Schematic representation of inspect area, measurement point, and number of averaging pixels along the y axis (S) using for LWR measurement with terminal pc software.
Line width of 14 CD-SEM measurements collected at the same location of the Si line after alignment (black lines). The grey solid line is the average of 14 line width curves.
(a) Average pattern profile reconstructed by CD-AFM after lithography step (b) LWR (3σ0, seg , k ) measured along the PR pattern height. The average LWR is obtained by the Eq. (16).
Experimental PSDs fitted by Eq. (10) with 8, 32, 128 frame number performed on the silicon line. Open squares stand for experimental data and the solid line represents the theoretical fit. Here, the wave number is 2πn/L, where n = 0,1,.,N − 1 and L is the length of the line equal to 2200 nm. Δy is the measurement interval (i.e., L/N) equal to 5.5 nm.
Frame-number dependence of 3σ0 , 3σtrue , and 3σnoise obtained on silicon lines calculated by PSD fitting method.
Frame-number dependence of 3σtrue , ξ, and α obtained on silicon lines calculated by PSD fitting method.
(a) Error χ and (b) 3σtrue , obtained by PSD fitting method as a function of the number of average PSDs N* for silicon line. The CD-SEM images are collected at 8 integration frame.
Comparison of 3σnoise extracted by the PSD fitting and 3σnoise , seg extracted by alignment/average methods as a function of the frame number performed on a silicon line.
3σ0 , 3σtrue , and 3σnoise obtained on silicon line with (a) alignment/average method and (b) PSD fitting method. Here, 71 sets of measurements (η) are preformed on different locations of the line. Each set presents 14 (n) measurements at the same location. The total number of SEM images is 994 (N*). In (a), 3σDisp is around 0.9 nm for 3σ0,seg and 3σtrue , seg and 0.4 nm for 3σnoise , seg . In (b), open squares stand for experimental data and the solid line represents the theoretical fit. The CD-SEM images are collected at 8 integration frame.
Experimental (open squares) and theoretical (lines) PSDs with 8 and 32 number of frames performed on a silicon line. The PSDs (2) are performed using similar conditions to PSDs (1) with an interval of 2 weeks between experiments.
Pattern height dependence of 3σ0,seg , 3σtrue , seg , and 3σnoise , seg obtained on location A of the silicon lines calculated by the alignment/average method using CD-AFM.
Experimental PSDs fitted by Eq. (10) with 4, 16, and 64 integration frames performed on photoresist patterns. Open squares stand for experimental data and the solid line represents the theoretical fit.
Frame-number dependence of 3σ0 , 3σtrue , 3σnoise obtained on photoresist lines calculated by PSD fitting method.
Frame-number dependence of 3σtrue , ξ, and α obtained on photoresist lines calculated by PSD fitting method.
Skeleton of line edges (right and left) obtained by CD-AFM on photoresist (a) before and (b) after 4 integration frame CD-SEM measurement. CD-AFM measurements are performed at the location A′, to overcome the roughness variation along the line. After SEM image capture, the CD-AFM measurement is performed again at location A′ by carefully excluding the top and down of the peripheral zone of the line edges which are impacted by the shrinkage phenomenon of the resist.
Pattern height dependence of 3σ0,seg , 3σtrue , seg , and 3σnoise obtained on location A′ of the photoresist lines calculated by alignment/average method using CD-AFM before CD-SEM measurements.
Skeleton of line edges obtained by Monte Carlo simulation. The location A of the silicon line and A′ of the photoresist is virtually reproduced by using Eq. (7) and fixed 3*σtrue , seg , ξ, and α in ACF: 3.4/√2 nm, 29 nm, and 0.89 for silicon line and 5.6/√2 nm, 17 nm, and 0.53 for photoresist line. For LWR calculation, 2 line edges are simulated for each location.
Impact of the CD-AFM tip diameter on the MC silicon and photoresist lines true LWR.
Estimation of silicon line LWR (3σtrue , seg ) of one segment on location A using CD-AFM and CD-SEM metrology tools. Alignment and scan size disable methods are used to estimate CD-AFM white noise while the CD-SEM noise is evaluated by alignment method.
3σ0,seg , 3σnoise , seg , and 3σtrue , seg of a photoresist line obtained by CD-AFM and CD-SEM. The LWR (3σ0,seg ) measured by CD-AFM and CD-SEM are performed at the same location A′. CD-AFM is performed before and after a 4 integration frame CD-SEM image. The noise of CD-SEM is extracted from 250 CD-SEM measurements on the same resist line (but other location) using PSD fitting method.
Article metrics loading...
Full text loading...