Line Edge Roughness, or LER, increasingly became a problem in semiconductor lithography as critical dimensions dropped below 28nm, blew past 14nm, and are now headed below 10nm. The problem is greater than process control because LER is still difficult to measure with CD SEMs because of white noise from the SEM itself. So you’re never sure a measurement indicates how bad the roughness in a line is, as the imaged roughness could be just an artifact of SEM noise and not real because the two lie in the same bathtub of data. So a process engineer could be throwing out the baby with the bath water. So what if you could separate the two … the SEM noise from the LER, that is … so you could get an accurate measurement of true variation in a line patterned on the silicon? That’s a problem that Chris Mack, legendary in lithography circles, has attacked and now has a solution with computational metrology. He’s got a new start-up, called Fractilia and we find what it’s all about today in this conversation between him and Dan Hutcheson.
To find out more about Chris’ new company, click here: https://www.fractilia.com/
To learn more about semiconductor lithography, check out Chris’s personal website here: http://www.lithoguru.com/scientist/lithobasics.html