Enhanced Defect Detection in After Develop Inspection with Machine Learning Disposition

McLaughlin, Matthew P.; Stamper, A. K.; Barber, Gabriel; Paduano, Janice; Mennell, Petra; Benn, Emerson; Linnane, Michael; Zwick, Justin; Khatumria, Chetan; Isaacson, Robert L.; Hoffman, Nathan; Menser, Clayton

doi:10.1109/asmc51741.2021.9435721

Cited by 4 publications

(1 citation statement)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This is rarely mentioned in the literature, probably because catastrophic whole wafer failure is a) unlikely in a reasonably mature process and b) would probably be caught by existing concrete metrology steps (e.g. [28]). Reliable die (rather than wafer) level yield prediction during wafer processing would allow a more fine-grained approach: if the die is shown to be bad after a step that prevents reworking, then further lithography steps could be skipped to relieve the pressure on litho machinery, the low throughput of which is problematic, especially for Extreme Ultraviolet (EUV).…”

Section: Introductionmentioning

confidence: 99%

Machine Learning on Multiplexed Optical Metrology Pattern Shift Response Targets to Predict Electrical Properties

Ashby,

Truffert,

Cerbu

et al. 2024

IEEE Trans. Semicond. Manufact.

View full text Add to dashboard Cite

Doing high throughput high accuracy metrology in small geometries is challenging. One approach is to build easily measurable proxy targets onto dies and make a predictive model based on those signals. We use optical Pattern Shift Response (PSR) proxy targets to build predictive models of the electrical characteristics of devices in the Back End Of Line (BEOL). Given the wide choice of PSR targets, we explore how to select combinations of them to maximise the utility of the features for building an accurate Machine Learning (ML) model; we call this approach Multiplexed Optical Metrology. We also explore the trade-off between chip area dedicated to targets and achievable accuracy. We run ML experiments using different selections of targets measured at different stages of BEOL processing: postlithography and post-Chemical-Mechanical-Planarisation (CMP). Our results show that a) reasonable predictive performance can be achieved for a reasonable area budget; b) ML model performance across target families varies significantly, thus justifying the need for careful selection of targets; c) longitudinal measurements of targets increases accuracy for no extra area penalty; d) increasing the the number of targets gives some improvement in accuracy for a dataset of this size, but relatively small compared to the increase in area budget needed.Ultimately we aim to do die-level yield prediction using these techniques. We discuss how collecting a larger dataset with appropriate yield information is the logical next step to achieving this.

show abstract

Section: Introductionmentioning

confidence: 99%