The technical objective of the work reported here is to assess whether radio-frequency (RF) measurements made on coplanar waveguide (CPW) test structures, which are replicated in conducting material on insulating substrates, could be employed to extract the critical dimension (CD) of the signal line using its center-to-center separation from the groundlines as a reference. The specific near-term objective is to assess whether this CPW-based CD-metrology has sensitivity and repeatability competitive with the other metrology techniques that are now used for chrome-on-glass (COG) photomasks. An affirmative answer is encouraging because advancing to a non-contact and non-vacuum implementation would then seem possible for this application. Our modeling of specific cases shows that, when the pitch of the replicated lines of the CPW is maintained constant, the sensitivity of its characteristic impedance to the CDs of the signal and ground lines is approximately 60 /µm. This is a potentially useful result. For the same implementation, the quantity C/ w has a value of approximately 45 (pF/m)/µm, which appears to be large enough to provide acceptable accuracy.
BACKGROUND AND OBJECTIVEIn photomask mask fabrication, critical dimension (CD) metrology is typically conducted by optical transmission, Scanning Electron Microscopy (SEM), and Atomic Force Microscopy (AFM) tools. These have different advantages and limitations. A metrology implementation that is not so widely known as these three techniques is electrical CD (ECD) metrology, a variation of which is the subject of this paper. There are prior reports on the extraction of geometrical information from the features of test structures replicated on chrome-on-glass (COG) masks by electrical means. For example, one approach was to probe electrical test structures that were patterned in chrome and were then tested in a dc mode. 1 In the cited report, an important innovation was providing for the electrical-length shortening of the bridge of a test structure that was configured as a micro-potentiometer. There were provisions in the test-structure design for the extraction of a parameter named L that characterized the micro-geometry of the intersection of two features. The V/I values of the two segments of the bridge of the micro-potentiometer were extracted with Kelvin voltage taps, which enabled managing the effect of contact resistance. 2 The measured resistance values of the bridge were then used to estimate the center-to-center separations of parallel features serving as voltage taps to the bridge more accurately than had been otherwise possible as a result of correcting the physical length of each segment of the bridge † Official contribution of the