Systems and methods for semiconductor chip hole geometry metrology

A system for training a model for measuring a geometric attribute of a hole structure, comprising:

• a memory;
• and
• at least one processor operatively coupled to the memory and configured to:
• provide a model for measuring a geometric attribute of a hole structure in a semiconductor chip;
• obtain a plurality of training samples each comprising a pair of one of optical spectrum signals and one of reference signals corresponding to a same hole structure,
• wherein:
• the optical spectrum signals are captured by an optical spectrometer, and the reference signals, each corresponding to one of the optical spectrum signals, are captured by an electron microscope;
• and
• each of the reference signals is labeled with a labeled geometric attribute of the hole structure;
• input one of the optical spectrum signals to estimate an estimated geometric attribute of the hole structure using the model;
• and
• adjust a parameter of the model based, at least in part, on a difference between the labeled geometric attribute and the estimated geometric attribute in each of the training samples.

A tangible computer-readable device having instructions stored thereon that, when executed by at least one computing device, cause the at least one computing device to perform operations comprising:

• providing a model for measuring a geometric attribute of a hole structure in a semiconductor chip;
• obtaining a plurality of training samples each comprising a pair of one of optical spectrum signals and one of reference signals corresponding to a same hole structure, wherein:
• the optical spectrum signals are captured by an optical spectrometer, and the reference signals, each corresponding to one of the optical spectrum signals, are captured by an electron microscope;
• and
• each of the reference signals is labeled with a labeled geometric attribute of the hole structure;
• input one of the optical spectrum signals to estimate an estimated geometric attribute of the hole structure using the model;
• and
• adjust a parameter of the model based, at least in part, on a difference between the labeled geometric attribute and the estimated geometric attribute in each of the training samples.