muirhead
Mentor Graphics Corp. (United States)

Proceedings Volume 10807, Photomask Japan 2018: XXV Symposium on Photomask and Next-Generation Lithography Mask Technology; 108070L (2018) Event: Photomask Japan 2018, 2018, Yokohama, Japan

Leave a Reply

(Note: This name will be displayed publicly)